What is the evidence that expanding the current newborn screening panel will provide clinical benefit to newborns? Health Technology Assessment Report prepared for The Newborn Screening Advisory Committee by Hilary Vallance (Chair) Carolyn Green PhD, Health Technology Assessment (HTA) Consultant with Working Group Members CF: George Davidson, Yolanda Lillquist, Hilary Vallance SCD: John Wu, Jason Ford, Greg Guilcher CAH: Dan Metzger, Kathleen De Waele TMS panel: Sylvia Stockler, Sandra Sirrs, Hilary Vallance Test Performance Parameters: Graham Sinclair, Hilary Vallance Research Support: Neelam Sandhu, Patricia Fortin Conflict of Interest: None declared

With funding from the BC Provincial Laboratory Coordinating Office

1

Executive Summary Should the current newborn screening (NBS) test panel for British Columbia (BC) and the Yukon be expanded to include more disorders? This report presents a health technology assessment (HTA) process undertaken to identify and synthesize emerging clinical research for the BC Newborn Screening Advisory Committee (NSAC) who are mandated to address this question. A realist approach was taken that combined research with contextual parameters to provide quantitative estimates of the anticipated numbers detected and health benefits for the BC newborn population for each candidate condition. Primary research was assembled using a tailored HTA framework and appraised by working groups that consisted of medical / scientific experts and HTA consulting support. The scope of this review included tandem mass spectrometry (TMS) panel (15+ disorders) as well as standalone screening tests for cystic fibrosis (CF) and congenital adrenal hyperplasia (CAH). Although Sickle cell disease (SCD) was reviewed, the full report is still in draft form and is not included here. Without documented benefit to newborns, there is little ethical justification for instituting new screening tests for the entire newborn population. The criterion, ‘health benefit to the newborns’ was, therefore, given primacy by the NSAC and provides the ethical foundation of the HTA review process. Currently in British Columbia newborn screening is offered to identify specific conditions in high-risk infants who appear otherwise healthy and for whom early treatment would improve health outcomes. The NBS laboratory at Children and Women’s Health Center tests approximately 42,000 newborns in BC and the Yukon annually. Up to 2003, three conditions were included in the routine testing panel: Phenylketonuria (PKU), Galactosemia and Congenital Hypothyroidism. Early detection and treatment of these disorders prevents mental retardation (PKU and hypothyroidism) and reduces the duration and severity of liver injury (Galactosemia). Screening for Medium Chain acyl-CoA dehydrogenase deficiency (MCAD) was added to the panel in 2003 and was made possible by obtaining tandem mass spectrometry (TMS) technology. Early detection and treatment of MCAD prevents sudden unexpected death in affected children. In October 2007, testing commenced for glutaric aciduria type I (GA-I) and long chain 3 hydroxyacyl-CoA dehydrogenate deficiency (LCHAD) as information and decisions emerged from the HTA process documented in this report. The cost implications of implementing recommendations of the NSAC arising from this HTA review are presented in a separate business case analysis. As TMS is already in use to screen for conditions on the current newborn screening panel, the resource implications are relatively minimal for the addition of conditions to the TMS panel (15+ disorders) utilizing strategies described herein to minimize false positive rates. However, new technology and operating costs would be required to institute newborn screening for CF, CAH and SCD. The key question of importance for the NSAC is: Will screening for more disorders improve health outcomes for affected children in BC and the Yukon? TMS could simultaneously detect 15+ other conditions; however, evidence of screening benefits is not as clear as for the conditions already on the panel. In 2005, NSAC reviewed two major reports that addressed and appraised the existing research evidence concerning TMS screening; one was produced by a university-based UK HTA group in 2004 and the second by a sub-committee of the American College of Medical Geneticists (ACMG), which released their review in early 2005.

i

The ACMG and UK assessments produced highly discrepant recommendations. The UK HTA recommended conservative screening by TMS for PKU and MCAD only. In contrast, the expert-based review process in the US recommended an expanded screening for 15+ conditions. European countries and Australia have broad newborn screening TMS test panels. In Canada, Saskatchewan and Ontario have for the most part followed the US guidelines. Alberta and the Maritimes NBS programs have a more restricted test panel. The other Canadian provinces are in various stages of decision-making and / or implementation regarding neonatal screening programs. The BC NSAC decided, on the basis of these discrepancies, to carefully examine the research underlying the UK and US reports before instituting a practice change affecting every birth in BC. Newborn screening for CF is now the standard of care in Australia and New Zealand, much of Europe, and many US states. Alberta has begun and Ontario is implementing, CF newborn screening. Therefore, examining the evidence and quantifying the expected impact of adding CF to the BC newborn screening panel was assessed and established as a priority by the BC Newborn Screening Advisory Committee. Notwithstanding the considerable political practice pressure to expand screening quickly, there is little ethical justification for identifying unsuspected conditions without evidence of health benefits to the newborn. The benefit to the newborn has not always been well quantified, nor has the evidence-base for each condition been documented sufficiently and explicitly to provide transparent appraisal for decision makers from parents to providers to policy makers or to the public.

Methods A transparent HTA framework was tailored for newborn screening. It leveraged primary research presented in well-resourced 2004 UK and US reviews by extracting core literature presented therein. This was then updated, reappraised locally and contextualized for BC with modest expenditure of resources. Working groups created knowledge claims based on the strongest evidence available. For each condition the strongest evidence was considered to be research studies providing direct or indirect measures of the health impact of screening on newborns obtained from study protocols with the best protection against bias. Local data and primary research were combined to estimate the impact on the health of newborns and other parameters for evaluation with standard screening program criteria. Though searches were systematically done this review is better characterized as a realist review. Given the generally low level of evidence detailed here-in, detailed critical appraisals were not conducted.

Findings The following statements represent an estimate of the impact of expanding the newborn screening panel to include each condition based on a combination of the best available research combined with features of the B.C. population and health system characteristics. They include an estimate of the number of cases that would likely be detected annually; a quantification of the expected benefits to the newborn; test performance characteristics and potential harms. Strength of evidence on newborn health impact is indicated by levels of evidence based on study design strength from level 1 randomized controlled trial to level 3b retrospective case series evidence. Statements summarizing key knowledge claims were formulated by working groups for every condition as follows:

ii

Conditions Requiring Additional Technology to Institute NBS in BC Cystic Fibrosis (CF): If CF testing is added to the routine newborn screening panel then approximately 3% of the approximately 42,000 newborns with the highest IRT values would go on to have molecular testing of the CFTR gene. Of these infants, an estimated 143 would be positive for at least one CF mutation. Following confirmatory testing with the conventional sweat test, 14 of the 143 screen positive newborns would be confirmed as cases and the remainder would be identified as carriers. The expected benefits are the prevention of one early death every 1.5 to 2.5 years (1 in 14 CF patients) and the prevention of chronic malnutrition/ stunted growth and chronic pulmonary infection leading to less treatment/fewer drugs and improved cognitive function in some of the 14 (level 1). Congenital adrenal hyperplasia (CAH) is caused by inherited defects in steroid biosynthesis,the most common of which is 21-hydroxylase deficiency. The resulting deficiency in adrenal hormones can lead to life threatening salt-wasting crises in the newborn period. Treatment when initiated early reduces morbidity (level 3a) and potentially mortality (level 3b). Approximately 3 newborns will be detected with CAH annually in BC. Approximately (~) 1 male per year would be expected to have the salt wasting form of CAH. CAH is screened traditionally using a 17-OH progesterone method. However, a high false positive rate exists due to large variation in 17-OHP levels with gestational age, birth weight and response to physiological stress in the newborn period. Even with birth weight specific cutoffs, the positive predictive value (PPV) of this method is ~ 0. 66% (~ 270 false positives for every true positive). A tandem MS-based 2nd tier CAH screening method involving measurement of steroids on the original blood spot has now been published by 2 independent laboratories. Using a first tier 17OHP cutoff (top 0.65%), the false positive rate is reduced from ~ 270 to < 26 false positives annually.

Conditions Detectable Using Existing and or Upgraded TMS Technology Glutaric aciduria type I (GA1): If GA-1 testing is added to the routine TMS newborn screening panel the expected benefits are the avoidance of an acute encephalitic crisis in some newborns that screen positive. (level 2 and level 3 evidence). This impact is likely to be realized between birth and age 3 but lifelong benefits would be realized through the avoidance of neurological disability. It is estimated that 2 newborn with GA1 would be detected every 5 years in BC (level 2c evidence) (100% test sensitivity and 17% positive predictive value) according to Lindner et al 2006. (level 2 evidence). During that period ~12 false positive test results would be generated. As for all disorders screened by NBS, all “positives” screens require appropriate and consistent information sharing with physicians and families and diagnostic algorithms to rule in or out disease. Propionic and Methylmalonic acidemia, Cobalamin A/B (PROP/MUT/CblA,B): If testing for this group of disorders is added to the routine TMS newborn screening panel the expected health outcomes in newborns will likely be poor even with the currently available treatment (level < 3c). Only 5% will have normal outcomes at age 10. One newborn will be detected with PROP/MUT/CblA,B every 1.5 years approximately (level 3c). It is likely that virtually all cases will be clinically ascertained regardless of screening. However, symptoms can be non-specific and diagnosis and treatment can be delayed. Benefits hypothesized may include reduced severity of first crisis, reduced time in hospital, and reduced need for critical care. The risk of doing harm through the detection of mild variants is low. By measuring C3 iii

acylcarnitine alone, positive predictive value is only ~ 1.5- 4%. With second-tier reflex testing measuring methylmalonic acid and methylcitrate the PPV is increased to ~ 26%. Overall, it can be justified to include PROP/MUT/CblA,B on the panel with the added benefit of having the opportunity to advance knowledge through prospective evaluation. Isovaleric acidemia (IVA): If IVA testing is added to the routine TMS newborn screening panel the expected benefits are that death may be prevented in 1 newborn with unsuspected IVA every 4 to 6 years (level 3c evidence) and treatment may lessen the severity and frequency of ketoacidotic episodes for 1 child every 4 to 6 years (level 3c). One newborn will be detected with IVA every 2 to 3 years approximately (level 3c). 3-Methylcrotonyl-CoA carboxylase deficiency (3MCC): Given the propensity to detect primarily benign variants of 3MCC (level 3c evidence) the BC NSAC recommended that 3MCC not be added to the current newborn screening panel as cases detected are not likely to benefit from early detection and treatment. One newborn would be detected with some variant of 3MCC every 2 years approximately if 3MCC were added to the routine TMS newborn screening panel. Of note, Germany has recently discontinued screening for 3MCC. Beta-Ketothiolase Deficiency (BKD): If BC commences screening for BKD there is the potential to avoid metabolic crises and subsequent death or chronic neurological deficit therefore it may be worthwhile to screen. However, the condition is so rare there is limited evidence. It may be worthwhile to detect this disorder if there are minimal false positives with the use of the C5:1 signal. However, test performance characteristics with this analyte are limited at this time and require evaluation. 3-Hydroxy-3-methylglutaryl-CoA Lyase Deficiency (HMG CoA Lyase): There has not been a case of HMG CoA Lyase diagnosed in BC (> 30 years). Therefore, detection with screening would be expected to be exceedingly rare. If there were a child born with HMG CoA Lyase deficiency, the newborn may or may not benefit from screening. There is a greater than 50% chance that the case would become clinically evident before test results were available. Survivors have a chance for normal development with treatment. It is not possible to screen for this disorder without also detecting cases of 3MCC as they both rely on the use of the C5OH analyte. Since 3MCC screening with C5OH measurement has not been recommended, it will not be possible to screen for HMG CoA lyase deficiency. Short chain acyl-CoA dehydrogenase deficiency (SCADD): The natural history of SCADD is poorly understood. It is the perspective of the NSAC that more data is required from larger screening programs before a recommendation can be made to screen for SCAD. Long chain 3 hydroxyacyl-CoA dehydrogenase deficiency (LCHAD): The expected benefits of LCHAD newborn screening are that one death will be prevented every 5 years (level 3b) though metabolic decompensation may occur despite treatment in 1/4 of patients and recurrent muscle pain in 1/3rd. Protein enhanced diets may lead to better metabolic control, reduced heart rate during exercise, more active lives and lower rates of obesity (level 1). One newborn will be detected with LCHAD every 2 years or more (level 3b evidence). Very Long Chain Acyl-CoA dehydrogenase deficiency (VLCAD): If VLCAD testing is added to the routine TMS newborn screening panel, the expected benefits are the prevention of metabolic crisis for most with prevention of organ damage and possibly death in some (level 3b or 4). One newborn will be detected with VLCAD annually with some risk of detection of mild variants. False negative results have been reported when newborn screening cards are iv

collected > 5 days after birth (level 3). This limitation will need to be considered in the development of collection procedures for expanded screening. Furthermore, infants who screen positive for VLCAD will usually require molecular +/- enzymatic confirmatory testing rather than reliance on repeat acylcarnitine testing alone to rule in or out VLCAD. Carnitine uptake disorder (CUD): The clinical course of patients with CUD appears to be highly variable. Evidence that treatment can prevent illness or death is limited. Asymptomatic family members are likely to be detected in follow up to abnormal newborn screening results. One newborn and one undiagnosed asymptomatic or possibly symptomatic mother would be detected annually. No decision regarding whether CUD should be added to the test panel has been made by NSAC at this time. Citrullinemia (CIT): If CIT testing is added to the routine TMS newborn screening panel the expected benefits are more likely to be achieved for late-onset forms compared to the early onset form of the disease. One newborn every 14 years or more may have reduced risk of premature death or neurological deficit (early onset form) (level 3b). One newborn every 5 years with a late-onset form of CIT is expected to have an improved health outcome with early diagnosis and appropriate treatment (level 3b) . Maple Syrup Urine disease (MSUD): If MSUD testing is added to the routine TMS newborn screening panel the expected benefits are that normal cognitive development (prevention of mental retardation) may be achieved with optimal outcomes predicted with detection within the first 1-2 weeks of life (level 3c). One newborn will be detected with MSUD every 7 to 10 years approximately (level 3c). Newborn screening for MSUD relies on the identification of elevated Leucine & Isoleucine (Leu/Ile). However, TMS can not differentiate between the isobaric amino acids Leucine, Isoleucine, allo-isoleucine (allo-ile) and hydroxyproline leading to a positive predictive value of ~5% and a risk of missing cases. A second-tier reflex test for MSUD with measurement of allo-isoleucine increases the PPV to ~ 100% while improving detection rates. Homocysteinuria (Hcy): If BC commences screening for Hcy the health benefit will include reduction in vascular complications, including stroke prevention, improved cognitive development / prevention of mental retardation, normal vision/ prevention of dislocated lenses, normal bone development/ prevention of scoliosis and osteoporosis and increased longevity (level 2). One newborn will be detected with Hcy every 3-4 years. Hcy is traditionally screened using methionine as a primary marker. Cut-offs set to maximize the sensitivity of this test lead to elevated false-positive rates and a low positive predictive value of 2%. Second tier reflex testing measuring homocysteine achieves a positive predictive value near 100% with only 0.3% of newborn screens being selected for second tier testing (level 1). Tyrosinemia (TyrI): If TyrI is added to the TMS newborn screening panel, early detection and treatment with nitisinone reduces the risk of liver failure, reduces the need for liver transplantation, and may reduce the risk of hepatocellular carcinoma. One newborn will be detected with some variant of TyrI every 3 years (level 3a). Elevated bloodspot tyrosine in the newborn period is nonspecific. It may indicate tyrosinemia Type I, II, or III and also infants with transient tyrosinemia of the newborn due to

v

liver immaturity. Accordingly, screening for tyrosinemia type I using only tyrosine as a marker can lead to significant false positive rates. Missed cases have also been reported. With direct measurement of succinylacetone as a primary screen for tryrosinemia, test sensitivity and specificity are ~ 100% for TyrI.

Quality of the Evidence When evaluated against a standard hierarchy of evidence, the evidence in support of newborn screening for specific conditions is relatively low with the exception of CF, for which there is evidence from randomized controlled trials (level 1). The available studies for inborn errors of metabolism and endocrine conditions are predominately uncontrolled case series (level 3). The lack of quality studies has resulted in part because these disorders are rare and TMS technology applied to newborn screening is a relatively new field of study. Studies with adequate control groups, standard testing and outcome measurement as well as adequate follow-up have not been done. Therefore, many uncertainties remain that can only be resolved in the context of rigorously evaluated population based newborn screening programs. More definitive studies are required and given the rarity of many of the conditions individually, the scale of the research and rapid advancement of knowledge, national and international partnerships would be ideal for higher quality research. Health policy decision-making often involves uncertainty under conditions of imperfect and incomplete information. The highest level of evidence was found in support of health benefit for CF newborn screening. In contrast, the available research cataloged in this report and reviewed by the BC working group for the TMS panel is indirect and has many methodological weaknesses. Nonetheless, it substantiates the modest clinical benefits that would be achieved if the TMS newborn screening panel were partially expanded with minimal laboratory cost impact given that labour and capital investment is already in place for MCAD screening. The decision to expand newborn screening programs should take into consideration contextual factors for British Columbia including opportunity costs. The opportunity costs are foregone benefits that could be achieved by using the same resources to support another health intervention to benefit population health. The cost implications of expanding the NBS panel are presented in a related business case analysis.

Conclusions This health technology assessment has reviewed and updated the clinical evidence contained in reviews conducted to inform policy in the US and UK. The discrepancy between recommendations was resolved through a consideration of the expected benefit to the newborn versus mitigated harms and costs in the context of British Columbia. Based on the findings of this HTA and related business case the NSAC recommended the addition of the following conditions to the current newborn screening panel: o Cystic Fibrosis (CF) o Congenital Adrenal Hyperplasia (CAH) o Sickle Cell disease (SCD)* Tandem Mass Spectrometry Panel (TMS) Primary Targets: o Glutaric aciduria type I (GAI) ** o Very Long Chain Acyl-CoA dehydrogenase deficiency (VLCAD) vi

o o o o

o o o o

Long chain 3 hydroxyacyl-CoA dehydrogenase deficiency (LCHAD)** Isovaleric acidemia (IVA) Maple Syrup Urine disease (MSUD) Methylmalonic acidemia (MUT) ƒ Methylmalonyl-CoA mutase deficiency ƒ Cobalamin A/B Propionic acidemia (PROP) Homocysteinuria (Hcy) Tyrosinemia (Tyr) Citrullinemia ƒ Argininosuccinate lyase deficiency ƒ Argininosuccinate synthase deficiency

Secondary targets (detected alongside other conditions but may not meet NBS criteria to be considered a primary target): o Trifunctional protein deficiency (TFP) o 2-Methylbutyryl-CoA dehydrogenase deficiency o Cobalamin C/D defects No decision has been made at this time regarding screening for: o Carnitine uptake disorder (CUD) The following conditions were reviewed and given the lack of evidence of health benefit and the potential for harms are not recommended for inclusion on the TMS test panel at this time: o 3-Methylcrotonyl-CoA carboxylase deficiency (3MCC) o 3-Hydroxy-3-methylglutaryl-CoA Lyase Deficiency (HMG CoA Lyase) o Short-chain acyl-CoA dehydrogenase deficiency (SCADD) o Beta-Ketothiolase Deficiency (BKD) * Not included in this version of HTA (in preparation) ** Added to NBS TMS panel October 2007

Out of scope The following conditions were not part of this HTA review: Carnitine palmitoyltransferase I and II (CPT1 and II) and Carnitine acylcarnitine translocase (CAT) deficiencies. They were considered a lower priority for review given that the health benefits of screening for them were unclear or controversial. As the evidence becomes clearer they may or may not be considered suitable for inclusion on the newborn screening panel. Ethical and psychosocial issues, such as informed consent, false diagnosis, parental anxiety, privacy, and confidentiality, warrant further consideration.

vii

Table of Contents EXECUTIVE SUMMARY ...............................................................................................I TABLE OF CONTENTS ............................................................................................VIII TABLES AND FIGURES ..............................................................................................IX NEWBORN SCREENING HEALTH TECHNOLOGY ASSESSMENT ................... 1 BACKGROUND.................................................................................................................. 1 METHODS FOR REVIEWING EVIDENCE ............................................................... 2 METHOD OVERVIEW ........................................................................................................ 2 A COMPREHENSIVE HTA FRAMEWORK TAILORED FOR NEWBORN SCREENING .............. 3 PRIORITIZATION PROCESS ................................................................................................ 4 SEARCH STRATEGY .......................................................................................................... 5 SELECTION CRITERIA ....................................................................................................... 6 * THESE SEARCHES WERE CONDUCTED IN LATE 2007 ...................................................... 9 DECISION CRITERIA ....................................................................................................... 10 APPRAISAL CRITERIA AND FORMS .................................................................................. 12 RESULTS ........................................................................................................................ 15 CYSTIC FIBROSIS........................................................................................................ 15 CONDITIONS DETECTABLE WITH TANDEM MASS SPECTROMETRY ...... 29 1. ORGANIC ACIDURIAS............................................................................................ 29 GUTARIC ACIDURIA TYPE I (GA-I)................................................................................. 29 PROPIONIC AND METHYLMALONIC ACIDURIA (PROP/MUT/CBLA,B) ........................ 36 ISOVALERIC ACIDEMIA (IVA) ........................................................................................ 42 3METHYLCROTONYL-COA CARBOXYLASE DEFICIENCY (3MCC) .................................. 46 BETA-KETOTHIOLASE DEFICIENCY (BKD).................................................................... 51 3-HYDROXY-3-METHYLGLUTARYL-COA LYASE DEFICIENCY (HMG) COA LYASE)..... 54 2. DISORDERS OF FATTY ACID OXIDATION AND CARNITINE METABOLISM............................................................................................................... 57 LCHAD AND TFP ......................................................................................................... 58 VLCAD......................................................................................................................... 64 SHORT CHAIN ACYL-COA DEHYDROGENASE DEFICIENCY (SCADD)............................. 74 3. DISORDERS OF AMINO ACIDS AND UREA CYCLE ....................................... 77 CITRULLINEMIA (CIT) ................................................................................................... 77 MAPLE SYRUP URINE DISEASE (MSUD)......................................................................... 82 HOMOCYSTEINURIA (HCY) ............................................................................................ 86 TYROSINEMIA TYPE I (TYRI).......................................................................................... 90 4. ENDOCRINOPATHIES DETECTED WITH STAND ALONE SCREENING TEST ................................................................................................................................ 95 CONGENITAL ADRENAL HYPERPLASIA (CAH) .............................................................. 95 viii

Tables and Figures FIGURE 1. COMPREHENSIVE HTA FRAMEWORK TAILORED FOR NEWBORN SCREENING...... 7 TABLE 1: SEARCH STRATEGIES BY DISEASE ........................................................................ 8 TABLE 1: SEARCH STRATEGIES BY DISEASE CONTINUED .................................................... 9 TABLE 2: BC NBS HTA CONDITION SPECIFIC DECISION CRITERIA FORM .......................... 11 TABLE 3. BC NBS HTA PROCESS DATA COLLECTION FORM .............................................. 12 TABLE 4: RECENT PRIMARY RESEARCH REVIEWED BY BC CF WORKING GROUP ................ 22 TABLE 5: ACMG BIBLIOGRAPHY OF CF RESEARCH ........................................................... 28 TABLE 6: RECENT PRIMARY RESEARCH REVIEWED BY BC GAI WORKING GROUP.............. 34 TEST PERFORMANCE .......................................................................................................... 37 TABLE 7: RECENT PRIMARY RESEARCH REVIEWED BY BC PROP/MUT/CBLA,B WORKING GROUP ........................................................................................................................ 41 TABLE 8: RECENT PRIMARY RESEARCH REVIEWED BY BC IVA WORKING GROUP.............. 45 TABLE 9: RECENT PRIMARY RESEARCH REVIEWED BY BC 3MCC WORKING GROUP .......... 49 TABLE 10: RECENT PRIMARY RESEARCH REVIEWED BY BC BKD WORKING GROUP .......... 52 TABLE 11: RECENT PRIMARY RESEARCH REVIEWED BY BC HMGCOA LYASE WORKING GROUP ........................................................................................................................ 55 TABLE 12: RECENT PRIMARY RESEARCH REVIEWED BY BC LCHAD WORKING GROUP ..... 61 TABLE 13: RECENT PRIMARY RESEARCH REVIEWED BY BC VLCAD WORKING GROUP ..... 67 TABLE 14: RECENT PRIMARY RESEARCH REVIEWED BY BC CUD WORKING GROUP .......... 72 TABLE 15: RECENT PRIMARY RESEARCH REVIEWED BY BC SCADD WORKING GROUP ..... 76 TABLE 16: RECENT PRIMARY RESEARCH REVIEWED BY BC CIT WORKING GROUP ............ 80 TEST PERFORMANCE .......................................................................................................... 82 TABLE 17: RECENT PRIMARY RESEARCH REVIEWED BY BC MCUD WORKING GROUP....... 85 TABLE 18: RECENT PRIMARY RESEARCH REVIEWED BY BC HCY WORKING GROUP ........... 89 TABLE 19: RECENT PRIMARY RESEARCH REVIEWED BY BC TYROSINEMIA WORKING GROUP ................................................................................................................................... 93 TABLE 20: RECENT PRIMARY RESEARCH REVIEWED BY BC CAH WORKING GROUP ........ 100 REFERENCES ............................................................................................................... 104 APPENDIX A: RECENT PRIMARY RESEARCH REVIEWED BY THE WORKING GROUP ON TEST PERFORMANCE PARAMETERS ................................................................................... 120 APPENDIX B: TMS SCREENING TEST PARAMETERS WITH FIRST AND SECOND TIER TESTING ................................................................................................................... 123

ix

Newborn Screening Health Technology Assessment Background information on newborn screening (NBS) is provided in the following section. Current practice in British Columbia (BC) is outlined first. The status of NBS in other provinces follows. The rationale for conducting a test panel review for BC policy makers is then presented. An overview of the discrepancies in recommendations and practice is then outlined. Finally, there is an examination of how differences in review methods, processes and standards may explain discordant reviews.

Background Newborn screening is offered to apparently well infants to identify those at high risk of a specific condition, for whom early treatment will improve health outcomes. The newborn screening (NBS) laboratory situated at Children’s & Women’s Health center currently tests approximately 42,000 newborns in BC and the Yukon for the following disorders: • • • •

Phenylketonuria (PKU) Galactosemia Congenital hypothyroidism, and Medium chain acyl-CoA dehydrogenase deficiency (MCAD)

Early detection and treatment prevents mental retardation (PKU and hypothyroidism) and reduces the duration and severity of liver injury (galactosemia). In 2003, tandem mass spectrometry (TMS) was introduced to screen for MCAD. Early detection and treatment of MCAD prevents sudden unexpected death in affected children. A review of the 1st 3 years experience in B.C. with MCAD screening, clinical outcomes and test performance, was recently accepted for publication [Can J Pub Health, in press]. In October 2007 screening commenced for glutaric aciduria type I (GA-I) and long chain 3 hydroxyacylCoA dehydrogenate deficiency (LCHAD) as information and decisions emerged from the HTA process documented in this report.

The Trend Towards Expanded Screening Programs All Canadian provinces with the exception of Manitoba have committed funds and are either fully implementing or currently implementing an expanded tandem mass spectrometry panel. With respect to Cystic fibrosis, Alberta implemented CF screening April 2007 and Ontario’s NBS program will add CF in 2008. BC currently is among the provinces with the fewest number of disorders screened in newborns.

Rationale for a Test Panel Review The Newborn Screening Advisory Committee periodically reviews the test panel and makes recommendations regarding the addition or removal of tests based on the Decision Criteria for Newborn Screening (see Table 2). In 2005, NSAC reviewed the available published health technology assessment (HTA) literature including the inborn errors of metabolism that can be detected by tandem mass spectrometry (TMS). The TMS test panel was reviewed by the United Kingdom HTA (UK-HTA) group in 2004.(1,2) and by a sub-committee of the American College of Medical Geneticists (ACMG) in 2005.(3) 1

These 2 reviews produced highly discrepant recommendations. The UK-HTA conducted a more rigorous review and recommended a similar TMS panel to the one currently utilized in BC plus cystic fibrosis screening while the ACMG in the US recommended screening for all possible conditions detectable by TMS (15+) plus several standalone tests (cystic fibrosis, congenital adrenal hyperplasia, biotinidase and galactosemia). Though the principles of screening promoted by both review groups were similar, the process of reviewing evidence was more rigorously applied in the UK. Neither review contained much information on test performance or a weighing of benefits to harms. Based on knowledge gaps identified in previous reviews and discrepancies in recommendations, a decision was made by NSAC to conduct a health technology assessment (HTA) with the assistance of a consultant expert and funding from the Provincial Lab Coordinating Office (PLCO). The quality of evidence was scored using a standardized approach (see Methods section). The scope of this review included the tandem mass spectrometry (TMS) panel and standalone tests for congenital adrenal hyperplasia, sickle cell disease and cystic fibrosis (CF). These 3 additional disorders were chosen because published reviews have consistently recommended newborn screening for these disorders. Therefore, it was timely to include them in this review. The NSAC findings are summarized in the third section of this report.

Methods for reviewing evidence Method overview A Health Technology Assessment (HTA) Framework was developed to provide an evidence base for the formulation of policy recommendations by the BC Newborn Screening Advisory Committee (BC NSAC). The goal was to develop a transparent and defensible HTA framework for developing policy recommendations for newborn screening in BC and to serve as a communication tool. Several features were considered by the committee to be paramount in designing the newborn screening HTA framework. The primary consideration of the BCNSAC was that the framework and eventual recommendations be transparent and defensible. ‘Health benefits for the newborn’ was identified as the key decision criteria and a defensible foundation for recommendations. Screening without a reasonable expectation of health benefit to the newborn is ethically fraught and thus is outside the mandate of the newborn screening program. Knowledge of a genetic disorder that may never cause a health problem and/or for which there is no effective treatment for example, is of doubtful benefit and may even cause harm. The key requirement in developing the tailored HTA Framework therefore, was to ensure the methodology adequately revealed the research evidence of clinical benefit for the newborn for each condition under consideration as an addition to the routine newborn screening panel.

2

Systematic review methodology as pioneered by review groups like the Cochrane Collaboration aims to apply a rigorous approach to reviewing clinical research. Systematic reviews are an important pillar of HTA though comprehensive HTA goes beyond systematic review to assemble all evidence required for health policy making. Systematic review methodology provides the standard that can be applied across health interventions. The topic of screening and in particular screening newborns for rare inborn errors of metabolism and other genetic conditions requires tailoring of these standards that were pioneered on therapeutic interventions. For example, the double blind placebo controlled clinical trial that is the standard for pharmaceutical interventions is not appropriate for detection of rare newly identified genetic conditions. In tailoring an HTA framework for newborn screening, another important principle was prudence—making the best use of material resources. Given the availability of comprehensive though discrepant reviews, the HTA framework was designed to make the most of the work of others by relying on their evidence library and updating this with more recent literature. While it was beyond the scope of the BC NSAC to oversee a comprehensive HTA nonetheless it was necessary to have access to the available evidence and to apply it to the BC context. Therefore a realist approach was used.1 Finally it was considered desirable to fully use and develop local capacity to conduct HTA and apply HTA to newborn screening decisions. The HTA framework therefore was conceptualized as a reusable template and process to coordinate local efforts. Finally the framework needed to be sufficiently transparent to serve as a communication tool amongst various stakeholder groups including BC parents, health care providers and health care system managers. The resulting framework presented in Figure 1 was designed to meet these requirements.

A Comprehensive HTA Framework Tailored for Newborn Screening The five stages of the HTA framework are presented along the left axis of Figure 1: these are, setting the HTA parameters; identifying research; creating an evidence library; synthesizing research and conducting a comprehensive HTA. The scope and parameters of the HTA were set with the active participation of the BC NSAC. This involved identifying the skills required for working groups to review the available evidence and inviting participation to obtain needed expertise. In addition, coordination and support for working groups was organized.

1 Realist approaches recognize that decision makers may need answers to more complex question than whether something works or not. Realist reviews recognize the importance of context and implementation and therefore look for empirical evidence to address a range of questions See: Pawson R; Greenhalgh T; Harvey G; Walshe K Realist review--a new method of systematic review designed for complex policy interventions. J.Health Serv.Res.Policy, 2005, 10 Suppl 1, 21-34.

3

Prioritization process The role of the committee was to prioritize candidate conditions for review. For the TMS panel, candidate conditions for priorization were sorted into the four categories of Obvious Benefits, Partial Benefits, Unclear or Controversial Benefits and No Benefit following an approach used by a European group. As the conditions for which there were obvious benefits were already on the routine TMS panel and the ones with no benefit were of low priority, this narrowed the categories of priority to the remaining two categories. Of these two categories, the ones for which partial benefit was considered possible were selected as a starting point. Those conditions of unclear or controversial benefit were identified as of second priority to be examined later given the potential for HTA to provide important insights. In addition to TMS, the other conditions that highly ranked in other reviews but that required standalone tests were also considered of priority. Standalone Tests Cystic fibrosis (CF) Congenital adrenal hyperplasia (CAH) Sickle cell disease2 (SCD) TMS panel (Partial health benefit considered possible) Glutaric aciduria type I (GAI) Citrullinemia (CIT) Long chain 3 hydroxyacyl-CoA dehydrogenate deficiency (LCHAD) Very long chain acyl-CoA dehydrogenase deficiency (VLCAD) Propionic acidemia / Methylmalonic acidemia+ (PROP/MUT) Maple syrup urine disease (MSUD) Isovaleric acidemia (IVA) The following 2 conditions were reviewed given that the emerging evidence was not favoring screening. 3-Methylcrotonyl-CoA carboxylase deficiency (3MCC) Short-chain acyl-CoA dehydrogenase deficiency (SCAD). The following 5 conditions on the TMS panel were undertaken towards the end of this HTA process, as there was the suggestion of either unclear health benefit or test performance issues: Beta-ketothiolase deficiency (BKD) 3-Hydroxy-3-methylglutaryl-CoA Lyase deficiency (HMGCoA Lyase) Carnitine uptake disorder (CUD) Homocysteinuria (Hcy) Tyrosinemia type I (TyrI)

2

Report not yet available.

4

Advances in screening approaches Towards the end of this HTA review process, evidence emerged that TMS testing performance can be improved by multi-analyte algorithms or with new 2nd tier tests applied only to the highest percentile of screening results. (4) The most recent test performance evidence was therefore reviewed by NSAC and this report updated in early 2008. In addition to the peer-reviewed published literature on test performance, information has also been derived from the United States, Region 4, Genetics collaborative HRSA project. Through a web-based data collection system, NBS programs from Canada, United States and other international participants provide data on confirmed true positive NBS cases, analyte reference ranges, percentiles and cut offs and performance metrics (detection rates, false positive rates, and positive predictive values). This data has been collated and analyzed by the principal investigator, Dr. Rinaldo, and then shared back to all participants.

Out of scope The following conditions were not part of this HTA review: Carnitine palmitoyltransferase I and II (CPT1 and II) and Carnitine acylcarnitine translocase (CAT) deficiencies. They were considered a lower priority for review given that the health benefits of screening for them were unclear or controversial. As the evidence becomes clearer they may or may not be considered suitable for inclusion on the newborn screening panel. Ethical and psychosocial issues, such as informed consent, false diagnosis, parental anxiety, privacy, and confidentiality, warrant further consideration. While the general expertise of working groups and the NSAC informed the formulation of policy recommendations, it is recognized that there is a potentially useful body of evidence contained in qualitative studies and policy analysis that would contribute useful insights to NBS policy formulation.

Search strategy The search strategies were designed to update and retrieve current research on newborn screening specific to conditions of interest with a focus on: mortality, complications, diagnosis, prevention & control, economics, epidemiology, and history where applicable and possible. Searches were broadened and subheadings were not used for diseases that were rare or for diseases that had few references. Limits applied include: humans, English language and dates covered are 2004 – Present (April 2006) inclusive. Primary databases searched were OVID Medline and Embase. In addition, Cinahl and PsychInfo were briefly scanned for unique references.

5

Selection criteria The criteria used to select research for inclusion were broad given the dearth of primary data for these largely rare conditions. Inclusion Criteria: Study design: All types of study design Subjects: Newborns, children or adults with or screened for conditions of interest Outcome: Any outcome of relevance to the health of newborns including data on test parameters Exclusion Criteria: Individual inborn errors and conditions other than phase 1 priority conditions; treatment trials, basic science studies to investigate gene mutation variation.

6

Figure 1. Comprehensive HTA Framework Tailored for Newborn Screening

7

Table 1: Search Strategies by Disease DISEASE Congenital Adrenal Hyperplasia (CAH)

Cystic Fibrosis (CF)

Glutaric Aciduria Type 1 (GAI) Isovaleric Acidemia (IA)

Maple Syrup Urine Disease (MSUD) Propionic Acidemia (PROP)

Citrullinemia

Sickle Cell Disease (SCD)

Very Long-Change Acyl-CoA Dehydrogenase Deficiency (VL CAD) Long-Chain 3 HydroxyacylCoA Dehydrogenase Deficiency (LCHAD) Note: These searches were combined as retrieval overlapped.

SEARCH STRATEGY

Adrenal Hyperplasia, Congenital/mo, co, di, pc, ep, hi [Mortality, Complications, Diagnosis, Prevention & Control, Epidemiology, History] limit 2 to (humans and English language and yr="2004-2006") Cystic Fibrosis/co, di, ep [Complication, Diagnosis, Epidemiology] limit 1 to (human and English language and yr="2004 - 2006") Glutaryl-CoA Dehydrogenase/ glutaric aciduria type 1.tw. glutaryl aciduria 1.de. Isovaleric Acid/ isovaleric acidemia/ep, di [Epidemiology, Diagnosis] isovaleric acidemia.mp. Maple Syrup Urine Disease/co, di, ep, hi [Complications, Diagnosis, Epidemiology, History] propionic acidaemia.mp. Propionic Acidemia/ep, di, dm, su, dt, th [Epidemiology, Diagnosis, Disease Management, Surgery, Drug Therapy, Therapy] Citrullinemia, argininosuccinate synthase, argininosuccinate synthase deficiency, argininosuccinate lyase, argininosuccinate lyase deficiency [Anemia, Sickle Cell/mo, co, di, pc, ec, ep, hi [Mortality, Complications, Diagnosis, Prevention & Control, Economics, Epidemiology, History]] limit 1 to (humans and English language and yr="2003 - 2006") Sickle Cell Anemia/co, pc, di, si, ep [Complication, Prevention, Diagnosis, Side Effect, Epidemiology] very long-chain acyl-CoA dehydrogenase deficiency.mp. VLCAD.mp. *"Acyl-CoA Dehydrogenase, Long-Chain"/ LCHAD.mp. *"3-Hydroxyacyl CoA Dehydrogenases"/ Acyl-CoA Dehydrogenase, Long-Chain"/

8

Table 1: Search Strategies by Disease Continued Short Chain acyl-CoA dehydrogenase deficiency Short chain acyl-CoA dehydrogenase deficiency (SCAD)* 3 methylcrotonyl CoA carboxylase, 3 methylcrotonyl 3-Methylcrotonyl CoA CoA carboxylase deficiency, multiple carboxylase carboxylase deficiency deficiency, holocarboxylase deficiency (3MCC)* Hereditary tyrosinemia type I, tyrosinemia, Tyrosinemia (Tyr-I)* hepatorenal tyrosinemia Homocysteinuria (Hcy)* Carnitine uptake disorder (CUD)* HMG CoA Lyase deficiency* Beta-ketothiolase deficiency*

Cystathionine beta-synthase deficiency, cystathionine synthase deficiency, homocysteinuria, homocystinuria Primary carnitine deficiency, carnitine uptake disorder, carnitine transporter defect Hydroxymethylglutaryl-CoA lyase deficiency Ketothiolase deficiency, beta-ketothiolase deficiency

* These searches were conducted in late 2007

9

Decision Criteria The decision criteria are distinct from those used to evaluate the quality of research. The following principles are essentially those developed by Wilson and Junger and have been in use since the 1960’s. (5) The decision criteria form (Table 2). based on these principles, was used to assemble information pertaining to conditions of priority in this HTA. The Condition 1. The condition should be an important health problem 2. The epidemiology and natural history of the condition should be adequately understood The Test 3. There should be a simple, safe, robust and validated screening test 4. The distribution of test values in the target population should be known and a suitable cut-off level defined and agreed 5. There should be an agreed policy on the further diagnostic investigation of individuals with a positive screening test result The Treatment 6. There should be an effective treatment or intervention for patients identified through early detection with evidence of early treatment leading to better outcomes than late treatment 7. There should be agreed evidence-based policies covering which individuals should be offered treatment and the appropriate treatment to be offered 8. Clinical management of the condition and patient outcomes should be optimised in all health care providers prior to participation in a screening program

The Screening Program 9. There should be evidence that the complete screening program (tests, diagnostic procedures, treatments/ interventions) is clinically, socially and ethically acceptable to health professionals and the public 10. The benefit from the screening program should outweigh the physical and psychological harm (caused by the test, diagnostic procedures and treatment) 11. The opportunity cost of the screening program (including testing, diagnosis and treatment, administration, training and quality assurance) should be economically balanced in relation to expenditure on medical care as a whole (i.e. value for money).

10

Table 2: BC NBS HTA Condition specific decision criteria form Criteria / Issue The condition: Incidence Phenotype at birth

Burden if untreated (Natural history of disease) The test

The treatment: Efficacy of existing treatment

Cost of treatment Availability of treatment Simplicity of therapy Cost per newborn screened / case detected Is there a satisfactory confirmatory test? Are there potential harms?

Familial and societal benefits of early intervention

Definition ¾ 1:5,000 ¾ 1:25,000 ¾ 1:50,000 ¾ 1: 75,000 ¾ < 1:100,000 Signs and symptoms clinically identifiable in the fist 48 hours after birth: Never < 25% of cases < 50% of cases < 75% of cases Always Morbidity: Severe, moderate, mild, minimal Mortality rate:

Level of Evidence

If there is more than one recognized test / test algorithm, compare Sensitivity, feasibility, robustness / costs / positive predictive value Category 1 Obvious health benefit with early therapeutic intervention. Category 2 Partial health benefits: some children may not benefit because of limited possibilities for treatment currently available. In addition usually treatment can only improve the condition but not avert all symptoms. Category 3 Unclear / Controversial: Potential health benefit for the individual is present but current knowledge on the course of the disease under early treatment is still limited. Category 4 No health benefit Is treatment limited? Is there consensus on treatment?

Yes / No Technology A versus B (if data available). False positive rate. Inadvertent detection of mild disease which may not benefit from therapy. Inadvertent detection of a 2nd disorder for which there is no treatment or no clear benefit to treatment. Potential for discrimination? Inadvertent carrier detection. Early knowledge of recurrence risks and availability of prenatal testing. Prevention of costly medical diagnostic investigations. Have costing studies been done?

11

Appraisal criteria and forms The form presented below guided condition specific NSAC working group members in their appraisal of primary research. It was adapted for use in newborn screening for rare disorders. The state of the science for rare disorders is less advanced than that for common conditions and risk factors. Rather than prospective collection of primary data, clinical/administrative datasets were frequently the sources of data for estimates of prevalence, incidence and prognosis under various screening and treatment protocols. Given the lack of research with design features that protect the study from bias, these standard criteria were more used as guidelines or prompts to help identify the susceptibility of the study to bias. For the most part the level of bias was taken as being implicit in the level of evidence represented by these case series.

Table 3. BC NBS HTA process data collection form Date:__________ Reviewer:____________________________1. Study ID Author/Date (e.g. Smith2001b) Title Study description Setting: Country, state, region

Participants: How detected? Eligibility criteria, age, diagnosis, denominator, ethnicity Study duration

Intervention: Differences in intensity intervention group versus control? Analysis: How was available data collected and analyzed

Outcome measure: All reported (including final, intermediate, process and cost outcomes)

12

Appraisal criteria for cohort studies (including case control conducted using clinical and administrative data sets) Is there sufficient description of the groups and the distribution of prognostic factors? Is the intervention/treatment reliably ascertained?

Was there adequate adjustment for the effects of these confounding variables? Was outcome assessment blind to exposure status?

What proportion of the cohort was followed-up?

Are the groups assembled at a similar point in their disease progression? Were the groups comparable on all-important confounding factors? Was a dose-response relationship between intervention and outcome demonstrated? Was follow-up long enough for the outcomes to occur? Were drop-out rates and reasons for drop-out similar across intervention and unexposed groups?

Hierarchy of Evidence: Where does this study fit? Criteria for screening and treatment studies I. Well-designed RCTs 2a. Well-designed controlled trials with pseudo-randomisation or no randomisation 2b. Well-designed cohort studies with low risk of confounding, bias or chance and a moderate probability that the relationship is causal i)Prospective with concurrent controls ii)Prospective with historical controls iii)Retrospective with concurrent controls 2c. Well-designed case-control (retrospective) studies with low risk of confounding, bias or chance and a moderate probability that the relationship is causal 3a. Evidence from case-control or cohort studies with a high risk of confounding, bias or chance and a significant risk that

13

the relationship is not causal 3b. Non-analytical studies, e.g.: case reports, uncontrolled case series 4. Expert opinions of respected authorities based on clinical experience, descriptive studies and reports of expert committees

Criteria for screening and treatment studies I. Well-designed RCTs 2a. Well-designed controlled trials with pseudo-randomisation or no randomisation 2b. Well-designed cohort studies with low risk of confounding, bias or chance and a moderate probability that the relationship is causal i)Prospective with concurrent controls ii)Prospective with historical controls iii)Retrospective with concurrent controls

14

Comment [DHW1]: Is this a major heading? If so it does not appear in your index.

RESULTS The evidence of each of condition reviewed for this health technology assessment is presented separately. After a brief introduction, a summary is provided of the knowledge claims formulated by working groups on their review of the body of clinical evidence available. These statements represent the most likely impact in terms of test results and health impact if each condition is added to the provincial newborn screening program. This is followed by a narrative description of the main studies used to formulate the recommendations and summary tables.

Cystic Fibrosis First distinguished in 1938, Cystic fibrosis (CF) is the most frequent life-threatening autosomal recessive condition among people of European descent, though it is also found in other populations. The incidence of CF is estimated at 1 in 2500 – 4500. In the UK, the incidence is about 1 per 2,500 live births with an incidence in BC of 1 per 3,673. (6) Patients with CF have a defect in the cystic fibrosis transmembrane conductance regulator (CFTR) gene. Newborns and toddlers with CF are often diagnosed after a long symptomatic interval during which they fail to thrive. Failure to thrive is due to respiratory symptoms with or without infections that result from an inability to clear secretions from respiratory airways as well as pancreatic insufficiency. The acquisition of chronic Pseudomona aeruginosa infection is linked to progressive lung deterioration and reduced life expectancy. (7) A small percentage of CF cases (10%) present neonatally with meconium ileus -- a clinically recognizable condition (because the first stool of the newborn is not able to be passed normally) leading to rapid diagnosis of CF. It is understood that newborn screening is unlikely to benefit this small sub-group of patients. A recent study on the natural history of CF in BC provides useful background information. (6) Steinraths et al, (in press), conducted a 12 year retrospective review of over 100 newly diagnosed CF patients in British Columbia (level 3b). The incidence of CF in British Columbia was found to be comparable to the rest of Canada: 1:3673 live births. The average delay in diagnosis of CF was 2.1 years which is consistent with a recent national survey. (6) At diagnosis the growth parameters were: • 46% of cases were below the 5th percentile for weight, • 34% of cases were below the 5th percentile for height. • 70% of CF patients were vitamin E deficient at the time of diagnosis with lower mean occipito-frontal head circumference (OFC) in the vitamin E deficient group

15

Screening protocols for CF vary around the world. The most commonly used screening method is measurement of immunoreactive –trypsinogen (IRT) on the conventional blood spot card. Values of IRT in CF patients are typically higher than in normal newborns yet the specificity of the IRT test alone is not high enough as a screening tool. More commonly, screening laboratories use a “2-tier” protocol involving an IRT primary screen cut-off and then either reflex CFTR mutation panel testing called “IRT-DNA” or repeat IRT called “IRT-IRT”. Two-tier approaches to screening increase specificity thereby reducing the false positive rate. The IRT-IRT approach has the advantage of not identifying CF carriers. However, the recall rate is high and is best suited to NBS programs that have a protocol for routine repeat card collection (2 cards requested on all newborns). In BC, there is no routine requirement for a 2nd card collection. Therefore, only the IRT-DNA approach was reviewed in detail. The most common mutation, delta-F508 accounts for approximately 70% of mutations in CF patients. Confirmation of diagnosis is either by the identification of 2 knowndeleterious mutations and / or sweat testing. The latter test is considered the ‘gold standard’ for CF diagnosis.

Screening approaches Comeau et al, 2004 conducted an analysis of population data to optimize a screening strategy for CF using the IRT / DNA approach. (19) Their algorithm was applied to the BC context to produce expected numbers of screening and follow-up tests that would result. As the application of this algorithm shows, when a newborn is identified with a positive IRT and one mutation in the CFTR gene, the family will be referred to the CF center for sweat testing and genetic counseling. An additional 4 CF cases will be detected and the remainder (~130) will be reassured that their child does not have CF but will be counseled with respect to the identification of carrier status. . . An additional impact of identifying 130 newborns each year in BC as carriers of the CF gene is a predicted cascade of testing in the extended families of each child that may generate as many as ~ 450 additional CF carriers. There are no immediate health impacts or benefits for the newly diagnosed carriers although this knowledge may lead to more informed family planning throughout the extended family and fewer CF births in the future. On the other hand there also may be harms associated with new knowledge that this HTA does not explicitly quantify. These may include stigmatization of newly detected carrier or a burden of anxiety for parents.

Setting a Threshold for CF Detection The higher the IRT, the more likely that an infant has CF when only 1 mutation is detected. Perinatal stress and other factors can increase IRT. Detecting only the Delta 508 defect through DNA testing detects 50% of cases compared to 75% with a multiple mutation panel. Using the latter approach, one of the advantages is faster referral to a CF center and reduced wait times for sweat testing. (19)

16

Parad & Comeau, 2005 outline the diagnostic dilemmas for BC in setting the IRT cutoff value for screening. (20) An IRT cutoff value can be chosen for use with single mutation testing that prompts DNA testing however a higher false negative result rate can be expected. Alternatively a mutation panel can be selected appropriate to the population to minimize false negative results -- this ideally only identifies infants in whom classic CF will develop. A failsafe approach is to refer infants with an extremely high IRT for sweat testing regardless of DNA results. However sweat testing is not a perfect test so repeat testing may be required. Particularly in the first month of life, too little sweat may be obtained to generate a reliable measurement necessitating repeat testing. It is expected that this will occur in approximately 5% of cases based on experience at BC Children’s hospital. If the sweat chlorine (Cl) value is in the borderline range (30-59 mEq/L) the natural history of CF is not well defined. A follow up protocol is needed to better understand the risks in this group. Approximately 11% of CF cases may fall in this range.

Knowledge Claims Formulated by BC Working Group on Cystic Fibrosis What is the evidence that newborns with CF will have improved health outcomes with early detection following newborn screening for CF? If BC commences screening for CF, the evidence reviewed in this HTA indicates that over a five-year period 2008/09 to 2012/13: • 15 newborns will be detected annually (1:3673 in 42,000 births) at an average annual testing cost of $235,589. • In 2008/9, per year would generate an estimated 41,717 initial IRT screening tests (1 per 41,717 estimated BC births) and approximately… • ~1,657 newborns with IRT > 96% (top 4%) would require further DNA testing. • Of these ~143 will be positive for at least one mutation ~1514 would have no observed mutation ~8 would have 2 CFTR mutations (consistent with CF) • ~143 babies will be referred for sweat testing • An additional 4 cases of CF will be identified through sweat testing • Therefore, a total of 12 of 143 infants referred for sweat testing (8%) will be diagnosed with CF3 • 131 of 143 (92%) will be found to be carriers (a potential harm) and these families will require counseling to mitigate harms associated with carrier detection • Cascade screening of extended family members can be expected and therefore additional CF carriers will likely be detected. This additional demand for testing will likely generate increased demand on counseling. There is the 3

Positive predictive value (PPV) is 8.5%. If you add in the really high IRT’s, the PPV increases to 9.4%

17

•

potential for identifying carrier couples and these couples would be offered prenatal testing services. Expected benefits for the newborn are: • 1 early death may be prevented every 1.5 to 2.5 years (4/234, 520) (level 1) (8)(9) • 10 of 14 may have Vitamin E / essential fatty acid deficiency prevented (level 1) (70% cases detected clinically in BC Vitamin E deficient)(6) and subsequent improvements in head circumference measures and cognitive function (10)(11) • Some of the 14 may have chronic malnutrition / stunted growth (height/weight) prevented (level 1)(12) • Some of the 14 in which pulmonary infections are recognized earlier may have a delay or eradication of chronic infections (level 2a)(13) however without screening BC has had 90% success with eradication. • Some of the 14 may require less treatment and fewer drugs as a result of the prevention of chronic pulmonary infections (level 2b)(14), • Some of the 14 may have better chest X-rays though there is conflicting evidence on whether chest X-ray is better or same. In the US there is no longer statistically significant difference after 10 years (level 1), in UK chest X-rays were consistently better in screened group compared to others with the same genotype (level 2b)

The Clinical Evidence Base The cumulative research of relevance to screening decision criteria on CF is large and much of it of a substantially higher quality than that for other candidate conditions for newborn screening. For this updated review, the baseline evidence base was taken to be the ACMG CF review which included an older Cochrane review in its review of evidence to 2004 (see table 5 for ACMG bibliography). (3) The ACMG produced a recommendation to include CF screening in the uniform newborn screening panel. The UK as well as some Canadian provinces and US states have adopted CF screening. An updated search was conducted July of 2006 (see methods section). This identified 132 relevant papers. The 48 potentially meeting selection criteria were retrieved. Of these 21 papers were presented to the working groups for review. The working group focused their resources on the best available evidence for estimating the impact of introducing CF to the BC newborn screening panel. The quality of research evidence for CF newborn screening was found to be of an overall higher caliber than for inborn errors of metabolism. CF is relatively more common and screening technologies have been available for many decades. The randomized controlled trial (RCT) is close to the top of the hierarchy of evidence being only superseded by systematic reviews of RCT evidence. There are two RCT’s available: the Wisconsin CF trial RCT with longitudinal data to 16 years (11) and the UK RCT by Doull et al, 2003. (8)

18

Early Mortality Over the last 30 years survival has improved largely without screening yet young children with CF are still at risk of death in childhood without early diagnosis. A UK study (Wales and West Midlands), randomized infants to be screened or unscreened over a 5-year period (1985-90) (level 2a evidence). (8) There were 4 early deaths among 59 patients with CF who were clinically ascertained (excluding those cases who presented with meconium ileus). There were no deaths in the screened cohort (n=74). These results are consistent with the other studies including the Wisconsin CF RCT (12) in which one in 15 patients in the clinical ascertained arm of the trial died with no early deaths in the screened group (level 1).

Growth and Nutrition The Wisconsin CF RCT provided evidence on the impact of screening on health outcomes related to growth and nutrition. (12) Between 1985 and 1994, newborns in Wisconsin were randomized to either ‘screened’ or control group. All newborns in the study were tested for CF but the results were kept from the control subjects until age 4. The results from this trial have been published in a series of papers over an extended follow-up period (level 1 evidence). In Farrell et al, 2001, children diagnosed through screening had significantly higher Z scores for height and weight compared with children in the control group. (9) Follow-up at 16 years showed height and weight differences persisted over the long term. The odds ratio for having a weight below the 10th percentile in the control group versus the screened group was 4.1 (95th % CI 1.6-10.4). The odds ratio for having a height below the 10th percentile was 4.6 (95th % CI 1.7 – 12.6). The evidence from further studies conducted in Italy and Australia suggest that a normal growth pattern can be achieved and maintained in most CF patients when treatment is started before they become malnourished (level 3b evidence). (15)(16)

Cognitive function Lower cognitive scores and smaller head circumference were correlated with indicators of malnutrition including low Vitamin E levels in CF patients in the Wisconsin trial. (12) Fat-soluble vitamin deficiencies such as vitamin E deficiency were common. (12) Vitamin E deficiency develops rapidly in CF patients who have pancreatic insufficiency and these deficiencies persist without Vitamin E supplementation.

Growth restriction of clinically ascertained CF cases in BC The BC study by Steinraths et al, (6) found that Vitamin E deficiency and restricted growth was prevalent amongst the 100 CF patients diagnosed without screening. Seventy percent of CF patients were Vitamin E deficient at the time of diagnosis. Amongst those patients who were Vitamin E deficient at diagnosis the mean occipito-frontal head circumference (OFC) was at the 32nd percentile (95% CI 24th-41st percentiles). This is significantly lower than OFC at diagnosis for those who were Vitamin E sufficient, whose mean OFC was at the 63rd percentile (95% CI 47th-78th percentiles). This local evidence documents that malnutrition exists amongst the BC CF population currently 19

diagnosed without screening and therefore suggests that malnutrition is likely to be reduced by NBS.

Lung damage and pulmonary function A retrospective cross sectional study by Sims et al of 7,294 CF cases included 73% of all registered cases in the UK in 2001/2002 (13) Cases were identified as screened or clinically ascertained and were matched by genotype (delta 508 homozygotes). Cases detected through screening had significantly better pulmonary scores than the clinically ascertained cases. When homozygous delta 508 cases were compared, there was a 50% lower prevalence of chronic Pseudomonas aeruginosa (PA) infection in the screened group. Clinically ascertained cases had worse chest x-ray scores, yet no difference was found in lung function (in agreement with the Wisconsin trial). The authors concluded that there was a benefit with screening in genetically homogeneous CF disease on the basis of improved anthropometric (growth), chest x-ray and morbidity scores within the context of modern center care. The Wisconsin CF RCT trial did not demonstrate the benefits in pulmonary function reported by Sims et al above. (11) Screened participants at the time of diagnosis had better chest X-ray scores and more NBS screened subjects had lung function tests within the normal limits at 7 years of age. However, over time X-rays worsened in the screened groups and long-term differences in lung function were statistically not significant between groups. A closer analysis of the groups as randomized revealed that despite adequate randomization, the screened group contained more CF cases that had pancreatic insufficiency, delta 508 genotypes and acquired PA infection at younger ages. Investigation revealed that more of these cases were treated at a clinic that integrated young and old people irrespective of their pulmonary status and therefore screening may have inadvertently facilitated the spread of PA at younger ages to screened cases. The validity and generalizability of the Wisconsin Trial findings of no benefit on measured pulmonary function is questionable.

Treatment intensity In a retrospective, cross sectional UK study, (14) screened and clinically ascertained cohorts were compared over three levels of treatment intensity; low intensity (inhaled therapies or oral antibiotics), medium intensity (nebulized therapies or oral corticosteroids) and high intensity (intraveneous antibiotic admin: at least one course of IV therapy per year). Screened cases received significantly fewer and less demanding therapies compared to the clinically ascertained cases. Clinically ascertained patients required a greater number of therapies to maintain a given lung function.

Consistency of evidence The Cochrane Collaboration and others have revealed the limitations of the evidence base for CF screening through rigorous critical appraisal. (17,18) Though randomization provided Wisconsin CF Trial with the advantage of the most rigorous study design, the outcome benefit may have been offset by earlier acquisition of PA infection as a result of inadvertent earlier exposure to infective agents. The knowledge and experience that has been gained in CF over the last 15 years has dated this pioneering study. For example, 20

infection control is better and enhanced treatment of malabsorption was universal by 1994 when enrollment in the Wisconsin trial ended. . The findings of the UK Doull et al, 2003 RCT may be more valid. (8) The UK study did not encounter the problem of high infection rates and so may be truer to the universal experience of screening. The numbers of deaths however were too small to reach statistical significance therefore the mortality advantage of screening is poorly supported by the available RCT evidence. With the available non-randomized analyses, it is problematic to simply compare a screened group with an unscreened group because the screened group is likely to include more mild CF phenotypes and therefore any measured outcome can be expected to be better just due to this confounding affect. Analysis of the clinical databases have been conducted in both the US and UK. The Sims et al, 2005 has the advantage of controlled for genotype to compensate for the excess in the number of mild CF cases. (13) It was also undertaken within the context of a modern care center. Moreover it checked for confounding factors such as social deprivation and corrected for the referral bias by matching screened and clinically ascertained patient by volume of referrals. In the absence of conclusive evidence, the issue of consistency of evidence is salient. Taken together these studies, which represent the highest level of available evidence, point to a consistent health benefit with screening. Concern of “over treating” screening detected cases is not supported by recent analysis.

21

Table 4: Recent primary research reviewed by BC CF working group

Author

Setting

Steinraths et al, (6)

British Columbia

Farrell et al, 2001 (9)

Wisconsin Screening Randomized Controlled Trial

Farrell et al, 2005 (11)

Wisconsin Screening Randomized Controlled Trial

Farrell et al, 1997 (12)

Wisconsin Screening Randomized Controlled Trial (Clinically detected group unblended after 4 years to pick up mild cases) Wisconsin Screening Randomized Controlled Trial

Koscik et al, 2004 (10)

Subjects

Outcomes

>100 detected without screening in a 12 year retrospective review 650 341 newborns randomized to screening or non screening

70% Vitamin E deficient at diagnosis. Of these the mean occipito-frontal head circumference was at the 32nd percentile (95% CI 24th-41st percentiles versus 63rd percentile (95% CI 47th-78th percentiles) among Vitamin. E sufficient.

650,341 newborns randomized between April 15, 1985, and June 30, 1994; 49 screened; 31 Control CF without meconium ileus who had pancreatic insufficiency 650,341 infants were screened; 325,171 early diagnosis group; 74 infants diagnosed with CF

Follow up at 16 years showed height and weight differences persistent long term • NBS cohort: better chest X-ray scores at diagnosis… deteriorated after 10 years due to confounders such as PA infection • Screened pts had significantly better length / height / weight / head circumference • malnutrition and growth failure prevented. • opportunity for better pulmonary outcomes though other factors can predominate over time and affect pulmonary prognosis

89 randomized subjects receiving early diagnosis and nutritional therapy

Better cognitive functioning inversely proportional with duration of vitamin E deficiency

Main outcome: enhanced long-term nutritional status

Malnutrition prevented • Higher height and weight percentiles • Higher head-circumference percentile (52nd, vs. 32nd) at diagnosis. • Significantly higher anthropometric indexes especially those with pancreatic insufficiency homozygous for delta F508 mutation.

22

Author

4 5

Setting

Subjects

Outcomes

Sims, McCormick et al, 2005a (13)

UK CF Database; Retrospective cross-sectional study 2001; 41 CF centers and 12 smaller CF clinics representing 73% of registered UK CF population

184 screened patients (98 homozygous 508) aged 1 to 9 years in 2002 compared with 950 matched CD4 controls (531 homozygous 508) in 3-year age groups (7294 patients)

Sims et al, 2005b(14)

As above with analysis by therapeutic intensity: low (inhaled therapies or oral antibiotics; medium (nebulized therapies or oral corticosteroids); 1 IV antibiotics/yr

As above Chronic PA: > 3 pos cultures in previous 12 mo. Intermittent PA defined as 1 or 2 pos cultures over same period

NBS segregates with better outcomes in patients as old as 6 years compared with age- and genotypematched control subjects who are clinically detected (CD) • improved median height • reduced morbidity • lower prevalence of transitory or chronic pulmonary infection/ better X-ray scores in NBS group with best scores in < age 4 • NBS group had 50% lower pulmonary scores in homozygous delta 508 subgroups compared with CD • No difference in weight with homozygous delta 508 subgroups • No difference mean or median FEB1 or FVC (lung function) for populations aged 5 years and older Authors report that erroneous conclusion can be reached that NBS is of no benefit if treatment intensity is not taken into account. • 70% PA free in NBS group compared to 60% PA free in CD group • Similar lung function can be achieved after NBS despite a significant lower number of long term therapies compared with CD controls • NBS pts < 6 years of age received significantly fewer and less demanding therapies • 7-9year olds: fewer NBS pts received IV Ab • NBS patients without PA, received fewer therapies but no differences were found between intermittently or chronically infected NBS and CD populations

Mastella et al, 2001 (15)

2 observational cohorts from 2 regions in Italy (Verona, Sicily) comparable for CF incidence, 5 CFTR genotypes, gender proportion and common treatment protocols

126 CF detected by screening with IRT compared with 152 patients clinically detected

• •

Malnutrition prevented Better survival and nutritional status

CD = clinically detected CFTR = cystic fibrosis (CF) transmembrane conductance regulator gene

23

Author

Australia: Observational Study

Comeau et al, 2004(19)

Massachusetts hospitals – a population-based approach

Parad & Comeau 2005(20)

6

Setting

Waters et al, 1999 (16)

Massachusetts CF NBS program

Subjects

Outcomes

57 with CF born before (1978 to mid 1981) and 60 CF born with screening (mid 1981 to 1984), followed to age 10 4-year cohort of 323, 506 infants born between February 1, 1999, and February 1, 2003, and screened for CF at 2 days Total number positive CF NBS = 1338, sweat data for 1214); total number CF true positives = 110 (# screened?)

Age and sex adjusted height and weight as well as pulmonary function tests (FVC/FEV) were higher in the screened vs. clinically ascertained group whereas there was no difference in x-ray scores at age 10. Authors conclude that screening may be important in determining subsequent clinical outcomes however there was no adjustment by CF genotype in CF

6

After review of data from 1st 9 months, IRT threshold was adjusted to 95th daily % for increased specificity. Use of multiple-CFTR-mutation testing improved sensitivity and post screening prediction of CF at the cost of increased referrals and carrier identification. Positive predictive value: 7.5% Carriers detected: 74% of the false positives or 904 / 1214 or 904 / 1338 (67%) Four problematic diagnostic categories generated by CF NBS were defined. In the absence of data on the natural history, careful follow-up was recommended

IRT = Immunoreactive trypsin a screening test for cystic fibrosis

24

Setting

Subjects

Outcomes

Scotet et al, 2001 (21)

Author

West Brittany, France (1992 to 1998)

160,019 tested with IRT IRT threshold was top 1.2% n=1964

“IRT levels of CF neonates were all located in the highest 0.5% of IRT values…” Recall rate only 0.5% Direct analysis of 3 exons of gene that contain most CFTR mutations 60 CF (prevalence 1:2667) and 213 carriers detected Carrier frequency 12.8% - a higher proportion of ‘mild’ mutations were observed

Sarles et al, 2005 (22)

5 French regions 2002 - Dec. 2003 Investigating using a combination of IRT and Pancreatitis associated protein (PAP) assays to avoid DNA analysis

van den Akkervan Marle et al, 2006(23)

Decision analysis of 4 approaches to neonatal CF screening in the Netherlands including cost and life-years gained

All newborns (n = 204,749) 1177 had elevated IRT at day 3 with the cut off 50 ng/mL until 2003 and 65 nm/mL after Sweat test 40 – 59 checked at age 3 6 and 12 months plus for symptoms Hypothetical cohort of 200, 000 neonates (annual births in the Netherlands) Birth prevalence of CF 1 in 3600

Doull et al, 2001 (8)

7

Randomized study of all births Wales and most of West Midlands 1985 – 1990 from a total population is 6.8 million accounting for 12% of UK CF births

230,076 infants were randomized to be screened/ and 234,510 to be unscreened

Threshold of 50 ng/mL for IRT and 1.8 ng/mL for pancreatitis associated protein (PAP) would confer 100% sensitivity with similar PPV as IRT / DNA approach but with the possible advantage of being easier to implement and cheaper. None of the newborns with IRT between 50 and 100 and PAP < 1.8 would require detection if it were agreed that borderline forms of disease should not be screened i.e. sweat chlorides in equivocal range.

Model predicts 55 children with CF per 200,000 newborns 10-20% diagnosed with meconium ileus 6% mortality during childhood Life expectancy of 45.8 years Initial IRT cut off: top daily 1% 7

IRT + DNA +DGGE emerged as the most cost effective of 4 strategies evaluated by this modeling exercise but DGGE is an emerging technology. 176 CF children were identified, 7 died in the first 5 years of life. Of these 7 deaths, 3 had presented with meconium ileus. All 4 of the low risk nonmeconium ileus deaths were in children randomized to the unscreened group. The authors conclude that newborn screening has the potential to decrease infant CF deaths, but if it is to be successful, identification and treatment must occur as soon as possible after birth

DGGE = denaturing gradient gel electrophoresis

25

Author Assael et al, 2002 (24) (level 3b)

Lai et al, 2005 (25) (level 3b)

McKay et al, 2005 (26) (level 3a)

McCormick et al, 2006 (27)

Setting

Subjects

Population-based retrospective cohort study analysis in northwestern Italy, population 5 million with NBS for CF from 1973 and universal from early 1980s. US CF Foundation Registry1986– 2000 analysis by mode of detection and symptoms (respiratory, nutritional or both) Australia – comparative follow-up at age 15 of screening versus non screened cohorts previously studied at 1, 5, and 10 years of age from comprehensive annual assessments

593 patients born 1938– 2000 followed by a single referral center Estimated incidence 1/2,650 live births annually.

Early recognition of CF was achieved in this region. Reported survival analysis over 3 consecutive decades revealed that improving survival diminished in last years of the study. Authors could not conclude that neonatal screening improves long-term survival if compared with diagnosis by symptoms in early infancy.

27,692 patients detected by 8 NBS, MI (pre or postnatal), Symptoms, Family History

Compared with screen-detected cases, cases detected with MI or by symptoms had significantly greater risk of shortened survival.

Follow up of 48 of 57 non screening detected cases (7 had died; 2 were lost to follow-up) versus 52 of 60 screen detected cases (4 had died; 2 transferred out of the country; 2 were lost to follow-up) 3,851 patients registered with the UK CF Database (diagnosed 1986–2003)

Screening detected cases displayed better Shwachman scores, X-ray and lung function at age 15

UK CF Database

Outcomes

Females presenting with respiratory symptoms alone were diagnosed 9 months later than males however no gender differences were observed for anthropometric, lung function, microbiological, supplemental feeding, or time since diagnosis for all patients or by genotype with presentation with respiratory symptoms alone.

8

Meconium ileus (MI) is a clinical condition leading to rapid diagnosis of CF in all cases because of intestinal blocks

26

Author

Setting

Armstrong et al, 2005 (28)

Australia NBS program

Accurso et al, 2005 (7) (level 3a)

US Cystic Fibrosis Foundation National Patient Registry for 2000 through 2002

Slieker et al, 2005 (29)

CF patients treated in the Netherlands on January 1, 2001, CF deaths between 1974 and 2000 in comparison to US CF registry data

Subjects

Outcomes

70 CF (aged 1.5–71 months) children detected by newborn screening and 19 (aged 2.0– 48 months) controls with chronic stridor Newly diagnosed cases: symptomatic diagnosis (n = 1760), prenatal diagnosis (n = 66), diagnosis by means of newborn screening (n = 256), or presentation with meconium ileus (n = 484) Actual CF live births/ deaths in the Netherlands compared with expected based on historical cohorts

Is the lower airway inflammation characteristic of CF initiated by the genetic defect of infection. Uninfected and control subjects had similar bronchoalveolar lavage fluid profiles while infected subjects showed elevated indices. The authors conclude that in CF, infection initiates and sustains airway inflammation. through the argument that the pulmonary sequelae of CF is potentially CF cases detected following symptomatic presentation have higher rates of complications / hospitalization (70%) throughout infancy, childhood, and adolescence compared with cases detected through newborn screening (29%).

The birth prevalence of CF in the Netherlands is lower than it was 30 years ago whereas survival has improved. The difference in survival between the Netherlands and the United States, as observed in the cohorts born > 20 years ago, has disappeared.

27

Table 5: ACMG bibliography of CF research • Ref. 1: NIH Consensus Development Conference Statement. Genetic testing for cystic fibrosis. April 14-16, 1997. Arch Intern Med 1999; 159: 1529-1539. •

Ref. 2: NNSGRC data reported by programs [Last updated April, 2004].

• Ref. 3: Moskowitz SM et al. CFTR-Related Disorders. [Genereviews last updated 824-04, www.geneclinics.org]. • Ref. 4: Rosenstein BJ, Cutting GR. The diagnosis of cystic fibrosis: a consensus statement. Cystic Fibrosis Foundation Consensus Panel. J Pediatr 1998; 132: 589-95. • Ref. 5: Kerem E et al. Clinical and genetic comparisons of patients with cystic fibrosis with or without meconium ileus. J Pediatr 1989; 114: 767-73. • Ref. 6: Doull IJ et al. Cystic fibrosis related deaths in infancy and the effect of newborn screening. Pediatr. Pulmonol 2001; 31: 363-366. • Ref. 7: Cystic Fibrosis Foundation: Patient registry 1998 Annual Data Report. Bethesda, MD (1999). • Ref. 8: Merelle ME et al. Newborn screening for cystic fibrosis (Cochrane Review). Cochrane Database Syst Rev 2001; 3: CD001402 • Ref. 9: Gregg RG et al. Newborn screening for cystic fibrosis in Wisconsin: Comparison of biochemical and molecular methods. Pediatrics 1997; 99: 819-24. • Ref. 10: Lee DS et al. Analysis of the cost of diagnosing cystic fibrosis with a newborn screening program. J Pediatr 2003; 142: 617-623. • Ref. 11: Farrell MH and Farrell PM. Newborn screening for cystic fibrosis: Ensuring more good than harm. J Pediatr 2003; 143: 707-712. • Ref. 12: Farrell P. Centers for Disease Control and Prevention: Newborn Screening for Cystic Fibrosis, November 20-21; 2003. http://www.cdc.gov/ncbddd/cf/meeting.htm. • Ref. 13: Campbell P. Centers for Disease Control and Prevention: Newborn Screening for Cystic Fibrosis, November 20-21, 2003. http://www.cdc.gov/incbddd/cf/meeting.htm.

28

Conditions detectable with tandem mass spectrometry 1. ORGANIC ACIDURIAS Glutaric aciduria type I (GA-I) Glutaric aciduria type I (GA-I) 9 is a rare disorder with a high risk of permanent neurological disability without early detection and treatment. First described in 1975, it is caused by a deficiency of glutaryl-CoA dehydrogenase (GCDH). GCDH is in the catabolic pathway of the amino acids, lysine and tyrptophan.(30) Most affected children appear normal at birth but are at risk of acute neurological crisis leading to severe permanent neurological (motor) disability.(31) The incidence in the US is 1:50,000 live births(31), though in BC the condition appears to be rarer.(32) Biochemically, GA-I is characterized by an accumulation of urine organic acids (glutaric acid and 3-hydroxyglutaric acid) and corresponding elevation of glutarylcarnitine in the blood. Identifying 3-hydroxyglutaric acid in the urine has been shown to be a highly sensitive marker for GA-I.(35) It is the elevation of glutarylcarnitine (C5DC) which can be detected in newborn screening blood spots by tandem mass spectrometry (LC/MS/MS).(33,34,35) Abnormal C5DC screening results require follow up testing and confirmation by urine organic acid analysis, enzyme assay and/or mutation identification. The clinical characteristics of GA-I can vary considerably, even between siblings with the same genotype.(34)(31) Most patients have an enlarged head at birth or shortly thereafter, though the brain itself is usually under-developed. Acute neurological deterioration typically occurs between 6 and 18 months of age, sometimes after a triggering event such as a fever.(31) The neurological outcome of these acute events is bilateral striatal damage and extrapyramidal signs which include dystonia, dyskinesia, choreoathetotic movements, dysarthria and spasticity. Morbidity and mortality is high in patients who have had a crisis. Different therapeutic protocols are used in different centres. The most beneficial approach appears to be low protein, lysine-restricted diets with carnitine supplementation.(36) Significantly, neurological damage was prevented (or otherwise did not occur) in 51 of the 65 patients reported up to 2004 that had been detected by newborn screening.(37) This compares with another group of 218 patients identified after symptom onset, where 168 demonstrated encephalopathic crises that resulted in moderate to severe disability.(36) The evidence for clinical benefit arising through early detection and treatment appears to be compelling. However, the results of GA-I newborn screening in the Ojibway-Cree population in Northern Canada are an exception to this trend. The predicted incidence of GA-I in BC may be less common than US incidence rates (1:50,000) based on a retrospective survey of cases detected in BC from 1979-1996 in which only 4 GA-I patients were identified in 785,400 births, or 1 case every 5 years at current birth rates.(32) 9

Alternate names are glutaric aciduria and glutaryl-CoA cehydrogenase deficiency (GCDH, EC.11.3.99.7)

29

Screening approaches GAI is detected by TMS using a C5DC level above a cut off value set by each laboratory. Lindner et al, 2004 report on GAI screening test performance.(33) although information on sensitivity and specificity is limited. Sensitivity appears to be 100% with a cut off at the 99.95th percentile. A false negative rate of zero is assumed based on the absence of reported clinically ascertained cases. There is some risk of under-ascertainment of missed cases. In Lindner 2006, analytical performance is further enhanced using various ratios of C5DC to other acylcarnitines. This improves the positive predictive value while optimizing detection rates. A positive predictive value of 17% can be achieved with a theoretical sensitivity of 100%. In the older literature, patients with GCDH deficiency have been reported as “missed” by newborn screening or by retrospective analysis of the newborn blood spots. It turns out that many of these false negatives were not genuine but were related to degradation of the C5DC signal on retrospective analysis of stored blood spot cards. In the few cases that were verified to be false negatives, patients were missed because in follow up to an initial positive blood spot acylcarnitine screening result, a repeat blood spot acylcarnitine profile was normal and the diagnosis was ruled out at that point. To avoid this problem, most labs now request urine for organic acid analysis in addition to a repeat blood spot card. Urinary 3-hydroxyglutaric acid is persistently elevated in affected cases and improves detection of this disorder.(35,45) Knowledge Claims Formulated by BC Working Group on GA-I What is the evidence that newborns with glutaric aciduria (GA-I) will have improved health outcomes with early detection following newborn screening with TM/TS? If BC commences screening for GA-I with TMS using C5DC with a cut off of 0.1 uM and ratios with other analytes, test performance values of ~100% sensitivity and 17% positive predictive value are achievable (Lindner, 2006 and BCWC data) The evidence reviewed in this HTA indicates that every 5 years approximately: ƒ 2 newborns will be detected with GA1. (level 2c evidence) • ~12 false positive test results will be generated • Acute encephalitic crisis will be avoided in some newborns who screening positive. (level 2c and level 3 evidence)10 • This benefit is likely to be realized between birth and age 3 during the time period when acute encephalitic crises are more frequent. Lifelong benefits are associated with the avoidance of serious complications associated with acute encephalitic crisis in GA-I patients such as neurological disability and death.

10

With a high risk confounding bias and moderate probability that the relationship is causal for acute encephalopathy events.

30

The Clinical Evidence Base BC Incidence BC analysis based on provincial laboratory records indicated that GA1 is rare based on clinical detection following symptomatic presentation.(32) Applegarth et al, 2000 found 3 cases among 196,350 births in BC over 20 years from 1969 to 1996.(32) This was revised to 4 after publication. Therefore, 1 case of GA1 would be detected every 4.7 years if routine newborn screening were initiated for GA1 (assuming the current birth rate of ~ 40,000 live births per year continues and the rate of detection with screening matches the rate with clinical detection). The numbers of cases symptom free in the absence of detection is low therefore the risk of screening leading to harm through unnecessary anxiety, stigmatization and over treatment of ‘silent cases’ is low. The estimated incidence for BC is at the low end of the range reported in a recent Finnish HTA. Autti-Ramo et al. report that a lowest possible incidence rate of 1:200,000 (1 every 5 years); a base estimate of 1:84,000 (1 every 2.1 years) and a highest possible incidence of 1:56,000 (1 every 1.4 years) based on 40,000 live births per year for BC.(38)

Efficacy of treatment Baseline Evidence from UK HTA report The UK HTA reported that the “strength and quality of the evidence for the effectiveness of dietary interventions for GAI were grade III.” (2) This finding was based on primary research presented by Monavari & Naughten, 2000 (39) and Bjugstad & Goodman, 2000 (30) Monavari & Naughten (39), 2000 compared 6 patients with GA1 detected through targeted screening of families with previously identified cases with 6 late-diagnosed symptomatic patients. Whereas 5 of 6 clinically detected patients had died and all suffered from dyskinetic cerebral palsy; 4 of 6 patients detected through familial screening were developing normally. (39) Those cases detected through familial screening may not have been of the same genotype as clinically detected cases. Also there was an absence of denominator data. A systematic review of 42 published research articles describing GA-I series conducted by Bjugstad & Goodman, 2000 found that dietary interventions given after the onset of GA1 symptoms were not effective in reversing neurological damage.(30) The authors concluded that symptomatic presentation of GA1 may therefore result in significant future impairments. Multiple regression analysis found that more that 50% of patients had onset of symptoms before age 8 months and almost all in whom symptoms developed had the symptoms within the first 3 years of life.(32) Only 6 of the 115 patients described had treatment before motor symptoms developed and all had a relatively normal development.

31

BC NSAC Working Group Update Thirty-five papers published between 2004 and early 2006 were identified by this BC NSAC update. A total of 17 published papers were initially identified as relevant for retrieval including primary studies revealed by the UK HTA(39) and Bjugstad & Goodman, 2000 (30) Six recent studies met inclusion criteria for evidence contributing to knowledge of treatment effectiveness and were appraised by the GAI working group.(36) (40-44). (See Table 2.) Three studies provided the strongest evidence of treatment benefit and were considered in depth (36,42,43). Kolker et al 2006 (level 2c) Kolker et al 2006 report onset and neurological sequelae on 279 GA1 cases from 35 metabolic centres around the world.(36) Kolker et al. 2006 was assessed by the working group as providing level 2c evidence for the benefit of treatment; that is, there is a high risk confounding bias and moderate probability that the relationship is causal for acute encephalopathy events. A strength of this analysis is that there were a relatively large numbers of cases included. This is difficult to compile for a rare disease. Of the 279 cases, 218 were detected clinically and 61 presymptomatically by neonatal screening (23), high-risk screening (24) and macrocephaly (14). Of the 218 symptomatic patients, 185 had encephalopathic crises. Seventy percent had a single crisis only which resulted in moderate to severe disability at a mean age of 9 months. Over 50% of the patients diagnosed before onset of symptoms remained well with follow-up of over 75 months. Timely detection in neurologically asymptomatic patients followed by treatment with L-carnitine and a lysine-restricted diet was found to be the best predictor of good outcome. (36) The working group acknowledged that many of the unanswered questions about the natural history of GA1 can only be answered if NBS is instituted. Subgroup analysis is not equivalent to controlled studies. Other limitations of the Kolker et al, study include an uncertain denominator and non-standardized treatments protocols. Finally as data collection was by questionnaire to 35 centers complete ascertainment of the cohort may not have been achieved. Naughten et al, 20 4 (42) (level 3a) The NSAC working group on GA1 appraised the Naughten et al, 2004 study as providing level 3a evidence. A strength is that there appears to be complete ascertainment of all GA1 patients detected in the Republic of Ireland over a 16-year period (42). In this cohort, 10 asymptomatic GAI cases were detected through high-risk screening, 1 in 10 had an acute encephalic crises, 6 in 10 had no neurological abnormality and 1 of 10 died. No striatal degeneration was detected in the high-risk screened group during follow-up of 14 years. By contrast, in the group of 11 symptomatic patients detected clinically, 10 of 11 developed movement disorders and 7 of 11 died. Though weak in study design and lacking many features that protect against bias, the working group found the sizable difference between groups compelling.

32

Strauss et al, 2003 (43) (level 3a) The NSAC working group on GA1 appraised the Strauss et al, 2003 study as providing level 3a evidence that early detection prevents early onset of neurological deficits. There were 37 Amish and 40 non-Amish patients among the 77 patients with GAI identified. The cohort is from Pennsylvania, USA where high-risk screening began in 1988 and general population screening in 1994. Strauss et al, 2003 report a basal ganglial injury rate of 35% in newborn screen detected cases compared to 94% in Amish and 85% in non-Amish patients detected clinically following onset of symptoms. The degree of motor disability among the newborn screening group was less with 65% with no symptoms, 9% with mild and 26% moderately to severely affected versus 74% with a moderate to severe degree of motor disability in the symptomatic group. Data collection was prospective and the study cohort was homogenous. Though treatment evolved over the 14-year study duration a small group of people were making treatment decisions, which would tend to increase the homogeneity of the study. Also this Pennsylvania cohort was not included in the international study reported by Kolker et al, 2006 therefore there is consistency in findings between the two studies. A limitation of the study was that some of the children detected were still quite young at the time of publication given that screening had started relatively recently (7/20 were under 1 year of age at publication). The author (Strauss) was contacted by a working group member and provided an update on these children. All 25 Amish patients were diagnosed as newborns and started on L-carnitine and either low protein diet (n=15) or a low protein diet supplemented with lysine free amino acids (n=10). The overall incidence of brain injury did not vary significantly by birth year or treatment strategy. 9/25 children (36%) developed brain lesions compared to historical pre-screening injury rate of 89% (n=57). This analysis strengthens the case for newborn screening yet as this more recent data is unpublished it is not available for detailed appraisal. An eventual ranking at level 3a is expected given that this is an extension of the previous study. Kyllerman et al 2004(41) (level 3a) The neurological outcome and survival rate of all 28 GA1 patients diagnosed in Norway, Sweden and Finland between 1975 and 2001 without mass screening were reported. Dystonia and spasticity were the main neurological signs. No neurological improvement following treatment was found in any patient with acute onset. Collins and Leonard 2000 (level 3a) In a letter to the editor, Collins and Leonard, 2000 refute the claim that the evidence presented by Monavari and Naughten(39)supports the effectiveness of dietary recommendations.(40) They present contrary evidence from 3 families with more than one affected child demonstrating highly variable outcomes in untreated GAI cases. They suggest that ‘diet is a significant imposition and all that may be needed is intensive treatment during intercurrent infections’.(40) This case series is selective and small and

33

therefore not useful for estimating population-based experience with screening. Nonetheless it is a useful critique. Kulkens et al 2005(44) (level 3b) Kulkens et al, 2005 reports 2 cases of neurological deterioration occurring after age 6 in an international cohort of 261 cases and 3 additional cases from the literature thereby demonstrating that GAI can present as a severe leukoencephalopathy in adulthood. Genotypes are the same as childhood onset cases. Given that this is an infrequent outcome in a rare condition, it is of marginal relevance to the decision on whether to extend newborn screening to include GA1. Table 6: Recent primary research reviewed by BC GAI working group

Author

Setting

Collins and Leonard, 2000(40)

Metabolic Unit, London, UK.

Strauss et al 2003(43)

Pennsylvania, USA High-risk screening from 1988 and NBS from 1994.

Kulkens et al Germany 2005(44)

Subjects

Outcomes

3 families with affected siblings

Variability in expression of GA-I deficiency with no dietary treatment: Family1: 1 severe movement disorder, 1 mild gait disturbance attending school Family2: 1 severe CP, 1 minimal symptoms attending school Family3: 2 died, 1 minimal problems 77 cases with GA1 • Incidence of basal ganglia injury representing 14-year clinical is 85% in non-Amish pts and experience in one state with 94% in retrospectively identified a high risk Amish population Amish children. (37Amish, 40 non-Amish. • The basal ganglia injury rate is 35% in the 20 Amish children diagnosed by neonatal screening and managed prospectively following early diagnosis • Cultural genetic background (i.e. Amish vs. non-Amish) not associated with severity of neurological outcome., 2 cases: a 65 year old man • Characteristically, manifestation and a 15 year old boy of neurologic disease in GCDH deficiency occurs during infancy or early childhood • This study demonstrated late-onset neurologic disease in 2 pts considered asymptomatic until age 35 years (pt 1) or 15 years (pt 2) • No difference in genotypes and biochemical phenotypes between pts with late onset and childhoodonset GCDH deficiency

34

Author

Setting

Subjects

Outcomes

Lindner et al Targeted newborn 2004(33) screening for GCDH of ~400 Oji-Cree newborns annually with DNA-based screening tests.

7 GCDH deficient • Screening high-risk groups with newborns (5 females and 2 a single or limited number of males) identified from 2200 mutations seem reliable screened (1:300 approaches to detecting GCDHprevalence) from May 1998 deficient pts before onset of through June 2003 neurological deterioration • North American, Australian and German neonatal screening programmes, enrolling approximately 2.5 million newborns worldwide with an overall mean frequency of 1:106 900 for GCDH deficiency

Tarini et al 2006(45)

• Newborns screened for universally mandated disorders screened with dried blood spots were counted

Review of US state screening programs using data from the National Newborn Screening and Genetics Resource Center regarding the screening practices for each of the 50 states, to determine the number of mandated disorders added to state newborn screening panels between 1995 and 2005 (consistent counting method)

• Average state mandated screening for 5 disorders in 1995 (range: 0–8 disorders); 1 decreased # of disorders screened; 2 remained the same; average state added 19 tests (median: 22 tests). 19 states added >29 disorders during the period and 5 states >39 • Assuming that an individual test for a disorder had a specificity of 99.995%, it is estimated that 2575 infants would have received false-positive results through screening with tandem mass spectrometry per year • If specificity was assumed to be 99.9%, then the number increased to 51 000 • Tandem mass spectrometry has allowed state newborn screening programs to increase dramatically the number of mandated disorders screened in the past decade

35

Author

Setting

Subjects

Tortorelli et al Mayo Clinic College • 26,797 urine samples 2005(35) of Medicine, were submitted to Rochester, MN, USA. laboratory for organic acid analysis • Urine acylcarnitine profiles were analyzed in 68 pts whose organic acid analysis revealed an abnormal excretion of glutaric acid and/or 3hydroxy glutaric acid • 14 of these 68 pts were diagnosed with GA1 (5 lacked significant glutaric aciduria, 9 had a normal plasma glutarylcarnitine concentration) • Controls: evaluation of 54 subjects with glutaric aciduria secondary to other causes (all had an abnormal glutaric acid excretion)

Outcomes • The excretion of glutarylcarnitine was significantly elevated in all GA1 pts (14-522 mmol/mol creatinine; reference range: <5.2) and in none of the controls with glutaric aciduria. Findings suggest that the urinary excretion of glutarylcarnitine is a specific biochemical marker of GA1 (useful in the work up of pts with suggestive clinical manifestations but without glutaric aciduria and normal plasma acylcarnitine profiles)

Propionic and Methylmalonic aciduria (PROP/MUT/CBLA,B) Biochemically and clinically related conditions Propionic and methylmalonic acidemias11 (PROP and MUT) comprise a group of disorders in branch chain amino acid metabolism. TMS screening, using C3, can detect PROP and different forms of MUT, but cannot distinguish between the genetic defects. It is clinically important to perform additional testing, such as urine organic acids, enzymatic and molecular studies, to determine the underlying defect as treatment and prognosis are different for each disorder.(46)(47) Given their close biochemical association, it is not surprising that MUT and PROP have similar clinical characteristics, as well as similar therapies.(47) First described in the late 1960s, they are among the most common organic acidemias, and the literature offers some insights into the natural history of disease.(48,49) Infants with MUT or PROP typically show progressive symptoms, from refusal to feed, vomiting, weight loss, generalized hypotonia and lethargy, seizures and coma, and finally brain damage or death; the most severe consequences can manifest within a few days of birth if not properly treated, or the disease may be of later onset.(46)(50) If early mortality is avoided, long-term complications frequently arise, including, 11

Also known as acidurias

36

cardiomyopathy, developmental delay, movement disorders, and severe organ damage. (50-52) Emergency treatment focuses on proper hydration, restricting protein intake, and supplementation with biotin for PROP and, as indicated, with vitamin B12 for MUT.(48) Other supplements may be provided, including carnitine and metronidazole.(47) Long term management may entail a low-protein, high-energy diet supplemented with amino acids.(50) Recently, liver transplantation has been posited as a treatment for individuals who do not respond to other therapy; such an approach may protect against developmental delay and the progress of cardiomyopathy.(50,52) Pre-symptomatic detection and treatment seems to improve prognosis for late onset forms of the disease. While the latter argument supports screening, the late-onset forms are less common than neonatal-presenting forms.(53) Most authorities would admit that more research is needed to confirm the clinical benefit of early detection. The incidence of MUT may fall between 1:115,000 and 1:169,000 live births; the reported frequency of PROP has ranged from 1:165,000 to 1:277,000.(47) A recent study of Japanese and other Asian cases detected by TMS showed that MUT occurred more frequently than all other organic acidemias combined.(53) It is estimated that 2 newborns with either MUT or PROP would be detected over a 3 year period in BC if NBS is instituted. PROP12 is caused by deficiency of the enzyme propionyl-CoA carboxylase (EC 6.4.1.3), which catalyzes a biochemical step leading to methylmalonyl-CoA. The MUT’s result from deficiency of methylmalonylCoA mutase or defects in the production of cobalamin (Vitamin B12). (47) Cobalamin A & B defects have isolated methylmalonic acidemia (MUT) while Cobalamin C and D defects have both MUT and homocystinuria. The CBL C type is the most common with about 250 known cases in the world. (54 - 56). Children with PROP, MUT or CBL disorders often present with acute illness as neonates or infants. Clinical signs and symptoms include failure to thrive, metabolic acidosis persistent ketotic episodes, hypoglycemia, hypotonia, hyperammonemia and neurological symptoms. Confirmation of diagnosis requires urine organic acid analysis and usually enzymatic testing in fibroblasts. Test Performance The C3 acylcarnitine provides a primary screen for MUT and PROP but suffers from poor specificity and sensitivity (PPV ~ 1.5-4%) (57)(58) Accordingly, several groups have designed second-tier follow-up tests for elevated C3, allowing them to set more conservative primary cut-offs to increase sensitivity while providing follow-up testing on the original newborn card without a need to recall the newborn.(4)(58) 2nd tier testing measures different analytes; MUT and methylcitrate by a column chromatographyMS/MS analysis. Using this approach, overall specificity is increased, leading to a PPV= 26% (4) with second tier testing on ~2% of screened newborns (~17 per week in BC predicted) 12

The gene defect is coded as McKusick 232000, 232050.

37

La Marca et al, 2007 is a retrospective analysis of cards from the NBS program in the Tuscany region of Italy. It represents level III evidence on the test parameter hierarchy. The evidence from Matern et al, 2007 on the other hand is population based and therefore representing level I evidence. Knowledge Claims Formulated by BC Working Group on PROP/MUT/CBLA,B What is the evidence that newborns with PROP/MUT/CBLA,B will have improved health outcomes with early detection following newborn screening with MS/MS? If BC commences screening for PROP/MUT/CBLA,B a positive predictive value of 26% can be achieved with second tier testing (level 1), the evidence reviewed in this HTA indicates that: • •

1 newborn will be detected with PROP or MUT every 1.5 years approximately (level 3c) at a cost of 17 false positives per week. Outcomes will be poor even with the currently available treatment (< level 3c) For newborns detected with PROP: 40% will die even with treatment (7 of 17) 65% have intermediate or severe neuromotor impairment 53% have intermediate or severe nutritional impairment 41% have intermediate or severe social impairment 29% have intermediate or severe mental impairment 12% have intermediate or severe psychological impairments only 5% can be expected to be normal at age 10 (59)

Including PROP/MUT/CBLA,B on the panel despite the low level of available evidence might be justified with prospective evaluation to advance knowledge given that: • The risk of doing harm through the detection of mild variants is low and they will all come to clinical attention eventually • A quality study has not been done and so claims of benefit or lack of benefit cannot be justified. • A long-term multicentre international study would be required to establish or disprove the benefits of screening. • Recent advances in treatment protocols have been based on better understanding of condition, however emerging data has not been analysed to date • Benefits hypothesized may include reduced severity of first crisis, reduced time in hospital, reduced need for critical care • Other provinces are going ahead though the evidence neither supports nor refutes institution of newborn screening. The Clinical Evidence Base

BC Incidence The estimated incidence for BC is 1:60,000. 38

Baseline Evidence from UK HTA report The evidence in support of screening for PROP/MUT/CBLA,Bs are from case series claiming success in treating children detected clinically in unscreened populations and the possibility that health outcomes could be further improved with earlier detection. The UK HTA report documented that “evidence for the effectiveness of dietary and or pharmacological interventions in patients with methylmalonic or propionic acidemia was limited to a few studies” and “ effectiveness of treatment ranged from low to high.” (1) (p 49,52.) The UK assessment of PROP rested on a retrospective analysis by van der Meer et al, 1996(59) as well as a 1997 HTA by Pollitt et al.(60) On the basis of this evidence Pandor et al, 2004 state that outcomes ‘appeared to be satisfactory in terms of survival, neurological and mental development’ and ‘general neurological outcomes were generally better for early-onset patients and most patients encountered growth retardation.’ (1) (p 49,52.) The van der Meer retrospective cohort data represented 20 years of experience at one hospital in Paris from the early 1970’s to 1990’s.(59) The 17 subjects with PROP were detected without screening. Early and late onset cases were compared and treatment regimes (which changed over time) described. It is a stretch for this study to be classified as Level 3c. Though it is retrospective, late onset cases do not provide an adequate control for the early onset cases due to the variability in the natural history of PROP. The findings may better represent the natural history of PROP rather than treatment effectiveness. The systematic review by Pollitt et al., 1997 concluded that “neonatal screening was unlikely to be of benefit to patients with the neonatal-onset forms of methylmalonic, propionic or isovaleric acidaemia; however, the later onset variants were more amenable to treatment” also that the disorders of PROP, MUT and IA are associated with significant mortality and morbidity.(60)p. 69 (1) p. 52. The level of evidence for the effectiveness of treatment was graded at IV. A 1994 study by van de Meer (61) used the same methodology for MET as the 1996 study of priopionic academia. This study therefore has the same limitations. The only other study relating to treatment effectiveness was a retrospective series of patients with PROP reported by Surtees et al. in 1992 (62) Late onset clinical presentation was compared to early onset which means that any difference is likely due to differences in condition severity. Only one patient in this cohort survived to 10 years of age. In summary, all primary data considered to 2004 from the UK HTA therefore ranks low on the hierarchy of evidence (Level 3C with poor quality study design).

39

BC NSAC Working Group Update Eighteen papers published between 2004 and early 2006 were identified by the BC NSAC sponsored update of propionic acidemia. A total of 9 published papers were identified as relevant for retrieval including 4 primary studies. Three 3 recent studies met inclusion criteria for evidence contributing to knowledge of treatment effectiveness. The systematic review by Pollitt et al., 1997 concluded that “neonatal screening was unlikely to be of benefit to patients with the neonatal-onset forms of methylmalonic, propionic or isovaleric acidaemia; however, the later onset variants were more amenable to treatment” also that the disorders of PROP, MUT and IA are associated with significant mortality and morbidity.(60)p. 69 (1) p. 52. The level of evidence for the effectiveness of treatment was graded at IV. A 1994 study by van de Meer (61) used the same methodology for methylmalonic acidaemia as the 1996 study of priopionic academia. This study therefore has the same limitations. The only other study relating to treatment effectiveness was a retrospective series of patients with PROP reported by Surtees et al. in 1992 (62) Late onset clinical presentation was compared to early onset which means that any difference is likely due to differences in condition severity. Only one patient in this cohort survived to 10 years of age. In summary, all primary data considered to 2004 from the UK HTA therefore ranks low on the hierarchy of evidence (Level 3C with poor quality study design).

40

Table 7: Recent primary research reviewed by BC PROP/MUT/CBLA,B working group

Author

Setting

Subjects

Outcomes

Dionisi-Vici et al Pts were from 2005(46) (
• Group of • Mortality rate was 51%, with a symptomatic pts significantly higher mortality (p < 0.01) in comprised 29 the neonatal-onset versus later-onset individuals diagnosed cases, in which no deaths were recorded and followed; 15 pts • About 40% of pts died during the initial had MUT, 13 had phase of the disease or before the second PROP and 1 had IVA year of life • 24 of the 29 pts • Normal neurodevelopment was observed (82%) symptoms in 60% of surviving pts at the first appeared within the evaluation before the 2nd year of life first 28 days of life • Through the following years, most of the (neonatal-onset) and affected children showed progressive in the remaining 5 cognitive deterioration or died. The cases symptoms mortality rate increased from 36% to 62%, appeared later (late- and a normal neurocognitive outcome was onset) recorded in only 27% of patients • Clinical • Unlike in neonatal-onset cases, the characteristics at natural history in the late-onset cases diagnosis, the clearly appears more favourable; survival neurocognitive and neurocognitive outcome was better, outcome and the with 4 out 5 pts showing a mortality rate in a normal/borderline development series of 18 pts (7 • NBS pts - neonatal mortality in this group IVA, 7 PROP and 4 of pts was significantly lower than in MUT) diagnosed by symptomatic pts (p < 0.03) neonatal screening in •MUT and PROP incidence: no significant Australia (11 pts) and differences between the two groups, 0.99 in Munich, Germany ± 0.35 per 100 000 births in the clinically (7 pts, 4 of whom diagnosed group compared to 1.04 ± 0.36 were prospectively in the screening group. diagnosed Sass et al Data were gathered • 49 pts with PROP • 32 of 38 newborns (84%) presented with 2004(63) (
41

Isovaleric acidemia (IVA) Isovaleric acidemia (IVA)13 is an inborn error of metabolism caused by a deficiency of isovaleryl-CoA dehydrogenase (IVD).(46) IVA may present in the first week of life, or later in childhood. There is a high mortality rates among the approximately 50% of cases where IVA becomes evident soon after birth. Recurrent vomiting, acidosis, coma and acute encephalopathy are life threatening to the newborn.(64) Prompt treatment can reduce the severity and frequency of ketoacidotic episodes. In the absence of screening, children diagnosed after the newborn period (the other 50% of IVA cases) are likely to have chronic symptoms, including failure to thrive and developmental delay.(65)(64) Therapy for IVA is aimed at preventing metabolic crisis; it includes monitoring the patient during illness or periods of fasting, a low protein diet, carnitine or glycine supplements, and medications to avoid the accumulation of toxic metabolites.(65) Early diagnosis allows for treatment to be initiated in a timely way, perhaps lessening the severity and frequency of ketoacidotic episodes.(64) While mortality appears to be higher when treatment is delayed, there is a need for additional long-term outcome studies of patients with IVA.(46) IVA was the first organic acidemia described in humans.(66) Specifically, IVA is an inborn error of leucine metabolism caused by a deficiency of IVD [MIM 243500, EC: 1.3.99.10] which catalyzes the 3rd step in the catabolism of leucine.(46) IVA results in the accumulation of isovaleryl-CoA derivatives and the biochemical diagnosis is based on the detection of isovalerylglycine and other metabolites in urine and of isovalerylcarnitine in plasma and whole blood. Test Performance TMS detection of elevated C5 acylcarnitine levels allows for presymptomatic diagnosis of IVA. NBS may identify mild or even asymptomatic variants of IVA as mutations are heterogeneous in IVA. The positive predictive value of an elevated C5 is high (low false positive rate). However, false positives due occur secondary to drugs (pivaloylcarnitine) and a rarer metabolic disorder: methylbutyrylglycinuria. Diagnostic confirmation can be provided though urine organic acid analysis +/- molecular and enzymatic testing.(46,65)

BC Incidence The incidence for this disorder varies worldwide, reported as anywhere from 1:62,500 (Germany)(67) to 1:365,000 live births.(USA)(57)(68) Ontario planners have suggested that IVA will affect one in every 100,000 to 200,000 babies born in the province. In BC, it is estimated that one case of IVA would be detected every 2 to 3 years if screening were to be instituted.

13

Also known asiIsovaleryl-CoA dehydrogenase deficiency

42

Knowledge Claims Formulated by BC Working Group on IVA What is the evidence that newborns with IVA will have improved health outcomes with early detection following newborn screening with TMS? If BC commences screening for IVA, the evidence reviewed in this HTA indicates that over a five-year period 2008/09 to 2012/13: • 1 newborn will be detected with IVA every 2 to 3 years approximately with some risk of detecting mild variants.(level 3c) Earlier detection will lead to the following benefits for newborns. • Death may be prevented in 1 newborn with unsuspected IVA every 4 to 6 years (level 3c evidence based on the assumption that 50% at risk for death early in the newborn period and all these have reduced mortality) • Treatment may lessen the severity and frequency of ketoacidotic episodes for 1 child every 4 to 6 years (level 3c evidence) based on the assumption that the other 50% all have reduced morbidity The Clinical Evidence Base Ensenauer et al, 2004(64) (level 3b) In this study, mutations in the IVA gene were determined in all patients who received a diagnosis by of IVA by NBS. A total of 19 patients were identified between the years 1998 and 2004; 12 from Germany and 7 from US (11 female, 8 male, 16 white, 2 arabic, 1 hispanic). At diagnosis, there was a broad range of C5 elevation ranging from mild to marked. Thirteen patients were found to have at least one copy of the 932CtoT [A282V] mutation; 5 patients were homozygous. Treatment for IVA was begun upon diagnosis. All patients (age range 2 months to 5 years – median age 1.9 years) were developing appropriately at the time of reporting. Correlation with C5 acylcarnitine concentration: Those patients carrying the 983CtoT either homozygous or compound heterozygous had C5 levels at least 1.2 above the cutoff value. However, C5 values were lower than those newborns with 2 different IVD mutations. Urine organic acid profiles were also less pronounced in those patients carrying at least one copy of 983CtoT. Two patients had been previously described. They were diagnosed at age 4 after metabolic work up for mild developmental delay, but no acute episodes of metabolic decompensation. It is uncertain whether the presence of 982CtoT explains their clinical phenotype or whether there is an underlying ascertainment bias. They may be at risk of having acute crises when exposed to common stressors such as fever or prolonged fasting. In an older sibling study, 24 asymptomatic siblings from 11 families were evaluated by urine orgs and plasma acylcarnitines. As a result 6 asymptomatic kids ages 3 to 11 were identified. These had normal development and remained asymptomatic during episodes 43

of common pediatric illnesses. Investigation of the families without at least one copy of 982CtoT did not disclose additional siblings with IVA. These results suggest that IVA with at least one copy of 982CtoT can exhibit a mild phenotype or be free of symptoms throughout childhood. Berry, 1987 (69) (level 3b) In this retrospective uncontrolled case series it is not clear how long each patient was followed. All patients were diagnosed at a metabolic center in Philadelphia after 1977. Nine patients with IVA were treated with low protein diet and supplemental glycine; 5 presented acutely and 4 chronic. Of 4 patients with the chronic phenotype, 3 whose treatment was delayed beyond 1st year are mentally retarded (MR). Two of 5 with acute phenotype are MR -- one was from a non-compliant family and the other had a stroke prior to diagnosis. No complications including no MR occurred in those patients who were treated from infancy. The small numbers tested decrease the validity of this finding, however, it appears that ketotic episodes can be reduced with therapy. Whether a reduction in ketotic episodes improves neurological outcome is less clear from this data. Other papers were excluded because they were case reports or because of insufficient data: Mehta et al, 1996 (one case of a boy diagnosed at age 5)(70); Rousson 1984 (11 reported cases with insufficient description to draw any useful conclusions)(71); Tokatili 1990 (natural history of initial presentation with description of acute versus chronic presentations) and Tanaka, 1990 (review).(72)

44

Table 8: Recent primary research reviewed by BC IVA working group

Author

Setting

Subjects

Outcomes

Ensenauer et al, US & Germany joint 2004 (64) study,

• 19 subjects (11 females and 8 males) and whose condition was detected through NBS led to the identification of one recurring mutation, 932C-->T (A282V), in 47% of mutant alleles

Mehta et al, 1996 (70)

1 5-year old boy

• 932CrT (A282V) mutation was found in 18 (47%) of 38 of mutant alleles from the individuals identified by NBS (5 of the subjects were homozygous for the mutation) • The C5 acylcarnitine concentration in NBS bloods spots of those carrying the 932CrT mutant allele, was significantly lower than in those of newborns with two different IVD gene mutations (P < .0019) • Individuals who harbored the 932CrT mutation also had an abnormal urinary excretion of isovalerylglycine; this was much less pronounced than in the group of symptomatic pts (n p 9) (P < .0001) • Plasma C5 acylcarnitine concentrations did not reliably differentiate between the pts carrying the 932CrT mutation and the symptomatic pt group (P < .2037) • Lack of early recognition of IVA may lead to severe psychomotor retardation or death • Diagnosis of IVA was confirmed by demonstrating markedly increased concentrations of isovalerylglycine (7058 mg/gm chromium), 3-hydroxybutyrate (l 57 mg/gm Cr), and fumarate (93 rag/gin Cr) in the urine (by gas chromotography-mass spectrometry); serum glycine concentration was 829 nmol/m (normal <424 nmol/m) • The lethargy and acidosis resolved with continuous administration of intravenous fluids including 5% dextrose and a 2 mmol/kg bolus of NaHCO3 • Subsequent developmental testing showed no evidence of delay and normal intelligence • Dietary improvements can achieve better outcomes

Pennsylvania, US; Case study

Dionisi-Vici et al Pts were from Bambino 2005(46) (
• Group of The incidence of IVA was significantly symptomatic pts higher in the screened population, 0.89 ± (mostly PROP / MA 0.49 per 100 000 births versus 0.20 ± 0.13 and 1 IVA) per 100,000 births (p < 0.0001) • compared to NBS • Neonatal screening could not detected group: 18 pts discriminate the early onset pts from those in all: 7 IVA, 7 PROP with a later onset or asymptomatic and 4 MUT

45

3Methylcrotonyl-CoA carboxylase deficiency (3MCC) 3-Methylcrotonyl-CoA carboxylase (3MCC) is a biotin dependent enzyme (MCC;EC6.4.1.4) in the leucine degradation pathway. The clinical course of 3MCC deficiency (3MCCD) is variable ranging from a more common mild or asymptomatic form to severe illness and death in infancy. A non-specific but severe phenotype has been described consisting of hypotonia, psychomotor retardation and seizures. The enzyme consists of 2 subunits (alpha and beta) encoded by 2 genes 3MCCA and 3MCCB. Prior to NBS, 3MCCD had been considered an extremely rare inborn error of metabolism. Expanded NBS revealed that 3MCCD is more frequent than assumed from clinically ascertained cases. 3MCCD is detected by an elevation in the C5OH acylcarnitine signal. Interestingly, C5OH may be elevated with other very rare inborn errors of metabolism: HMG CoA Lyase deficiency, Beta-Ketothiolase deficiency (BKD) and 3-Methylglutaconyl CoA hydratase. These additional disorders were considered ‘secondary targets’ of screening by the US ACMG primarily due to the rarity of the conditions. For instance, in the close to 1 million babies screened in North Carolina between 1997 and 2005, 26 cases of 3MCCD were detected and 2 cases of BKD and no cases of the other ‘secondary target’ disorders. Based on the very low incidence figures of other disorders detectable by the C5OH signal and limited literature (sporadic case reports), the following review is focused on 3MCC deficiency only, as a potential primary target for screening. Knowledge Claims Formulated by BC Working Group on 3MCC What is the evidence that newborns with 3MCC will have improved health outcomes with early detection following newborn screening with TMS? If BC commences screening for 3MCC, the evidence reviewed in this HTA indicates that: • 1 newborn will be detected with some variant of 3MCC every 2 years approximately. (level 3C) However, • 1 death could potentially be prevented every 200 years if the current birth rate is maintained and given an estimated risk of 1 death from 3MCC for every 8.5 million births. • 1 adverse event may be prevented every 50 years with an estimated risk of 1 or 2% of all cases being at risk of adverse outcomes. • 1 child every 20 years would have minor symptoms prevented. Given the propensity to detect primarily benign variants of 3MCC the BC NSAC recommended that 3MCC not be added to the current newborn screening panel as cases detected are not likely to benefit from early detection and treatment.

46

The Clinical Evidence Base Stadler et al, 2006 (73) (level 3c) Between 1999- 2004, 677,852 newborns were screened for 3MCC in Bavaria and South Germany. Eight neonates were detected and then confirmed to have 3MCC deficiency giving an incidence of 1:84,700 (1:12,500 to 1:196,000 95% confidence interval). Based on the findings in this study which indicated mild variants were being detected combined with little evidence of benefit conferred by early detection, the newborn screening program in Bavaria / Germany was stopped screening for 3MCCD. Repeat blood spot acylcarnitines, urine organic acids and 3MCC enzyme activity in fibroblasts confirmed the diagnosis. MCCA and MCCB mutation analysis was also performed. The requirement for a 2nd card (recall rate) was 0.03% with a positive predictive value of 3.7% using both the C5OH signal alone and as a ratio to C3 (C50H:C3 ratio). Four mothers were diagnosed with 3MCCD as a result of investigation of transient C5OH elevation in newborns that was retrospectively attributed to the mothers having the disease. Family studies following the diagnosis of the index cases revealed 2 elder asymptomatic siblings with 3MCCD. The study also included 3MCCD cases ascertained by high-risk screening (clinically ascertained rather than NBS ascertained). Thirty different mutations were identified; 15 each in MCCA and MCCB. Only 8 of the 30 mutations had been previously described. This paper also includes an analysis of clinical phenotypes of reported individuals in the literature. A summary of published data is as follows: 31 case reports included confirmation of diagnosis enzymatically. Six were diagnosed by family investigation. Twenty-one published case reports did not include detailed clinical characterization and were not reviewed. Four 3MCCD cases died (33 days, 26 mo, 40 mo, 11 mo). Cause of death included cerebral edema and cardio-respiratory failure. The lethality rate is approximately 10% from these reported cases. Nine 3MCCD cases didn’t show any clinical signs (27%) while a large subgroup developed mild mostly neurological symptoms. Onset of symptoms was from 1 day of life to 4.7 years of age. Hypoglycemias, mild imbalance of blood gases and sporadic hyperammoniemia have been reported. A frequent finding in single case reports is impaired consciousness with development of coma. Neurological symptoms like hypotonia, psychomotor retardation or seizures are prominent (30%). There is one case report of cardiomyopathy and 4 with brain atrophy, 2 with leukodytstropy and a single case fatty liver / muscle. Regarding clinical phenotypes, the potential for ascertainment bias in these reports is high. 3MCC has a much higher incidence with newborns screening. Outcome of the newborn screening cases has been excellent; all healthy and growing normally up to age 6.5 years. 47

Treatment consisted of mild protein restriction, carnitine supplementation and prevention of catabolic episodes. 9 mothers and 4 elder siblings – asymptomatic or mild unspecific symptoms (2 moms) were identified. The authors concluded that if NBS was done for over 20 years: 170,000,000 children born in US and Europe would have been screened. From published data: 4 individuals would be predicted to die of 3MCC. Assuming that some cases would be missed due to non-reporting or non-recognition (5 fold rate), approximately 20 deaths could be attributed to 3MCCD over this time period with a maximum risk of death of 1 in 8,500,000 (1: 8.5 million). Twenty-seven children were reported to have minor clinical symptoms. Assuming 5-fold higher rate, 135 cases would be expected to have some expression of clinical symptoms with a maximum rate of 1 in 1,200,000. Compared with mean birth prevalence in NBS (1:84,000) the authors hypothesize that the proportion of children who might develop minor symptoms would be around 10% and only less than 1 or 2% might have a risk for severe adverse outcome. Had NBS been implemented 20 years ago it might have identified more than 2000 MCC cases with a potentially benign outcome while 50,000 infants would require a 2nd card to rule out disease (false positives (0.03%). The authors also show that mutations are heterogeneous and do not predict clinical outcome. Biochemical markers and enzymatic testing are also of no help in predicting clinical outcome. The authors suggest that part of the explanation lies in the functional reserve in the Leucine catabolic pathway. Based on the findings in this study, the NBS program in Bavaria / Germany stopped screening for 3MCCD. Dantas et al, 2005 (74) NBS-detected cases from Germany, Australia, US, Turkey and Taiwan were compared with clinically ascertained cases from Europe. No clear correlation between genotype and phenotype was found in the MCCA and B genes. The estimated incidence through newborn screening was 1 in 50,000 with no reported confidence interval. Koeberl et al, 2003 (75) Twenty children were detected with high C5OH in the North Carolina NBS program.: ƒ 10 cases had 3OH isovaleric acid and 3methylcrotonylglycine detected in urine. ƒ 4 cases were found to have transient C5OH elevation and subsequent testing revealed that it was of maternal origin. ƒ 8 cases had probable 3MCCD and 7 of these cases had enzymatic confirmation. All infants were healthy and were growing and developing normally when they were referred for follow-up testing. The estimated incidence was 1:64.000 (no confidence interval reported). Ficicioglu and Payan, 2006 (76) This is a case report of a 19-month-old girl with 3MCC detected by NBS. She was treated for 12 months and then lost to follow up. Later she presented acutely with severe

48

metabolic acidosis, hypolgycmeia and low plasma carnitine. She responded to treatment and is developmentally normal. Table 9: Recent primary research reviewed by BC 3MCC working group

Author Stadler et al, 2006 (73)

Dantas et al, 2005 (74)

Setting

Subjects

Outcomes

Bavaria, Germany • Screened 677,852 • NBS revealed 8 newborns (5 females, (Australia and Austria neonates for 25 3 males) with biochemically confirmed collaboration) conditions, including MCCD were detected, corresponding to MCCD; 17 biochemical a frequency of 1:84,700 (1:12,500 to and clinical phenotypes 1:196,000; 95% confidence interval) in based on elevated the population investigated (677,852 concentrations of 3newborns) and 6 relatives with MCCD, hydroxyisovalerylcarnitine suggesting a higher frequency than (3-HIVA-C) previously assumed (1:84,700) • Genotypes of MCCA • A strikingly heterogeneous spectrum of (MCCC1) and MCCB 22 novel and 8 reported mutations was (MCCC2) were assessed discovered in identified newborns, • Comparative analysis of case reports their relatives, and in with NBS data implied that only few individuals (n=17) from individuals (<10%) develop symptoms. other regions, and In addition, none of the symptoms correlated to biochemical reported so far can clearly be attributed and clinical phenotypes to MCCD Of the 28 probands investigated, 10 (36%) had mutations in MCCA and 18 (64%) in MCCB 28 probands from 23 families were analyzed. Switzerland (US 28 MCC-deficient • Of 28 probands studied, 9 were Collaboration) probands, 19 of whom diagnosed because of clinical were asymptomatic symptoms, while 19 were asymptomatic newborns detected by newborns detected by TMS based TMS newborn screening, newborn screening and 9 presented with • Of the 28 probands investigated, 10 clinical symptoms (36%) had mutations in MCCA and 18 (64%) in MCCB including missense, nonsense, frameshift and splice site mutations, and show that three of the missense mutations result in severely decreased MCC activity when expressed in MCC deficient cell lines • Data demonstrate no clear correlation between genotype and phenotype suggesting that factors other than the genotype at the MCC loci have a major influence on the phenotype of MCC deficiency

49

Author

Setting

Ficicioglu & Payan, 2006 (76)

Case study; Pennsylvania, USA

Koeberl et al 2003 (75)

North Carolina NBS program

Subjects

Outcomes

• 1 Nineteen-month-old • At 19 months, the child was admitted to girl with a 3a local hospital with lethargy, methylcrotonyl-coenzyme hypoglycemia, and acidosis, which A carboxylase deficiency developed during a febrile upper • A protein-restricted diet respiratory infection (2 g/kg per day) and • She was treated with intravenous 10% carnitine supplementation dextrose water and 0.45 normal saline in (50 mg/kg per day) were addition to intravenous carnitine (100 prescribed mg/kg per day) and bicarbonate • A low-protein diet was restarted, and the same dose of carnitine orally was continued • The patient is currently healthy and has a normal development for her age • All newborns with • 10 of 20 infants had significant isolated abnormal C5OH concentrations of both 3on TMS newborn hydroxyisovaleric acid and screening who were 3-methylcrotonylglycine in their urine, referred for follow-up suggestive of 3MCC deficiency testing • 20 infants with 2 4 of 10 abnormal metabolites were of consecutive elevated maternal origin C5OH evaluated for 8 cases of 3MCC confirmed 3MCC 1 of 7 infants had only marginally decreased 3-MCC activity in lymphocytes but deficient 3-MCC in fibroblasts. Authors conclude that TMS will detect additional inborn errors of metabolism, such as 3MCC deficiency

50

Beta-Ketothiolase Deficiency (BKD) Beta-ketothiolase deficiency (BKD), also known as mitochondrial acetoacetyl-CoA thiolase deficiency [EC 2.3.1.9], involves isoleucine breakdown and ketone body metabolism. The major clinical manifestations of this disorder are intermittent ketoacidosis, vomiting, lethargy leading to coma that may be precipitated by intercurrent illness. Chronic neurological disability with progressive loss of sensory and motor functions, seizures and mental retardation can occur. The mechanism of brain injury is unclear. At least some patients appear to remain in good health between acidotic episodes. BKD is a very rare inborn error of metabolism (~ 1 in 500,000 live births). (79) Frazier et al, 2006 report 2 cases detected in 1 Million babies screened.(57) Patients have a characteristic urine organic acid profile with 2-methyl, 3-hydroxybutyate and tiglylglycine. In blood, tiglylcarnitine (C5:1) and 2-methyl 3-hydroxybutyrylcarnitine (C5OH) are variably elevated. The C5:1 signal can detect some cases of BKD. (79)

Knowledge Claims Formulated by BC Working Group on BKD What is the evidence that newborns with BKD will have improved health outcomes with early detection following newborn screening with TMS? If BC commences screening for BKD, the evidence reviewed in this HTA indicates that: • 1 newborn will be detected with BKD every 10 to 12 years or longer. (level 3c) (32,57) • In that one newborn it may be possible to avoid metabolic crises and death or chronic neurological deficits but the condition is so infrequent that there is a lack of good data to support treatment effectiveness. • It may be worthwhile to detect this disorder if there are minimal false positives with the use of the C5;1 signal. However, test performance characteristics with this analyte are limited at this time and would require evaluation. The Clinical Evidence Base Ozand et al, 1994(80) (level 3a) Ozand et al, 1994 report a small uncontrolled case series of 4 clinically ascertained cases revealing organic academia with associated encephalopathy. In 3 cases MRI of brain showed an increased T2 signal intensity within posterior lateral part of putamen bilaterally. Treatment was with mild protein restriction to limit isoleucine, avoidance of fasting and carnitine supplementation. Frazier et al, 2006 (57) (level 3a) Frazier DM et al 2006 report the results of tandem mass spectrometry screening of 944,078 newborns between 1997 and 2005 in North Carolina.(57) Two cases of BKD were diagnosed using the C5:1 analyte as the primary marker (+/- an elevation in C5OH). During this time period, one case of BKD was clinically ascertained and had a C5:1 on newborn blood spot within normal limits (false negative). This child was found to have an elevated C5:1 only during ketoacidotic episodes. The positive predictive value of C5:1 for BKD is not reported in the paper.

51

Fukao et al, 2003 (81) (level 3a) Fukao et al, 2003 report a small uncontrolled case series of 5 cases. Four were symptomatic with onset at 9 to 23 month. Only 1 patient had a value above a blood spot acylcarnitine C5:1 cut of 0.49. All were above C5OH/C2 cut-off of 0.032. This illustrates the challenges in determining a cut off value and therefore for screening. Table 10: Recent primary research reviewed by BC BKD working group

Author

Setting

Chace et al, 2003(82)

Pennsylvania, USA

Fukao et al, 2001(83)

Gifu, Japan

Fukao et al, 2003(81)

Gifu, Japan

Subjects

Outcomes

Narrative literature review (verbatim) 2-methyl-3hydroxybutyrylcarnitine is increased in BKD. BKD may also be characterized by an additional increase at mass 300 that represents tiglylcarnitine The molecular basis of Genotype does not predict clinical BKD [was determined] in severity and mutant sibs can have 26 patients at the levels different clinical phenotypes; there were of skin fibroblast enzyme no consistent differences in clinical activity, protein integrity, severity between patients with nulland DNA nucleotide conferring or residual-conferring sequence genotypes; only the absence of or a low urinary excretion of tiglyglycine during ketoacidosis correlated with a mild genotype. In general, BKD has a favorable outcome and 23 of 26 patients developed normally; and two have developmental delay one died during the first ketoacidotic episode. The median age at onset for the first ketoacidotic episode is 15 months (range 3 days to 48 months). The frequency of attacks falls with age; 11 patients had only one episode and 3 patients had none. Authors conclude that clinical are avoidable with early diagnosis, appropriate management of ketoacidosis, and modest protein restriction. (verbatim) two groups [identified]: those with null mutations in either allele (group 1) and those with mutation(s) retaining some residual enzyme activity in at least one of two mutant alleles (group II). BKD (group II) can be misdiagnosed as normal if these analyses are performed under nonepisodic conditions and possibly during the newborn screening for inborn errors of metabolism.

52

Author

Setting

Ozand et al, 1994(80)

Riyadh, Saudi Arabia

Mnastiri et al, 1999(84)

Monastir, Tunisia

Sewell et al, 2005(87)

Frankfurt, Germany

Pasquali et al, 2006(85)

Salt Lake City, Utah, USA

Subjects

Outcomes

4 clinically ascertained patients

2 abnormal EEGs Reports organic academia with associated encephalopathy 3 patients with brain MRI changes to the putamen 4 clinically ascertained (verbatim) [cases] were all males with a cases mean age at diagnosis of 10 months, their parents were first cousins, and they had a family history of early sibling death. They all developed a severe ketoacidosis with pH values below 7.1 after a febrile illness or gastroenteritis. They exhibited neurological symptoms with lethargy and coma in three cases; two developed acute renal failure and needed peritoneal dialysis. Two boys died after their second decompensation and another had slight psychomotor retardation. Case report of a 25-year- (verbatim) [MS/MS] of maternal old woman pregnant for plasma demonstrated tiglylcarnitine the first time in the 32nd From then on she received carnitine week of gestation. She supplementation (100 mg/kg per day)…. had been diagnosed with …patients with MAAT deficiency mitochondrial have a relatively good prognosis acetoacetyl-CoA thiolase provided acute episodes of ketoacidosis deficiency at age 8. and dehydration are adequately treated. Long-term dietary treatment includes protein restriction ; although this was not strictly followed by our patient, she had nevertheless voluntarily reduced her protein intake. During her last weeks of pregnancy, no changes in the already elevated urinary excretion of MHB and TG could be discerned. … Narrative literature review Newborn screening by MS/MS can identify infants with BKD with severe mutations, but it might fail to detect infants with the ‘‘milder’’ mutation(s). The outcome of BKD is favorable with early diagnosis, dietary therapy, and appropriate treatment of ketoacidosis

53

3-Hydroxy-3-methylglutaryl-CoA Lyase Deficiency (HMG) CoA Lyase) 3-Hydroxy-3-methylglutaryl (HMG) CoA lyase [E.C.4.2.3.4] catalyzes the last step of ketogenesis as well as leucine degradation. HMG CoA lyase deficiency is a very rare metabolic disorder that usually presents with acute episodes of hypoglycemia and metabolic acidosis in early infancy. A major factor in the causation of hypoglycemia is impairment in ketone production. The pattern of organic aciduria is excretion of 3hydroxyisovaleric, 3-methylglutaric, 3-methylglutaconic and 3-hydroxy3-methylgltuaric acids that are compounds of the leucine catabolic pathway. Treatment consists of supportive measures during acute crisis (correction of fluid deficits, IV glucose and ventilation if necessary). Long-term management includes avoidance of fasting, and a high carbohydrate, low protein diet. In an earlier review, NSAC recommended not screening for 3MCC deficiency, the most common cause of an elevated C5OH on the acylcarnitine panel. In the very rare disease, HMG-CoA lyase deficiency, both C5OH and C6DC are elevated. However, C6DC is not reliably measured in blood spots and the use of this signal is not sensitive or specific for HMG CoA lyase deficiency. For technical reasons, it will therefore not be possible to easily screen for this rare disorder.

Knowledge Claims Formulated by BC Working Group on HMGCoA Lyase What is the evidence that newborns with HMGCoA Lyase will have improved health outcomes with early detection following newborn screening with TMS? If BC commenced screening for HMG deficiency, the evidence reviewed in this HTA indicates that: • 1 newborn will be detected with HMG deficiency infrequently. No case has ever been detected clinically in BC. The large German screening series did not detect a case. Still cases are not rare in some populations (Saudi Arabia for example) (level 3c) • If there was a birth with a serious HMG deficiency in BC the neonate may or may not benefit from screening. • There is a greater than 50% chance the case may become clinically evident before screening test results are available. Survivors have a chance for normal development with treatment. • In some cases metabolic crises may be prevented in an unsuspected case and normal development enhanced through treatment. The Clinical Evidence Base Gibson et al, 1988 (88) (level 3b) In this classic paper, HMG is described as a defect of leucine catabolism and final step of ketogenesis. On the basis of the evidence of the 18 first reported and clinically 54

ascertained cases, Gibson et al. report that HMG often presents as a life threatening illness in neonatal period with severe hypoglycemia, metabolic acidosis, lactic acidosis, hepatomegaly, lethargy and coma. About half of the patients present in 1st week of life with the second half later and usually during 1st year of life. Ketosis is not a part of the clinical picture. HMG is treated with a high carbohydrate and low protein diet supplemented with carnitine and the avoidance of fasting. Many newborns with this condition are developmentally normal. Kayser 2006(89) (level 4) This review from the US National Institute of Health Human Genome Research Institute provides an overview of the disorders of ketone production and use. It provides the knowledge claim that the abnormal cerebral white matter foci seen on brain MRI do not appear to be progressive demyelination. This paper reports that ‘dried blood spots from newborns evaluated by tandem mass spectroscopy show elevations in 3methylglutarylcarnitine, providing a means for newborn screening for HMG deficiency.’ (89) p282 Ozand et al, 1991(90) (level 3a) Of the 11 infants whose clinical course is detailed in this case series from Saudi Arabia, 2/3 presented in neonatal period. Therefore these cases are likely to be clinically evident before screening results are available. Normal development was noted in 4 cases. Table 11: Recent primary research reviewed by BC HMGCoA Lyase working group

Author

Setting

Al-Sayed et al, Saudi Arabia 2006(91) Bischof et al, 2004(92)

Tubingen, Germany

Cardoso ML 2004(93)

Porto, Portugal

Subjects

Outcomes

34 3HMG dried blood spots

3 mutations account for >94% of the pathogenic mutations underlying 3HMG in Saudi Arabia 36-year-old woman with (verbatim) The patient improved seizures, recurrent clinically on oral L-carnitine substitution. metabolic disturbances, This treatable condition can remain and severe unrecognized in adults and should be leukoencephalopathy considered a potential cause of leukoencephalopathy. Molecular study of the (verbatim)The relatively high frequency HMGCL gene in 11 of the "common" HMGCL Portuguese cases originated from the mutation makes useful the development Northern area of Portugal of a rapid and specific molecular confirmation of new cases with HL deficiency in [Portugal].

55

Author Cipriano et al, 2007(94)

Gibson et al, 1988 (88)

Mitchell et al, 1998(95)

Ozand et al, 1991(90)

Setting

Subjects

Outcomes

London, ON, Canada

A decision-analytic model (verbatim)The incremental costs of was constructed from screening ranged from $222,000 (HMGsecondary sources and CoA lyase deficiency) to $142,500,000 expert opinion estimates (glutaric acidemia type II) per LY gained. of the clinical parameters. Screening for a bundle of diseases The London Health including PKU and the 14 most costSciences Centre Case effective diseases to screen for cost less Costing Initiative, the than $70,000 per LY gained, and the Ontario Health Insurance incremental cost-effectiveness of adding Plan Schedule, and the each of the 14 diseases to the bundle Ontario Drug Benefits was less than $100,000 per LY gained. plan formulary were used The incremental cost of adding the 15th to estimate costs. most cost-effective disease was $309,400 per LY gained. Literature review plus Summary of 18 reported The defect and cardinal clinical features practice cases patients are described. Reported health outcomes ranged from normal development to severe mental retardation and death. Saudi Arabia Study of genetic [Authors] developed a screening mutations in 9 Saudi and procedure for HL missense mutations 32 non-Saudi HLthat yield residual activity at levels deficient probands comparable to those obtained using purified HL peptides. 11 infants with HMG Devastating acidotic attack within first Saudi Arabia deficiency day of life for 7 cases and by 3rd day for 2 cases Normal development for 4 surviving infants diagnosed and treated early.

Stacey et al, 1985(96)

Harrow, London, UK

4 month old twins with 3- One twin was asymptomatic, the other hydroxy-3-methylglutaric symptomatic. Dietary treatment aciduria due to HMG CoA improved symptoms but did not prevent lyase deficiency episodes similar to Reye’s syndrome in both.

Yylmaz et al, 2006(97)

Istanbul, Turkey

3.5-year-old previously healthy male who was admitted with complex partial seizures and was diagnosed as having 3hydroxy-3-methylglutaryl coenzyme A lyase deficiency

Medical history unremarkable, development and neurological exam were normal. Prominent corticospinal tract and pontine involvement with focal cerebral white matter changes on MRI not previously reported.

56

2. DISORDERS OF FATTY ACID OXIDATION AND CARNITINE METABOLISM Very-long chain acyl CoA dehydrogenase deficiency (VLCAD) / Long chain 3 hydroxyacyl-CoA dehydrogenate deficiency (LCHAD)/ Short chain acyl-CoA dehydrogenase deficiency (SCADD) / Carnitine uptake disorder MCAD and the fatty acid oxidation disorders (FAOD’s) in general Fatty acid14 oxidation plays a major role in energy production especially during periods of fasting. More than 20 enzymes and transport proteins are involved in the metabolism of fatty acids. Carnitine is involved in the transport of fatty acids into the mitochondria where they are broken down or oxidized. A number of enzyme defects in this pathway have been described.(98) Medium chain acyl-CoA dehydrogenase deficiency (MCAD) is the most common fatty acid oxidation disorder and is currently included in the BC newborn screening panel. The literature on very-long chain acyl CoA dehydrogenase deficiency (VLCAD) / Long-chain acyl-CoA dehydrogenase deficiency (LCHAD) / Short chain acyl-CoA dehydrogenase deficiency (SCADD) and carnitine uptake disorder (CUD) is reviewed in this section. Clinically, fatty acid oxidation disorders (FAODs) each have their own unique signs, symptoms, mortality rates and long-term health problems. Common to all of these disorders, fasting and other physiologic stressors such as infection or exercise often induce clinical symptoms. Early detection and treatment improves health outcomes for these disorders to varying degrees. MCAD is a fatty acid disorder for which the available evidence supports screening and has been extensively reviewed by others including the UK HTA 2004. MCAD may present with low blood sugar or sudden unexpected death. The mortality rate prior to the commencement of newborn screening programs for MCAD was approximately 25%. Newborn screening for MCAD has been successful in reducing the risk of death to < 1%. There are no long-term chronic problems that have been recognized so far with children diagnosed and treated with MCAD from birth. For instance, children with MCAD do not suffer from cardiomyopathy, which is common to other fatty acid oxidation disorders. MCAD is detected by blood spot acylcarnitine analysis performed by tandem mass spectrometry (TMS). An elevation in C8 and C8:C10 ratio in blood spot card is highly sensitive and specific for this disorder.

14

Acids obtained through the breakdown of fats that can be used by body cells as energy.

57

Many other fatty acid oxidation disorders can also be detected by TMS and each have their own typical acylcarnitine profiles. The following review covers 3 other common FAOD’s; LCHAD including trifunctional protein deficiency, VLCAD and CUD.

LCHAD and TFP LCHAD is an autosomal recessive disorder first described in 1990.(99) LCHAD is part of an enzyme complex that catalyzes three steps of long chain fatty acid oxidation— hence, the name trifunctional protein (TFP).15 TFP in turn is encoded by two different genes; an alpha and beta subunit.(100) The mature enzyme is an octamer made up of 4 alpha and 4 beta subunits. Various gene defects either cause isolated LCHAD deficiency or a more generalized condition called TFP deficiency, which involves loss of all three enzymes in the TFP complex.(98) The biochemical hallmark of LCHAD deficiency is the accumulation of long-chain 3-hydroxyacylcarnitines and fatty acids in the blood (as well as dicarboxylic acids in the urine). The disease presents most prominently with fasting or illness-induced hypoketotic hypoglycemia, along with a combination of chronic non-specific symptoms that include cardiomyopathy, severe liver disease, muscle cramps, pigmentary retinopathy (leading to impaired vision), and peripheral neuropathy.(101) The early onset forms of the disease tend to be more severe, often resulting in coma and death. (98) Screening can be achieved through TMS measurement of the C16OH analyte, but this approach does not discriminate between LCHAD and the more severe trifunctional protein deficiency (TFP). Confirmation of diagnosis is achieved by enzymatic and / or molecular testing which differentiates between isolated LCHAD and TFP. Patients with beta subunit defects usually have a defect in all 3 enzymatic components (trifunctional protein deficiency). There is a common mutation in the alpha subunit (1528G>C), which causes isolated LCHAD deficiency. Treatment has traditionally involved avoiding fasting (usually through a diet high in carbohydrates) and limiting long-chain fatty acid intake (sometimes in favour of medium-chain fatty acid supplements).(101) Individuals with LCHAD deficiency may experience intensified symptoms after periods of physical exertion, a phenomenon that sometimes leads to exercise avoidance and obesity. Investigations are underway that focus on enhanced physical activity and short-term higher protein diets to facilitate metabolic and especially weight control in children with LCHAD deficiency.(102,103) While the condition occurs more frequently than other TFP-related conditions, LCHAD deficiency is still rare; for instance, the incidence in BC is about one case every 2 years. Even though metabolic decompensation is still observed in some individuals receiving treatment, screening may be justified based on the sensitivity and positive predictive value of TMS screening, the reduction in mortality at least for isolated LCHAD deficiency, and the improved quality of life for many patients.(104) Furthermore, wherever TMS is already being used for the screening of other disorders, extension of 15

Also known as MTP, for mitochondrial trifunctional protein

58

newborn screening to include LCHAD deficiency is possible with minimal additional costs for reagents.(105) There is considerable clinical overlap between isolated LCHAD deficiency and TFP deficiency. Some researchers have studied the 2 together while others have studied them separately. Where a distinction is made, the evidence will be presented as either LCHAD or TFP. Of note, the acylcarnitine profile that detects LCHAD and TFP are the same; an elevation in C16OH is the most sensitive marker for the detection of long chain fatty acid oxidation defects. Knowledge Claims Formulated by BC Working Group on LCHAD What is the evidence that newborns with LCHAD will have improved health outcomes with early detection following newborn screening with TM/MS? If BC commences screening for LCHAD, the evidence reviewed in this HTA indicates that over a five-year period 2008/09 to 2012/13: •

1 newborn will be detected with LCHAD every 2 years or more. (level 3b evidence)

Earlier detection will lead to the following benefits for newborns. • Death will be prevented in 1 newborn with unsuspected LCHAD every 5 years (level 3b evidence) • Despite detection and treatment, metabolic decompensation may occur in 1/4 patients and recurrent muscle pain in 1/3rd (level 3b evidence) • Protein enhanced diets may lead to better metabolic control, reduced heart rate during exercise, more active lives and lower rates of obesity (level 1 evidence) The Clinical Evidence Base Does early detection (newborn screening) prevent mortality and morbidity in LCHAD? Many case reports and uncontrolled case series were identified but no cohort studies or randomized control trials were identified that would supply conclusive evidence of benefit. Specifically, there are no published studies that compared health outcomes in clinically ascertained versus newborn-screened groups. However, two key papers provide some evidence that mortality may be reduced with earlier detection of LCHAD cases but with limited evidence that newborn screening can reduce morbidity: Natural history of LCHAD Den Boer et al 2002 (99) (50 patients, level 3b) The largest series of clinically-ascertained patients with isolated LCHAD deficiency was published by Den Boer et al, 2002.(82, 99) The 1528G>C mutation was found in 84/98 alleles tested; 36 homozygous and 12 heterozygous. Interestingly, 15% of pregnancies were complicated by HELLP syndrome16 and 2 were complicated with acute fatty liver of 16

HELPP stands for hemloysis elevated liver enzymes and low platelets which are the problems of this syndrome associated with pregnancy induced hypertension.http://healthcare.utah.edu/healthinfo/pediatric/hrpregnant/hellp.htm

59

pregnancy. Age of onset was 1 day to 26 months with a mean of 5.8 months. Seventyeight percent of patients presented acutely with hypoketotic hypoglycemia. Twenty-two percent of patients presented with a more chronic disorder; liver disease, failure to thrive, feeding difficulties and or hypotonia. Retrospective analysis revealed that most patients who presented with hypoglycemia had other symptoms first (failure to thrive, feeding difficulties, hypotonia). Nineteen of 50 patients had died (38%); 14 before a diagnosis was made and 5 within the 1st 3 months after diagnosis. Twenty-nine of the 31 surviving patients were reported to be generally in good clinical condition at a median follow up 3.4 years (range 0.5 – 11 years). Morbidity in this group is still high with recurrent metabolic crises and episodes of rhabdomyolysis even with treatment. The authors concluded that treatment could reduce mortality in LCHAD deficiency. Sander et al, 2005 (105)(7 patients, level 3b) Sander et al. 2005, described a series of LCHAD patients detected by newborn screening.(105) Follow up ranged from less than 1 year to 5 years. Seven cases of LCHAD were detected in 1.2 million births. Outcomes were as follows: 1 case of LCHAD had no clinical symptoms (3 years of follow up). All other LCHAD patients have had some clinical manifestations ranging from hypoglycemic episodes, cardiomyopathy, episodes of rhabdomyolysis, mild developmental and motor delay. All 7 children were alive at the time of publication. These results suggest that newborn screening can prevent mortality but may not prevent other clinical manifestations. Olpin et al, 2005 (106) (3 patients level, 3b) Olpin et al, 2005 reports one case of isolated LCHAD with failure to thrive (FTT) during 1st 6 months of life who presented with acute life threatening crisis at 6 months, cardiomyopathy and 2 cardiac arrests. Episodes of decompensation (high CK) requiring readmission occurred but this child remains physically active and developing with normal IQ and only mild nonprogressive pigmentary retinopathy at 14 years. Tyni et al, 1997 (107) (13 patients, level 3b) Tyni et al reports a series of 13 LCHAD cases (homozygous for G1528C) all with histories of hypoglycemia, cardiomyopathy, muscle hypotonia and hepatomegaly during the 1st 2 years of life.(107) All but one died before age 2. Seven had recurrent metabolic crises, 6 had a steadily progressive course, 2 had cholestatic liver disease, 1 had peripheral neuropathy, and 6 had retinopathy with focal pigmentary aggregations. Only 1 patient who has had dietary treatment for 9 years was alive at 14 years. Morris et al, 1997 (108) (9 patients, level 4) Morris et al report a case series of 9 patients with LCHAD in a letter to the editor.(108) They report all are alive with mean age of 3.5 years (range 1 to 8 years). Two families had previously unexplained neonatal deaths. All cases but one had hypoglycemic encephalopathy. One had neurological problems. Four cases had cardiomyopathy that resolved in the 3rd and improved in the 4th. Three children had single episodes of rhabdomyolysis. Two older children had mild funal pigmentary changes. Treatment was 60

with low fat, high carb diets supplemented with MCT and essential fatty acids. Fasting was rigorously avoided with the importance of overnight feeds emphasized. Randomized Controlled Trials of Dietary Supplementation The highest quality evidence (level 1) is of the physiological affects of variations in dietary treatment tested in randomized controlled trials. Gillingham, Purnell et al, 2006 (102) (9 patients with LCHAD or TFP, level 1) Gillingham, Purnell et al 2006 conducted a randomized cross over design trial of higher dietary protein intake on energy balance and metabolic control in children age 7 to 14 with LCHAD or TFP.(102) During 6 day trial as inpatients, the dietary innovation resulted in lower energy consumption an resting energy expenditure was higher compared to high carbohydrate diets. Longer-term trials are needed to ascertain whether higher protein diets will reduce rates of overweight and obesity. Gillingham, Scott et al, 2006 (103) (9 patients with LCHAD, level 1) Gillingham, Scott et al 2006 conducted a randomized cross over design trial of protein supplementation versus orange juice immediately before exercise in children with LCHAD.(103) With supplementation, steady state heart rate decreased with better metabolic control. These findings demonstrate that coordinated MCT supplementation before exercise is a strategy that may lead to increased activity with better metabolic control in children with LCHAD. Table 12: Recent primary research reviewed by BC LCHAD working group

Author Gillingham, Scott et al 2006(103)

Gillingham, Purnell, 2006

Setting

Subjects

Outcomes

Subjects were recruited from a previous study; randomized cross-over study of the ability to tolerate exercise with and without MCT supplementation

Inclusion criteria were • Coordinating medium-chain triglyceride a confirmed diagnosis supplementation with periods of increased of LCHAD or TFP activity may improve the metabolic control deficiency and seven of children with LCHAD and TFP years of age or older. deficiency following exercise 9 subjects, age 7–14 • Cumulative long-chain 3hydroxyacylcarnitines were 30% lower and were enrolled. hydroxybutyrate was three-fold higher after the Subjects were fed a MCT-pretreated exercise test compared to the controlled diet consisting of 10% of test with orange juice alone. total energy from LCFA and 11% of total energy from MCT with varying amounts of carbohydrate and protein Oregon: same cohort as To determine short High protein diet was safe, well tolerated above term safety and and associated with lowered energy efficacy of high intake and higher energy expenditure protein diets in LCHAD compared to high carb diet; / TFP subjects - longterm studies are needed to determine if high protein diet will reduce rates of obesity

61

Author

Setting

Gillingham, Oregon USA 2005 (what ref? Not – check online/ database – not in biblio1 02)

den Boer et al, 2002(99)

1 clinic in Amsterdam, the Netherlands

Olpin et al, 2005(106)

Children's Hospital, Sheffield, UK

Subjects Randomized, crossover design of the result of optimal dietary therapy on visual function; 9 subjects (3 males, 6 females) with LCHAD or TFP deficiency were recruited from a previous cohort for this study, age 7–14 were enrolled; duration: patients were followed / treated for 3 – 5 years Physician survey reporting on 61 unselected pts with LCHAD diagnosed in 1 clinic in Amsterdam (82% return rate) • Mean age of clinical presentation was 5.8 months (range: 1 day26 months)

Outcomes Optimal dietary therapy as indicated by low plasma 3-hydroxyacylcarnitine and high plasma DHA concentrations was associated with retention of retinal function and visual acuity in children with LCHAD or TFP deficiency.

• 15% of the pts presented in the neonatal period; 78% presented with hypoketotic hypoglycemia (classical features of a fatty acid oxidation disorder); 22% presented with chronic problems, consisting of failure to thrive, feeding difficulties, cholestatic liver disease and/or hypotonia • 82% also suffered from a combination of chronic nonspecific symptoms before metabolic crises • 38% mortality rate all before or within 3 months after diagnosis • 94% survivors reported as “in good clinical condition” Authors conclude that survival can be improved by prompt diagnosis, but morbidity remains high despite current therapeutic regimes 9 cases span the full Biochemical, clinical and mutation data spectrum of disease (8 document a wide clinical spectrum of with general TFP disease ranging from severe deficiency and 1 pt neonatal/infantile cardiomyopathy and with isolated early death to mild chronic progressive LCHAD deficiency) sensorimotor polyneuropathy with episodic rhabdomyolysis. (latter is probably under diagnosed given difficulty in diagnosis). The mild clinical presentation and relative difficulty in diagnosis suggest that this form of TFP is probably under diagnosed.

62

Author Sander et al, 2005(105)

Tyni et al, 1997(107)

Morris et al, 1997(108)

Setting

Subjects

Outcomes

7 newborns with isolated LCHAD developed • Screening for satisfactorily during an observation period of up LCHAD has been included in the routine to 64 months - 4 with typical mutation (1528 G>C) German NBS panel 4 children did not survive, two of them showing • Elevated levels of complete deficiency of TFP and two showing C(16)-OH and deficiency of long-chain 3-ketoacyl-CoA C(18:1)-OH, and, in thiolase. some cases, of Authors conclude: Despite the rarity of the C(14:1) and/or C(14)- disease, screening for TFP deficiencies is OH were found in 11 justified based on the following criteria: of the 1.2 million blood improved quality of life for pts with isolated samples LCHAD deficiency, absence of • stigmatisation of babies showing mild variants without necessity of treatment, no significant increase of the total number of false positive screening results, no false negative results to their knowledge. Finally, extension of analysis to TFP deficiencies is achieved with only negligible additional costs for screening laboratories already using TMS Finland, Sweden, UK & • 13 pts diagnosed • 7 pts had a relapsing-remitting disease, Netherlands joint 1991 to 1995 with and they died after two or three venture LCHAD (G1528C deterioration phases. Between mutation): exacerbations the pts were more or less , 12 with diagnosis on free of symptoms. In the remaining six pts post mortem the disease was progressive. Only one pt • is alive at the age of 14 years • Reductions in the enzyme activities of cultured skin fibroblasts did not correlate with the severity of the clinical outcome • Early institution of dietary management may alter the otherwise invariably poor prognosis Letter to Editor Re: Tyni 9 unrelated pts with Aggressive dietary treatment (e.g. very et al LCHAD deficiency, all low-fat, high-carbohydrate diet with fatty alive, mean age of 3.5 acid supplements) may achieve better years (range 1 to 8 outcomes (Tyni et al) years)

German Screening Commission

63

VLCAD Very long-chain acyl-CoA dehydrogenase (VLCAD) deficiency is another autosomal recessive disorder of fatty acid oxidation that was discovered in 1992. Early presentation is the most frequent subtype, and it often involves the most severe expression of the disease; the characteristics are similar to LCHAD deficiency, including cardiomyopathy, muscle weakness, and lethargy. Individuals who are diagnosed later in childhood, or even after childhood usually experience milder forms of metabolic disturbance.(109) TMS can be used to screen for VLCAD deficiency, yielding results of high specificity and sensitivity. Further tests, such as enzyme assays and DNA analysis, are necessary to precisely diagnose VLCAD deficiency, differentiating it from other fatty acid oxidation disorders.(110) The importance of early screening for VLCAD deficiency is underlined by the fact that blood samples may be unreliable for diagnostic purposes if drawn from a child more than four days old; furthermore, many genetically affected individuals are asymptomatic until a severe and sometimes fatal metabolic crisis. Treatment is similar to that for other fatty acid oxidation disorders, namely, avoiding prolonged fasting and following a modified diet that restricts fat intake and perhaps includes substitution of medium-chain triglycerides. Incidence worldwide is estimated to be anywhere from 1:50,000 to 1:120,000 live births; thus, in BC, no more than one case a year would be expected.(111) Despite such rarity, timely detection and treatment may be useful in preventing the onset of arrhythmias, heart failure, episodes of rhabdomyolysis, hypoglycemia and death.(112) VLCAD catalyzes the initial step of mitochondrial beta-oxidation of long chain fatty acids with a chain length of 14-18 carbons. In VLCAD deficiency, C14 to C18 accumulate before the enzymatic block, with C14:1 carnitine as the disease specific marker. The apparent incidence has increased since the introduction of newborn screening (NBS) suggesting that milder variants that may or may not be clinically ascertained are being detected by NBS. The clinical phenotype of clinically ascertained cases is variable. Early onset forms present with cardiomyopathy, life threatening arrhythmias, Reye-like symptoms and hypoketotic hypoglycemia. Milder phenotypes with episodic myoglobinuria (rhabdomyolysis) have also been reported in young adults. The majority of cases detected through newborn screening are asymptomatic at time of diagnosis and remain asymptomatic during follow up with preventative dietary measures such as avoidance of fasting long chain fat restriction and supplementation of MCT. It is not yet apparent which patients diagnosed by newborn screening would have become symptomatic if screening had not been performed. Screening for VLCAD: A blood spot C14:1 > 1 umol/L on a card collected on day 2 or 3 is highly predictive of VLCAD (positive predictive value > 50%). Repeat blood spots 64

(day 5-7) may have normal C14:1 levels in confirmed cases. Therefore, all initially positive screens require follow up with enzymology and molecular testing. There is some genotype / phenotype correlation in VLCAD. Cases detected through newborn screening appear to be milder than clinically ascertained cases. Knowledge Claims Formulated by BC Working Group on VLCAD What is the evidence that newborns with VLCAD will have improved health outcomes with early detection following newborn screening with TM/MS? If BC commences screening for VLCAD, the evidence reviewed in this HTA indicates that over a five-year period 2008/09 to 2012/13: • 1 newborn will be detected with VLCAD annually with some risk of detection of mild variants and false negatives if 2nd test not done within 4 days after birth. (level 3) Earlier detection will lead to the following benefits for newborns. • Prevention of metabolic crisis for most with prevention of organ damage and possibly death in some (level 3b or 4). The Clinical Evidence Base Newborn screening cohort studies are the highest level of evidence available to determine the benefit of early detection of VLCAD deficiencies. The 2006 study by Boneh et al and 2006 study by Leibig et al (109,110) provide the highest quality evidence of screening benefit. Given the short follow-up and detection of mild variants that may not have been of clinical significance, these studies have little protection against systematic bias. Boneh et al, 2006(109) (6 patients, level 3) Boneh et al,2006 report on the pitfalls of newborn screening for VLCAD deficiency. (109) Six VLCAD subjects detected by newborn screening in Australia among 189,000 infants screened between 2002 and 2005 yielding a prevalence of 1 in 31,500 with a 95% CI of 1:15,300 to 1:90,900. C14:1 C14 and C14:C10 ratio all increased on 1st card and decline on 2nd card. All babies with abnormal 1st screen were followed. The C14:1 to C10 ratio was also diagnostic and extended up to 100 hours after birth and appeared to be somewhat more reliable in 2nd sample than C14:1 alone. Note that VLCAD may be missed by newborn screening if the 2nd sample is drawn after day 4 of life. Reliance on a repeat blood spot to confirm initial findings of VLCAD when the baby is older than 4 days may lead to false negatives. All newborns identified in present study had at least one mutation that may lead to residual enzyme activity. Some individuals with milder forms may have had unrecognized episodes and some may remain symptom free until adulthood or throughout life unless challenged by a metabolic stress.

65

One patient was symptomatic at time of diagnosis with hypoglycemia and cardiovascular collapse without cardiomyopathy on day 4 of life. His acylcarnitine levels were abnormal on 2nd day. All other patients have remained free of symptoms with normal urine organic acid and subsequent blood spot acylcarnitine profiles up to age 6 months. Treatment was initially high MCT formula with carnitine. Change to free breastfeeding, monitoring carnitine levels and supplementing as required and a sick day regime similar to that used in MCAD. Older kids are on a low fat diet. There has not been an episode of metabolic decompensation documented under that regime even when babies have been sick with viral illness etc. Growth and development have been normal though very short follow up time. Liebig et al, 2006 (110) (11 patients, level 3) Liebig reports on 1 million neonates screened on day 3 of life. Eleven had elevated C14:1 and enzyme and molecular done. On repeat sampling, 8/11 had abnormal profile (day 57). This cohort may not be from the whole of Germany. Seven of the 11 had decreased palmitoylCoA oxidation, 2 had residual activities of 48,44%. Two disease causing mutations were detected in 6/7 remaining neonates, 1 had only 1 mutation. Nine were asymptomatic at time of diagnosis. Patient 0 suffered from CHF and surgery. Patient 5 had hypoglycemia at 3 months of age. The conclusions of this study were that C14:1 > 1 umol/L on day 3 correctly identifies VLCAD:

elevated C14:1 < 1 umol/L does not allow sufficient discrimination. By day 5-7: normalization of C14:1 as caloric intake increases after birth. The authors recommend that elevated C14;1 on newborn screening should always result in further diagnostic workup including measurement of fatty acid oxidation flux and DNA even if the 2nd acylcarnitne sample is normal.

66

Table 13: Recent primary research reviewed by BC VLCAD working group

Author

Setting

Subjects

Outcomes

Boneh et al, 2006(109)

Victoria State, Australia 6 VLCADD subjects were detected by screening 189,000 infants between Feb. 2002 and April 2005 (yielding a prevalence of 1:31,500 with a 95% confidence interval of 1:15,300 to 1:90,900). Newborn screening samples were normally collected at 48–72 h of age

• All confirmed VLCADD subjects had significantly increased levels of C14:1, C14, and C14:1/C10 ratio in the initial newborn screening sample • C14, C14:1, and C14:1/C10 values generally declined in the second sample taken after 130 h • Timing of blood sampling for newborn screening and performance of mutation analysis to avoid false-negative diagnoses of VLCADD in asymptomatic newborn babies is important

Liebig et al, 2006(110)

Germany, Netherlands and US joint study. Patients from Dusseldorf, Germany

Estimated 1,000,000 • Screening on day 3 of life revealed neonates screened for elevated C14:1-carnitine in all 11 VLCADD, 11 neonates, suggesting VLCADD presented with an • All neonates excluding one were elevated C14:1asymptomatic at the time of diagnosis carnitine. • In 6 of 7 pts with significantly reduced palmitoyl-CoA oxidation in lymphocytes suggested VLCADD, 2 disease-causing mutations in the VLCAD gene were delineated

Hoffman et al, 2006(113)

New England Medical Center, Boston, MA, USA

2 cases with adult onset at age 20 who survived basic military training before presentation

• With adult onset ,the presenting symptom is usually myalgia after long fast or vigorous exercise with possible progression of severe rhabdomyolysis to renal failure, cardiomyopathy, and even death

Doi et al, 2000(114)

Japan; Case study

Treatment with medium chain triglycerides and L-carnitine improved pt's hepatomegaly and cardiomegaly

Cox et al, 1998(115)

Boston, MA, USA

• Six-year old boy with VLCADD diagnosis was supported by the results of acyl-CoA dehydrogenase activity for C8 and C16 fatty acids in skin fibroblasts • 5-year old girl compared with 3 normal control subjects and 7 cases with VLCAD

• Nutritional supplementation enabled normal growth processes with weight at 75% to 90%, height at 90% to 95%, and head circumference at 75% to 90% • Heart size normalized over 24 months • 3 hospitalizations for presumed viral gastroenteritis accompanied by decreased oral intake, dehydration, and irritability • fatigue with exercise and occasional myalgias relieved by rest, carbohydrate intake, and/or ibuprofen • Normal neuropsychosocial development

67

Author

Setting

Subjects

Outcomes

Sluysmans et al, Belgium & France joint • Boy born in 1990 1997(116) study and sibling fetus both diagnosed with VLCADD s

• First sibling survived neonatal seizure, lethargy, and metabolic acidosis with a recurrence at age 10 months that lead to diagnosis • Prenatal diagnosis of VLCADD in sibling fetus suggests that prenatal diagnosis is feasible and may forestalls neonatal injury through close clinical and metabolic monitoring • Dietary therapy improved outcome Brown-Harrison USA 2 unrelated parents Of 4 children 2 had VLCADD and 2 were et al, 1996(117) who had 4 children, 2 normal. One VLCADD child died and the of which had early other survived with good outcome onset of significant apparently due to dietary therapy metabolic acidosis Cairns et al, Belfast City Hospital, 41-year old man Adult presentation of VLCADD has a 2000(118) Northern Ireland presenting with dark milder phenotype. Rhabdomylosis brown urine, extreme following more than usual level of exercise tiredness, pain and in the fasting state and failure to maintain heaviness in the legs an adequate fluid intake Engbers et al, Netherlands 1 -year old boy • Rhabdomyolysis may occur in the 2005(119) presented with absence of hypoglycaemia in young hypoglycemia in first infants week of life which lead • CK levels should be monitored in VLCAD to diagnosis during periods of prolonged fasting glucose and even acylcarnitine profiles • Dietary therapy improved outcomes Touma et al Child born to first cousin • 1 male infant with • VLCAD enzyme activity was severely 2001(120) parents severe VLCADD decreased to 2% of control levels acute presentation • Dietary management consisted of with cardiomyopathy skimmed milk supplemented with medium chain triglycerides and L-carnitine. • All clinical manifestations of the disorder including cardiomyopathy were completely resolved by appropriate dietary treatment Voermans et al Netherlands 71-year-old female • ATP production fails during prolonged 2005(121) presented/diagnosed exercise or fasting at age 51 with • Dantrolene sodium influences the exercise intolerance calcium influx (limits it) in muscle and thus due to VLCADD and limits excitation–contraction coupling, was treated with which prevents rhabdomyolysis in dantrolene sodium VLCADD Wood et al, USA • This case demonstrates that VLCADD 2001(112) could have been detected from the pt’s Male infant presenting own neonatal heel-stick sample; clinically with presymptomatic diagnosis would have VLCADD with avoided part of a lengthy and intensive retrospective analysis prediagnosis hospitalization with an of NBS blood spot estimated cost of $400 000 screening card • Fatty acid oxidation defects, including VLCAD deficiency, may account for as many as 5% of sudden infant deaths

68

Author

Setting

Subjects

Outcomes

Vianey-Saban et France al, 1998(122)

30 VLCADD cases in 2 phenotypes observed and described: 27 families a ‘severe’ presentation characterized by an early onset of symptoms, with hypertrophic cardiomyopathy and a high incidence of death, and a ‘mild’ form with hypoketotic hypoglycaemia, resembling MCAD. • Correlation with the two phenotypes provides an opportunity for accurate prognosis and early implementation of appropriate treatment. • Prenatal diagnosis of this life-threatening disorder was successfully performed in seven pregnancies in six of those families by assay of trophoblasts or amniocytes. Andresen et al, Denmark, UK, Investigated 58 • Clear genotype-phenotype relationship is 1999(123) Netherlands, France & different mutations in in sharp contrast to what has been Switzerland 55 unrelated VLCAD observed in MCADD, in which no pts cases correlation between genotype and representing all known phenotype can be established clinical phenotypes • 58 different mutations—only 18 of which and correlated the are present in 11 pt—identified in 110 mutation type with the mutant alleles from 55 unrelated pts. clinical phenotype.

69

Carnitine Uptake Disorder (CUD) Carnitine serves as a cofactor in the transport of long chain fatty acids across the mitochondrial membrane. It is within the mitochondria that fatty acids are oxidized to produce energy for the cell. Carnitine is thus an essential component of normal fatty acid oxidation. Although carnitine can be synthesized in the body, it predominantly comes from the diet. Carnitine is freely filtered by the kidney and is then reabsorbed at the proximal renal tubule by a specific high affinity transporter (carnitine transporter, OCTN2, OMIM 212140). The carnitine transporter is also found on the membranes of small intestine and skeletal muscle. A defect in the transporter leads to loss or ‘wasting’ of carnitine in the urine and resultant deficiency of carnitine in the blood and tissues. This deficiency prevents the body from using fats that are important energy reserves. CUD can cause sudden crises during periods without food as can occur during illness. Meat and dairy as well as nuts and seeds are the common sources of carnitine available through most diets.

Knowledge Claims Formulated by BC Working Group on CUD What is the evidence that newborns with CUD will have improved health outcomes with early detection following newborn screening with TMS? ƒ

1 newborn would be detected with symptomatic CUD every year.

ƒ

The clinical course of these children would be expected to be variable with or without detection with risk of serious illness or sudden death and including increased hypoketoic glycemia, cardiomyopathy (level 3b) Treatment may not prevent unexpected death. (level 3b) Carriers would be detected within families. (level 3b) One mother with symptomatic CUD would be detected once every two years. . (level 3b) One mother with asymptomatic CUD would be detected once every two years.

ƒ ƒ ƒ ƒ

The Clinical Evidence Base Schimmenti et al, 2007 (124)(level 3b) This is a retrospective case series of mothers identified by a low free carnitine after NBS of their infant. Six mothers were detected in 161,420 births spanning July 2004-Sept 2006 for a frequency of 1:40:000.The diagnosis was confirmed by carnitine transport (% of control) in fibroblasts. Of 6 mothers detected, 3 were without symptoms and 3 had symptoms: Case 4) fasting intolerance, fatigue, Case 5) decreased stamina, non-specific T wave abnormality, Case 6) ventricular arrhythmias, prolonged QT interval. This data supports knowledge claims that patients with CUD can be asymptomatic. Also there may be risk of illness

70

(cardiomyopathy / arrhythmia / hypoglycemic episodes / sudden unexpected death). The proportion of cases that are symptomatic but undiagnosed is unknown. Vijay et al, 2006 (125)(level 3b) Four mothers with CUD were identified after their infants were found to have low carnitine blood levels (free carnitine < 2 umol/L) through NBS program: either cord blood 2003-05 (approximately 12,000 screens) or heel prick from 2004-5 (approximately 62,000 screened). The incidence of mothers with CUD was ~ 1:15,500. All were asymptomatic with normal echocardiogram and no evidence of cardiomyopathy. The diagnosis was confirmed in fibroblasts. All were from consanguineous17 families. Wilken et al, 2001 (126) (level 3b) Wilcken et al, 2001 report one case of CUD detected through an NBS and a retrospective review of 5 clinically ascertained cases. The NBS was done on day 4 with low free carnitine confirmed on repeat. The echo of the infant detected was normal versus cardiac abnormalities with 3 of 5 cases. The authors do not discuss the analytical problems of testing or the appropriateness of the testing interval post clinical ascertainment. Boles et al, 1998 (127) (level 3b)In this retrospective analysis based on samples from a Brain and Tissue bank for Developmental Disorders for Maryland state, 27 cases representing 5 fatty acid oxidation disorders and compared the results with retrospective blinded 418 cases of sudden infant death. Following analysis there were 4 cases of CUD identified based on very low total carnitine in liver. This was not a firm diagnosis as there was no enzymatic confirmation. Cederbaum, 2003 (128) (level 3b) Cederbaum provide a follow-up report on a case previously reported as being well on treatment who suddenly collapsed and died. This was despite compliance with medication. A mild cardiomegaly had previously been reported on x-ray. This case supports the hypothesis that treatment may not prevent sudden unexpected death. It also illustrated the vulnerability of individuals and necessity for cardiac vigilance. Koizumi et al, 1999 (129)(level 3a) This population based study in 973 adult white collar workers in on prefecture of Japan estimated a carrier frequency of approximately 1% with an incidence of 1:40,000. Serum free carnitine was measured with 14 had consistently low free carnitine < 28 uM females 38 uM male. Nine subjects had mutations. Lamhonwah et al, 2002 (130)(level 3b) Lamhonwah et al, 2002 reports on 11 index cases clinically detected. There is likely to be ascertainment bias in this uncontrolled case series. The most common presentation is progressive cardiomyopathy in late infancy or early childhood. Episodic hypoglycaemia and hypoketotic coma were identified which tend to occur at a younger age. Fasting stress, infection and carnitine deficient diets were identified as exogenous factors in 17

consanguineous refers to intermarriage or mating with relatives

71

exacerbating carnitine deficiency. In addition to an L-carnitine therapy valporic acid was used for drug therapy. This paper supports the knowledge claim that genotype phenotype correlation is not likely. As well, exogenous factors are likely important. Wang et al, 2001 (131)(level 3b) Wang et al, 2001 report on 4 families with CUD. Heterogeneous mutations in the SLC22A5 gene cause primary carnitine deficiency. No correlation with clinical phenotype was found. Lund et al, 2007 (132)(level 3b) In this study in an isolated population (Faroe Island) with an increased incidence of CUD (1:15,000), 11 individuals from 5 families were analysed. Symptomatic patients responded favourably to L: Carnitine but only if treated early. Clinical manifestations were variable and the authors conclude that diagnosis following NBS with MS/MS will lead to good outcomes. Table 14: Recent primary research reviewed by BC CUD working group

Author Boles et al, 1998 (127)

Setting

Subjects

Outcomes

State of Maryland, USA with data from the statewide Brain and Tissue Bank for Developmental Disorders

27 cases with proven Support for the view that 5% of fatty acid disorders (19 sudden infant death caused by initially diagnosed with unrecognized fatty acid disorder. sudden infant death 4 cases of CUD syndrome) compared with 34 cases of accidents and abuse in negative control group (from 418 cases of sudden death in the first year of life in the dataset) Cederbaum et Los Angeles, USA 1 male previously Patient collapsed and died as he was al, 2003 (128) reported as well preparing to go to work and feeling well. He was found in asystole by paramedics and could not be revived. He was known to be ‘meticulous about taking carnitine.’ Koizumi et al Akita, Japan 14 with CUD among Estimate 1% heterozygotes in this 1999 (129) 973 ‘white collar’ region with incidence of primary workers systemic carnitine deficiency of 1 in 40,000 births. Predisposed to late onset benign cardiac hypertrophy compared with wild types. 3 deceased siblings identified as homozygous for the mutation.

72

Author

Setting

Subjects

Outcomes

Lamhonwah et Toronto, Canada al, 2002 (130)

11 cases in which CUD Clinical history characterized by was clinically potentially lethal, progressive ascertained infantile-onset cardomyopathy, weakness, and recurrent hypoglycaemic hypoketotic encephalopathy. Legge et al, Christchurch, New Case report of a 6 yr Post - mortem retrospective analysis 1985 (133) Zealand old girl with prolonged of tissue and plasma samples were illness and sudden used to diagnose CUD. Changes in death. liver, heart and brain are described in this seminal paper. Lund et al, Faroe Island 11 patients with CUD Natural history ranged from sudden 2007 (132) from 5 families death to asymptomatic individuals. All responded to L:-carnitine or biotin but only if treated early. Schimmenti et Utah NBS program, Mothers of infants 3 mothers asymptomatic al, 2007 (124) USA with low free carnitine 3 mothers symptomatic: after NBS case 4) fasting intolerance, fatigue case 5) decreased stamina, nonspecific T wave abnormality case 6: ventricular arrhythmias, prolonged QT interval (p. 442) Incidence in mothers (1 in 40,000) Vijay et al, UK Screening All without symptoms and none with 4 women identified 2006 (125) programs in through their infants cardiomyopathy. Bradford Royal with low carnitine Infirmary and West concentrations Yorkshire region identified by NBS 1st umbilical cord Feb ’03Aug ‘05 2nd – heel-prick at 7 days Mar ’04 – Aug ‘05 Wang et al, European – Swiss, Four families Four mutations illustrate 2001 (131) Iranian, heterogeneous mutations but differences in presentation were observed in children with identical mutations.

73

Author

Setting

Wilcken et al, New Wales, 2001 (126) Australia

Subjects

Outcomes

1 case detected through Two cases clinically detected would NBS compared with have been missed with previous clinically identified to threshold level. investigate the 3 of 5 among clinically detected sensitivity of MS/TS cases had cardiac involvement. A carrier rate of 1:240 was based on extrapolation. Detection of this one infant through NBS was considered useful.

Short chain acyl-CoA dehydrogenase deficiency (SCADD) Short chain acyl-CoA dehydrogenase deficiency (SCADD) is a rare disease. Only 19 cases were described in the world literature prior to newborn screening. Classic SCAD is defined as marked ethylmalonic aciduria (EMA) and methylsuccinic aciduria with marked elevation in whole blood or plasma C4 acylcarnitine and decreased SCAD activity in fibroblasts with pathogenic mutations. Less than ½ of the reported 19 cases fit these criteria. Variant SCAD (larger group) is characterized by intermittent EMA and variably decreased SCAD activity with an SCAD gene susceptibility mutation. There are 2 known variants; 625G>A and 511C>T. The SCAD variant has been described in patients who are homozygous for either the 625G>A or 511C>T and who also are heterozygous for an inactivating mutation in SCAD gene. Clinical abnormalities show marked heterogeneity, which raises concern of ascertainment bias. Elevated C4 in the acylcarnitine profile can signal either SCAD or Isobutyryl-CoA dehydrogenase (IBD) deficiency. The natural history of IBD is not well understood. In addition, it is not clear how to treat IBD. The C4 cut off levels may play a role in detecting variant forms of SCAD versus the classic forms.

Recommendation of the NSAC on SCAD The natural history of SCAD deficiency is not well understood. Although there is theoretical treatment for SCAD deficiency, there is no evidence that early detection and treatment reduces morbidity and / or mortality. The SCAD variants, which are thought to be benign, are frequently detected by some NBS programs (California) while others appear to be able to set cut offs to avoid the detection of SCAD variants. Overall there is insufficient evidence of health benefit to recommend screening for SCAD deficiency at this time.

74

Koeberl et al, 2003 (78) There has been NBS screening for SCADD since 1999 in North Carolina. Two of 3 infants diagnosed with classic SCADD remained asymptomatic on treatment. The 3rd presented with seizures and cerebral infarct at 10 weeks of age. All 3 have inactivating mutations and highly elevated plasma C4. The 4th case has IBCD deficiency. The authors make the following valid point in their discussion. The natural history of patients with SCADD variants is uncertain. Detection of the variant can be avoided by setting the C4 threshold high enough. In North Carolina, no SCAD variants were detected using a cutoff value of the mean + 4 SD. The claim that classic SCADD is a clinically well-described entity is disputable. The author cites 3 references to support this and claims that there is effective treatment for SCADD deficiency. However, the 3 cited references do not contain any information about effectiveness of treatment for SCAD deficiency. Bok et al, 2003 (134) Eight members of a family were investigated following diagnosis of a SCAD infant at 2 months to a mother who had HELLP syndrome during pregnancy. The index patient was premature at 29 weeks and presented with heptomegaly and jaundice. The mother was heterozygous for 1138C>T. The index child had 1138C>T (homozygous). The authors conclude that there is a link between classic SCADD and HELLP. The father also had slight increase EMA, slightly high C4 yet “deficient” SCAD in fibroblasts and a variant genotype; 1138C>T and 625C>A. At age 3 the boy is growing along the 3rd percentile with normal mental development, liver function, and slight hypotonia. Treatment was not described. Corydon et al, 2001(135) Corydon et al, 2001 report the California experience with 755,698 births in newborn screening pilot program. There were 18 SCADD cases detected for an incidence of 1 per 20,000 (positive predictive value of 22%). No clinical information was provided. Sass et al, 2004 (136) Sass et al, 2004 report on 2 children with Isobutyryl-CoA dehydrogenase deficiency -- a defect in the valine catabolic pathway. Both were detected through NBS – one German and one Turkish. One boy is normal. An 8-month-old girl has muscular hypotonia and slight developmental delay. It is not yet certain if IBD causes significant morbidity in affected children.

75

Table 15: Recent primary research reviewed by BC SCADD working group

Author

Setting

Koeberl et al, 2003 (78)

North Carolina, USA

Bok et al, 2003 (134)

Netherlands & Denmark

Subjects

Outcomes

3 infants with SCAD and All three infants had putatively one with isobutyryl-CoA inactivating mutations in both alleles of dehydogenase deficiency the SCAD gene were detected on the • It is possible to ascertain at least some basis of elevated C4SCAD variants by newborn screening if carnitine by state NBS the cutoff for C4-carnitine is set too low program • The plasma C4-carnitine was • significantly elevated (p < 0.001) by greater than 14-fold in the three infants with classic SCAD deficiency compared with the variant SCAD patients, for whom plasma C4-carnitine concentrations were within normal range • A family with SCAD • Homozygosity for an inactivating SCAD deficiency mutation does not necessarily result in disease • Two siblings had a homozygous inactivating 1138C>T mutation, whereas the father was compound heterozygous for this mutation and the common 625G>A polymorphism. There was a good correlation between the type of SCAD mutation, the residual SCAD enzyme activity, and the levels of urinary ethylmalonic acid and plasma C4carnitine in each of the eight family members • None of the family members had hypotonia, developmental delay, or episodes of ketotic hypoglycemia

76

3. DISORDERS OF AMINO ACIDS AND UREA CYCLE Citrullinemia (CIT) Citrullinemia (CIT) is caused by one of 2 enzyme defects in the urea cycle which converts the toxic metabolite ammonia to the non-toxic metabolite, urea. About 1/3rd of patients with CIT present acutely ill within the first week of life while the other 2/3rd present later with chronic and persistent symptoms.(137) It is the late-onset chronic cases that are most likely to benefit from NBS as there is an asymptomatic phase that provides the necessary time for newborn screening to be completed and pre-symptomatic treatment initiated to prevent adverse health outcomes.(138) CIT is usually fatal in the approximately 1/3 of cases that present neonatally during the first week of life unless there is immediate medical attention. Neonatal-onset CIT presents with lethargy progressing rapidly to coma as a result of high ammonia and high citrulline levels. The clinical outcome and particularly the neurological outcome depend on the degree and duration of hyperammonemia prior to stabilization with effective treatment.(139) While early intervention is critical, the condition has usually already become clinically evident by the time newborn screening results are available and so the window of opportunity for NBS to provide benefit is short.(140) The defect in the urea cycle pathway that causes CIT is most commonly traceable to argininosuccinate synthetase (ASS),[ EC 6.3.4.5.] deficiency (MIM 2215700) and as described above, it is the neonatal form of ASS deficiency that is usually fatal. Citrulline can be detected simultaneously along with other analytes by TMS. Citrulline is reasonably sensitive and highly specific for the detection of ASS deficiency, as well as for the less common urea cycle disorder known as argininosuccinate lyase (ASL) deficiency. The latter defect has a similar clinical and biochemical phenotype to ASS deficiency.(141) The two disorders are heterogeneous at the genetic level, with over 50 mutations reported for the neonatal and late-onset forms.(138) There is a broad clinical spectrum in patients with varying degrees of residual enzyme activity. At least some late onset cases could be very mild forms of citrullinemia that may not require treatment, at least under stable conditions.

BC Incidence Incidence of ASS deficiency and related conditions is approximately 1:100,000 live births. A study of 610,000 blood samples in Germany revealed 4 cases of classical neonatal onset citrullinemia, or an incidence of 1:152,500 live births.(140) Including mild and/or persistent cases, combined incidence of the two underlying disorders may approach 1:50,000, or about two cases every five years in BC.(139)

77

Knowledge Claims Formulated by BC Working Group on CIT What is the evidence that newborns with CIT C will have improved health outcomes with early detection following newborn screening with TMS? If BC commences screening for CIT, the evidence reviewed in this HTA indicates that: • 2 newborns will be detected with some variant of CIT every 5 years approximately. (level 3b) Earlier detection will lead to the following benefits for newborns. • •

1 newborn every 14 years or more may have their risk of premature death or neurological deficit reduced if screening test results precede clinical manifestation of acute neonatal onset (level 3b). 1.3 newborns every 5 years with a late-onset form of CIT are expected to have an improved health outcome with early diagnosis and appropriate treatment (level 3b)

The Clinical Evidence Base Sander et al, 2003 (140)(level 3b) This is a retrospective review of citrullinemia cases detected by the NBS program in Germany since 1999. In total 610,000 neonates were screened, 17 samples had citrulline levels above 1.5 mg/dL or 85.5 umol/L. Three cases also had elevated arginosuccinate and were confirmed later to have ASL deficiency. These 3 cases are not described further in the paper. In follow up, there was a persistent increase in citrulline in 12 cases (7 / 12 Turkish) plus the 3 ASL deficient cases (total n=15). Ten of these cases had molecular confirmation. o 4 presented with typical symptoms of neonatal onset citrullinemia type I during 1st week of life o 2 / 4 already had a diagnosis by the time NBS result came in. Both of these infants were in a coma. o 2 / 4 had a normal neurological outcome at time of reporting (one NBS detected and one clinically ascertained). The remaining 8 cases were well at diagnosis and upon follow up. 5/8 have had Larginine administration, reduced protein intake and have been clinically stable. Patient 6 with persistent citrullinemia has had no treatment at all. Patient 7 had protein restriction for 12 weeks but has received unrestricted diet since and has developed normally despite persistent high Cit levels. Patient 9 received breast milk and no meds. In summary, NBS for citrullinemia may prevent severe clinical symptoms and sequelae in some patients. It is uncertain whether NBS results can be made available to provide for earlier detection in the 1/3 of patients that present neonatally. Most of the health benefit would appear to be for the later onset cases. However, at least some of these late onset 78

cases could be very mild forms of citrullinemia that may not require treatment at least under stable conditions. This paper also includes a review and discussion of the natural history of citrullinemia. At least compared to later-onset clinically ascertained cases, NBS appears to offer an opportunity to institute earlier treatment and may prevent long-term neurological sequaele related to persistent hyperammonemia. Nassogne et al, 2005 (138) (level 3c) Nassogne et al, 2005 conducted a retrospective review of 217 patients with urea cycle defects including 121 with neonatal forms and 96 with later onset forms over a period of 25+ years (1972 – 2000). There are 33 ASS and 20 ASL deficient cases reported. Changes in clinical care over time are not accounted for in this paper. A high mortality rate (60%) was the main finding with poor neurological outcome in neonatal onset cases. The neurological outcome appears to be more favourable in ASS compared to ASL deficiency. Among the 33 patients with ASS defects, 26 presented neonatally versus 7 with late onset. Of the 7 with late onset further to ASS defects, 3 presented with acute neurological signs, 2 with psychomotor retardation, 2 with chronic recurrent neurological signs. Among the 20 with ASL defects, 14 presented neonatally versus 6 with late onset. Of the late onset with ASL defects, 3 presented with acute and 3 with chronic recurrent neurological signs. Choi et al, 2006 (142) (level 3b) Choi et al, 2006 report on instances of metabolic stroke in 2 infants with citrullinemia; 1 at 53 days and the other at 3 months. Although the author note that CIT should be included in the differential diagnosis of stroke, another implication is that these strokes may have been prevented if they had been detected with NBS because the occurred at 3 months of age in one and 53 days of life in the other thereby implying that they were potentially preventable. Wilcken, 2004 (143)(level 4) This review and editorial on mild ASL/ mild citrullinemia type I and II includes the case report of a boy with mild ASL who was quite well when first seen at 14 days of age but had elevated glutamine 1313 umol/L and ammonia of 167. This level of chronic hyperammonemia could cause intellectual deficit however with appropriate treatment this boy has normal development at age 4 thus supporting the case for early detection and treatment before adverse events occur in well children. Smith et al, 2005 (144) (3 case reports) Smith et al, 2005 include in a paper on urea cycle disorders, the case report of a child with ASL presenting at 3 years and 9 months with trichorrhexis nodosa and alopecia, seizures, delayed gross motor skills. Following diagnosis of ASL deficiency and

79

treatment, his clinical condition improved thereby lending support for early detection and treatment. Table 16: Recent primary research reviewed by BC CIT working group

Author Sander et al, 2003 (140)

Setting Germany

Nassogne et al, Necker-Enfants 2005 (138) Malades Hospital, France

Subjects

Outcomes

• 17 samples exceeding the level of 1.5 • 610,000 blood samples mg/dl; hypercitrullinaemia was identified were analysed in one in 15 newborns screening centre for • Follow-up studies revealed persistent citrulline as a pilot study hypercitrullinaemia in 12 of the for neonatal screening newborns (3 cases were using tandem mass argininosuccinate lyase deficiency) spectrometry (01/03/99 to • Classical citrullinaemia has an 30/06/02) incidence of 1:152,500; including mild cases, the calculated incidence was 1:50,800 • Cases with a urea cycle defect diagnosed between 1972 and 2000 with files available for chart review

• 217 patients with urea cycle defects were included; 121 (55%) presented with a neonatal-onset form (during the first 28 days of life) and 96 (45%) had a latepresenting form. An OTC defect was the most common (69%) followed by the ASS deficiencies (15%), the ASL deficiencies (9%), and finally CPS deficiency (6.5%) • Early-onset defects have a high mortality rate of 84% for all the defects combined; • Long-term neurological outcome or these pts nevertheless remains poor; ASS deficiency, neurological outcome appears more favourable and pts with neonatal onset of ASL deficiency show a poor neurological outcome, especially with mental retardation • 96 pts had late-onset symptomatology; in 55 pts, the diagnosis was finally made after a long history of chronic symptoms; in 41 pts, the diagnosis was made from the onset of symptoms; 45% of boys with late-onset OTC deficiency died at the time of diagnosis. • Long-term neurological outcome shows that about 43% of patients have moderate to severe neurological sequelae • Of the patients still alive, 35% have moderate neurological damage • Patients with late-onset forms may present at any age and carry a 28% mortality rate and a subsequent risk of subsequent disabilities

80

Author Choi et al, 2006 (142)

Wilcken et al, 2003 (143)

Smith et al, 2005 (144)

Setting Seoul, Korea

Subjects

Outcomes

2 infant females, Age: 3 months and 53 days presenting with ‘metabolic stroke’

• Pt 1 was drowsy and had increased muscle tone and hyperreflexia in both right upper and lower extremities. The plasma ammonia concentration was greater than 400 μmol/L. • Patient 2 was brought to the emergency room at the age of 53 days because of generalized tonic seizures and impaired consciousness. Her height and weight were both below the 3rd centile. She was mentally comatose and had increased muscle tone and hyperreflexia in both extremities. The plasma concentration of ammonia on admission was 535 μmol/L Serum amino acid analyses revealed a markedly elevated citrulline concentration at 1376.0 nmol/ml, with normal concentrations of glutamine and arginine. • Urea cycle disorders should be included in the differential diagnosis of unexplained stroke during infancy and childhood Open discussion N/A RECOMMENDATIONS on how to during a session of a Newborn screening by manage urea cycle disorders: symposium held in tandem mass • Prompt referral should take Sydney in 2003 spectrometry makes precedence over everything else entitled ‘‘New possible the diagnosis of • Acute hospital management: (1) Developments in Urea UCDs (citrullinemia) in measure the plasma ammonia level; (2) Cycle Disorders’’ neonates start therapy with intravenous ammonia scavenging drugs; and (3) make preparation for dialysis • Nutrition Management: (1) a reduced protein intake, (2) giving essential amino acid supplements, (3) supplementation with minerals, vitamins, and trace elements, and (4) monitoring growth, hair, skin, nails, and biochemical indices of nutrition Uncontrolled case 3 pts (8 month girl, 10 • Dietary protein restriction improves pt study; US joint study month boy, 3.75 years outcomes old boy) • Although most commonly associated with infancy, UCDs present outside the neonatal period, frequently in childhood, in the majority of patients

81

Maple syrup urine disease (MSUD) Maple syrup urine disease (MSUD) is caused by a deficiency of the enzyme, branch chain alpha-keto acid dehydrogenase (BCKD), which is in the catabolic pathway of 3 branch chain amino acids (BCCA’s); leucine, valine and isoleucine. The accumulation of these 3

BCAAs is neurotoxic at high concentrations. The “classic” form presents earliest and with the most severe symptoms and consequences, including mental retardation and neurological disability.(145-147) The disorder can remain asymptomatic during infancy or even early childhood; however, stressors such as infection and or periods of fasting can provoke metabolic intoxication and rapid cerebral deterioration at any time.(148) Detection and diagnosis of classic MSUD, followed by timely treatment, is considered critical for avoiding permanent brain damage, cognitive impairment, and even death.(149-153) The efficacy of branched chain amino acid (BCAA) restricted dietary regimen for this disease is well established. Delayed diagnosis, prolonged elevation of BCAAs and the resultant extended periods of encephalopathy lead to poor neurological outcomes. Other effective therapies include thiamine and other forms of supplementation as well as liver transplantation. Test Performance Newborn screening for MSUD relies on the identification of elevated Leucine & Isoleucine (Leu/Ile) in dried blood spots. However, tandem mass spectrometry can not differentiate between the isobaric amino acids Leucine, Isoleucine, allo-isoleucine (alloile) and hydroxyproline leading to a high false positive rate and a positive predictive value of ~5%.(154) False negatives have also been reported due to very subtle Leu/Ile elevations.(155) A second-tier test for MSUD has been developed using a column chromatography approach to separate appropriate isobaric amino acids (Leu, Ile, allo-ile, OHPro) and allow for quantitation of allo-ile, a much more specific marker for MSUD. (156) Obelsbee et al, 2008 data is from known and control cases therefore representing level II evidence. While no cases of MSUD have been identified by this approach, the false positive rate was dropped to zero while employing the second tier test on 0.1% of newborn screens. In BC, approximately 42 second-tier tests are expected per year (less than 1 sample per week).

BC Incidence Great variation in incidence is seen, with the highest occurrence found in groups that tend to practice consanguineous marriage; global incidence is about 1:185,000 live births, with a somewhat lower rate among Caucasians of European descent.(145) This equates to a single case in BC every 7 to 10 years; real world data over almost 30 years confirms this estimate, as only two patients with MSUD were diagnosed in that period.(32) The benefits of mass screening may be greater in groups with higher incidence (e.g., certain 82

Arab or Mennonite populations); contemporary migration may increase the concern in regions where cases have been very rare in the past.(157)

Knowledge Claims Formulated by BC Working Group on MSUD What is the evidence that newborns with MSUD will have improved health outcomes with early detection following newborn screening with TMS? If BC commences screening for MSUD with second tier testing using a column chromatography approach and second tier testing, the evidence reviewed in this HTA indicates that: • •

1 newborn will be detected with MSUD every 7 to 10 years approximately (level 3c) 42 false positives per year would require second tier retesting (level II) .

Earlier detection will lead to the following benefits for newborns. ƒ

Normal cognitive development/prevention of mental retardation with optimal outcomes predicted with detection within the first week of life (level 3c).

The Clinical Evidence Base Yoshino et al, 1999 (158) (level 2b) Yoshino et al., 1999 collected data for this controlled case series by sending a questionnaire to each pediatrician treating any of 42 MSUD patients enrolled in a specialized Japanese dataset. The response rate was 33% or 13 patients covering 14 episodes of metabolic decompensation. Nine of these episodes of metabolic decompensations developed neonatally. In the other 5 episodes of metabolic decompensation onset was delayed until infancy or later. In neonatal cases, plasma leucine > 40 m/100mL or duration of altered alertness > 10 days lead to poor neurological outcome. Episodes of metabolic decompensation in infancy or after did not affect or only minimally affected the neurological outcome. The recommended therapeutic goals to improve neurological outcome are therefore to shorten the duration of the altered level of consciousness and to minimize the peak plasma leucine level as much as possible Hilliges et al, 1993 (159) (level 2c) Hilliges et al, 1993 compared the intellectual performance of kids with MSUD to matched controls with PKU treated early or normal children.(159) In this controlled case series 27 of the children were known to have MSUD in Germany from 1988: 22 were tested, 16 had the classical form, 3 were severe variants and 3 mild cases. One examiner performed psychometric tests. The mean IQ was 74, range 50 – 103. The authors concluded that the length of time after birth that plasma leucine concentration remained > 1 mmol/L and quality of long term metabolic control were important influences on IQ.

83

Le Roux et al, 2006 (151) (level 3b) Le Roux et al, 2006 examined the neuropsychometric outcome predictors for adults with MSUD by assessing 14 adults with MSUD and correlating these with retrospective data from their clinical history. The sample included 10 female and 4 male adults (25 years of age, range 18 – 43). Three of the 14 had intermediate MSUD. One examiner did all IQ assessments. The mean IQ was 76 (range 54-113) and verbal IQ 82 range 55-110. Mean age at diagnosis was 10 days ( 5-21 days) in the classical group and 3.2 years (range 2 – 4.4) in intermediate group. There was no correlation between IQ and mean leucine. IQ tended to be higher in those with an earlier diagnosis in those with classical MSUD. These findings led the authors to conclude that: Age at diagnosis appears to be a predictor of IQ even in adults while long term blood leucine does not correlate with IQ. (151) p 201. Simon et al, 2006 (150)(level 3a) In this case control series Simon et al, 2006 compared 10 infants with classical MSUD detected by TMS based NBS to the ‘youngest’ available 10 clinically diagnosed patents. Cases were screened between 10 infants with classical MSUD detected by screening 1999 to 2005. Infants with MSUD were ascertained by ‘regular’ inquiries at screening labs and pediatric departments that specialized in metabolic diseases in Germany and Austria. Details on lab data clinical course and management were entered into questionnaires that were completed by the attending physicians or were extracted from the patients’ medical files. It was found that MSUD diagnosis in NBS screened kids was confirmed earlier than in patients diagnosed on clinical grounds (7 days versus 12 days). A positive correlation was found between the age at diagnosis and the plasma leucine concentration at start of treatment. In the screened group, 1-screened patient presented with impaired consciousness, 2 patients had experienced seizures, 5 patients experienced mild encephalopathic symptoms and 2 were completely asymptomatic. By contrast, consciousness was impaired in all clinically diagnosed patients and 3 of them had seizures. Morton et al, 2002 (148) (level 3a) In this select cohort, 18 neonates with MSUD were identified between 12 and 24 hours of age. All were at risk newborns with older siblings or at high risk given their (Mennonite) ethnic group. The Leu : alanine ratio was used as a screen. None of the infants identified before 3 days of age and treated, became ill during neonatal period. Sixteen of 18 were managed without hospitalization. At follow up, despite infection-triggered metabolic decompensation, developmental outcomes were uniformly good. Four developed life threatening cerebral edema but all recovered with treatment

84

The authors conclude that classical MSUD can be managed to allow a benign neonatal course, normal growth and development and low hospitalization rates however metabolic decompensation can occur and requires prompt treatment.(148) Zytkovicz et al, 2001 (154) (level 2b) In this cohort analysis of NBS of 160,000 newborns, NBS for MSUD was achieved by measuring Leu / Isoleu signal. A positive predictive value (PPV) of 5% was achieved while the Leu / phe ratio had a PPV of 12%. Table 17: Recent primary research reviewed by BC MCUD working group

Author

Setting

Yoshino et al, 1999 (158)

Controlled case study Japan

Hilliges et al,(159) 1993

Controlled case study; Germany

Subjects

Outcomes

• Responses received on 14 of 42 patients (33.3%)(3 males, 11 females)

• 9 of the 14 pts developed metabolic decompensation in the neonatal period. • In the other 5, onset of disease was delayed until infancy or later • Neonatal onset group, a pretreatment level of plasma leucine > 40mg/100mL or a duration of altered level of alertness longer than 10 days was associated with poor neurological outcome • Therapeutic measures implied included: IV glucose, electrolyte solution, exchange transfusion, peritoneal dialysis, thiamine • 5 of the 9 pts developed neurological sequelae • Episodes of metabolic decompensation in infancy or after did not affect or minimally affected neurological outcome • Of the 27 known • The mean IQ score of the MSUD MSUD pts (16 subjects was far below the lower limit of classical, 3 severe normal intelligence whereas that of the variant and 3 mild PKU group was in the normal range variant forms) were • Only 6 children with MSUD, 3 with tested classical, 1 with a severe variant, and 2 • 2 control groups with mild variant forms had normal IQs matched by age, sex, • The postnatal time period during which nationality, and socio- plasma leucine levels remained > 1 economic status were mmol/1 ranged from 1 to 67 days and tested; 1 group correlated inversely with the IQ (r=0.57; P consisted of normal = 0.019 for all MSUD pts, and r = 0.51; P = children, the other one 0.049 for classic MSUD) of early treated • Neonatal screening on the 5th day of life, phenylketonuria as is done in Germany, cannot be helpful (PKU) patients in identifying classical MSUD patients in the presymptomatic or early symptomatic state • To improve the prognosis of the intellectual outcome, the metabolic disorder has to be identified at the very beginning of the neurological deterioration based solely on clinical signs

85

Homocysteinuria (Hcy) Homocysteinuria (Hcy) is an autosomal recessive disorder caused by a deficiency of the enzyme cystathionine beta-synthase ([EC4.2.1.22). The amino acid, homocysteine, accumulates in blood and body tissues and leads to a multitude of complications. The natural history (n=629) of the condition was described by Mudd and colleagues in 1985. (160) Individuals with Hcy appear normal at birth. If untreated, patients may present with dislocated lenses, thromboembolic events, mental retardation, scoliosis and osteoporosis. Approximately 50% of patients respond to pyridoxine (B6) therapy alone with normalization or significant improvement in homocysteine levels. Other therapeutic strategies are required for the non-pyridoxine form of the disease. If left untreated, most non-pyridoxine responsive patients, will have dislocation of the lens by age 10 and clinically detected thromboembolic events (stroke) by age 15. Without treatment, approximately ¼ will not survive to age 30 (Mudd et al 1985).(160) Mental retardation is frequent in clinically ascertained cases; only ~ 22% of B6-responders had IQ of 90 or above compared to 4% of B6-non-responders.

Test Performance Hcy is traditionally screened using methionine as a primary marker. Cut-offs set to maximize the sensitivity of this test lead to elevated false-positive rates and a low positive predictive value of 2%.(161) Second tier testing has been developed for Hcy, measuring homocysteine directly in dried bloodspots as part of a panel of second tier analytes. (The 2nd tier test also includes methylmalonic acid and methylcitric as a follow-up for C3 elevations).(4) The methodology requires a second, larger punch from the original newborn card, which is extracted and separated using column chromatography prior to MS/MS detection. The protocol defines a conservative Methionine cutoff to maximize sensitivity, followed by a weekly second tier testing on all samples above the cutoff. Using the 2-tier approach a positive predictive value of 100% has been obtained with only 0.3% of newborn screens being selected for second tier testing.(4) This represent level 1 evidence of test performance. In BC this would represent approximately 150 second tier tests annually (approx. 3 per week) based on 42,000 births per year. Knowledge Claims Formulated by BC Working Group on Hcy What is the evidence that newborns with Hcy will have improved health outcomes with early detection following newborn screening with TMS? If BC commences screening for Hcy using the second tier approach, sensitivity and specificity are expected to approach 100%. If this is achieved the evidence reviewed in this HTA indicates that:

86

• 1 newborn will be detected with Hcy every 3 – 4 years approximately (Matern et al, 2007) • With second tier testing applied to an extra 150 tests per year, false positives will be negligible Earlier detection will lead to the following benefits for newborns. ƒ Normal vision/ prevention of dislocated lenses ƒ Normal or near normal neurological development/prevention of stroke and other thromboembolic complications ƒ Normal bone development/prevention of scoliosis and osteoporosis ƒ Normal or near normal cognitive development/prevention of mental retardation (level 2). ƒ Increased longevity/ prevention of death by age 30

The Clinical Evidence Base Yap et al, 2001 (162) (level 2) The controlled case series with historical controls reported by Yap et al, 2001 provides the best currently available evidence of a reduction in vascular complications with early detection of Hcy. Data on 158 Hcy cases from Ireland, Australia, Netherlands and the United Kingdom were compared with historical controls from a study of 629 untreated patients. Three different treatment regimes were used by all centers: pyridoxine +/- folate, methionine restriction and betaine. The number of vascular events in treated patients was significantly reduced (P< 0.0001). Without treatment 112 vascular events would have been expected. Only 17 vascular events occurred in 12 patients (relative risk of 0.09). Treatment regimens lower plasma homocysteine significantly and reduce cardiovascular risk in Hcy. Yap et al, 1998 (level 2) (163) Yap et al, 1998 report a controlled case series with historical controls. Newborn screening started in 1971 in Ireland where there is a higher incidence of Hcy (1 in 65,000). Over 25 years, 1.6 million newborns were screened and 24 pyridoxine nonresponsive cases detected. Dietary treatment was started by 6 weeks of age. There were no thromboembolic events in any of the 25 screened cases. Myopia occurred in cases with higher homocysteine levels. Two cases that were not compliant with dietary recommendations had complications of ectopia lentis. There was one case of osteoporosis and 2 with mental handicap. Of 4 cases missed on screening, 3 presented with ectopia lentis after age 2. This study provides evidence in support of the knowledge claim that Hcy (Pyridoxine non-responsive forms) can be detected by the measurement of methionine in blood spots. Pyridoxine responsive forms of the disease will be missed using this approach. Early detection of Hcy and reduction in homocysteine with dietary modification and medications reduces the risk of thromboembolic events. (level 3)

87

Sokolova et al, 2001 (164) Sokolova et al, 2001 present data from the European screening program. The authors present an analysis suggesting that 75% of homocystineuric patients are currently missed with biochemical screening approaches. Global incidence is reported to range from 1:58,000 to 1:1,000,000. The carrier frequency in Danish newborns for c.833T>C is 1.4% with a predicted incidence of 1:20,500.(165) c.833T>C is neutralized when it occurs on a hapotype with a nearby insertion of 68 bp. Data reported is from a “selective biochemical screening program of 4.8 million newborns”. The incidence was compared to the square of the estimated population frequency of pathogenic alleles. The estimated incidence of 1:83:000 is about 4 times higher than the incidence observed by biochemical screening. This suggests that up to ¾ of homocystineuric patients are being currently missed. Therefore, in Central Europe homocysteinuria due to CBS deficiency may be largely unrecognized. Chace et al, 1996 (166)(level 3b) A 1996 paper by Chace et al, reports that using Met / Leu plus Ile ratios by MS/MS increases specificity. It is not possible, however, to derive sensitivity, specificity and the positive predictive value from the data presented in this analytical paper. Chien et al, 2005 (167) In Taiwan over the last more than 20 years methionine levels (cut off 134 umol/L) have been measured in dried blood spots as part of a national newborn screening program. In 1.7 million newborns there were 17 cases of hypermethioninemia with only 1 case of homocysteinuria. In more than half of the 16 cases of isolated hypermethioninemia had the MAT1A gene defect causing methionine adenosyltransferase deficiency. Moat et al, 2004 (168) Moat et al, 2004 report on the spectrum of mutations in the CBS gene in clinically ascertained cases (536 patient alleles with 130 different mutations). In UK, and US 29% of patients had the I278T mutation. The incidence of G307S was found to be 21% in the UK and 8% in the US. Perschmitt, Simmons & Levy, 1999 (161) (level 3) This study provides support for the knowledge claim that homocystinuria (pyridoxine non-responsive form) can be detected by the measurement of methionine in blood spots. Pyridoxine responsive forms of the disease will be missed. Presented is 32 years of screening in New England by bacterial inhibition assay (pre TMS). For the first 23.5 years, the cut off for blood methionine was 134 umol/L for 2.2 million infants resulting in 8 cases detected for an incidence of 1:275,000. In 1990, the cut off changed to 67 umol/L. Among 1.1 million infants screened, 7 were diagnosed for an incidence of 1:157,000. Only one B12 responsive case was detected (by lower cut off). This case was detected by routine collation at 5 weeks of age. The false positive rate went from 0.006 to 0.03 percent (40 / 135,000), which appears to be within an acceptable range. 88

Refsum et al, 2004 (169) In 1999, genotyping was performed on 1133 random samples from the Norwegian newborn screening program. Six different mutations among Norwegian families with CBS deficiency were studied. The prevalence was calculated based on the assumption of Hardy-Weinberg equilibrium. CBS heterozygosity for one of the 6 mutations was 2.5% for an estimated incidence of 1:6400. Pyridoxine responsive phenotype was found in 90% of cases. Table 18: Recent primary research reviewed by BC Hcy working group

Author

Setting

Sokolova 2001 Czech and Slovak republics

Subjects

Outcomes

Prevalence compared in The authors conclude that sensitive NBS screened populations by be considered for Hcy because it is two methods sufficiently common. Analytical paper reporting (verbatim conclusions) the ratio of a new method for methionine to leucine plus isoleucine diagnosis of determined by tandem mass homocystinuria from spectrometry will successfully detect dried blood spots. hypermethioninemias with very low rates for false positives and false negatives.

Chace et al, 1996(166)

Pennsylvania, USA

Chien et al, 2005 (167) Febriani et al, 2004 (170)

Taiwan

1.7 million newborns

Hiroshima, Japan

Analytic paper

Levy and Albers, 2000 (171)

Harvard Medical School, Boston, Massachusetts, USA

Review of current status Discussion of the implications of “tandem mass spectrometry (MS-MS) and the introduction of molecular (DNA) testing to increase the specificity of testing for several disorders, thereby reducing false-positive rates.” 14 unrelated UK and 38 The authors conclude that the spectrum unrelated US cases of mutations observed in the UK and US more similar to northern Europe than Ireland. Review of 32 years of A cutoff of 1mg/dl for blood methionine newborn screening with was recommended to identify infants methionine who have only slightly elevated concentrations and to reduce false negatives 4,992 samples from Total homocysteine is a common but not routine NBS specific finding in newborns. Screening should be further evaluated in regions with high prevalence.

Moat et al, 2004 Cardiff, UK (168) Peterschmitt 1999 (161)

New England, USA

Refsum et al, 2004(172)

Oxford, UK

17 cases of hpermethioninemia with only 1 case of homocysteinuria High performance liquid chromatography (HPLC) method with fluorescence detection to measure total homocysteine (tHcy) in DBS might be useful for newborn screening

89

Author

Setting

Yap et al, 1998 Dublin, Ireland (163) Yap et al, 2001 Dublin, Ireland (173)

Subjects

Outcomes

Reporting on 25 years of Of four cases missed on screening, 3 newborn screening for presented with ectopia lentis after age 2. Hcy There were no thromboembolic events in 25 cases detected through screening. 23 pyridoxine-non Data suggests that early treatment with responsive individuals good biochemical control prevents assessed for mental mental retardation and thromboembolic functioning events

Tyrosinemia type I (TyrI) The tyrosinemias are inborn errors of metabolism that results from deficiencies in the enzymes that break down the amino acid tyrosine. (79) Hereditary tyrosinemia type I (TyrI) is an autosomal recessive disorder caused by a deficiency of fumarylacetoacetase (FAH); the last enzyme in the tyrosine catabolic pathway. As a result of the block in the pathway, highly reactive compounds (maleylacetoacetate and fumarylacetoacetate) accumulate and are thought to be responsible for the liver and kidney injury characteristic of this disorder. The accumulated fumarylacetoacetate is converted into succinylacetone (SA), which is readily detected in blood and urine. As will be seen below in the review, it is this marker, SA, that has been recently shown to be highly sensitive and specific for the purposes of newborn screening. Clinically tyrosinemia is characterized by progressive liver damage with liver failure, a high risk of hepatocellular carcinoma and renal tubular dysfunction, which often results in rickets. Some patients have episodes of neurological crisis causes by a secondary block in the porphyrin pathway.(79) TyrI varies in severity.(79) In the acute form (infantile or early onset) failure to thrive, ascites, edema, bleeding diathesis, hypoglycemia, vomiting, diarrhea occur before 6 months of age. Without adequate treatment, progressive liver failures, severe coagulopathy, renal tubular defects and death occur by 6-8 months of age. In milder later onset forms of the disorder, symptoms usually start after 6 months of age and are associated with nodular liver cirrhosis, renal tubular nephropathy, and rickets leading to growth failure. A major advance occurred in the treatment of tyrosinemia with the discovery of the compound nitisinone or NTBC which blocks the tyrosine pathway at the 2nd step and thereby prevents the formation of the toxic metabolites lower down in the pathway. NTBC has proved useful in improving liver disease. However, liver transplantation remains an important treatment modality for some patients.

90

Test Performance Elevated bloodspot tyrosine in the newborn period is nonspecific. It may indicate tyrosinemia Type I, II, or III and also infants with transient tyrosinemia of the newborn due to liver immaturity. Accordingly, screening for tyrosinemia type I using only tyrosine as a marker can lead to significant false positive rates. False negatives have also been identified, suggesting decreased sensitivity of tyrosine as a marker.(82) Second tier testing using bloodspot succinylacetone (a specific marker of TyrI) has significantly increased specificity of tyrosinemia type I screening (100%) but the test remains dependent on a primary tyrosine cutoff and concerns over sensitivity still exist.(174) Recently, a methodology has been developed to integrate succinylacetone as a primary screen for tryrosinemia. The residual dried blood spot, following methanol extraction of the acylcarnitines and other amino acids is extracted using a water:acetonitrile mixture including hydrazine. The hydrazine cyclizes and stabilizes the succinylacetone extract which is combined with the original acylcarnitine/amino acid mixture for analysis as a single MS/MS injection (Sander et al. 2006). This approach has proven to be 100% sensitive and specific so far (61,000 samples), leading to a positive predictive value (PPV) of 100%.(175) Level I evidence on test performance data is provided by Turgeon et al, 2008 with validation from retrospective samples and prospective screening as well as the prospective study of La Marca et al, 2008.(176,177)

Knowledge Claims Formulated by BC Working Group on Tyrosinemia What is the evidence that newborns with Hereditary Tyrosinemia type I (TyrI) will have improved health outcomes with early detection following newborn screening with TMS? If BC commences screening for TyrI using measurement of succinylacetone as a primary screening approach, a sensitivity approaching 100% and PPV of 100% may be achievable. The evidence reviewed in this HTA indicates that: • 1 newborn will be detected with some variant of TyrI every 3 years approximately. (level 3a) (32) Earlier detection will lead to the following benefits for newborns. • Early treatment with nitisinone may prevent the development of hepatocellular carcinoma (cancer) for most though 10% will not respond to treatment (level 3a) • Death through liver or renal failure may be averted with treatment for most (level 3b) • Liver transplantation may provide enhanced quality of life for non responders (level 3b)

91

The Clinical Evidence Base Applegarth et al(32) 2000 (level 3a) This retrospective analysis based on the records of the Biochemical Diseases Laboratory at Women’ and Children’s hospital in BC report 8 cases of TyrI every 25 years for a detection rate of approximately 1 case every 3 years given the current annual birth rate of about 42,000. Ashorn et al, 2006(178) (level 4) The 2006 review by Ashorn et al., outlines the natural history and treatment of TyrI. Though no primary evidence is provided in the paper, it supports the knowledge claim that TyrI is a treatable metabolic disorder and that early detection provides an opportunity to treat at an earlier stage of liver disease. Although TyrI is rare, it is the most common of the 3 known diseases caused by defects in tyrosine metabolism. Ashorn et al report that TyrI has an incidence of 1/100,00 – 1/200,00 with an incidence as high as 1/2000 in populations of Northeastern Quebec. Pharmacologic treatment with nitisinone (NTBC), an inhibitor of tyrosine catabolic pathway improves liver function. However longer follow up periods are needed to establish the role of this drug in protecting patients from end stage organ involvement and hepatocellular carcinoma. Early Liver transplantation reduces mortality. It is considered to avoid the development of liver malignancy. Mohan et al, 1999 (179) (level 3b) The retrospective analysis of Mohan et al, 1999 is used to support the knowledge claim that orthotopic liver transplantation (OLT) corrects the TyrI metabolic disorder. The report on 17 TyrI cases presenting between 1989 and 1997 including 7 cases representing clinical management in the pre NTBC period and 10 post NTBC since 1992. Two of the 10 on NTBC required liver transplantation. The authors support the following indications for OLT: poor response to NTBC and hepatic dysplasia (risk of malignancy). As well they report an incidence of 1:100,000 to 1;120,000. The authors discuss the ‘founder effect’ of TYR in Quebec and Finland, X262X. Sander et al, 2006 (175)(level 3a) This analysis based on a newborn screening cohort supports the knowledge claim that direct measurement of succinylacetone (SA) in blood spots is a highly specific and sensitive method for the detection of Tyrosinemia. SA was extracted from blood spots. Of 61,344 samples 99.6% had < 5 umol/L SA concentrations. Using a cut off of 10 umol/L there were no false positives. Two patients with SA levels of 152 and 271 umol/L and tyrosine concentrations of 54 and 129 umol/L were found to have TyrI. Allard et al, 2004(180) (level 3a) Allard et al, 2007 analysed succinylacetone (SA) extracted differently from the approach used by Sander et al, 2006. They also demonstrate a high sensitivity and specificity for HT-1 with their approach. 92

Magera et al 2006(174) (level 3a) Magera et al report on the Mayo Clinics use of the SA method of testing. There were no detected cases of in 124,780 newborns tested. This indicates how rare the condition is in some populations. De Braekeller et al, 1990(181) (level 3a) De Braekeller et al report a case series of TyrI detected in Canada from 1970 to 1988 that supports the knowledge claim that Tyrosinemia type I is common in the Saguenay Lac St Jean region of Quebec. The majority of the cases were from the northeastern part of Quebec. The prevalence was 1:1,846 among live births with a total of 98 cases detected. In the Saguenay-Lac-St Jean region of Quebec the carrier rate is 1/20. Table 19: Recent primary research reviewed by BC Tyrosinemia working group

Author Allard et al, 2004(180)

Ashorn et al, 2006(178)

Setting

Subjects

Outcomes

New England Newborn Succinylacetone (SA) SA is a simple method that can be used Screening program levels in 31,999 dried for screening HT in newborns. Based on blood spot from this limited experience, a 2 umol/L SA unaffected newborns was cutoff could result in up to 100% compared with clinically sensitivity and specificity ascertained patients with hepatorenal tyrosinemia Finland Narrative literature review Conclusions (verbatim) TyrI seriously affects hepatic and renal function in the majority of patients, although milder forms also occur. Management is based on dietary intervention and supportive care. The new pharmacologic agent, nitisinone, offers an improved tool to treat tyrosinemia and even to prevent the development of HCC [hepatocellular carcinoma]. Until more knowledge on the effect of nitisinone on the risk for HCC has been gained, early liver transplantation has to be considered to avoid the development of liver malignancy in these patients.

93

Author Chace et al, 2003(82)

Setting Pennsylvania, USA

De Braekeller et Quebec Network of al, 1990(181) Genetic Medicine, Chicoutimi, Quebec, Canada

Subjects

Outcomes

Narrative literature review Regarding Tyrosine (verbatim) of TM/TM for multianalyte Unfortunately, not all cases of screening of newborns tyrosinemia will exhibit a significantly increased tyrosine concentration during the neonatal period. Several falsenegative cases of TyrI have been reported in laboratories relying on tyrosine analysis alone that collect samples on day 2 of life. Succinylacetone analysis may be the only reliable method for detection of type I. In addition to potential false-negative rates, many positive results (often considered false-positive results) arise from the high incidence of TNT in the general newborn population. 98 tyrosinemic children Prevalence at birth estimated at 1/1,846 mostly from the and carrier rate of 1/20 inhabitants in northeastern part of SLSJ. Quebec in the SaguenayLac-St-Jean region Results support a founder effect. (SLSJ). 3 control groups matched to tyrosiemic obligate-carrier couples generated using population register of SLSJ Description of use of Measured SA newborn bloodspot normal TM/TM methods for SA range (n¼152) is 0–0.30 lmol/L. quantification are Bloodspots from children with described hepatorenal tyrosinemia type 1, and kept at room temperature for up to 7 years, afforded SA concentrations of 0.9–5.7 umol/L. (verbatim)

Johnson et al 2007(182)

South Australia and Munich Germany

Magera et al 2006(174)

Mayo Clinic, Rochester To assess the effectiveness new LC–MS/MS based method for the of a two-tier screening MN, USA determination of SA in DBS. This assay approach for TYR 1 we has the potential to significantly reduce applied [our approach] to the number of false positive results in our newborn screening newborn screening for HT- 1 and can program over the last 15 also be used for the laboratory follow months. up of patients treated for HT-1.

94

Author Mohan et al, 1999(179)

Sander et al, 2006

Schulze et al, 2001(183)

Setting

Subjects

Outcomes

Birmingham Children’s 17 children with HT-1 8 of 17 treated with orthotopic liver Hospital, United presenting to this centre transplantation Kingdom Cirrhosis in 8, hepatic dysplasia in 6, hepatocellular carcinoma in 1 Plasma tyrosine and AFP normalized 2 deaths 1 post transplant lymphoproliferative disease resolved 3 renal dysfunction with fall in GFR Good quality of life with median followup of 6.7 years. Germany 61344 dried blood spot Conclusions (verbatim) Tandem mass samples routinely spectrometric quantification of SA collected as part newborn directly from residual blood spots is a screening program for useful method for the early detection of inborn errors of hepatorenal tyrosinemia in newbornmetabolism screening programs. Heidelberg, Germany Report of a method for Conclusions (verbatim) The confirming HT-1 in dried spectrophotometric microassay for blood sports ALA-D is a simple and sensitive test for HT. This represents a basis for further examination of its general reliability as a confirmatory test if tyrosine is found to be increased.

4. ENDOCRINOPATHIES DETECTED WITH STAND ALONE SCREENING TEST Congenital Adrenal Hyperplasia (CAH) Congenital Adrenal Hyperplasia (CAH)18 is caused by one of several defects in the synthesis pathway of cortisol. The most common cause of CAH is a defect in the steroid 21-hydroxylase (21-OHD) enzyme leading to impaired synthesis of cortisol from cholesterol. Insufficient cortisol leads to elevated levels of the hormone adrenocorticotropic hormone (ACTA). Excess ACTA in turn induces over activity of the steroid producing cells of the adrenal gland leading to virilization (masculinization of the genitals) that is evident in females. There is a range of clinical severity in 21-OHD deficiency. A complete deficiency of the enzyme leads to inadequate cortisol and aldosterone production. Deficiency of aldosterone results in salt-wasting, which is often life threatening. Salt wasting leads to 18

Congenital Adrenal Hyperplasia (CYP21A2, EC:1.14.99.10) / synonyms18, Adrenal Virilism, Adrenogenital Syndrome, Corticosterone Methloxidase Deficiency Type I, Hydroxylase Deficiency

95

dehydration and acute metabolic crisis requiring intensive medical intervention. This form of CAH is termed ‘classic’ and is distinguished from ‘simple virilizing’ CAH and late onset or ‘non-classic’ forms of CAH. Approximately 2/3 of children affected with 21-OHD deficiency have the salt-wasting type of CAH. Females are usually readily recognized at birth with ambiguous genitalia. Females with simple virilizing CAH and ambiguous genitalia may even be raised as males without proper diagnosis but this is rare in contemporary health care systems. For the most part, the potential health benefit associated with newborn screening is primarily for males with the salt-wasting form of CAH. In BC the incidence of CAH is ~1 in 14,700, Therefore, there are approximately 3 cases per year of which 1 male per year is expected to have the salt-wasting form. Treatment of CAH includes glucocorticoid replacement therapy that needs to be increased during periods of stress. Individuals with the salt wasting form require treatment with 9 α-fludrohydrocortisone and often sodium chloride. Bilateral adrenalectomy may be indicated for some individuals with poor control with hormonal replacement therapy. Screening for CAH CAH is screened traditionally using a 17-OH progesterone FIA method, however, a high false positive rate exists due large variations in 17-OHP levels with weight, low specificity for the FIA method itself, and elevations in response to physiological stress in the newborn period. Using the FIA method and birth weight specific cutoffs, the PPV of this method can only reach 0.7-1.0% (~140 false positives for every true positive).(184,185) Both the Mayo group and a German screening lab have introduced a tandem MS-based second tier CAH screening method involving chromatographic separation of steroids extracted from a large punch of the blood spot. Both groups reported similar first tier FIA 17OHP cutoffs and 0.65% positive rates on screens of over 200,000 newborns. The Mayo protocol used a ratio of (17OHP+Androstenedione)/cortisol as their diagnostic marker and found a PPV of 7.2%, a 10-fold increase in the PPV.(4) The German group utilized a diagnostic marker of (17OHP+21-deoxycortisol)/cortisol and reported a PPV of 100%.(184) The apparent increase in specificity using 21-deoxycortisol appears remarkable and the primary authors have been contacted to question this apparent discrepancy in findings. Based on the BC birth rate and a similar first tier positive rate (0.65%), we would expect to do ~275 second tier tests per year (~6 per week) to identify 2-3 cases per year (No false positives by Janzen’s numbers and 26 false positives by the Mayo numbers; estimate ~12). (184)(4) The available research represents level 1 evidence for test performance.

96

Knowledge Claims Formulated by BC Working Group on CAH What is the evidence that newborns with congenital adrenal hyperplasia (CAH) will have improved health outcomes with early detection following newborn screening? If CAH testing is added to the routine newborn screening panel the expected benefits for the newborn are: Approximately 3 newborns will be detected with CAH annually. With birth weightadjusted 17OHP reference ranges the false positive rate is ~ 270 / year. There is concern that the high numbers of false positives will generate parental anxiety, increase labreporting resources to handle the case load and increased costs of educating family physicians. Two-tier screening methodologies (17OHP followed by sterol profiling by TMS of the highest values) reduce the false positive rate to < 26 per year and therefore screening with CAH should not be implemented until the capacity for second tier testing is obtained.. Expected benefits for the true positive newborns are: • Up to 1 male newborn per year would be detected earlier (drop mean age of diagnosis of up to 2 weeks) (level 3a) • Up to 1 male newborn per year would have salt wasting detected before becoming ill or earlier in the course of illness with possibility of decreased morbidity in the neonatal period • Long-term reduction in morbidity and mortality is claimed but data is lacking (level 4 expert opinion) Timely diagnosis is critical to the success of the screening program. Therefore, CAH screening would require improvements to surveillance and transportation system of blood spot cards from the approximately 100 birthing hospitals across the province. It is also important to establish a baseline to which outcome data can be compared. The pediatric endocrine group at Children’s are conducting a retrospective chart review to document clinical and laboratory indicators that reflect the degree of illness at the time of clinical ascertainment as well as the number of days from birth to diagnosis. This data will also be collected prospectively for the newborn-screened group and comparison between the 2 groups will determine if screening reduces morbidity and the interval between birth and diagnosis.

The Clinical Evidence Base BC Incidence The unpublished BC estimates on the incidence of CAH detected without screening are about 1 in 14,700 (68/1,000,000) based on an analysis of the clinical records of patients.19 This is comparable to international experience in the absence of screening. With a birth rate of approximately 42,000 this represents on average 2.9 births per year. The BC 19

Data provided by Drs Dan Metzger and Ralph Rothstein

97

cohort has equal numbers of male and females. Salt-wasting CAH was four times more common that simple virilizing CAH. With the absence of screening there is not good data on how frequently late onset CAH leads to avoidable consequences in BC.

Baseline Evidence from US ACMG report CAH was not reviewed in the UK HTA that focused on inborn errors of metabolism. The expert consensus reported by the US ACMG was that there was ‘clear evidence that early intervention optimizes outcome’ and that screening had the ‘potential to prevent most negative consequences’.(3) (Appendix 1, p. 143). Claims of treatment efficacy were supported by the following citations.(186-189) Only 2 were references to primary research studies(188) (189); one was to a textbook (186) and the other to conference proceedings unattainable through the Canadian library system. These statements are level 4 expert opinions given that they are not well supported by the research cited. Pang et al, 1988 (189) This collaborative retrospective pooling and analysis combined experience of screening programs in 6 countries (France, Italy, Japan, New Zealand, Scotland, USA) for classical congenital adrenal hyperplasia due to 21-hydroxylase deficiency between 1980 and 1988. A worldwide incidence rate of 1:14,199 live births for homozygous patients was identified. A greater than normal incidence of CAH was found in Yupik Eskimos of southwestern Alaska (1:282) and the people of La Reunion, France (1:2,141). Incidence is estimated at 1:14,199 for homozygous pts, 1:60 for heterozygous pts; 1:11,909 (homozyous) and 1:55 (heterozygous) for Caucasian patients. Incidence for the salt wasting form of CAH was 1:18,850 (58:1,093,310) compared with 1:57,543 (19:1,093,310) for CAH in the simple virilizing form; therefore salt-wasting CAH was 3X more common. Estimated incidence of CAH in Italy & France (1:10,866) was higher than in Scotland (1:17,098), New Zealand (1:14,500); Japan (1:15,800) no difference from the Caucasian populations.

Virilization of Females A serious consequence of missing CAH in female newborns is that females could be raised as males. In the BC experience, even in the absence of population screening for CAH, a problem with female newborns with signs of virilization being missed and raised as males is negligible. Citing expert reviews, the US ACMG also reported that: “female masculinization begins in the prenatal period so not all sequelae are avoided; normal height may not be reached when treated.” (186,187) Lajic et al, 1998(188) Lajic et al., 1998 followed 44 at-risk for CAH pregnancies monitored from 1985 to 1995 in Scandinavia.(188) Of these 44 pregnancies; 37 total were treated short term (approximately during the 6–12th weeks of pregnancy), because the fetus was either unaffected (n = 28) or an affected male (n = 9); 4 pregnancies were short term (unaffected fetuses) resulted in spontaneous abortions after chorionic villous sampling, and 1 pregnancy was terminated (an affected male); 7 of the 44 cases were treated from the 6th 98

week of pregnancy until term and resulted in 5 females with predicted severe CAH, 1 with mild CAH, and 1 unaffected female. The five girls with severe CAH showed little virilization compared with elder affected sisters. Results for the majority were comparable to matched controls. Adverse events such as failure to thrive and delayed psychomotor development amongst treated infants led the authors to call for more research. BC NSAC Working Group Update Forty-four papers published between 2004 and early 2006 were identified by the BC NSAC sponsored update. This is in addition to the above 2 papers provided by the US ACMG report. A total of 31 published papers were identified as potentially relevant for retrieval including primary studies. Seven recent studies met inclusion criteria for evidence contributing to knowledge of treatment effectiveness (see Table 2). Three studies provided the strongest evidence of newborn benefit and were therefore the focus of the BC Working Group. (190,191)(192). Brosnan et al, 1999 (190) (level 3a) This retrospective cohort study published in 1999 collected incidence and morbidity data on patients with CAH in 3 states over a 5-year period (190). The authors then compared the unscreened population in Arkansas and Oklahoma (n = 400,118) with the screened population in Texas (1,13,378). Differences in the populations and health care systems of screening versus non screening states may accounted for the findings therefore a significant risk of bias was noted by the working group. A benefit for male newborns with salt-wasting CAH was identified. The median age at diagnosis dropped from 26 to 12 days (z-2.49; P - .01). (190) The age gap to diagnosis could be closed by efficient delivery of cards to testing laboratory and communication of results as well as through screening per se. The BC Working group noted it may be difficult to match the 12 day median time to diagnosis even with screening given the dispersed geography of British Columbia and related transportation challenges. Reduced time to diagnosis was also associated with reduced morbidity and shorter lengths of hospitalization. The overall incidence was similar between groups statistically with 0.73 additional males per 100,000 picked up in the screened state. No mortality benefit was identified. Steigert et al, 2002 (191) (level 3c) Steigert et al, 2002 report an uncontrolled cohort analysis conducted on 50% of all Swiss newborns screened between 1993 and June 2001 using a dried blood spot testing method. (191) CAH was detected in 31 of 333,221 newborns screened ( 1 : 10,749). The working group noted that this is a lower incidence that found in BC (1: 15,000) without screening. In 15 CAH had not been detected clinically and so was considered a benefit given that treatment began in the first week of life and none went on to a salt-wasting crisis. Test parameters were reported. 99

Van Vliet, 2005 (192)(level 4) Van Vliet reports that many jurisdictions declining to screen for CAH because of the high ratio of false to true positives.(192) This review by a Canadian clinical expert outlines the process requirements to minimize false positives including continued auditing to ensure rapid response time. Table 20: Recent primary research reviewed by BC CAH working group

Author Lajic et al, 1998(188)

Setting Stockholm, Sweden

Subjects

Outcomes

• 43 fetuses at risk of • Of the entire cohort of 44 prenatally being affected with treated fetuses, there were two couples severe CAH were where the mother or the father was treated with DEX to affected with mild, nonvirilizing CAH prevent prenatal • DEX treatment did not have any negative virilization of affected effect on fetal growth, as treated children females were born normal in size compared to the • The complete cohort matched controls; treated CAH-affected of 44 prentatally boys demonstrated increased birth lengths treated fetuses compared to untreated siblings (P > 0.05, consisted of 28 by t test). unaffected fetuses • More recently recorded heights and treated short term, 9 weights for prenatally treated CAHCAH affected males affected boys and girls and their untreated treated short term, 6 siblings did not differ significantly (P > CAH-affected females 0.05, by t test), although among the girls treated full term, and 1 there are two cases (aged 5.5 and 7 yr) unaffected female whose growth is +/-2 SD below the mean treated full term • Developmental milestones were normal for the majority of the prenatally treated children. However, some adverse events were observed among cases treated both short term and full term

100

Author Pang et al, 1988 (189)

Setting NBS programs in US, Scotland, France, New Zealand and Japan were assessed

Subjects

Outcomes

1,093,310 newborns were screened between 1980 and 1988 of whom 77 had CAH

• In 4 of 5 populations, overall incidence was higher than previously reported for both forms (75% vs., 50-60%), suggesting improved case detection by NBS • The direct benefit of congenital adrenal hyperplasia screening was the prevention of adrenal crisis and its sequelae, the reversal of incorrect sex assignment, and early diagnosis of girls having CAH • Newborn screening for this disorder was shown to be cost effective when its cost is compared with the lifetime tax contribution of a productive citizen

Donohoue et al, Copenhagen, Denmark 14 control subjects 2001(186)

Among 14 control subjects, the intragenic restriction patterns of the 21-OHA and 21OHB genes are invariant. The few restriction fragment length polymorphisms (RFLPs) found in some controls result from polymorphic restriction sites outside the 21-OH genes. In pts with CAH, several different mechanisms for mutation of the 21-OHB gene have been described: deletion of the unique sequences of the 21-OHB gene, conversion of the unique sequences of the 21-OHB gene to those of 21-OHA, and mutations of 21-OHB which do not result in a detectable alteration of restriction pattern (e.g., point mutations). Duplication of the 21-OHA gene has been found in some pts with attenuated CAH; however, the significance of this finding remains unclear

New et al, 2004(193)

With hormonal and molecular genetic screening, previously undiagnosed pts may be identified and can therefore receive glucocorticoid treatment, which has been shown to reverse symptoms within 3 months.

Department of Review by US clinical Pediatrics, Mount Sinai expert. School of Medicine, NY, USA

101

Author

Setting

Subjects

Outcomes • Incidence of diagnosis per 100,000 newborns in unscreened cohort (5.75) and in the screened cohort (6.26) was similar (RR 0.92; CI 95%, 0.58-1.44) • Unscreened group had 0.73 fewer male newborns with salt-wasting CAH diagnosed per 100,000 newborns (RR 0.73; CI 95%, 0.35-1.56) • Median age at diagnosis was 26 days for male newborns with salt-wasting CAH in the unscreened cohort vs.12 days in the screened cohort (z=2.49; P = 0.1); Males with simple-virilizing & non-classical CAH were only detected in the screened group • There was not a statistically significant (P = .73) increase in the diagnosis of saltwasting CAH in the screened cohort • Male newborns benefited as a result of significantly (P = .01) earlier diagnosis, reduced morbidity, and shorter lengths of hospitalization

Brosnan et al, 1999 (190)

Arkansas, Oklahoma, and Houston Texas USA;

An unscreened population in Arkansas and Oklahoma (n = 400,118) was compared with a screened population in Texas (n = 1,613,378) during a 5year period in a retrospective cohort comparison

Steigert et al 2002(191)

Zurich Screening Program, Switzerland

•. 31 newborns had • 31 of 333,221 newborns had CAH; of CAH, 30 detected by these, 30 had 21-OHD, and 1 had 11screening OHD • 333,221 newborns • Incidence of 21-OHD CAH was thus 1 in were screened for 11,108 CAH during 01/01/93 • Screening detected 30 of the 31 CAH and 31/05/01; 162,512 pts; sensitivity = 97% (48.77%) were boys, • Rate of false positive screening tests and 170,709 (51.23%) (recalls where diagnosis of CAH was were girls wrongly considered) was 30 of 333,221 • A preterm rate of (0.009%); of the 30 false positive recalls, 5.69% (314,261 term 24 were performed in preterm infants vs.18,960 preterm (84%) infants out of 333,221 • Specificity of the screening was 99.99% newborns); sex ratio and the positive predictive value of 50% in the preterm infants • 15 of 31 (48%) pts profited from was 54.28% males vs. screening, as CAH had not been 45.72% females recognized clinically • The timely availability of screening results made therapy possible within the first week of life in most cases and helped in preventing salt-wasting crisis in all pts

102

Author van Vliet 2004(194)

Setting Quebec, Canada

Subjects Review by Canadian clinical expert

Outcomes The justification to start a CAH screening program should be based on local evaluation of the clinical epidemiology (especially the sex ratio) of the condition, and the success of the program in achieving predetermined goals should be assessed at regular intervals

103

REFERENCES (1) Pandor A, Eastham J, Beverley C, Chilcott J, Paisley S. Clinical effectiveness and cost-effectiveness of neonatal screening for inborn errors of metabolism using tandem mass spectrometry: a systematic review. Health Technol. Assess. 2004 Mar;8(12):iii, 1121. (2) Pandor A, Eastham J, Beverley C, Chilcott J, Paisley S. Clinical effectiveness and cost-effectiveness of neonatal screening for inborn errors of metabolism using tandem mass spectrometry: a systematic review (Appendices). Health Technol Assess 2004 March;8(12):122-300. (3) American College of Medical Genetics. Newborn screening: toward a uniform screening panel and system—report for public comment. Last accessed 3/2007 March 2005;Available at: http://mchb.hrsa.gov/screening. (4) Matern D, Tortorelli S, Oglesbee D, Gavrilov D, Rinaldo P. Reduction of the falsepositive rate in newborn screening by implementation of MS. J Inherit Metab Dis 30(4):585-92; 2007. (5) Wilson JM, Jungner G. Principles and Practice of Screening for Disease. 1968;WHO Public Health Paper Number 34. (6) Steinraths M, Vallance H, Davidson A. Clinical Ascertainment and Diagnosis of Cystic Fibrosis in British Columbia: 1993-2005. Canadian Medical Association Journal (in process). (7) - Accurso FJ, FAU - Sontag,Marci K., - Sontag MK, FAU - Wagener,Jeffrey S., Wagener JS. - Complications associated with symptomatic diagnosis in infants with cystic fibrosis. - J Pediatr.2005 Sep;147(3 Suppl):S37-41. (8) Doull IJ, Ryley HC, Ryley HC, Weller P, Weller P, Goodchild MC, et al. Cystic fibrosis-related deaths in infancy and the effect of newborn screening. Pediatr Pulmonol.2001 May;31(5):363-6. (9) Farrell PM, Kosorok MR, Kosorok MR, Rock MJ, Rock MJ, Laxova A, et al. Early diagnosis of cystic fibrosis through neonatal screening prevents severe malnutrition and improves long-term growth. Wisconsin Cystic Fibrosis Neonatal Screening Study Group. Pediatrics.2001 Jan;107(1):1-13. (10) Koscik RL, Farrell, PM, Farrell PM, Kosorok MR, Kosorok MR, Zaremba KM, et al. Cognitive function of children with cystic fibrosis: deleterious effect of early malnutrition. Pediatrics.2004 Jun;113(6):1549-58. (11) Farrell PM, Lai HJ, Li Z, Kosorok MR, et al. Evidence on improved outcomes with early diagnosis of cystic fibrosis through neonatal screening: enough is enough! - J Pediatr.2005 Sep;147(3 Suppl):S30-6. 104

(12) Farrell PM, Kosorok MR, Laxova A, Shen G, et al. Nutritional benefits of neonatal screening for cystic fibrosis. Wisconsin Cystic Fibrosis Neonatal Screening Study Group. N Engl J Med.1997 Oct 2;337(14):963-9. (13) Sims EJ, McCormick J, Mehta G, Mehta A, et al. Neonatal screening for cystic fibrosis is beneficial even in the context of modern treatment. J Pediatr.2005 Sep;147(3 Suppl):S42-6. (14) Sims EJ, McCormick J, Mehta G, Mehta A, et al. - Newborn screening for cystic fibrosis is associated with reduced treatment intensity. - J Pediatr.2005 Sep;147(3):30611. (15) Mastella G, Zanolla L, Castellani C, Altieri S, et al. Neonatal screening for cystic fibrosis: long-term clinical balance. Pancreatology.2001;1(5):531-7. (16) - Waters DL, Wilcken B, Irwing L, Van Asperen P, et al. Clinical outcomes of newborn screening for cystic fibrosis. - Arch Dis Child Fetal Neonatal Ed.1999 Jan;80(1):F1-7. (17) Merelle ME, Nagelkerke AF, Lees CM, Dezateux C. Newborn screening for cystic fibrosis. Cochrane Database Syst Rev. 2001;(3). (18) Dankert-Roelse JE, Merelle ME. Review of outcomes of neonatal screening for cystic fibrosis versus non-screening in Europe. J Pediatr.2005 Sep;147(3 Suppl):S15-20. (19) Comeau AM, Parad RB, Dorkin HL, Dovey M, et al. Population-based newborn screening for genetic disorders when multiple mutation DNA testing is incorporated: a cystic fibrosis newborn screening model demonstrating increased sensitivity but more carrier detections. Pediatrics.2004 Jun;113(6):1573-81. (20) Parad RB, Comeau AM. Diagnostic dilemmas resulting from the immunoreactive trypsinogen/DNA cystic fibrosis newborn screening algorithm. J Pediatr.2005 Sep;147(3 Suppl):S78-82. (21) Scotet V, De Braekeleer M, Audrézet MP, Lodé L, Verlingue C, Quéré I, et al. Prevalence of CFTR mutations in hypertrypsinaemia detected through neonatal screening for cystic fibrosis. Clin Genet. 2001 Jan;59(1):42-7. (22) Sarles J, Berthezene P, Le Louarn C, Somma C, et al. Combining immunoreactive trypsinogen and pancreatitis-associated protein assays, a method of newborn screening for cystic fibrosis that avoids DNA analysis. J Pediatr.2005 Sep;147(3):302-5. (23) van den Akker, van Marle ME, Dankert HM, Verkerk PH, Dankert-Roelse, JE, et al. Cost-effectiveness of 4 neonatal screening strategies for cystic fibrosis. Pediatrics.2006 Sep;118(3):896-905.

105

(24) Assael BM, Castellani C, Ocampo MB, Iansa P, et al. - Epidemiology and survival analysis of cystic fibrosis in an area of intense neonatal screening over 30 years. Am J Epidemiol 2002 Sep 1;156(5):397-401. (25) Lai HJ, Cheng Y, Farrell PM. The survival advantage of patients with cystic fibrosis diagnosed through neonatal screening: evidence from the United States Cystic Fibrosis Foundation registry data. J Pediatr.2005 Sep;147(3 Suppl):S57-63. (26) McKay KO, Waters DL, Gaskin KJ. The influence of newborn screening for cystic fibrosis on pulmonary outcomes in new South Wales. J Pediatr.2005 Sep;147(3 Suppl):S47-50. (27) - McCormick J, Sims EJ, Mehta A. - Delayed diagnosis of females with respiratory presentation of cystic fibrosis did not segregate with poorer clinical outcome. J Clin Epidemiol.2006 Mar;59(3):315-22. (28) - Armstrong DS, Hook SM, Jamsen KM, Nixon GM, et al. Lower airway inflammation in infants with cystic fibrosis detected by newborn screening. Pediatr Pulmonol.2005 Dec;40(6):500-10. (29) Slieker MG, Uiterwaal CS, Sinaasappel M, Heijerman HGM, et al. Birth prevalence and survival in cystic fibrosis: a national cohort study in the Netherlands. Chest.2005 Oct;128(4):2309-15. (30) Bjugstad KB, Goodman SI, Freed CR. Age at symptom onset predicts severity of motor impairment and clinical outcome of glutaric acidemia type 1. J Pediatr.2000 Nov;137(5):681-6. (31) Hedlund GL, Longo N, Pasquali M. Glutaric acidemia type 1. Am J Med Genet Part C Semin Med Genet 2006: 142C:86–94. . (32) Applegarth DA, Toone JR, Lowry RB. Incidence of inborn errors of metabolism in British Columbia, 1969-1996. Pediatrics.2000 Jan;105(1):e10. (33) Lindner M, Kolker S, Schulze A, Christensen E, et al. Neonatal screening for glutaryl-CoA dehydrogenase deficiency. J Inherit Metab Dis.2004;27(6):851-9. (34) Christensen E, Ribes A, Merinero B, Zschocke J, et al. Correlation of genotype and phenotype in glutaryl-CoA dehydrogenase deficiency. J Inherit Metab Dis.2004;27(6):861-8. (35) Tortorelli S, Hahn SH, Cowan TM, Brewster TG, et al. The urinary excretion of glutarylcarnitine is an informative tool in the biochemical diagnosis of glutaric acidemia type I. Mol Genet Metab.2005 Feb;84(2):137-43.Epub 2004 Nov 19. (36) Kolker S, Garbade SF, Greenberg CR, Leonard JV, Saudubray JM, Ribes A, et al. Natural history, outcome, and treatment efficacy in children and adults with glutaryl-CoA dehydrogenase deficiency. Pediatr Res. 2006 Jun;59(6):840-7 106

(37) Kolker S, Greenberg CR, Lindner M, Muller E, et al. Emergency treatment in glutaryl-CoA dehydrogenase deficiency. J Inherit Metab Dis.2004;27(6):893-902. (38) Autti-Ramo I, Makela M, Sintonen H, Koskinen H, et al. Expanding screening for rare metabolic disease in the newborn: an analysis of costs, effect and ethical consequences for decision-making in Finland. Acta Paediatr.2005 Aug;94(8):1126-36. (39) Monavari AA, Naughten ER. Prevention of cerebral palsy in glutaric aciduria type 1 by dietary management. Arch Dis Child.2000 Jan;82(1):67-70. (40) Collins J, Leonard JV. Natural history of glutaric aciduria type 1. Arch Dis Child.2000 Jul;83(1):87. (41) Kyllerman M, Skjeldal O, Christensen E, Hagberg G, et al. Long-term follow-up, neurological outcome and survival rate in 28 Nordic patients with glutaric aciduria type 1. Eur J Paediatr Neurol.2004;8(3):121-9. (42) Naughten ER, Mayne PD, Monavari AA, Goodman SI, et al. Glutaric aciduria type I: outcome in the Republic of Ireland. J Inherit Metab Dis 2004;27(6):917-20. (43) Strauss KA, Puffenberger EG, Robinson DL, Morton DH, et al. Type I glutaric aciduria, part 1: natural history of 77 patients. Am J Med Genet C Semin Med Genet.2003 Aug 15;121(1):38-52. (44) Kulkens S, Harting I, Sauer S, Zschocke J, et al. Late-onset neurologic disease in glutaryl-CoA dehydrogenase deficiency. Neurology.2005 Jun 28;64(12):2142-4. (45) Tarini BA, Christakis DA, Welch HG. State newborn screening in the tandem mass spectrometry era: more tests, more false-positive results. Pediatrics.2006 Aug;118(2):448-56. (46) Dionisi-Vici C, Deodato F, Roschinger W, Rhead W, et al. 'Classical' organic acidurias, propionic aciduria, methylmalonic aciduria and isovaleric aciduria: long-term outcome and effects of expanded newborn screening using tandem mass spectrometry. J Inherit Metab Dis.2006 Apr-Jun;29(2-3):383-9. (47) Deodato F, Boenzi S, Santorelli FM, Dionisi-Vici C, et al. Methylmalonic and propionic aciduria. Am J Med Genet C Semin Med Genet.2006 May 15;142(2):104-12. (48) de Baulny HO, Benoist JF, Rigal O, Touati G, et al. Methylmalonic and propionic acidaemias: management and outcome. J Inherit Metab Dis.2005;28(3):415-23. (49) Chandler RJ, Venditti CP. Genetic and genomic systems to study methylmalonic acidemia. Mol Genet Metab; 2005; 86 (1-2):34-43. (50) Barshes NR, Vanatta JM, Patel AJ, Carter BA, O'Mahony CA, Karpen SJ, et al. Evaluation and management of patients with propionic acidemia undergoing liver transplantation: a comprehensive review. Pediatr Transplant 2006;10(7)773-81. 107

(51) Baumgartner D, Scholl-Burgi S, Sass JO, Sperl W, Schweigmann U, Stein JI, et al. Prolonged QTc intervals and decreased left ventricular contractility in patients with propionic acidemia. J Pediatr 2007; 150(2):192-7. (52) Meyburg J, Hoffmann GF. Liver transplantation for inborn errors of metabolism. Transplantation 2005;80(1 Suppl):S135-7. (53) Hori D, Hasegawa Y, Kimura M, Yang Y, et al. Clinical onset and prognosis of Asian children with organic acidemias, as detected by analysis of urinary organic acids using GC/MS, instead of mass screening. Brain Dev 2005 Jan;27(1):39-45. (54) Lerner-Ellis JP, Tirone JC, Pawelek PD, Dore C, Atkinson JL, Watkins D, et al. Identification of the gene responsible for methylmalonic aciduria and homocystinuria, Nat Genet 2006; 38(1):93-100. (55) Whitehead VM. Acquired and inherited disorders of cobalamin and folate in children. Br J Haematol 2006;134(2):125-36. (56) Martinez MA, Rincon A, Desviat LR, Merinero B, Ugarte M, Perez B. Genetic analysis of three genes causing isolated methylmalonic acidemia: identification of 21 novel allelic variants. Mol Genet Metab 2005, 84:317-325. (57) Frazier DM, Millington DS, McCandless SE, Koeberl DD, Weavil SD, Chaing SH, et al. The tandem mass spectrometry newborn screening experience in North Carolina: 1997-2005. J Inherit Metab Dis 2006;29(1):76-85. (58) la Marca G, Malvagia S, Pasquini E, Innocenti M, Donati MA, Zammarchi E. Rapid 2nd-tier test for measurement of 3-OH-propionic and methylmalonic acids on dried blood spots: reducing the false-positive rate for propionylcarnitine during expanded newborn screening by liquid chromatography-tandem mass spectrometry. Clin Chem 2007; 53(7):1364-9. (59) van der Meer SB, Poggi F, Spada M, Bonnefont JP, et al. Clinical outcome and longterm management of 17 patients with propionic acidaemia. Eur J Pediatr.1996 Mar;155(3):205-10. (60) Pollitt RJ, MJ. Treating rare inborn errors of metabolism. - Lancet.2004 Dec 1831;364(9452):2158-60. (61) van der Meer SB, Poggi F, Spada M, Bonnefont JP, et al. Clinical outcome of longterm management of patients with vitamin B12-unresponsive methylmalonic acidemia. J Pediatr.1994 Dec;125(6 Pt 1):903-8. (62) Surtees RA, Matthews EE, Leonard JV. Neurologic outcome of propionic acidemia. Pediatr Neurol 1992;8(5):333-7. (63) Sass JO, Hofmann M, Skladal D, Mayatepek E, et al. Propionic acidemia revisited: a workshop report. Clin Pediatr 2004 Nov-Dec;43(9):837-43. 108

(64) Ensenauer R, Vockley J, Willard JM, Huey JC, et al. A common mutation is associated with a mild, potentially asymptomatic phenotype in patients with isovaleric acidemia diagnosed by newborn screening. Am J Hum Genet.2004 Dec;75(6):1136-42. (65) Vockley J, Ensenauer R. Isovaleric acidemia: new aspects of genetic and phenotypic heterogeneity. Am J Med Genet C Semin Med Genet 2006;15;142(2):95-103. (66) Watson MS. Current status of newborn screening: decision-making about the conditions to include in screening programs. Med Retard Dev Disabil Res Rev 2006;12(4)230-5. (67) Schulze A, Lindner M, Kohlmuller D, Olgemoller K, Mayatepek E, Hoffmann GF. Expanded newborn screening for inborn errors of metabolism by electrospray ionizationtandem mass spectrometry: results, outcome, and implications. Pediatrics 2003;111(6 pt 1):1399-406. (68) Lin WD, Wang CH, Lee CC, Lai CC, Tsai Y, Tsai FJ. Genetic mutation profile of isovaleric acidemia patients in Taiwan. Cytogenet Genome Res 2007;119(3-4):291-6. (69) Berry HK. Screening for mucopolysaccharide disorders with the Berry spot test. Clin Biochem 1987;20(5):365-71. (70) Mehta KC, Zsolway K, Osterhoudt KC, Krantz I, Henretig FM, Kaplan P. Lessons from the late diagnosis of isovaleric acidemia in a five-year-old boy. J Pediatr 1996;129(2):309-10. (71) Rousson R, Guibaud P. Long term outcome of organic acidurias: survey of 105 French cases (1967-1983). J Inherit Metab Dis 1984;7 Suppl 1:10-2. (72) Tanaka K. Isovaleric acidemia: personal history, clinical survey and study of the molecular basis. Prog Clin Biol Res 1990;321:273-90. (73) Stadler SC, Polanetz R, Maier EM, Heidenreich SC, Niederer B, Mayerhofer PU, et al. Newborn screening for 3-methylcrotonyl-CoA carboxylase deficiency: population heterogeneity of MCCA and MCCB mutations and impact on risk assessment. Hum Mutat 2006; 27(8):748-59. (74) Dantas MF, Suormala T, Randolph A, Coelho D, Fowler B, Valle D, et al. 3Methylcrotonyl-CoA carboxylase deficiency: mutation analysis in 28 probands, 9 symptomatic and 19 detected by newborn screening. Hum Mutat 2005; 26(2):164. (75) Koeberl DD, Millington DS, Smith WE, Weavil SD, Muenzer J, McCandless SE, et al. Evaluation of 3-methylcrotonyl-CoA carboxylase deficiency detected by tandem mass spectrometry newborn screening. J Inherit Metab Dis 2003;26(1):25-35. (76) Ficicioglu C, Payan I. 3-Methylcrotonyl-CoA carboxylase deficiency: metabolic decompensation in a noncompliant child detected through newborn screening. Pediatrics 2006;118(6):2555-6. 109

(77) Tamamori A, Fujimoto A, Okano Y, Kobayashi K, Saheki T, Tagami Y, et al. Effects of citrin deficiency in the perinatal period: feasibility of newborn mass screening for citrin deficiency. Pediatr Res 2004; 56(4):608-14. (78) Koeberl DD, Young SP, Gregersen NS, Vockley J, Smith WE, Benjamin DK Jr, et al. Rare disorders of metabolism with elevated butyryl- and isobutyryl-carnitine detected by tandem mass spectrometry newborn screening. Pediatr Res 2003;54(2):219-23. (79) Scriver C editor. The Metabolic & Molecular Bases of Inherited Disease. New York: McGraw-Hill; 2001. (80) Ozand PT, Rashed M, Gascon GG, al Odaib A, Shums A, Nester M, et al. 3Ketothiolase deficiency: a review and four new patients with neurologic symptoms. Brain Dev 1994;16 Suppl:38-45. (81) Fukao T, Zhang GX, Sakura N, Kubo T, Yamaga H, Hazama A, et al. The mitochondrial acetoacetyl-CoA thiolase (T2) deficiency in Japanese patients: urinary organic acid and blood acylcarnitine profiles under stable conditions have subtle abnormalities in T2-deficient patients with some residual T2 activity. J Inherit Metab Dis 2003;26(5):423-31. (82) Chace DH, Kalas TANaylor EW. Use of tandem mass spectrometry for multianalyte screening of dried blood specimens from newborns. Clin Chem.2003 Nov;49(11):1797817. (83) Fukao T, Scriver CR, Kondo N, t2 Collaborative Working Group. The clinical phenotype and outcome of mitochondrial acetoacetyl-CoA thiolase deficiency (betaketothiolase or T2 deficiency) in 26 enzymatically proved and mutation-defined patients. Mol Genet Metab 2001;72(2);109-14. (84) Monastiri K, Amri F, Limam K, Kaabachi N, Guediche MN. beta-Ketothiolase (2methylacetoacetyl-CoA thiolase) deficiency: a frequent disease in Tunisia? J Inherit Metab Dis 1999;22(8):932-3. (85) Pasquali M, Monsen G, Richardson L, Alston M, Longo N. Biochemical findings in common inborn errors of metabolism. Am J Med Genet C Semin Med Genet 2006;142(2):64-76. (86) Rosa RB, Schuck PF, de Assis DR, Latini A, Dalcin KB, Ribeiro CA, et al. Inhibition of energy metabolism by 2-methylacetoacetate and 2-methyl-3hydroxybutyrate in cerebral cortex of developing rats. J Inherit Metab Dis 2005;28(4):501-15. (87) Sewell AC, Herwig J, Wiegratz I, Lehnert W, Niederhoff H, Song XQ, et al. Mitochondrial acetoacetyl-CoA thiolase (beta-ketothiolase) deficiency and pregnancy. J Inherit Metab Dis 1998;21(4):441-2.

110

(88) Gibson KM, Breuer J, Nyhan WL. 3-Hydroxy-3-methylglutaryl-coenzyme A lyase deficiency: review of 18 reported patients. Eur J Pediatr 1988;148(3):180-6. (89) Kayer MA. Disorders of ketone production and utilization. Mol.Genet.Metab. 2006, 87, 4, 281-283. (90) Ozand PT, al Aqeel A, Gascon G, Brismar J, Thomas E, Gleispach H. 3-Hydroxy-3methylglutaryl-coenzyme A (HMG-CoA) lyase deficiency in Saudi Arabia. J Inherit Metab Dis 1991;14(2):174-88. (91) Al-Sayed M, Imtiaz F, Alsmadi OA, Rashed MS, Meyer BF. Mutations underlying 3-hydroxy-3-methylglutaryl CoA lyase deficiency in the Saudi population. BMC Med Genet 2006;16(7):86. (92) Bischof F, Nagele T, Wanders RJ, Trefz FK, Melms A. 3-hydroxy-3-methylglutarylCoA lyase deficiency in an adult with leukoencephalopathy. Ann Neurol 2004;56(5):72730. (93) Cardoso ML, Rodrigues MR, Leao E, Martins E, Diogo L, Rodrigues E, et al. The E37X is a common HMGCL mutation in Portuguese patients with 3-hydroxy-3methylglutaric CoA lyase deficiency. Mol Genet Metab 2004;82(4):334-8. (94) Cipriano LE, Rupar CA, Zaric GS. The cost-effectiveness of expanding newborn screening for up to 21 inherited metabolic disorders using tandem mass spectrometry: results from a decision-analytic model. Value Health 2007;10(2):83-97. (95) Mitchell GA, Ozand PT, Robert MF, Ashmarina L, Roberts J, Gibson KM, et al. HMG CoA lyase deficiency: identification of five causal point mutations in codons 41 and 42, including a frequent Saudi Arabian mutation, R41Q. Am J Hum Genet 1998;62(2):295-300. (96) Stacey TE, de Sousa C, Tracey BM, Whitelaw A, Mistry J, Timbrell P, et al. Dizygotic twins with 3-hydroxy-3-methylglutaric aciduria; unusual presentation, family studies and dietary management. Eur J Pediatr 1985;144(2):177-81. (97) Yylmaz Y, Ozdemir N, Ekinci G, Baykal T, Kocaman C. Corticospinal tract involvement in a patient with 3-HMG coenzyme A lyase deficiency. Pediatr Neurol 2006;35(2):139-41. (98) Spiekerkoetter U, Khuchua Z, Yue Z, Bennett MJ, Strauss AW. General mitochondrial trifunctional protein (TFP) deficiency as a result of either alpha- or betasubunit mutations exhibits similar phenotypes because mutations in either subunit alter TFP complex expression and subunit turnover. Pediatr Res 2004;55(2):190-6. (99) den Boer ME, Wanders RJ, Morris AA, et al. Long-chain 3-hydroxyacyl-CoA dehydrogenase deficiency: clinical presentation and follow-up of 50 patients. Pediatrics.2002 Jan;109(1):99-104.

111

(100) Orii KE, Orii KO, Souri M, Orii T, Kondo N, Hashimoto T, et al. Genes for the human mitochondrial trifunctional protein alpha- and beta-subunits are divergently transcribed from a common promoter region. J Biol Chem 1999;274(12):8077-84. (101) Gillingham MB, Connor WE, Matern D, Rinaldo P, et al. Optimal dietary therapy of long-chain 3-hydroxyacyl-CoA dehydrogenase deficiency. Mol Genet Metab.2003 Jun;79(2):114-23. (102) - Gillingham MB, FAU - Purnell,Jonathan Q., - Purnell JQ, FAU - Jordan J, Jordan J, FAU - Stadler D, et al. - Effects of higher dietary protein intake on energy balance and metabolic control in children with long-chain 3-hydroxy acyl-CoA dehydrogenase (LCHAD) or trifunctional protein (TFP) deficiency. Mol Genet Metab.2007 Jan;90(1):64-9. (103) Gillingham MB, Scott B, Elliott D, Harding CO, et al. Metabolic control during exercise with and without medium-chain triglycerides (MCT) in children with long-chain 3-hydroxy acyl-CoA dehydrogenase (LCHAD) or trifunctional protein (TFP) deficiency. Mol Genet Metab.2006 Sep-Oct;89(1-2):58-63. (104) Choi JH, Yoon HR, Kim GH, Park SJ, Shin YL, Yoo HW. Identification of novel mutations of the HADHA and HADHB genes in patients with mitochondrial trifunctional protein deficiency. Int J Mol Med 2007;19(1):81-7. (105) Sander J, Sander S, Steuerwald U, Janzen N, et al. Neonatal screening for defects of the mitochondrial trifunctional protein. Mol Genet Metab.2005 Jun;85(2):108-14. (106) Olpin SE, Clark S, Andresen BS, Bischoff C, et al. Biochemical, clinical and molecular findings in LCHAD and general mitochondrial trifunctional protein deficiency. J Inherit Metab Dis.2005;28(4):533-44. (107) Tyni T, Palotie A, Viinikka L, Valanne L, et al. Long-chain 3-hydroxyacylcoenzyme A dehydrogenase deficiency with the G1528C mutation: clinical presentation of thirteen patients. J Pediatr.1997 Jan;130(1):67-76. (108) Morris AA, Clayton PT, Surtees RA, Leonard JV, et al. Clinical outcomes in longchain 3-hydroxyacyl-coenzyme A dehydrogenase deficiency. J Pediatr.1997 Dec;131(6):938. (109) Boneh A, BS, Gregersen N, Ibrahim M, et al. VLCAD deficiency: pitfalls in newborn screening and confirmation of diagnosis by mutation analysis. Mol Genet Metab.2006 Jun;88(2):166-70. (110) Liebig M, Schymik I, Mueller M, Wendel U, et al. Neonatal screening for very long-chain acyl-coA dehydrogenase deficiency: enzymatic and molecular evaluation of neonates with elevated C14:1-carnitine levels. Pediatrics.2006 Sep;118(3):1065-9.

112

(111) Schymik I, Liebig M, Mueller M, Wendel U, et al. Pitfalls of neonatal screening for very-long-chain acyl-CoA dehydrogenase deficiency using tandem mass spectrometry. J Pediatr.2006 Jul;149(1):128-30. (112) Wood JC, Magera MJ, Rinaldo P, Seashore MR, et al. Diagnosis of very long chain acyl-dehydrogenase deficiency from an infant's newborn screening card. Pediatrics.2001 Jul;108(1):E19. (113) Hoffman JD, Steiner RD, Paradise L, Harding CO, et al. Rhabdomyolysis in the military: recognizing late-onset very long-chain acyl Co-A dehydrogenase deficiency. Mil Med.2006 Jul;171(7):657-8. (114) Doi T, Abo W, Tateno M, Hayashi K, et al. Milder childhood form of very longchain acyl-CoA dehydrogenase deficiency in a 6-year-old Japanese boy. Eur J Pediatr.2000 Dec;159(12):908-11. (115) Cox GF, Souri M, Aoyama T, Rockenmacher S, et al. Reversal of severe hypertrophic cardiomyopathy and excellent neuropsychologic outcome in very-longchain acyl-coenzyme A dehydrogenase deficiency. J Pediatr.1998 Aug;133(2):247-53. (116) Sluysmans T, Tuerlinckx D, Hubinont C, Verellen-Dumoulin C, et al. Very long chain acyl-coenzyme A dehydrogenase deficiency in two siblings: evolution after prenatal diagnosis and prompt management. J Pediatr.1997 Sep;131(3):444-6. (117) Brown-Harrison MC, Nada MA, Sprecher H, Vianey-Saban C, et al. Very long chain acyl-CoA dehydrogenase deficiency: successful treatment of acute cardiomyopathy. Biochem Mol Med.1996 Jun;58(1):59-65. (118) Cairns AP, O'Donoghue PM, Patterson VH, Brown JH, et al. Very-long-chain acylcoenzyme A dehydrogenase deficiency--a new cause of myoglobinuric acute renal failure. Nephrol Dial Transplant.2000 Aug;15(8):1232-4. (119) Engbers HM, Dorland L, de Sain MG, Eskes PF, et al. Rhabdomyolysis in earlyonset very long-chain acyl-CoA dehydrogenase deficiency despite normal glucose after fasting. J Inherit Metab Dis.2005;28(6):1151-2. (120) Touma EH, Rashed MS, Vianey-Saban C, Sakr A, et al. A severe genotype with favourable outcome in very long chain acyl-CoA dehydrogenase deficiency. Arch Dis Child.2001 Jan;84(1):58-60. (121) Voermans NC, Poels PJ, Kluijtmans LA, van Engelen BG, et al. The effect of dantrolene sodium in Very Long Chain Acyl-CoA Dehydrogenase Deficiency. Neuromuscul Disord.2005 Dec;15(12):844-6. (122) Vianey-Saban C, Divry P, Brivet M, - Nada M, et al. Mitochondrial very-longchain acyl-coenzyme A dehydrogenase deficiency: clinical characteristics and diagnostic considerations in 30 patients. Clin Chim Acta.1998 Jan 12;269(1):43-62.

113

(123) Andresen BS, Olpin S, Poorthuis BJ, Scholte HR, et al. Clear correlation of genotype with disease phenotype in very-long-chain acyl-CoA dehydrogenase deficiency. Am J Hum Genet.1999 Feb;64(2):479-94. (124) Schimmenti LA, Crombez EA, Schwahn BC, Heese BA, Wood TC, Schroer RJ, et al. Expanded newborn screening identifies maternal primary carnitine deficiency. Mol Genet Metab. 2007 Apr;90(4):441-5. (125) Vijay S, Patterson A, Olpin S, Henderson MJ, Clark S, Day C, et al. Carnitine transporter defect: diagnosis in asymptomatic adult women following analysis of acylcarnitines in their newborn infants. J Inherit Metab Dis. 2006 Oct;29(5):627-30. (126) Wilcken B, Wiley V, Sim KG, Carpenter K. Carnitine transporter defect diagnosed by newborn screening with electrospray tandem mass spectrometry. J Pediatr. 2001 Apr;138(4):581-4. (127) Boles RG, Buck EA, Blitzer MG, Platt MS, Cowan TM, Martin SK, et al. Retrospective biochemical screening of fatty acid oxidation disorders in postmortem livers of 418 cases of sudden death in the first year of life. J Pediatr. 1998 Jun;132(6):924-33. (128) Cederbaum SD, Koo-McCoy S, Tein I, Hsu BY, Ganguly A, Vilain E, et al. Addendum to 'Carnitine membrane transporter deficiency: a long-term follow up and OCTN2 mutation in the first documented case of primary carnitine deficiency."[Mol. Genet. Metab. 77 (2002) 195–201.] Mol.Genet.Metab. 2003;78(1):82. (129) Koizumi A, Nozaki J, Ohura T, Kayo T, Wada Y, Nezu J, et al. Genetic epidemiology of the carnitine transporter OCTN2 gene in a Japanese population and phenotypic characterization in Japanese pedigrees with primary systemic carnitine deficiency. Hum Mol Genet. 1999 Nov;8(12):2247-54. (130) Lamhonwah AM, Olpin SE, Pollitt RJ, Vianey-Saban C, Divry P, Guffon N, et al. Novel OCTN2 mutations: no genotype-phenotype correlations: early carnitine therapy prevents cardiomyopathy. Am J Med Genet. 2002 Aug 15;111(3):271-84. (131) Wang Y, Korman SH, Ye J, Gargus JJ, Gutman A, Taroni F, et al. Phenotype and genotype variation in primary carnitine deficiency. Genet Med. 2001 Nov-Dec;3(6):38792. (132) Lund AM, Joensen F, Hougaard DM, Jensen LK, Christensen E, Christensen M, et al. Carnitine transporter and holocarboxylase synthetase deficiencies in The Faroe Islands. J Inherit Metab Dis. 2007 Jun;30(3):341-9. (133) Legge M. Systemic carnitine deficiency as the cause of a prolonged illness and sudden death in a six-year-old child. J Inherit Metab Dis. 1985;8(3):159. (134) Bok LA, Vreken P, Wijburg FA, Wanders RJ, Gregersen N, Corydon MJ, et al. Short-chain Acyl-CoA dehydrogenase deficiency: studies in a large family adding to the complexity of the disorder. Pediatrics. 2003 Nov;112(5):1152-5. 114

(135) Corydon MJ, Vockley J, Rinaldo P, Rhead WJ, Kjeldsen M, Winter V, et al. Role of common gene variations in the molecular pathogenesis of short-chain acyl-CoA dehydrogenase deficiency. Pediatr Res. 2001 Jan;49(1):18-23. (136) Sass JO, Sander S, Zschocke J. Isobutyryl-CoA dehydrogenase deficiency: isobutyrylglycinuria and ACAD8 gene mutations in two infants. J Inherit Metab Dis. 2004;27(6):741-5. (137) Dillon BJ, Prieto VG, Curley SA, Ensor CM, Holtsberg FW, Bomalaski JS, et al. Incidence and distribution of argininosuccinate synthetase deficiency in human cancers: a method for identifying cancers sensitive to arginine deprivation. Cancer. 2004 Feb 15;100(4):826-33. (138) Nassogne MC, Heron B, Touati G, Rabier D, Saudubray JM. Urea cycle defects: management and outcome. J Inherit Metab Dis. 2005;28(3):407-14. (139) Haberle J, Pauli S, Linnebank M, Kleijer WJ, Bakker HD, Wanders RJ, et al. Structure of the human argininosuccinate synthetase gene and an improved system for molecular diagnostics in patients with classical and mild citrullinemia. Hum Genet. 2002 Apr;110(4):327-33. (140) Sander J, Janzen N, Sander S, Steuerwald U, Das AM, Scholl S, et al. Neonatal screening for citrullinaemia. Eur J Pediatr. 2003 Jun;162(6):417-20. (141) Kleijer WJ, Garritsen VH, van der Sterre ML, Berning C, Haberle J, Huijmans JG. Prenatal diagnosis of citrullinemia and argininosuccinic aciduria: evidence for a transmission ratio distortion in citrullinemia. Prenat Diagn. 2006 Mar;26(3):242-7 (142) Choi JH, Kim H, Yoo HW. Two cases of citrullinaemia presenting with stroke. J Inherit Metab Dis. 2006 Feb;29(1):182-3. (143) Wilcken B. Problems in the management of urea cycle disorders. Mol Genet Metab. 2004 Apr;81 Suppl 1:S86-91. (144) Smith W, Kishnani PS, Lee B, Singh RH, Rhead WJ, Sniderman King L, et al. Urea cycle disorders: clinical presentation outside the newborn period. Crit Care Clin. 2005 Oct;21(4 Suppl):S9-17. (145) Mitsubuchi H, Owada M, Endo F. Markers associated with inborn errors of metabolism of branched-chain amino acids and their relevance to upper levels of intake in healthy people: an implication from clinical and molecular investigations on maple syrup urine disease. J Nutr. 2005 Jun;135(6 Suppl):1565S-70S. (146) Chuang DT, Chuang JL, Wynn RM. Lessons from genetic disorders of branchedchain amino acid metabolism. J Nutr. 2006 Jan;136(1 Suppl):243S-9S. (147) Simon E, Flaschker N, Schadewaldt P, Langenbeck U, Wendel U. Variant maple syrup urine disease (MSUD)--the entire spectrum. J Inherit Metab Dis. 2006 Dec;29(6):716-24.

115

(148) Morton DH, Strauss KA, Robinson DL, Puffenberger EG, Kelley RI. Diagnosis and treatment of maple syrup disease: a study of 36 patients. Pediatrics. 2002 Jun;109(6):999-1008. (149) Hoffmann B, Helbling C, Schadewaldt P, Wendel U. Impact of longitudinal plasma leucine levels on the intellectual outcome in patients with classic MSUD. Pediatr Res. 2006 Jan;59(1):17-20. (150) Simon E, Fingerhut R, Baumkotter J, Konstantopoulou V, Ratschmann R, Wendel U. Maple syrup urine disease: favourable effect of early diagnosis by newborn screening on the neonatal course of the disease. J Inherit Metab Dis. 2006 Aug;29(4):532-7. (151) le Roux C, Murphy E, Hallam P, Lilburn M, Orlowska D, Lee P. Neuropsychometric outcome predictors for adults with maple syrup urine disease. J Inherit Metab Dis. 2006 Feb;29(1):201-2. (152) Harris RA, Joshi M, Jeoung NH. Mechanisms responsible for regulation of branched-chain amino acid catabolism. Biochem Biophys Res Commun. 2004 Jan 9;313(2):391-6. (153) Rodriguez-Pombo P, Navarrete R, Merinero B, Gomez-Puertas P, Ugarte M. Mutational spectrum of maple syrup urine disease in Spain Hum Mutat. 2006 Jul;27(7):715. (154) Zytkovicz TH, Fitzgerald EF, Marsden D, Larson CA, et al. Tandem mass spectrometric analysis for amino, organic, and fatty acid disorders in newborn dried blood spots: a two-year summary from the New England Newborn Screening Program. Clin Chem.2001 Nov;47(11):1945-55. (155) Bhattacharya K, Khalili V, Wiley V, Carpenter K, Wilcken B. Newborn screening may fail to identify intermediate forms of maple syrup urine disease. J Inherit Metab Dis.2006 Aug;29(4):586. (156) Oglesbee D, Sanders KA, Lacey JM, Magera MJ, Casetta B, Strauss KA, et al. Second-Tier Test for Quantification of Alloisoleucine and Branched-Chain Amino Acids in Dried Blood Spots to Improve Newborn Screening for Maple Syrup Urine Disease (MSUD). Clin Chem. 2008 Mar;54(3):542-9. (157) Robinson D, Drumm LA. Maple syrup disease: a standard of nursing care. Pediatr Nurs. 2001 May-Jun;27(3):255-8, 261-4, 270. (158) Yoshino M, Aoki K, Akeda H, Hashimoto K, Ikeda T, Inoue F, et al. Management of acute metabolic decompensation in maple syrup urine disease: a multi-center study. Pediatr Int. 1999 Apr;41(2):132-7. (159) Hilliges C, Awiszus D, Wendel U. Intellectual performance of children with maple syrup urine disease. Eur J Pediatr. 1993 Feb;152(2):144-7. (160) Mudd SH, Skovby F, Levy HL, Pettigrew KD, Wilcken B, Pyeritz RE, et al. The natural history of homocystinuria due to cystathionine beta-synthase deficiency. Am J Hum Genet. 1985 Jan;37(1):1-31. 116

(161) Peterschmitt MJ, Simmons JR, Levy HL. Reduction of false negative results in screening of newborns for homocystinuria. N Engl J Med. 1999 Nov 18;341(21):1572-6. (162) Yap S, Boers GH, Wilcken B, Wilcken DE, Brenton DP, Lee PJ, et al. Vascular outcome in patients with homocystinuria due to cystathionine beta-synthase deficiency treated chronically: a multicenter observational study. Arterioscler Thromb Vasc Biol. 2001 Dec;21(12):2080-5. (163) Yap S, Naughten E. Homocystinuria due to cystathionine beta-synthase deficiency in Ireland: 25 years' experience of a newborn screened and treated population with reference to clinical outcome and biochemical control. J Inherit Metab Dis. 1998 Oct;21(7):738-47. (164) Sokolova J, Janosikova B, Terwilliger JD, Freiberger T, Kraus JP, Kozich V. Cystathionine beta-synthase deficiency in Central Europe: discrepancy between biochemical and molecular genetic screening for homocystinuric alleles. Hum Mutat. 2001 Dec;18(6):548-9. (165) Gaustadnes M, Ingerslev J, Rutiger N. Prevalence of congenital homocystinuria in Denmark. N Engl J Med. 1999 May 13;340(19):1513. (166) Chace DH, Hillman SL, Millington DS, Kahler SG, Adam BW, Levy HL. Rapid diagnosis of homocystinuria and other hypermethioninemias from newborns' blood spots by tandem mass spectrometry. Clin Chem. 1996 Mar;42(3):349-55. (167) Chien YH, Chiang SC, Huang A, Hwu WL. Spectrum of hypermethioninemia in neonatal screening. Early Hum Dev. 2005 Jun;81(6):529-33. (168) Moat SJ, Bao L, Fowler B, Bonham JR, Walter JH, Kraus JP. The molecular basis of cystathionine beta-synthase (CBS) deficiency in UK and US patients with homocystinuria. Hum Mutat. 2004 Feb;23(2):206. (169) Refsum H, Fredriksen A, Meyer K, Ueland PM, Kase BF. Birth prevalence of homocystinuria. J Pediatr. 2004 Jun;144(6):830-2. (170) Febriani AD, Sakamoto A, Ono H, Sakura N, Ueda K, Yoshii C, et al. Determination of total homocysteine in dried blood spots using high performance liquid chromatography for homocystinuria newborn screening. Pediatr Int. 2004 Feb;46(1):5-9. (171) Levy HL, Albers S. Genetic screening of newborns. Annu Rev Genomics Hum Genet. 2000;1:139-77. (172) Refsum H, Grindflek AW, Ueland PM, Fredriksen A, Meyer K, Ulvik A, et al. Screening for serum total homocysteine in newborn children. Clin Chem. 2004 Oct;50(10):1769-84. (173) Yap S, Rushe H, Howard PM, Naughten ER. The intellectual abilities of earlytreated individuals with pyridoxine-nonresponsive homocystinuria due to cystathionine beta-synthase deficiency. J Inherit Metab Dis. 2001 Aug;24(4):437-47. (174) Magera MJ, Gunawardena ND, Hahn SH, Tortorelli S, Mitchell GA, Goodman SI, et al. Quantitative determination of succinylacetone in dried blood spots for newborn screening of tyrosinemia type I. Mol Genet Metab. 2006 May;88(1):16-21. 117

(175) Sander J, Janzen N, Peter M, Sander S, Steuerwald U, Holtkamp U, et al. Newborn screening for hepatorenal tyrosinemia: Tandem mass spectrometric quantification of succinylacetone J Clin Endocrinol Metab. 2007 Jul;92(7):2581-9. (176) Turgeon C, Magera MJ, Allard P, Tortorelli S, Gavrilov D, Oglesbee D, et al. Combined Newborn Screening for Succinylacetone, Amino Acids, and Acylcarnitines in Dried Blood Spots. Clin Chem. 2008 Feb 15 [Epub ahead of print] (177) la Marca G, Malvagia S, Pasquini E, Innocenti M, Fernandez MR, Donati MA, et al. The inclusion of succinylacetone as marker for tyrosinemia type I in expanded newborn screening programs. Rapid Commun Mass Spectrom. 2008;22(6):812-8. (178) Ashorn M, Pitkanen S, Salo MK, Heikinheimo M. Current strategies for the treatment of hereditary tyrosinemia type I. Paediatr Drugs. 2006;8(1):47-54. (179) Mohan N, McKiernan P, Preece MA, Green A, Buckels J, Mayer AD, et al. Indications and outcome of liver transplantation in tyrosinaemia type 1. Eur J Pediatr. 1999 Dec;158 Suppl 2:S49-54. (180) Allard P, Grenier A, Korson MS, Zytkovicz TH. Newborn screening for hepatorenal tyrosinemia by tandem mass spectrometry: analysis of succinylacetone extracted from dried blood spots. Clin Biochem. 2004 Nov;37(11):1010-5. (181) De Braekeleer M, Larochelle J. Genetic epidemiology of hereditary tyrosinemia in Quebec and in Saguenay-Lac-St-Jean. Am J Hum Genet. 1990 Aug;47(2):302-7. (182) Johnson DW, Gerace R, Ranieri E, Trinh MU, Fingerhut R. Analysis of succinylacetone, as a Girard T derivative, in urine and dried bloodspots by flow injection electrospray ionization tandem mass spectrometry. Rapid Commun Mass Spectrom. 2007;21(1):59-63. (183) Schulze A, Frommhold D, Hoffmann GF, Mayatepek E. Spectrophotometric microassay for delta-aminolevulinate dehydratase in dried-blood spots as confirmation for hereditary tyrosinemia type I. Clin Chem. 2001 Aug;47(8):1424-9. (184) Janzen N, Peter M, Sander S, Steuerwald U, Terhardt M, Holtkamp U, et al. Newborn screening for congenital adrenal hyperplasia: additional steroid profile using liquid chromatography-tandem mass spectrometry. J Clin Endocrinol Metab. 2007 Jul;92(7):2581-9. (185) Minutti CZ, Lacey JM, Magera MJ, Hahn SH, et al. Steroid profiling by tandem mass spectrometry improves the positive predictive value of newborn screening for congenital adrenal hyperplasia. J Clin Endocrinol Metab.2004 Aug;89(8):3687-93. (186) Donohoue P, et al. Congenital Adrenal Hyperplasia. In: Scriver CR, et al., editors. The Metabolic and Molecular Bases of Inherited Disease. 8th ed. New York: McGrawHill; 2001. p. 1971-2005.

118

(187) New MI & Wilson RC. Steroidal disorders in children: congenital adrenal hyperplasia and apparent mineralocorticoid excess. Proc Nat Acad Sci 1999:1279012797. (188) Lajic S, Wedell A, Bui TH, Ritzen EM, et al. Long-term somatic follow-up of prenatally treated children with congenital adrenal hyperplasia. J Clin Endocrinol Metab.1998 Nov;83(11):3872-80. (189) Pang SY, Wallace MA, Hofman L, Thuline HC, et al. Worldwide experience in newborn screening for classical congenital adrenal hyperplasia due to 21-hydroxylase deficiency. Pediatrics.1988 Jun;81(6):866-74. (190) Brosnan PG, Brosnan CA, Kemp SF, Domek DB, et al. Effect of newborn screening for congenital adrenal hyperplasia. Arch Pediatr Adolesc Med.1999 Dec;153(12):1272-8. (191) Steigert M, Schoenle EJ, Biason-Lauber A, Torresani T, et al. High reliability of neonatal screening for congenital adrenal hyperplasia in Switzerland. J Clin Endocrinol Metab.2002 Sep;87(9):4106-10. (192) Van Vliet G. Neonatal endocrinopathies: Should we change the screening methods? Ital J Pediatr 2005;31:209-212. (193) New MI. An update of congenital adrenal hyperplasia. Ann.N.Y.Acad.Sci., 2004, 1038, 14-43. (194) Van Vliet G, Czernichow P. Screening for neonatal endocrinopathies: rationale, methods and results. Semin.Neonatol. 2004, 9, 1, 75-85.

119

APPENDIX A: Recent primary research reviewed by the working group on test performance parameters

Author

Setting

Schulze et al, Heidelberg, 2003(67) Germany

Subjects

Outcomes

• Specimens of dried blood • Overall prevalence of a metabolic disorder spots were collected from identified on newborn screening was 1 in 2400 infants between 04/1998 newborns (95% confidence interval (CI) and 09/2001 1:2900–1:2000) • 250 000 neonates were • There were 65 newborns with amino investigated for 23 inborn acidemias (prevalence 1:3800, 95% CI errors of metabolism by 1:5100–1:3100) electrospray ionization• 24 had FAO disorders (1:10 400, 95% CI MS/MS; 98% were white 1:17 400–1:7400) • In 106 newborns, inborn • 17 with organic acidurias (1:14 700, 95% 17 errors of metabolism were with organic acidurias (1:14 700, 95% confirmed • 26 infants were judged positive on newborn screening and confirmed by recall, but a definite diagnosis remained questionable either because the diagnosis is difficult to achieve (N = 17) or because they were lost to follow-up (N = 9) • 825 tests (0.33%) were classified as false positive, yielding an overall specificity of 99.67%; specificity for amino acidemias and FAO disorders was 99.90% in each, and for organic acidurias it was 99.87% • Positive Predictive Value of the overall MS/MS screening was found to be 11.31%, reflecting high diagnostic specificity of the method; highest positive predictive value was obtained for amino acidemias (20.12%), followed by FAO disorders (8.57%) and organic acidurias (5.09%) • Diagnositc sensitivity for amino acidemias was 94.2%, 100% for FAO disorders and organic acidurias; overall sensitivity of MS/MS screening was 100% for classic forms of all disorders and 92.6% for variant forms • Of the 106 positive screened infants, 97 infants (92%) were found asymptomatic, the rest were subdivided into groups of disorders, this comprised 63 of 65 patients (97%) with amino acidemias, 22 of 24 pts (92%) with FAO disorders, and 12 of 17 pts (71%) with organic acidurias • 61 (58%) of the 106 newborns first diagnosed by screening, needing treatment, and remaining asymptomatic can be assumed to have benefited from screening; these were 32 pts (49%) with amino acidemias, 22 (92%) with FAO disorders, and 7 (41%) with organic acidurias

120

Author Frazier et al, 2006(57)

Setting

Subjects

Outcomes

North Carolina, • Newborn blood samples • Between 28/0797 and 28/07/05, 944,078 USA were obtained by heel infants were screened and 219 diagnoses of prick and spotted on disorders were confirmed on newborns (overall Overview of Scheicher and Schuell 903 incidence of 1:4300) NBS screening filter paper sampling cards • Disorders of phenylalanine metabolism experience • The mean age at the represented 79% of the 62 cases of time of sampling was 39 h aminoacidopathies; the additional use of the • The average number of PHE:TYR ratio (≥3) greatly reduced the high newborns screened per false positive rate of using PHE alone year was 118 010 • Among the organic acidaemias, 3-MCC • Between 1999 and 2004, deficiency was the most frequently detected, the racial and ethnic with 26 cases breakdown was 73% • MCADD was confirmed in 73 newborns in this caucasian, 23% African 8-year period, making it the fatty acid oxidation American, 2.5% Asian and defect most commonly detected by the NC 1.5% Native American. MS/MS NBS Program; Caucasian incidence Hispanic infants 1:10,000, African Americans 1:30,000; represented approximately Hispanic 1:20,000 12% of the all newborns • 6 false-negative cases in total • In addition to MS/MS • For PPV calculations for this report, false analysis, the samples positives were only those samples that had were screened for either an initial diagnostic or two borderline galactosaemia (McKusick elevated screens, and negative confirmatory 230,400), primary testing. True positives were those with both hypothyroidism (McKusick elevated screens and positive confirmatory 274,200), congenital testing adrenal hyperplasia • For 2003 and 2004, the PPV was 53% (McKusick 300,300), and • For those samples with two borderline haemoglobinopathies elevations the PPV was 33%, while for those (McKusick 603,903, with a single diagnostic elevation the PPV was 141,900, etc). From 60% December 2004, the • These results are higher than the samples were also tested approximately 10% PPV reported by NBS for biotinidase deficiency programmes who did confirmatory testing on (McKusick 253,260) all newborns who had initial elevated screening results

121

Author

Setting

Stadler et al, 2006 (73)

Bavaria, Germany (Australia and Austria collaboration)

Subjects

Outcomes

• Screened 677,852 • NBS revealed 8 newborns (5 females, 3 neonates for 25 males) with biochemically confirmed MCCD conditions, including were detected, corresponding to a frequency of MCCD; 17 biochemical 1:84,700 (1:12,500 to 1:196,000; 95% and clinical phenotypes confidence interval) in based on elevated the population investigated (677,852 concentrations of 3newborns) and 6 relatives with MCCD, hydroxyisovalerylcarnitine suggesting a higher frequency than previously (3-HIVA-C) assumed (1:84,700) • Genotypes of MCCA • A strikingly heterogeneous spectrum of 22 (MCCC1) and MCCB novel and 8 reported mutations was discovered (MCCC2) were assessed • Comparative analysis of case reports with in identified newborns, NBS data implied that only few individuals their relatives, and in (<10%) develop symptoms. In addition, none of individuals (n=17) from the symptoms reported so far can clearly be other regions, and attributed to MCCD correlated to biochemical and clinical phenotypes Of the 28 probands investigated, 10 (36%) had mutations in MCCA and 18 (64%) in MCCB 28 probands from 23 families were analyzed. • Acylcarnitines in dried blood spots obtained for NBS were analyzed as butyl-esters on a triple quadruple tandem mass spectrometer with a TurboIon Spray source • Mutation analysis was also performed

122

APPENDIX B: TMS Screening Test Parameters with First and Second Tier Testing

Disorder PKU MCAD GALT CH

Analyte(s) Phe (Phe/Tyr) C8 (C8/C10) GALT TSH

GAI

C5DC (ratios)

LCHAD

C16OH (ratios)

IVA

C5 (C5/C3)

VLCAD CIT

C14:1(ratios) Cit

MMA (+PA)

C3 (ratios)

Hcys

Cutoff (uM) Sensitivity FP/42,000 TP/42,000 99% 4 5 180 (2.2) 100% 4 3 0.3 (2) Non-det 100% 1 2 20 99% 29 14

PPV 56% 43% 67% 33%

Predicted Incidence 1/12,000 1/12,000 1/23,000 1/4,000

0.1

100%

5

1

17% 1/100,000

0.15

99%

0

1

100% 1/50,000

0.5 (0.4)

100%

3

1

25% 1/75,000

0.75 40

98% 99%

0.2 2

1.1 1

87% 1/40,000 33% 1/41,000

5.6

~80%

75

1.1

1.4% 1/40,000

Met (Met/Phe)

62 (1.35)

95%

5

0.3

5.7% 1/150,000

MSUD

Leu (Leu/Phe)

295 (5)

95%

110

0.2

0.18% 1/185,000

CAH

17-OHP

38 (nM)

100%

273

1.8

0.66% 1/23,000

Tyrosine Tyr I Overall Rate per 42,000

202

50%

670 1181 2.81%

0.4 32.9 0.08%

0.06% 1/100,000 3%

Source BCCH In-house data and cutoffs BCCH In-house data and cutoffs BCCH In-house data and cutoffs BCCH In-house data and cutoffs BCCH 2007 data applied to cutoff algorithms from Linder (2006) BCCH 2007 data applied to Region 4 Collaborative cutoffs BCCH 2007 data applied to Region 4 Collaborative cutoffs Frazier et al (2006) with numbers adjusted to BC birth rate Frazier et al (2006) with numbers adjusted to BC birth rate BCCH 2007 data applied to Region 4 Collaborative cutoffs BCCH 2007 data applied to Region 4 Collaborative cutoffs Minutti et al (2004) with numbers adjusted to BC birth rate BCCH 2007 data applied to Region 4 Collaborative cutoffs

123

Second Tier MMA (+PA) Hcys MSUD

MMA + MetCit Hcys Allo-Ile

100% 100% 100%

14.7 0 0

3.8 0.3 0.2

21% 1/16,000 100% 1/150,000 100% 1/185,000

17OHP+Andro/C 26 1.8 6.5% 1/23,000 CAH 100% SUAC 0 0.4 100.0% 1/100,000 Tyr I 100% Overall 92 36 28% Rate (%) 0.22% 0.09% Second tier testing increases the total PPV to 28% from 3% while increasing the sensitivity for MMA, Hcys, Tyr I, and MSUD to ~100%

Matern et al. (2007) Matern et al. (2007) Oglesbee et al. (2008) Minutti et al. (2004) with numbers adjusted to BC birth rate Magera et al. (2006 )

CF top 3%

98%

139

14

9% 1/3700

CF

IRT/DNA

CAH Note:

Janzen et al (2007) used 17OHP+21-DOC/C as a marker in a study similar to Minutti et al (2004) and found a PPV of 100%. Under this scenario our total PPV would increase to 35%

BCCH NBS Business plan

Disorder Legend: (PKU) Phenylketonuria; (CH) Congenital hypothyroidism; (LCHAD) Long chain acylCoA dehydrogenase deficiency; (IVA) Isovaleric acidemia; (VLCFA) Very long chain Acyl-CoA dehydrogenase deficiency; Cit) Citrullinemia; (MMA) Methylmalonic acidemia; (PA) Propionic acidemia; (Hcys) Homocysteinuria; (MSUD) Maple Syrup Urine disease; (CAH) Congenital adrenal hyperplasia; (TyrI) Tyrosinemia type I; (CF) Cystic fibrosis

124