Clin Perinatol 29 (2002) 395 – 414

Animal models of long-term consequences of early exposure to repetitive pain C. Celeste Johnston, RN, DEda,*, Claire-Dominique Walker, PhDb, Kristina Boyer, RN, MSc(A)a b

a School of Nursing, McGill University, 3506 University St., Montreal, QC H3A 2A7, Canada Department of Psychiatry, McGill University, 3506 University St., Montreal, QC H3A 2A7, Canada

The purpose of this article is to present to clinicians paradigms, findings, and issues related to animal studies of the long-term consequences of early exposure to repetitive pain so that the studies are comprehensible and can be critiqued for relevance to clinical situations. It is well documented that neonates and infants in neonatal intensive care units (NICUs) experience repeated painful events [1– 3]. There are also emerging data that indicate there may be long-term consequences of repeated pain, particularly in preterm neonates [4 –7]. There appear, however, to be some conflicting results. For example, Fitzgerald et al [8] reported that repeated heel lance in neonates results in heightened responsiveness within 72 hours of heel tissue injury, comparable to the study of Taddio et al [9] who used a single painful procedure. In this report [9], it was found that unanesthetized circumcision in the first few days of life results in increased responsiveness to pain at 2 months of age. In contrast, there are reports that an increased number of minor but invasive procedures in infants born at 28 weeks postconceptional age results in decreased behavioral responsiveness to pain when they reach 32 weeks postconceptional age [10], and that lack of sufficient recovery time between procedures also results in decreased responsiveness [11]. Grunau and colleagues found that infants exposed to repeated pain are less stable physiologically [6], but that treating pain during the NICU stay results in more robust responsiveness [12]. There are explanations that can be offered for these apparent discrepancies that include the gestational and postconceptional age of the infant at the time of exposure to pain, the intensity of the painful event (eg, circumcision is presumably more painful than heel lance), the timing of the specified outcome, the type of the specified outcome, as well as co-interventions such as use of

* Corresponding author. E-mail address: [email protected] (C.C. Johnston). 0095-5108/02/$ – see front matter D 2002, Elsevier Science (USA). All rights reserved. PII: S 0 0 9 5 - 5 1 0 8 ( 0 2 ) 0 0 0 2 0 - 9

396

C.C. Johnston et al. / Clin Perinatol 29 (2002) 395–414

nonpharmacologic pain-relieving strategies [13] and medications [12]. Although it is critical to obtain an emerging model of mechanisms controlling the effects of repeated pain on physiologic and behavioral outcomes in human infants, systematic studies are only possible to carry out in animals.

Why use animals in pain studies of infants? All human neonate studies are observational. The designs may be such that certain factors such as age are controlled for through matching or case-control methods, but for obvious ethical reasons, the studies must remain observational. There is little control over the timing, frequency, or intensity of the painful experiences the infants undergo because they are directly related to clinical care. Even the outcomes are related to clinical care (eg, responsiveness to later painful experiences such as immunization). The mechanisms underlying the observed outcomes can only be determined theoretically. With existing methods, there can be no empirical evidence to support hypothesized mechanisms that might be responsible for the observed changes. Infants who show decreased response to painful events several weeks following repeated painful events are different behaviorally than infants who have not experienced pain and who have more robust response [10], but what has transpired neurologically to result in these behavioral changes is unknown. It is important to understand the underlying mechanisms of long-term consequences of early exposure to repetitive pain for several reasons. Analgesics that are commonly used with infants have several effects—some of which are beneficial—such as better postoperative recovery [14] and better neurobehavioral development [15]; other effects are potentially detrimental, such as slow hepatic clearance with rebound [16,17]. By understanding the underlying mechanisms that link repeated pain to long-term sequelae, existing analgesics could be used more effectively and more targeted analgesics could be developed. If the underlying mechanisms are better understood, then other interventions aimed not just at decreasing pain but also at promoting development may be introduced with more confidence in conjunction with interventions aimed at decreasing the pain. Finally, in providing clear controlled data at a more mechanistic level, greater credence is attributed to observational human infant data. Thus, investigators interested in understanding the underlying mechanisms of the effects of repeated pain must turn to animal models. In using animal models, multiple mechanisms and effects can be studied simultaneously. Effect sizes tend to be large, as many extraneous variables can be controlled. Littermates are from specific strains of rodents, so that genetic variability is diminished. Animals are kept in well-controlled environments with less variability than a hospital intensive care unit, which decreases the sample size that is needed for a given effect. Because animal lifespan is short, from a human perspective, long-term effects can be seen in a relatively short time. For example, a rat is considered to be a fully mature adult at 60 postnatal days (P60).

C.C. Johnston et al. / Clin Perinatol 29 (2002) 395–414

397

Numerous sequential studies can be conducted rapidly, so many reports using animal models include several studies.

Why is the rat pup used to study consequences of neonatal pain? All reported studies have used rats to study pain response in preterm neonates. Although rodents are common and relatively inexpensive compared with other animals, there are specific reasons for selecting the rat as a species to study. Because the rat is born fairly immature compared with other mammals (sheep, guinea pig, and so forth) and, in particular, to the human infant at term, the postnatal period of the rat might best approximate some of the developmental patterns seen in human infants during the third trimester of gestation [18,19]. On this basis, it has been suggested that at approximately P10, the brain of the rat is at the stage of rapid brain growth, accelerated synaptogenesis, myelinization, and astrocyte proliferation characteristic of the human newborn at term. When electroencephalographic activity and traces reflecting active and quiet sleep changes are considered, the human infant of between 35 and 37 weeks’ gestation corresponds well to the rat of P10 to P13. Thus, the first week of a rat pups’ life can be used to model human preterm development [20]. Weaning in rats occurs at P21. At P35, the rat is considered to be an adolescent; at P60, a full adult [21]. It should be emphasized, however, that development of various brain regions within a single species is asynchronous and, thus, estimates of developmental stage of the brain based on generalized whole brain morphologic or functional measures are too simplistic. Because human behavior is more complex than rodent behavior, the extent to which results from animal studies can inform clinicians is limited. This is not to say that animal research is not important but the clinical relevance needs to remain tied to specific questions that can be validly answered by animal studies. The specific questions relate to the effects of peripheral injury of differing types and magnitude on the central nervous system (CNS), how long the effects last, how widespread the changes are (peripheral, spinal, supraspinal), and what mechanisms can block the change. Although animal studies have long provided the basis of understanding mechanisms of pain, it is only recently that long-term consequences of neonatal pain have been the subject of animal studies. Important earlier work with young animals has provided ways of studying pain in the rapidly developing fetuses. The paradigms that are used in studying long-term effects are presented below.

What are the paradigms used in animal studies and how do they relate to clinical experience? Several paradigms are used to expose animals to pain, which vary in the intensity and duration of effect. These include acute pain such as single or

398

C.C. Johnston et al. / Clin Perinatol 29 (2002) 395–414

repeated needlestick, inflammatory pain, cutaneous tissue injury, and nerve ligation. Some paradigms (eg, inflammatory pain) are used either as a stimulus for exposure or as an outcome. In all of the paradigms of acute or repeated exposure to pain, pups must be removed from their mothers during the procedure. This separation might introduce important confounds in the analysis of the data, as will be discussed later. Needlestick in neonates To mimic repeated heel lances in infants during their stay in the NICU, repeated pain can be induced in pups by needlestick using 28 or 25 G needles that are quickly inserted into either the dorsal or plantar surface of the hind paw [22 –24]. The animal often vocalizes at the time of the insertion, but there is rarely visible swelling or discoloration caused by this procedure. Thus, it is considered to be acute pain, not inflammatory pain. When needlesticks are inflicted four times a day over a week, there may be inflammation developing; however, no studies to date have clearly reported measures of inflammation other than casual observation of redness or swelling in this model. Whether or not human neonates who receive repeated heel lances experience only acute pain or, additionally, experience inflammatory pain will depend on the frequency of the heel lances, the type of lancet, the skills of the person making the lance, and the amount of blood required. Inflammatory pain Inflammatory pain can be induced through the injection of formalin [25], carrageenan [26], capsaicin [27], bee venom [28], or complete Freund’s adjuvant (CFA) [29]. Formalin produces pain that lasts approximately half an hour in young animals and an hour in adults [30,31], as determined by the occurrence of pain behavior. Carrageenan, on the other hand, produces inflammation up to 2 weeks following a single injection [32]. CFA and bee venom are stronger inflammatory agents, resulting in persistent inflammatory pain. In particular, CFA injection can result in long-term activation of immune responses, mimicking autoimmune disease [33]. In terms of relevance to clinical pain from repeated procedures, intravenous infiltrations, repeated tape removals, and the milder inflammatory agents such as repeated needlestick or formalin are probably the most adequate models. Tissue injury This procedure is described as a 2 mm  2 mm cut made in the surface skin at a single time point during development, which clearly causes peripheral nerve damage [34,35]. This model is used to examine nerve regeneration following tissue injury. This can be considered equivalent to incisions made in human neonates (eg, chest tube insertions).

C.C. Johnston et al. / Clin Perinatol 29 (2002) 395–414

399

Nerve ligation Nerve ligation [36] is a more extreme method of testing for the effect of tissue damage. In this paradigm, a nerve is surgically severed. This type of damage is more relevant for neuropathic pain, a pain modality that has not been reported in infants.

What are the most commonly used outcome measures to assess the consequences of early pain? In examining the effects of early exposure to repetitive pain, various outcomes are used, including tests of pain sensitivity or mechanical sensitivity, evaluation of stress behaviors, or determination of neural substrates susceptible to modulate pain responses on fixed postmortem tissues. Among these outcomes, the most common is pain in response to different sorts of pain stimuli: thermal, mechanical, or inflammatory. The responses may be indicators of spinal or supraspinal involvement. Reflexive responses such as paw withdrawal in a thermal stimulation are indicative of spinal cord involvement, whereas more complex responses such as licking paw, or self-grooming after formalin injection involve supraspinal mechanisms. The implication for clinical practice is important because behavioral differences seen in survivors of NICUs involve almost always supraspinal and CNS-mediated mechanisms. Tests of sensitivity Thermal sensitivity Thermal sensitivity is commonly used in animal research because the various tests are relatively easy to administer, do not require sophisticated equipment, and the intensity of the stimulus can be well calibrated. A brief description of the hot plate test, the tail flick test, and the Hargreaves test is presented: the hot plate test involves placing the rat’s paws—one or all of them—on a hot (52
C) metal surface and measuring the latency for the animal to lift the paw (in general, the rear paw) from the hot plate. The time to paw withdrawal is referred to as the latency or threshold, with shorter times reflecting heightened pain sensitivity. The tail flick test is similar to the hot plate test, but in this instance, the animal’s tail is immersed in hot (40
C) water. Latency to remove the tail from the water (flick) is taken as pain threshold. The Hargreaves test is a variant of the hot plate test but has a better sensitivity because not all four paws are exposed to the heat. In this case, the animal is placed in a Plexiglas box and a light beam (50
C) is shone on the paw until it is withdrawn [37]. All of these responses are reflexive, and all of these tests require that the animal be accustomed to the testing environment, that is, the Plexiglas box for the Hargreaves test or the hot plate apparatus. Testing is less reliable with young animals (< P7), in that they must be able to stand relatively still so that the reflex is able to be clearly differentiated from other random movements.

400

C.C. Johnston et al. / Clin Perinatol 29 (2002) 395–414

Mechanical sensitivity The cutaneous flexor reflex (CFR) is elicited by mechanical stimulation through applying graded von Frey hair filaments against specific areas of the body with enough pressure to make the filament bend. When used on the foot (either plantar or dorsal surface), the response is dorsiflexion. A von Frey hair filament weight threshold at which the response is first elicited can thus be determined. Lower weights eliciting the reflex are reported as lower thresholds and indicate greater excitability or sensitization of the mechanical response. The filaments range from 0.08 g to 280 g, with weight intervals of 1.7 g so that differences in threshold are expressed in grams and yield continuous data. Fitzgerald et al’s [20] seminal work comparing the developmental changes in response to CFR between humans and rats provided a model of mechanical sensitivity that appears to be exquisitely sensitive to tissue damage [38]. Although this measure seems to be sensitive to tissue damage [34], it does not seem to be sensitive to inflammatory pain [32]. The CFR is recruiting pathways involved with sensory inputs and, although not a painful stimulus in itself, this test has been demonstrated to have validity to determine sensitivity to analgesics in adult rats [39,40]. In addition, the pathways involved in eliciting the response to CFR are the same as those involved in pain perception. In adult rats, the CFR is elicited by triggering C fibers, but in younger animals, it can be elicited by exciting A fibers, in particular, A-b fibers [38]. This observation suggests that in young animals, pathways additional to those modulating pain signals are specifically recruited in the CFR test. Electrophysiologic monitoring of sensory neurons resulting from brush touch and pinch has also been reported [29], with greater activity reflecting greater sensitivity. Both measures of mechanical sensitivity—the CFR and electrophysiologic monitoring—are reflexive. Inflammatory pain The formalin test has become well established as a model response to inflammatory pain and is particularly interesting in studies of human infant pain for two reasons. First, it involves supraspinal mechanisms, so any changes seen in response to this test will implicate higher CNS mechanisms. Second, the method of testing is well validated with different ages, including neonatal rats [30,31,41]. In this paradigm, formalin of either 2.5%, 5%, or 10% is injected into the plantar surface of the hind paw. The animal is then immediately placed into a Plexiglas box with a mirror under it so the paws can be well visualized. Every minute, specific pain behaviors and general behaviors are scored on well-defined scales [41,42]. The specific pain behaviors in descending order of severity are licking or shaking the paw, lifting the paw, and protecting the paw. In rats older than 35 days, there is a biphasic response to formalin injection, with an initial phase occurring immediately after the injection, followed by a quiescent period of normal behavior (absence of pain behavior) lasting approximately 10 minutes, and a second phase during which the animal exhibits pain behavior. This second phase is more sensitive to interventions compared with the initial phase of pain behavior. Neonatal rats aged 15 days post partum show only the first phase, with

C.C. Johnston et al. / Clin Perinatol 29 (2002) 395–414

401

no delayed pain responses [41]. This suggests that mechanisms mediating pain responses in developing rats are still maturing. Other behaviors such as locomotion and rearing (active, mobile) can be compared with lying or sitting immobile as indices of general arousal or behavioral activation. Stress behaviors Some studies have looked beyond pain behaviors specifically and included behaviors related to anxiety along with levels of stress hormones because pain is a stressor [43 – 45] and because an underlying assumption about long-term consequences of early exposure to repetitive pain is that it would lead to heightened vigilence and exaggerated responsiveness to stressors. Paradigms typically used to test for stress responsiveness are used to evaluate the outcomes of repeated neonatal pain. The stress paradigms that have been used in studies of developmental pain in animals include psychologic stressors such as the air puff startle, exposure to an open field test, and social discrimination and more physical stressors such as brief exposure to ether vapors. Usually, secretion of stress hormones like glucocorticoids or corticotropin following exposure to these paradigms reflect activation of brain centers processing sensory, emotional, and physical inputs. A large body of literature produced in the last 15 years has demonstrated the long-term effects of fetal and neonatal stress on stress responsiveness and coping strategies in adult animals (for review, see [18]). Only a few studies, however, have determined the effects of pain as a stressor on further stress responsiveness at neonatal, juvenile, and adult age. In the air puff startle test, the animal is subjected to a sudden puff of air that is directed to its face, causing the animal to startle. Stress hormones are determined before and immediately after (5 minutes) exposure to this psychologic stressor. In the open field test, the animal is placed in a brightly lit open arena divided into squares and with no bedding to contrast with the home cage. Rats are allowed to explore the novel environment for determined amounts of time, and the exploratory trajectory and stress hormones are measured prior to and after exploration. The latency to explore the center of the field is an indication of how secure the animal is in its new environment. In the social discrimination test, a new animal is introduced to the cage of the subject animal. The new animal is removed and then reintroduced an hour later. Animals that have heightened anxiety will have retained the odor of the new animal, indicating that they are still vigilant about the revisiting of the strange animal. In the ether test, the animal is exposed for a brief period of time (1– 3 minutes) to ether vapors, which produces anesthesia. Hormonal responses to this stressor are generally high compared with the psychologic stressors described above. These four stress paradigms test the animal’s ability to respond to new challenges, and the resulting physiologic responses (hormonal) that are measured indicate whether potential dysregulation of the CNS is observed in animals exposed to repeated pain during the neonatal period. Behavioral responses to these stressors provide information about the changes in arousal level (vigilence) or anxiety that occur as a result of the neonatal manipulations. Another behavioral indicator of stress is represented by the propensity of rats to

402

C.C. Johnston et al. / Clin Perinatol 29 (2002) 395–414

Table 1 Animal studies related to long-term consequences of early exposure to repetitive pain Authors

Year

Sample

Purpose

Pain model

Alvares et al [32]

2000

Sprague-Dawley rats (P0)

Isolate factors contributing to hypersensitivity following peripheral injury

1. Hind paw injury of 2 mm  2 mm incision or anesthetized controls 2. Carrageenan injection or vehicle control into plantar surface of hind paw

Anand et al [22]

1999

Sprague-Dawley rats (P0 – P7)

Examine long-term effects of repeated heelsticks

25 G needlestick one, two, or four times daily (N1, N2, N4) or touched one, two, or four times daily (T1, T2, T4)

Bhutta et al [25]

2001

Long-Evans Measure long-term rat pups (P1 – P7) behavioral effects of prolonged inflammatory pain in neonatal pups

Daily injections of 10% formalin into paws: . all four paws (P1) . forepaws (P2,P4,P6) . hind paws (P3,P5,P7)

C.C. Johnston et al. / Clin Perinatol 29 (2002) 395–414

Paradigm

Outcomes

. Inflammatory . pain at birth and at 6 to 8 wk . . Wounded versus nonwounded tissue . Treatment with anti-NGF

RIA of cutaneous innervation Mechanical threshold (vFh)

Results

Conclusions

. Reinnervation begins at 3 d, aberrant growth at 5 d, and hyperinnervation at 7 d . No effect of anti-NGF . Inflammation following single injection of carageean  2 wk, but no effect on pain threshold . At 6 to 8 wk, response to inflammatory pain same in controls and carageean-treated . N2 weighed less than T2 . HP latencies decreased in N4 compared with T4 at P16 and P22 . N4 had greater preference for alcohol, especially in females . N4 socially anxious . Fos expression higher in T4 30 min after HP exposure

No lasting effects found from inflammation given at birth lasting 2 weeks’ duration

. Weight at P8, P15, and P21 . Pain threshold (HP) at P16 and P22 . Alcohol preference over 24 h  3 . Defensive withdrawal in open field test . Social discrimination . ACTH response to air puff startle . Fos expression in somatosensory cortex Control versus . Body weight (P67) . TF latencies longest in morphine . Alcohol pain versus group, significant preference morphine for males only versus pain plus (P66 – P68) . HP latencies . Pain thresholds morphine greatest for pain (TF: P74 – P79; group; difference HP: P94 – P95) in effect of morphine . Locomotor seen only in males activity (P81 – P83; after amphetamine . Controls of both sexes P90 – P92) preferred ethanol Touch versus pain

403

Comments

Long-term consequences of acute repetitive pain result in decreased pain thresholds

Repetitive acute pain in neonatal period increases sensitivity later

Long-term consequences of inflammatory pain result in increased thresholds, decreased alcohol preference, locomotion, and weight gain

Inflammatory pain in neonatal period may cause greater depressive changes than repetitive acute pain

(continued on next page)

404

C.C. Johnston et al. / Clin Perinatol 29 (2002) 395–414

Table 1 (continued ) Authors

Year

Sample

Purpose

Pain model

deLima et al [35]

1999

Sprague-Dawley rats (P0)

Test effect of spinal block on hyper-innervation following skin wound

Spinal anesthesia of 0.25% bupivacaine versus saline prior to injury and at 6 h and 24 h post injury versus no injury controls

Johnston and Walker [24]

2000

Sprague-Dawley rats (P0 – P7)

Measure long-term behavioral effects of repeated needlestick

. 28 G needle four times daily (P1 – P8) (HP) . Removed from dam four times daily (handled) . Unhandled

Lee and Chung [36]

1996

Sprague-Dawley rats (P7, P14, P21)

Determine whether peripheral nerve injury results in neuropathic pain

Ligation of L4,L5, and L6 at 1 wk, 2 wk, or 3 wk; sham surgery on controls

Liu et al [27]

in press Long-Evans rat pups (P14 or P21)

. Hourly injections Search for higher (forebrain) mechanisms of 0.01% capsaicin, or saline to explain long-term . Touched or behavioral conseuntouched controls quences of early exposure to pain

C.C. Johnston et al. / Clin Perinatol 29 (2002) 395–414

Paradigm

Outcomes

405

Results

Conclusions

Comments

. Controls showed more locomotion than other groups in males only . Body weight was greater in controls or morphine groups Hyperinnervation and hypersensitivity compared with controls in wounded pups, which was equal in those treated with spinal block or saline . Unhandled: higher thresholds and less pain response than both H and HP . Maternal grooming highest in HP group . Younger pups (P15) show less pain behaviors . No sex differences observed Hypersensitivity greatest when injury made at 3 wk

Morphine alone or given prior to inflammatory pain increases threshold latencies but does not effect locomotion Spinal analgesia is not effective in preventing long-term consequences of tissue injury

Different alcohol preference response [22]

Tissue injury

. RIA of cutaneous innervation . Mechanical threshold vFh at P7

Acute repeated pain versus maternal separation

. Inflammatory pain response (formalin test) . Thermal threshold (Hargreaves test) . At P15,P35, and P60

Nerve injury

. vFh of two weights applied  10 . % withdrawals from P14 to 5 wk post ligation

Inflammatory pain

Uncoupling of . Lower thermal . Basal and forebrain opioid threshold in forskolin-stimulated receptors may capsaicin group AC activity be mechanism . AC activity . Inhibition underlying greater in of AC by increased capsaicin group DAMGO nociception . Inhibition of . Naloxone AC activity by reversibility DAMGO reduced . Pertussis in capcaisin group sensitivity and was not modified . Thermal by naloxone threshold

Clinically relevant to practice

The effect of four needlesticks may be blunted by maternal grooming Separation may be as important as pain

Only center to account for separation from mother

Critical periods exist and may be related to development of C fibers.

Different use of vFh; testing ‘‘impossible’’ before P15, thus there may be neuropathic pain earlier than 2 wk Novel mechanisms studied and identified

(continued on next page)

406

C.C. Johnston et al. / Clin Perinatol 29 (2002) 395–414

Table 1 (continued ) Authors

Year

Sample

Purpose

Rahman et al [26]

1997

Male rat pups(P1, P7) (strain not reported)

Effects of morphine on Carrageenan (in hind inflammatory pain paw) plus morphine (s/c) versus carrageenan plus saline versus saline plus morphine versus saline plus saline.

Reynolds and 1995 Fitzgerald [34]

Sprague-Dawley rat pups (P0 – P21)

Describe response of local nerve terminals over time to neonatal skin injury

2 mm circle incised in dorsal surface of hind paw at P0, P7, P14, or P21

Ruda et al [29]

Sprague-Dawley male rat pups(P0, P1, P3, P14)

Examine impact of neonatal tissue injury and pain on development of nociceptive neuronal circuitry

CFA versus saline, versus untouched at P0, P1, P3, and P14

2000

Pain model

C.C. Johnston et al. / Clin Perinatol 29 (2002) 395–414

Paradigm

Outcomes

1. Inflammatory . pain treated with morphine versus control 2. Morphine in presence of pain or not

Single-unit extracellular recordings from C, A-delta, and A-beta fibers in dorsal horn neurons in response to stimuli (touch, pinch) at 8 wk alone and with intrathecal morphine in incremental doses . Windup and input of dorsal horn neurons calculated

Tissue injury . Nerve sprouting plus inflam(C and A-delta mation at fibers) determined different ages by immunocytoversus controls. chemistry at injury and 7, 14, 21, 42 d or 12 wk post injury . Mechanical threshold (vFh)

Persistent hind paw inflammation

. Sciatic nerve afferents and DRG population (NGF) via immunohistochemistry (CGRP labeling) . Behavioral response to thermal pain (Hargreaves test)

407

Results

Conclusions

Comments

. Morphine at P1 without pain reduced the effect of morphine at 8 wk . Inflammatory pain alone has no effect on morphine effectiveness but does diminish the reversal effect of naloxone . Morphine in the presence of inflammatory pain does not change later efficacy of morphine . P0 wounds lead to greatest hyperinnervation . P14 and P21 similar to adult wound . At P7, sprouting shows recovery by day 21, but P0 never shows recovery . Sympathetic neurons remain normal, but A-delta and C fibers are over hyperinnervated wound area . Sensory thresholds lowered by P5 in pups injured at P0 . Increased hyperinnervation in sciatic afferents . Increased number of large fibre DRG axons in P1 CFAtreated animals . Initial thermal pain latencies equivalent, but when CFA

Morphine given in the absence of pain in early life is less effective in inhibiting pain later

Possible direct clinical implications

Important The younger study the age at the documenting time of injury, susceptibility the more to long-term profound the hyperinnervation consequences of tissue of pain fibers; this can be seen injury in in the reflection youngest population of decreased sensory thresholds

Persistent inflammatory pain in neonatal period has widespread long-lasting effects, greater for the younger ages

First study to show changes in CNS CFA can cause severe tissue necrosis

(continued on next page)

408

C.C. Johnston et al. / Clin Perinatol 29 (2002) 395–414

Table 1 (continued ) Authors

Year

Sample

Purpose

Pain model

Walker et al [23]

1998

Sprague-Dawley rats (P1 – P7)

Examine long-term stress response to repeated neonatal pain

28 G needle once daily (P1 – P8) (HP), removed from dam once daily handled, unhandled

Abbreviations: AC, adenylate cyclase; ACTH, corticotropin; CFA, complete Freund’s adjuvant; CGRP, calcitonin gene-related protein; DAMGO, [D-Ala(z), N-Me-Phe(4), Gly(5)-ol]-enkephalin; DRG, dorsal root ganglia; HP, hot plate test; NGF, nerve growth factor; P, postnatal day; RIA, radioimmunoassay; TF, tail flick test; vFh, von Frey hair filament.

consume alcohol. Rats usually do not drink alcohol unless under conditions of chronic stress or as consequences of prenatal/postnatal stress. In this test, animals are offered a choice of water or alcohol to drink and their preference for alcohol is recorded over 24 hours. Animals who prefer alcohol are considered abnormal. Examining neural structures through immunocytochemistry and immunohistochemistry There exists an enormous body of literature on the approaches to studying neural, including neurotransmitter, changes in response to painful stimuli; however, a complete review is beyond the purposes of this article. The paradigms of pain tested range from the most severe (nerve ligation) to acute inflammatory. The interest to clinicians in these studies is that they can elucidate the extreme plasticity of the nervous systems in early life [43,46 –49]. Normal innervation moves from functionally polymodal neurons to more specific functioning and, finally, apoptosis or programmed cell death [47,50]. In regeneration after nerve damage, nerve growth overcompensates for the damage and sprouts abnormally, resulting in hyperinnervation both in the surrounding tissue and in the dorsal root

C.C. Johnston et al. / Clin Perinatol 29 (2002) 395–414

Paradigm

Acute pain repeated versus maternal separation on stress response

Outcomes

Results

Conclusions

. Inflammatory pain (formalin) . Dorsal horn response to touch and pinch at 8 to 12 wk

administered to adults, decreased thresholds in P1 CFA group . Earlier onset of late-phase formalin response in CFAtreated group . Greater response in neurons to touch and pinch in CFA-treated group of all ages . Unhandled: shortest open field latency . Unhandled: lowest cortisol response to ether test . Open field latency longer in HP group . Maternal grooming significant covariate, blunts difference between handled and HP group

Inhibits normal neuronal pruning of nerves and thus decreases inhibitory mechanisms

. Anxiety tests (open field test, air puff startle, ether test) . Cortisol at P15

409

Comments

Maternal grooming blunts stress response due to repeated pain

ganglion [46,51]. The growth pattern can be observed through staining or marking of specific fibers via immunocytochemical methods. Targeting specific receptors such as opiate receptors or amino acid receptors (eg, N-methyl-Daspartate, g-aminobutyric acid) is also possible (see [48,52]). After tissue is sliced, frozen, and labeled, there can be an actual count of receptors or of neurons. Comparisons are made with controls, and significant differences can be visualized (eg, see [53]). Depending on differences in the presence of various receptors, gene expression, neurons, or neuropeptides, the implications of the findings will vary. What is clear is that hyperinnervation is abnormal and is associated with increased pain behaviors such as allodynia or hypersensitivity [54 –56].

What do animal studies tell us about the long-term consequences of repeated pain? Table 1 presents animal studies related to long-term consequences of early exposure to repetitive pain. As can be noted from this short summary of results,

410

C.C. Johnston et al. / Clin Perinatol 29 (2002) 395–414

few findings are consistent across studies, possibly because different pain paradigms, outcomes, and ages were employed. The finding that tissue injury results in neural reorganization in the periphery or spinal cord, however, stands out as being relatively consistent across experimental paradigms. Specifically, abnormal sprouting of primary afferents, hyperinnervation of wounded cutaneous tissue [35,53], and hyperexcitability of dorsal root ganglia [26] have been observed in paradigms involving tissue injury. Although it would be helpful to have these findings replicated by several different laboratories, the use of the tissue injury model with varying the outcome measures would strengthen the validity and consistency of this finding. This neural reorganization may lead to behavioral consequences because tissue injury also results in behavioral hypersensitivity [22], although this was only partly supported in another study where changes in maternal behavior due to the experimental procedure were taken into account [24]. Because much of the development of the CNS occurs postnatally in the rat, the age at which painful procedures or tissue injury takes place is critical in order to determine the magnitude, duration, and/or direction of the effects. Indeed, it appears that there is heightened sensitivity to tissue injury at earlier ages compared with adulthood [57]. It would be logical, therefore, to hypothesize that heightened sensitivity at earlier developmental stages would make younger infants more susceptible to long-term consequences. The study by Reynolds and Fitzgerald [34] supported this idea, although this hypothesis has been challenged by Lee and Chung [36], whose study showed greater potential for long-term effects of nerve ligation when it occurred between the second and third postnatal week versus after 1 week [36]. It has to be noted, however, that important differences in the effect of nerve ligation as opposed to cutaneous tissue injury could explain these conflicting results. Results on studies of early exposure to inflammatory pain are also somewhat conflicting. It seems that in contrast to acute tissue injury (ie, nerve ligation), which leads to hyper-responsiveness to pain, long-term effects of early inflammatory pain may lead to decreased responsiveness [25], no change [32], or increased sensitivity [29]. The differences between the various studies might be related to the agent used to produce inflammation: Bhutta et al [25] used daily formalin injections, Alvares et al [32] used a single dose of carrageenan, and Ruda et al [29] used CFA. One issue to consider when reviewing animal studies is how closely the paradigm mimics the clinical situation or findings. Until we have access to more precise indicators of the degree of inflammation in infants or animals, it will remain difficult to interpret the results of animal studies that use different agents to produce varying degrees of inflammation and even more difficult to match the degree of inflammation found in the clinical setting. Because repeated injury such as heelstick can result in inflammation, the need to quantify inflammation precisely is particularly important in order to know whether the clinical situation we wish to mimic is acute pain, tissue injury, or inflammatory pain. The paradigm of pain studied may also be affected by the various strains of animal used. Indeed, strains vary in their thermal sensitivity after CFA-induced

C.C. Johnston et al. / Clin Perinatol 29 (2002) 395–414

411

arthritis, as shown by Lariviere et al [58] for Lewis and Fisher rats, and research out of Mogil and colleagues’ laboratory demonstrates clear differences between mice strains in sensitivity to different types of pain [59 – 61]. Although there is no evidence to date that there are genetic effects in human infant responses, the evidence of strain-related differential responses to pain make it imperative that the same-strain rodent be used across studies. The studies on longterm consequences of pain in rat neonates have mostly used Sprague-Dawley or Long-Evans rats, and one study [26] does not report strain. There are no reports on potential differences in pain response between Sprague-Dawley and Long-Evans rats. Another critical factor to take into account when designing studies on neonatal pain is the introduction of significant changes in maternal behavior. Earlier data from the authors’ laboratory have demonstrated that pups subjected to pain emit more distress vocalizations (ultrasonic vocalizations) than pups subjected to handling only. These ultrasonic vocalizations serve to alert the mother and direct her behavior toward the pups displaying distress and, in fact, pups subjected to pain receive greater amounts of maternal grooming compared with handled pups [24]. Maternal care (and pup grooming more specifically) is well recognized to play a critical role in the physiologic and behavioral development of the young [62]. Elegant work by Blass et al [63] demonstrated the blunting effect of maternal behavior on pain response in infant rats, implying that changes in maternal behavior induced by the pain procedure might be a critical factor in altering pain sensitivity. Although the human infant is somewhat less dependent than the rat on sensory stimulation for basic functions such as urination, studies conducted in Romanian orphanages demonstrate the long-term sequelae of sensory deprivation [64]. In studies of pain in neonatal rats, the effects of maternal separation and changes in pup grooming was accounted for only in studies from the authors’ laboratory [23,65]. Given the robust finding that pain induced significantly more grooming of the pups and reduced stress responses in the open field, it would be important to take this variable into account in future studies on pain in neonatal rats. Future animal research should use existing paradigms to determine whether findings are replicated. Efforts should be made to standardize methods so that comparisons can be made across studies. Negative results are as important as results with significant differences, particularly in an emerging field.

Summary Although animal models will never match the complexity of human systems, a number of basic mechanisms can be accessed only by using animal models. Results from studies using animal models of pain can give insight into basic mechanisms underlying long-term consequences of pain and provide sufficient data to generate hypotheses to be tested in human infants. Interaction between clinicians and basic scientists, with an understanding of the domain in

412

C.C. Johnston et al. / Clin Perinatol 29 (2002) 395–414

which each group is working, is critical to the meshing of efforts from these domains. With collaboration between these groups, more relevant research can be conducted that can lead to the decrease in pain and its consequences in neonates.

References [1] Anand KJ, Selankio JD. SOPAIN Study Group. Routine analgesia practices in 109 neonatal intensive care units (NICUs) [abstract]. Pediatr Res 1996;39:192A. [2] Johnston CC, Collinge JM, Henderson S, Anand KJS. A cross-sectional survey of pain and analgesia in Canadian neonatal intensive care units. Clin J Pain 1997;13:1 – 5. [3] Porter FL, Wolf CM, Gold J, Lotsoff D, Miller JP. Pain and pain management in newborn infants: a survey of physicians and nurses. Pediatrics 1997;100:626. [4] Grunau R, Whitfield MF, Petrie JH. Pain sensitivity and temperament in extremely lowbirthweight premature toddlers and preterm and fullterm controls. Pain 1994;58:341. [5] Grunau RE. Long-term consequences of pain in human neonates. In: Anand KJS, Stevens BJ, McGrath PJ, editors. Pain in neonates. Amsterdam: Elsevier; 2000. p. 55. [6] Oberlander TF, Grunau RE, Whitfield MF, Fitzgerald C, Pitfield S, Saul JP. Biobehavioral pain responses in former extremely low birth weight infants at four months’ corrected age. Pediatrics 2000;105:e6. [7] Porter FL, Grunau RE, Anand KJ. Long-term effects of pain in infants [review]. J Dev Behav Pediatr 1999;20:253. [8] Fitzgerald M, Millard C, McIntosh N. Cutaneous hypersensitivity following peripheral tissue damage in newborn infants and its reversal with topical anaesthesia. Pain 1989;39:31. [9] Taddio A, Goldbach M, Ipp M, Stevens BJ, Koren G. Effect of neonatal circumcision on pain responses during vaccination in boys. Lancet 1995;345:291. [10] Johnston CC, Stevens BJ. Experience in a neonatal intensive care unit affects pain response. Pediatrics 1996;98:925. [11] Johnston CC, Stevens BJ, Franck LS, Jack A, Stremler RL, Platt R. Factors explaining lack of response to tissue damage in preterm neonates. J Obstet Gynecol Neotatal Nurs 1999;28:587. [12] Grunau RVE, Oberlander TF, Whitfied MF, Fitzgerald C, Lee SK. Demographic and therapeutic determinants of pain reactivity in very low birth weight neonates at 32 weeks’ postconceptional age. Pediatrics 2001;107:105 – 11. [13] Stevens BJ, Johnston C, Franck L, Petryshen P, Jack A, Foster R. The efficacy of developmentally sensitive interventions and sucrose for relieving procedural pain in very low birth weight neonates. Nurs Res 1999;98:35. [14] Anand KJS, Hickey PR. Halothanemorphine compared with highdose sufentanil for anesthesia and postoperative analgesia in neonatal cardiac surgery. New Eng J Med 1992;326:1. [15] Anand KJ, Barton BA, McIntosh N, Lagercrantz H, Pelausa Young TE, Vasa R. Analgesia and sedation in preterm neonates who require ventilatory support: results from the NOPAIN trial. Neonatal Outcome and Prolonged Analgesia in Neonates [published erratum appears in Arch Pediatr Adolesc Med 1999;153(8):895]. Arch Pediatr Adolesc Med 1999;153:331. [16] Lynn AM, Slattery JT. Morphine pharmacokinetics in early infancy. Anaesthesia 1987;66:136. [17] Koren G. Postoperative morphine infusion in newborn infants: assessment of disposition. J Pediatr 1985;107:963. [18] Walker CD, Plotsky PM. Corticosteroids, stress, and development. In: Pfaff D, Arnold A, Farbach S, Rubin R, editors. Hormones, brain, and behavior. San Diego (CA): Academic Press; 2002. p. 487. [19] Narsinghani U, Anand KJS. Developmental neurobiology of pain in neonatal rats. Lab Anim 2000;29:27. [20] Fitzgerald M, Shaw A, MacIntosh N. The postnatal development of the cutaneous flexor reflex:

C.C. Johnston et al. / Clin Perinatol 29 (2002) 395–414

[21]

[22] [23] [24]

[25]

[26]

[27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44]

413

a comparative study in premature infants and newborn rat pups. Dev Med Child Neurol 1988; 30:520. Dobbing J. Nutritional growth restriction and the nervous system. In: Davison AN, Thompson RHS, editors. The molecular basis of neuropathology. London: Edward Arnold Co.; 1981. p. 221 – 33. Anand KJS, Coskun V, Thrivikraman KV, Nemeroff CV, Plotsky PM. Long-term behavioral effects of repetitive pain in neonatal rat pups. Physiol Behav 1999;66:627. Walker CD, Rontoar L, Oates M, Sherrard A, Johnston CC. Stress response in neonatal rats following repeated exposure to pain [abstract]. Pain Res Manage 1998;3:58 – 9. Johnston CC, Walker CD. The effects of exposure to repeated minor pain during the neonatal period on long-term inflammatory and thermal pain responses. Abstract from the Canadian Pain Society Meeting. Banff, Alberta, Canada; 2000. Bhutta AT, Rovnaghi C, Simpson PM, Gossett JM, Scalzo FM, Anand KJ. Interactions of inflammatory pain and morphine in infant rats: long-term behavioral effects. Physiol Behav 2001;73:51. Rahman W, Fitzgerald M, Aynsley-Green A, Dickenson AH. The effects of neonatal exposure to inflammation and/or morphine on neuronal responses and morphine analgesia in adult rats. In: Jensen TS, Turner JA, Wiesenfeld-Hallin Z, editors. Proceedings of the Eighth World Congress on Pain. Seattle (WA): IASP Press; 1997. p. 783. Liu JL, Rovnaghi C, Anand KJS. Novel mechanisms for hyperalgesia in 21-day-old rats: uncoupling of forebrain opioid receptors. Anaesthesiology, in press. Lariviere WR, Melzack R. The bee venom test: a new tonic-pain test [see comments]. Pain 1996; 66:271. Ruda MA, Ling Q-D, Hohmann AG, Peng BP, Tachibana T. Altered nociceptive neuronal circuits after neonatal peripheral inflammation. Science 2000;289:628. Abbott FV, Franklin KBJ, Westbrook RF. The formalin test: scoring properties of first and second pain responses in rats. Pain 1995;60:91. Guy ER, Abbott FV. The behavioral response to formalin in preweanling rats. Pain 1992; 51:81. Alvares D, Torsney C, Beland B, Reynolds M, Fitzgerald M. Modelling the prolonged effects of neonatal pain. Prog Brain Res 2000;129:365. Billiau A, Matthys P. Modes of action of Freund’s adjuvants in experimental models of autoimmune diseases [review]. J Leukoc Biol 2001;70:849. Reynolds ML, Fitzgerald M. Long-term sensory hyperinnervation following neonatal skin wounds. J Comp Neurol 1995;358:487. de Lima J, Alvares D, Hatch DJ, Fitzgerald M. Sensory hyperinnervation after neonatal skin wounding: effect of bupivacaine sciatic nerve block. Br J Anaesth 1999;83:662. Lee DH, Chung JM. Neuropathic pain in neonatal rats. Neurosci Lett 1996;209:140. Hargreaves K, Dubner R, Brown F, Flores C, Joris J. A new and sensitive method for measuring thermal nociception in cutaneous hyperalgesia. Pain 1988;32:77. Fitzgerald M, Jennings E. The postnatal development of spinal sensory processing [review]. Proc Natl Acad Sci USA 1999;96:7719. Willer J-C. Studies on pain. Effects of morphine on a spinal nociceptive fexion reflex and related pain sensation in man. Brain Res 1985;331:105. Willer JC. Nociceptive flexion reflexes as a tool for pain research in man. Adv Neurol 1983; 39:809. Teng CJ, Abbott FV. The formalin test: a dose-response analysis at three developmental stages. Pain 1998;76:337. Abbott FV, Ocvirk R, Najafee R, Franklin KB. Improving the efficiency of the formalin test. Pain 1999;83:561 – 9. Anand KJ. Effects of perinatal pain and stress [review]. Prog Brain Res 2000;122:117. Anand KJS, Hickey PR. Pain and its effects in the human neonate and fetus. New Eng J Med 1987;19:1321.

414

C.C. Johnston et al. / Clin Perinatol 29 (2002) 395–414

[45] Anand KJS, Carr DB. The neuroanatomy, neurophysiology, and neurochemistry of pain,stress, and analgesia in newborns and children. Pediatr Clin N Am 1989;36:795. [46] Fitzgerald M, Woolf CJ, Shortland P. Collateral sprouting of the central terminals of cutaneous primary afferent neurons in the rat spinal cord: pattern, morphology, and influence of targets. J Comp Neurol 1990;300:370. [47] Anand KJ, Scalzo FM. Can adverse neonatal experiences alter brain development and subsequent behavior? [review]. Biol Neonate 2000;77:69. [48] Hofer M, Constantine-Paton M. Regulation of N-methyl-D-aspartate (NMDA) receptor function during the rearrangement of developing neuronal connections [review]. Prog Brain Res 1994; 102:277. [49] NIMH. Proceedings from the National Institute of Mental Health Conference. Advancing research on developmental plasticity: integrating the behavioral science and the neuroscience of mental health. Chantilly, VA, May 12 – 15, 1996. [50] Coggeshall RE, Pover CM, Fitzgerald M. Dorsal root ganglion cell death and surviving cell numbers in relation to the development of sensory innervation in the rat hindlimb. Brain Res Dev Brain Res 1994;82:193. [51] Fitzgerald M, Shortland P. The effect of neonatal peripheral nerve section on the somadendritic growth of sensory projection cells in the rat spinal cord. Brain Res 1988;470:129. [52] Hori Y, Kanda K. Developmental alterations in NMDA receptor-mediated currents in neonatal rat spinal motor neuron [Ca2+] elevation in substantia gelatinosa neurons of neonatal rat spinal cord. Neurosci Lett 1996;205:99. [53] Reynolds M, Alvares D, Middleton J, Fitzgerald M. Neonatally wounded skin induces NGFindependent sensory neurite outgrowth in vitro. Brain Res Dev Brain Res 1997;102:275. [54] Ma QP, Allchorne AJ, Woolf CJ. Morphine, the NMDA receptor antagonist MK801 and the tachykinin NK1 receptor antagonist RP67580 attenuate the development of inflammationinduced progressive tactile hypersensitivity. Pain 1998;77:49. [55] Ma QP, Woolf CJ. The progressive tactile hyperalgesia induced by peripheral inflammation is nerve growth factor dependent. Neuroreport 1997;8:807. [56] Woolf CJ. An overview of the mechanisms of hyperalgesia [review]. Pulm Pharmacol 1995; 8:161. [57] Fitzgerald M, Beggs S. The neurobiology of pain: developmental aspects [review]. Neuroscientist 2001;7:426 – 57. [58] Lariviere WR, Chesler EJ, Mogil JS. Transgenic studies of pain and analgesia: mutation or background genotype? [review]. J Pharmacol Exp Ther 2001;297:467 – 73. [59] Lariviere WR, Wilson SG, Laughlin TM, Kokaueff A, West EE, Adhikari SM, et al. Heritability of nociception. III. Genetic relationships among commonly used assays of nociception and hypersensitivity. Pain 2002;97:75 – 86. [60] Mogil JS, Wilson SG, Bon K, Lee SE, Chung K, Raber P, et al. Heritability of nociception II. ‘‘Types’’ of nociception revealed by genetic correlation analysis. Pain 1999;80:83. [61] Mogil JS, Wilson SG, Bon K, Lee SE, Chung K, Raber P, et al. Heritability of nociception I: responses of 11 inbred mouse strains on 12 measures of nociception. Pain 1999;80:67. [62] Levine S, Haltmeyer GC, Karas GG, Denenberg VH. Physiological and behavioral effects of infantile stimulation. Physiol Behav 1967;2:55. [63] Blass EM, Shide DJ, Zaw-Mon C, Sorrentino J. Mother as shield: differential effects of contact and nursing on pain responsivity in infant rats evidence for nonopiod mediation. Behav Neurosci 1995;109:342. [64] Chugani HT, Behen ME, Muzik O, Juhasz C, Nagy F, Chugani DC. Local brain functional activity following early deprivation: a study of postinstitutionalized Romanian orphans. Neuroimage 2001;14:1290 – 301. [65] Walker CD. Chemical sympathectomy and maternal separation affect neonatal stress responses and adrenal sensitivity to ACTH. Am J Physiol 1995;268:R1281.