The Role of Activity in Developing Pain Pathways

Proceedings of the 10th World Congress on Pain, Progress in Pain Research and Management, Vol. 24, Edited by Jonathan O. Dostrovsky, Daniel B. Carr, and Martin Kaltzenburg, IASP Press, Seattle, © 2003.

Maria Fitzgerald and Suellen Walker

Department of Anatomy and Developmental Biology,
University College London, London, United Kingdom

The postnatal period is a critical time in the development of spinal sensory systems. It is a time of structural and functional reorganization of sensory connections accompanied by marked changes in expression of molecules, receptors, and channels associated with sensory transmission. (Alvares and Fitzgerald 1999; Fitzgerald and Jennings 1999). In addition, it is becoming increasingly evident that these postnatal events are dependent upon neural activity and that synaptic development requires defined patterns of afferent inpt (Ben-Ari 2002; Debski and Kline 2002; Fox 2002). Abnormal or excessive activity related to pain and injury in early life may therefore change the course of development and cause long-term changes in somatosensory and pain processing (Alvares et al. 2000; Anand 2000). Clinical studies suggest that early pain related to surgical and procedural intervention during intensive pain management of premature neonates can have long-term consequences upon pain behavior and perception in later life. (Porter et al. 1999; Grunau 2000). This chapter discusses clinical and laboratory evidence for these changes and outlines possible underlying mechanisms.


Preterm infants in intensive care can receive many invasive procedures, and adequate levels of analgesia are frequently hard to gauge (Anand and Porter 1998). Even the youngest preterm infant will display clear responses to noxious stimuli and tissue injury (Johnston et al. 1995; Fitzgerald and DeLima 2000). Several studies have addressed whether early pain experience, if excessive or repeated, may alter future pain responses. These responses have been divided into shorter-term changes (days and weeks) and longer term changes (months and years).


Several studies have investigated lasting pain and sensitivity in and around an area of injury in infants. They have been shown that even very young preterm babies are capable of displaying a prolonged cutaneous sensitization or hyperalgesia for days or weeks, when exposed to repeated painful stimulation such as heel lances (Fitzgerald et al. 1989). In addition, secondary hyperalgesia can be observed, for instance in the contralateral limb following local ischemic injury (Andrews and Fitzgerald 1994). The area around an abdominal surgical wound shows a similar enhanced cutaneous sensitivity in the postoperative period (Andrews and Fitzgerald 2002). A recent study shows that reactions (evidenced by grimacing or crying) to the pain caused by venipuncture on the forearm are also increased in full-term infants who have undergone repeated heel lances in the previous 24-36 hours, compared to control infants (Taddio et al. 2002). This effect, which extends to increased grimacing or crying even during non-noxious skin cleansing, suggests a spread of central sensitization well outside the area of direct injury.

Other studies have asked a rather different question: Does the overall experience of several weeks of intensive care alter pain responsiveness? These studies do not demonstrate an enhanced response. In fact, facial expression and autonomic and other biobehavioral measures show that pain responses in infants of 32 weeks postconceptional age who have undergone 4 weeks of repeated invasive procedures can be blunted compared to those of age-matched controls. (Johnston and Stevens 1996; Grunau et al. 2001).

The debate on how best to measure pain responses in such studies continues (Stevens and Franck 2001), and it is very clear that there is great individuality of response, even at very young ages (Franck et al. 2000); Morison et al. 2001). One advantage of measuring cutaneous sensitization and hyperalgesia is that they are relatively simple spinally mediated responses with direct parallels with laboratory studies (Fitzgerald and DeLima 2001). Interestingly, it appears that below 32 weeks, grimacing (facial action coding) and autonomic responses may also be mediated subcortically, because there is no difference in these responses to heel lance in intact and severely brain-injured infants (Oberlander et al. 2002).


While understanding the persistent effects of early pain over days and weeks is extremely important for adequate pain management and recovery, our interest here really lies in the potential for such stimuli to cause changes in responses lasting into childhood, adolescences, or adulthood. Very few data are available in this area, and what little information we have is hard to interpret. Grunau (2000) emphasized the complexity of such studies and the need for careful design and interpretation. As yet no studies have been directed toward sensitivity in and around the site of the injury itself as the child matures.

One important study shows that boys who have been circumcised at birth show increased pain responses to vaccinations at 4-6 months compared to those who have not (Taddio et al. 1995). In a follow-up, prospective study on 87 infant boys, uncircumcised infants were found to have the lowest pain scores at vaccination 4-6 months later, followed by those circumcised after treatment with lidocaine-prilocaine cream (EMLA), while those circumcised after placebo cream showed the greatest responses (Taddio et al. 1997).

There is evidence of persistent hypersensitivity following infant surgery. A follow-up study of infants 3 months after corrective surgery for unilateral hydronephrosis showed that the majority still displayed increased abdominal sensitivity compared with control infants of the same age (Andrews et al. 2003).

Again, the situation is somewhat different when examining preterm infants and the effects of multiple invasive procedures in intensive care. The blunting of pain responses after several weeks of intensive care, as described above, appears to have disappeared by the time the infants are 4 months old. Biobehavioral pain responses to blood collection by finger lance at 4 months were similar overall between former low birth weight infants and term-born controls (Oberlander et al. 2002). This finding appears inconsistent with the pioneering study of 195 toddlers at 18 months of varying birth weights from 480 g to over 2500 g, in which parents perceived the lowest birthweight groups to have the lowest pain responsiveness and where, unlike the case for toddlers of greater birth weight, there was no relationship between temperament and pain perception (Grunau et al. 1994a). Also in older children of 4.5 years, “somatization,” the occurrence of numerous pains that cannot be accounted for medically, was significantly greater in the lowest birth weight children. (Grunau et al. 1994b), although this finding was not observed at age 8—10 years (Grunau et al. 1998). Interestingly, these older children do rate pictures of painful events as more painful than do their peers of normal birth weight (Grunau et al. 1998).

All of these longer-term studies are hard to design due to confounding factors such as gestational age at birth (Grunau 1994a), length of intensive care stay (Johnston and Stevens 1996), intensity of the stimulus (Porter et al. 1999a), therapeutic management (Grunau et al. 2001), and parenting style (Grunau 1994b). In addition, children born preterm can have reduced cortical growth (Ajayi-Obe et al. 2000), reduced cognitive test scores, and increased incidence of attention deficit and hyperactivity disorder and other behaviors (Bhutta et al. 2002.) The older the child gets, the harder the studies are to conduct because learned patterns of behavior within families are major determinant of perceived sensitivity to pain (MacGregor et al. 1997).

Changes, if they occur, may be hard to detect in human subjects because the very plasticity that we are investigating may act elsewhere in the nervous system to compensate for the adverse effects of altered or excessive inputs. For instance, in contrast to the intense pain that occurs in adults, no chronic long-term pain follows brachial plexus avulsion at birth, and there is excellent restoration of sensory function and localization of restored sensation in avulsed spinal root dermatones following surgical repair (Anand and Birch 2002). This finding is consistent with the observation in rat pups that despite profound alteration of plantar hindpaw innervation induced by early nerve transection, cutaneous nociceptive impulses maintain an essentially normal spatial organization (Holmberg and Schouenborg 1996). In these cases of early nerve damage, an impressive reorganization must have taken place in the central nervous system to limit the sensory deficit.



A key step toward evaluating the impact of early pain upon future somatosensory processing is to establish animal models where tissue injury or repeated noxious stimuli applied at infancy lead to changes in adult sensory behavior. Such models have been reported, but the effects depend critically upon the nature of the injury.

Exposure to repetitive needle pricks to the paw four times a day in rat pups from post-natal (P) days P0 to P7 produces hypealgesia, in the form of reduced hot-plate latencies, at day P16, but this does not last into adulthood (Anand et al. 1999). The neonatally injured animals do, however, show an increased preference for alcohol and manifest other behavior changes as adults. On the other hand, a more severe injury of repeated 10% formalin injections into paws from P1 to P7 leads to hypoalgesia in adulthood, in the form of increased hot-plate and tail-flick latency. Preemptive morphine treatment before the first injection of formalin ameliorates the effects in males only, but the situation is complicated by the fact that morphine alone at birth increases adult tail-flick latencies more severely than does an injury. In this model, neonatal pain and morphine decrease alcohol preference (Bhutta et al. 2001). The difference of these results may be explained by the fact that formalin is a more damaging stimulus that may lead to some sensory neuron death (Tsujino et al. 2000), whereas repeated needle prick will presumably produce only local inflammation.

The longterm effects of an injection of inflammatory agents at birth is a subject of some controversy and clearly depends on the agent and dose used. Nociceptive thresholds fall in neonatal animals within a few hours of an inflammatory lesion (Jiang and Gebhart 1998; Marsh et al. 1999), but the question here is whether the effects last beyond the resolution of the peripheral damage. One report describes hindpaw injection of 0.25% carrageenan in newborn rat pups causing mechanical and heat hypoalgesia in 60-day-old adults while increasing the hyperalgesia produced by an injection of complete Freund’s adjuvant (CFA) (Lidlow et al. 2001). These effects were reversed if the neonatal inflammation was accompanied by four injections of sciatic bupivacaine nerve block, lasting up to 9 hours after the inflammation. However, in two other separate blinded studies, no changes in mechanical or heat responses were observed in adults following neonatal carrageenan injection (Alvares et al. 2000); Walker and Fitzgerald 2002). Despite the profound and long-lasting inflammation (14 days) produced by a single injection of 10 µL 2% carrageenan in the neonatal hindpaw, no difference was found in the three groups (neonatal carrageenan, saline, and anesthetic only) in mechanical or heat thresholds (either between left and right paws or between inflamed and control groups) at any stage tested. In addition, the reapplication of 2% carrageenan or CFA in these rats, when they had reached maturity, caused normal inflammatory, hypealgesic, and allodynic responses that did not differ from that of controls. A stronger inflammatory agent (25 µL CFA) injected into the hindpaw at birth also leaves adult baseline thermal withdrawal latencies unchanged, but following a second challenge with CFA, hyperalgesia increases very slightly and the time course but not the magnitude of the formalin response is altered (Ruda et al. 2000). The volume and doses of inflammatory agents used in infant rats for such studies must be considered with care. Injection of 25 µL CFA in a newborn rat leaves the paw swollen into adulthood (Walker and Fitzgerald 2001) and thus is clearly not an injury that resolves itself in the neonatal period.

Another type of injury that has been investigated is a full-thickness skin wound in the hind paw of the newborn rat pup, which heals rapidly. A long-lasting hypersensitivity, in the form of lowered von Frey mechanical threshold, persists in the previously injured region (Reynolds and Fitzgerald 1995; DeLima et al. 1999). The effect is apparent after one week and persists for at least six weeks (DeLima et al. 1999); it only occurs if the wound is made in the postnatal period (Reynolds et al. 1997). In this case, local sciatic nerve block with bupovacaine for the first 24 hours after wounding did not affect the onset or magnitude of mechanical hypersensitivity.


While changes in animal behavior are important, they are prey to almost as many confounding factors as human studies. There is considerable advantage in looking for direct effects of early injury upon the development of neural connections. Analysis at the cellular level can reveal changes in sensory connections that are not evident in behavioral tests.

Early peripheral inflammation has a transient influence on the postnatal development of rat primary sensory neuron subtypes (Beland and Fitzgerald 2001). This phenomenon has been studied using calcitonin gene-related peptide (CGRP) to label peptide-containing and IB4 to label non-peptide-containing nociceptive neurons and using NF200 to label larger non-nociceptive neurons. Following carrageenan inflammation in rats on day P1, the normal rise in IB4-positive binding on the dorsal root ganglion occurs earlier but is the same as controls by day P21. The CGRP-positive population increases at 2 and 6 days after carrageenan, because of an increase in both small CGRP/IB4 and larger CGRP/NF200 double labeled cells, but again, is normal by 3 weeks. The effects are different from adult inflammation, where carrageenan causes a transient increase in CGRP/IB4 cells only (Beland and Fitzgerald 2001).

There is a clear acute effect (within 2-5 hours) of carrageenan inflammation upon the properties of dorsal horn cells, including increased spontaneous activity, evoked responses, and A-fiber sensitization, although the pattern of effects is age dependent (Torsey and Fitzgerald 2002). These short-term effects do not last beyond the acute inflammation, however, and in parallel with the behavioral results described above, the dorsal horn properties 6 weeks later do not differ from controls (C. Torsney and M. Fitgerald, unpublished transcript).

CFA injections on the neonate, on the other hand, do cause long-lasting changes in spinal circuitry that can be observed in the adult. These consist of expanded central C-fiber terminal fields and CGRP expression and increased fos-like immunoreactivity, a measure of neuronal activity in dorsal horn neurons (Ruda et al. 2000; Tachibana et al. 2001). These central changes must be viewed in the context of very severe damage caused by large volumes of inflammatory agents that may have neuropathic and systemic consequences and that may produce an inflammatory response that lasts into adulthood (Walker and Fitzgerald 2002).

Of all the early models of pain, skin wounding at birth has some of the longest-lasting peripheral and central consequences upon sensory connections. The wound heals rapidly, but the sensory nerve terminals in the area show a profound sprouting response, which long outlasts the injury (at least 12 weeks in the rat) (Reynolds and Fitzgerald 1995); De Lima et al. 1999; Alvares et al. 2000). The effect is most dramatic when wounds are performed at birth and decreases progressively with age at wounding. This is a sensory A- and C-fiber nerve response with no sympathetic involvement (Reynolds and Fitzgerald 1995). In addition, skin wounding at birth leads to a long-lasting expansion of dorsal horn cell receptive fields, which is clearly observed at six weeks (C. Torsney and M. Fitzgerald, unpublished observations).

The only other infant model that leads to central changes of this kind is neonatal colonic distension or irritation with mustard oil between postnatal days 8 and 21, which leads to sensitization of the abdominal withdrawal reflex and heightens the responses of viscerosensitive neurons during colon distension in adult rates. This model appears to lead to chronic visceral hypersensitivity in the adult in the absence of identifiable peripheral pathology and only occurs if the colon is distended in young animals (Al-Chaer et al. 2000).


The mechanisms by which early experience alters somatosensory processing are likely to involve activity-dependent changes in the developing nervous system. The influence of sensory experience upon the formation of somatosensory synaptic connections is well established in the rodent trigeminal system, where alternations in whisker stimulation doing a critical period of postnatal development result in receptive field reorganization in the brainstem, thalamus, and cortex (O’Leary et al. 1994; Fox 2002; Kaas and Catania 2002). This period of plasticity is transitory, usually encompassing only a short time soon after the onset of the sensory stimulus, which is immediately postnatal for whisker barrel formation (Fox 1992), and after eye opening in relation to the visual cortex (Berardi et al. 2000). If the source of activity is altered during this critical period, normal patterns of connectivity are disrupted. A common mechanism has been proposed whereby synaptic connections are strengthened when pre- and postsynaptic activity is correlated, which those connections that are uncorrelated being weakened and eliminated (Feldman et al. 1999; Sanes and Yamagata 1999). The molecular basis of this mechanism is thought to involve the induction of N-methyl d-aspartate (NMDA)-dependent long-term potentiation and depression (Fox 2002). In support of this theory, normal sensory connectivity patterns are disrupted in NMDA-R1-receptor knockout mice (Isawato et al. 1997). While the limits of sensory connections are dependent upon afferent terminal patterns, functional somatotopic maps are determined by the size and pattern of receptive fields of individual target neurons, and the construction of these respective fields in the developing visual system is also NMDA-dependent (Huang and Pallas 2001).


Spinal cord dorsal horn somatosensory maps also undergo postnatal refinement over a critical postnatal period. Primary afferent A fibers extend more superficially to laminae I and II of the dorsal horn at early postnatal stages, with a subsequent gradual withdrawal down to lamina III or below over the first three postnatal weeks (Fitzgerald et al. 1994; Beggs et al. 2002). This withdrawal is accompanied by a progressive reduction in A-fiber input to the susstantia gelatinosa (Park et al. 1999; Nakatsuka et al. 2000) and a gradual reduction in the cutaneous receptive field size of dorsal horn cells (Fitzgerald and Jennings 1999; Torsney and Fitzgerald 2002). The postnatal refinement of sensory inputs in the spinal dorsal horn is likely to contribute to sensory processing and may underlie the increase in cutaneous mechanical reflex thresholds that occurs between birth and adulthood (Fitzgerald 1999; Fitzgerald and Jennings 1999).

The postnatal withdrawal of A-fiber terminals from the substantia gelatinosa may be the result of a competitive process because neonatal destruction of C fibers with capsaicin prevents the effect (Torsney et al. 2000). Recently we have shown that chronic, local exposure of the dorsal horn of the lumbar spinal cord to the NMDA antagonist MK801 from birth prevents the normal functional and structural reorganization of A-fiber connections (Beggs et al. 2002). Dorsal horn cells in spinal MK801-treated animals, investigated at 8 weeks of age by “in vivo” electrophysiological recording, had significantly larger cutaneous mechanoreceptive fields and greater A-fiber-evoked responses than did vehicle-treated controls. C-fiber-evoked responses were unaffected. Chronic application of MK801 also prevented the normal structural reorganization of A-fiber terminals in the spinal cord. The postnatal withdrawal of superficially projecting A-fiber primary afferents to deeper laminae did not occur in treated animals, although C-fiber afferent terminals and cell density in the dorsal horn were apparently unaffected. Spinal MK801-treated animals also had significantly reduced behavioral reflex thresholds to mechanical stimulation of the hindpaw compared to naïve and vehicle treated animals, whereas noxious heat thresholds remained unaffected. The results indicate that the normal postnatal structural and functional development of A-fiber sensory connectivity within the spinal cord is an activity dependent process requiring NMDA-receptor activation (Beggs et al. 2002).

The finding that blockade of NMDA activation of the dorsal horn neurons during a critical period of development selectively alters the normal development of spinal sensory connections demonstrates the plasticity and vulnerability of the system. It seems likely, therefore, that altered patterns of C-Fiber excitation resulting from local injury in neonates will modify synaptic connectivity within the central nervous system via an NMDA-dependent mechanism and so alter the normal maturation of sensory pathways.


In common with other areas of the central nervous system, synaptic development of spinal sensory connections is experience or activity dependent. Evidence from both animal and human studies shows that alterations in the patterns of sensory activity that can arise from tissue injury and pain in early life may disrupt normal synaptic organization within the somatosensory system. While these studies are incomplete and more investigation is needed in this area, the potential clinical importance of neonatal plasticity in pain development is clear.


Suellen Walker was an IASP John J. Bohica Fellow. Support from the Medical Research Council and Children Nationwide is gratefully acknowledged.


Correspondence to: Maria Fitzgerald, PhD, Department of Anatomy and Developmental Biology, University College London, Gower Street, London WCIE 6BT, United Kingdom. Email:

CIRP logo Note:

* This note appears to have been inadvertently omitted from the article. It has been supplied by the Circumcision Reference Library.


The Circumcision Information and Resource Pages are a not-for-profit educational resource and library. IntactiWiki hosts this website but is not responsible for the content of this site. CIRP makes documents available without charge, for informational purposes only. The contents of this site are not intended to replace the professional medical or legal advice of a licensed practitioner.

Top   © 1996-2024 | Please visit our sponsor and host: IntactiWiki.