- Email: [email protected]
Long-term effects of neonatal surgery on adulthood pain behavior
Pain 113 (2005) 347–353 www.elsevier.com/locate/pain
Long-term effects of neonatal surgery on adulthood pain behavior Wendy F. Sternberg*, Laura Scorr, Lauren D. Smith, Caroline G. Ridgway, Molly Stout Department of Psychology, Haverford College, 370 Lancaster Avenue, Haverford, PA 19041, USA Received 4 August 2004; received in revised form 3 November 2004; accepted 15 November 2004
Abstract The long-term consequences of neonatal noxious stimulation on adulthood pain behavior were investigated in male and female mice. On the day of birth, mouse pups were exposed to a laparotomy under cold anesthesia followed by an analgesic dose of morphine (10 mg/kg) postoperatively, or a saline control. An additional group of subjects was exposed to the non-noxious aspects of the surgical procedure (cold exposure, separation from the dam, injection) comprising a ‘sham’ surgery control group, whereas another group of control subjects was administered an injection of saline or morphine, but was otherwise undisturbed. Behavioral observations of the pups immediately following the procedure indicated that the laparotomy produced increased distress vocalizations in the ultrasonic range (40 kHz) compared to both groups of control subjects. During 90 min observations periods following the surgery and 1-week later, maternal care did not vary among treatment conditions. In adulthood, offspring were tested for nociceptive sensitivity on the hot-plate (HP; 53 8C), tail-withdrawal (TW; 50 8C) and acetic acid abdominal constriction test (AC). On both the TW and the AC tests, neonatal surgery decreased pain behavior relative to both groups of control subjects, an effect that was reversed by post-operative morphine treatment. On the HP test, both groups of subjects exposed to the stressful aspects of neonatal surgery (laparotomy or sham surgery) exhibited decreased pain behavior in adulthood. These findings suggest that early exposure to noxious and/or stressful stimuli may induce long-lasting changes in pain behavior, perhaps mediated by alterations in the stress-axis and antinociceptive circuitry. q 2004 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved. Keywords: Neonatal surgery; Tail-withdrawal; Hot-plate; Abdominal constriction
1. Introduction Pain pathways are formed and functional at birth in humans and in laboratory animals (Anand, 2000, 2001; Fitzgerald and Beggs, 2001; Fitzgerald and Gibson, 1984; Garg et al., 2003; Johnston et al., 1993; McLaughlin et al., 1990). Clinicians observe behavioral and physiological responses to routine noxious stimuli (e.g. needle sticks) in human infants (Grunau et al., 1990; McCulloch et al., 1995), and neonatal intensive care often involves painful invasive procedures. Thus, neonatal pain experience may contribute to variability in adult sensitivity to noxious stimuli. In rats, early injury leads to hyperinnervation of the injured area that persists into adulthood. Unilateral application of Complete Freund’s Adjuvant (CFA) to * Corresponding author. Tel.: C1 610 896 1237; fax: C1 610 896 4963. E-mail address: [email protected] (W.F. Sternberg).
the hindpaw causes long-lasting increases in adult pain behavior following re-inflammation, with increased density and lowered activation thresholds of ipsilateral superficial dorsal horn terminals (Peng et al., 2003; Ruda et al., 2000; Tachibana et al., 2001). In this model, basal thermal nociception is unaffected by the neonatal noxious stimulus (Ruda et al., 2000), although ipsilateral dorsal horn neurons display heightened responsiveness to noxious thermal stimuli applied to the affected paw in adulthood (Peng et al., 2003). In contrast, formalin injection applied to all four paws daily for the first post-natal week leads to longlasting thermal hyposensitivity of the hindpaw (Anand et al., 1999; Bhutta et al., 2001). Carrageenan inflammation of the hindpaw similarly leads to a long-lasting lowering of sensitivity to thermal noxious stimuli in the affected paw, but also causes hypersensitivity to thermal stimuli in the presence of ongoing inflammation following adulthood exposure (Lidow et al., 2001). These disparate findings
0304-3959/$20.00 q 2004 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.pain.2004.11.013
348
W.F. Sternberg et al. / Pain 113 (2005) 347–353
have led some to conclude that neonatal inflammation (of moderate intensity) leads to both basal hyposensitivity in the affected region, and hypersensitivity to re-inflammation or injury in adulthood (Lidow, 2002; Ren et al., 2004). These findings extend to a non-inflammatory noxious stimulus (neonatal footshock), which similarly leads to a diminished response to noxious thermal stimuli applied in adulthood (Bernardi et al., 1986; Shimada et al., 1990). Few investigations, however, have assessed overall nociceptive sensitivity (i.e. away from the site of injury) in adults exposed to locally applied neonatal noxious stimuli. It has been observed that some of the procedures that are used to model neonatal pain experience are unusually severe in their intensity and chronicity (Lidow, 2002). Although some clinically relevant neonatal procedures have been employed in animal studies (e.g. neonatal skin wounding, short-duration inflammation, repeated needle sticks), a more realistic animal model of neonatal pain experience could prove informative for understanding long-lasting alterations. The purpose of the current study was three-fold. First, we wished to assess the long-term alterations in pain behavior following neonatal abdominal surgery, a clinically relevant noxious stimulus. Second, we assessed adulthood alterations in nociceptive sensitivity at body loci other than the site of injury. Lastly, we attempted to extend the literature on long-term effects of early noxious stimulation from the rat (the species used in all previous studies) to the laboratory mouse, since we expect future investigations to be concerned with the interaction between early noxious stimulation and genetic profile.
2. Methods 2.1. Subjects The subjects in this study were the offspring of breeding pairs of outbred CD-1w mice, bred in our laboratory (breeders obtained from Harlan Sprague–Dawley, Indianapolis, IN). Animals were housed in a light- (12:12 h, lights on at 08:00) and temperature (20 8C)-controlled environment with free access to food (Harlan Teklad 8604) and tap water. Following weaning at 3 weeks of age, subjects were housed with same-sex littermates in groups of 2–6 in polypropylene cages until adult testing (7–12 weeks of age). All procedures were reviewed and approved by the Haverford College Animal Care and Use Committee. 2.2. Mating procedures Male and female CD-1w breeders were obtained commercially. Upon arrival, female mice were monitored for estrus cyclicity via daily vaginal smears for 8 days. Examination of cell cytology was used to predict which mice were sexually receptive (predominance of cornified epithelial cells). When determined to be receptive, the female was placed in a cage with two males to maximize likelihood
of conception. Conception was verified by observation of a mucus plug. Pregnant mice were housed singly until parturition. 2.3. Surgery The surgical procedure used as a noxious stimulus in the present study was a simple laparotomy. Pups were anesthetized using cryoanesthesia (Phifer and Terry, 1986)—each pup assigned to the surgical condition was covered in crushed ice for 2 min, then removed and placed atop a bed of ice for the duration of the surgical procedure. An abdominal incision was made extending from umbilicus to genitalia. A curved tip forceps was inserted into the incision and rotated. Sutures were placed using chromic gut (6–0 needle). Pups were then placed under a heat lamp on a warming bed to recover. 2.4. Behavioral observations Following the neonatal surgical procedure (or its control, described below) male and female pups undergoing the same procedure were returned to a single dam. In order to determine whether maternal behavior was responsible for any potential group differences in adulthood, maternal behavior was observed for a 90 min period immediately following the neonatal procedure, and again 1 week later. The dam was removed from the cage, the pups were scattered around the home cage and rolled in the bedding, and then the dam was re-introduced. During the first 15 min following reunion, the latency to gathering of the first pup and gathering 80% of the litter were recorded, and the number of 40 kHZ ultrasonic vocalizations (USVs) made by the litter as a whole was recorded using a bat detector (Ultrasonic Advice, London, UK). USVs were counted by an observer listening through headphones and using an electronic switch attached to a digital counting device. At the end of the 15-min period, pups who were not gathered were placed in the nest and a 90-min observation period ensued during which the duration of the following maternal behaviors were timed: nesting (crouching atop the litter of suckling pups); grooming self; grooming pups; eating/drinking; wandering/lying alone. The day of neonatal manipulation was identified as D0 (always performed on post-natal day 0 or 1); a follow up observation including the 15min gathering period and 90-min maternal behavior observation was conducted on D6 or D7 (6 or 7 days after the neonatal manipulation). These observation parameters were based on pilot studies that indicated the peak period of activity following disruption of the home cage. The 1 week follow-up was included to determine if effects of the procedure outlasted the period of discomfort induced by the surgery (as assessed in pilot studies). 2.5. Drugs Morphine sulfate (MOR; Sigma) was dissolved in saline and administered subcutaneously (10 mg/kg) under gathered loose skin on the back in an injection volume of 20 ml. Isovolumetric injections of saline (SAL) were used as a control for the injection procedure. This dose of morphine was chosen based on our previous observations of the analgesic effects of 10 mg/kg morphine in newborn mice on standard algesiometric assays (Sternberg et al., 2004).
W.F. Sternberg et al. / Pain 113 (2005) 347–353
2.6. Neonatal procedure All procedures took place between 09:00 and 14:00 h (during the light cycle). Two litters born within 24 h of each other were tested simultaneously. The pups in the two litters were removed from the dams and combined. The entire set of pups (combined across litter) was sexed, and half of the pups of each sex were placed in separate piles. Each pile was randomly assigned to one of the following two surgical conditions: surgery (laparotomy, as described above) or sham surgery (described below). All pups manipulated on a given day were exposed to the same drug condition (SAL, MOR), although the assignment of drug across days was random. Thus, each testing day produced two groups of subjects with an equal sex ratio: surgery/sham surgery, each of which was exposed to the same drug treatment. During the surgical procedure (lasting from 75 to 110 min), the mothers were habituated to the observation environment. The procedure took place as follows: one pup from the surgery group and a same-sex subject from the sham surgery group were selected simultaneously. Both members of a given pair of pups were placed in a container of crushed ice (for cryoanesthesia induction), and removed when both pups were immobile and failed to respond to tail-pinch. The pup assigned to the surgery group underwent the laparotomy procedure. The pup assigned to the sham surgery group was kept on the surface of a bed of crushed ice for the same duration as the matched pup in the surgery group. When the surgery was complete, both pups were removed from the ice and left to warm on a heating pad under a heat lamp. When both pups were warmed (at least 25 8C as determined by an infrared surface thermometer, Kant model 600C), had regained a pink color and were observed to move on their own, they were administered the assigned injectant. When both sets of pups (surgery and sham surgery) were fully recovered from the anesthesia, each group (pups undergoing the same treatment) was returned to a single dam. Maternal behavior was assessed in both dams simultaneously (by observers blind to the pup treatment group) for the 90 min period following reunion. Thus, for each set of ‘surgery’ pups, there was a matched group of ‘sham surgery’ pups, with length of exposure to the ice, duration of separation from the dam, sex composition, and number of pups returned to the dam matched between the groups. Furthermore, although each set of pups returned to a given dam underwent the same experimental procedure, each set consisted of male and female pups combined across two birth litters. Thus, genetic differences between the experimental groups were minimized. Separate litters were bred and were injected with either SAL or MOR (but not separated from the dam) on the day of birth, and were kept with the birth mothers to serve as injection-only controls. Thus, within each drug condition, subjects were either exposed to surgery (SURG), or served in one of two control groups: those exposed to anesthesia and separation from the dam for the same duration, and administered the same drug as a set of pups in the SURG group Table 1 Number of animals surviving to adulthood in each of the surgical/control conditions SURG
Male Female
SHAM
349
(SHAM), or those simply injected with the drug (INJ). The number of male and female subjects surviving to adulthood in each of the neonatal conditions is outlined in Table 1.
3. Adulthood pain measurement Male and female mice were subjected to three adulthood nociceptive tests: tail-withdrawal (TW), hot-plate (HP), abdominal constriction (AC), in that order, with at least 1 week intervening between testing sessions by experimenters blind to the neonatal treatment group. All of the animals surviving to adulthood were included in the HP/TW tests. Due to the increased amount of time required for the AC test compared to the TW and HP tests, only a subset of animals were tested in the AC test. No fewer than five subjects (randomly chosen across the 3–4 litters in each group) in any sex/surgical condition/drug condition group contributed data to the AC test. 3.1. HP test The hot-plate apparatus (Model 39C, IITC, Woodland Hills, CA) consists of a temperature-controlled metal surface (53 8C) with a cylindrical Plexiglas enclosure placed on the surface. Mice were placed inside the cylinder and a timer was started when all four paws were in contact with the surface. Latency to hind paw lick or shake was recorded, and the animal quickly removed from the surface. A 60-s cut-off latency was imposed on non-responsive animals. 3.2. TW test A hot-water circulator (Fisher Isotemp) with water maintained at a temperature of 53 8C was used for the TW test. Mice were lightly restrained by wrapping in a towel, and the distal end of the tail was immersed in the water. Latency to vigorously withdraw the tail from the hot water was recorded. A 15-s cut-off was employed. Two readings were taken for each test; latency was the average of the two readings. 3.3. AC test Subjects were habituated to Plexiglas cylinders for 20 min, after which they were given an intraperitoneal injection of 0.9% acetic acid (10 ml/kg injection volume). Abdominal constrictions were then observed for 30 min following injection. A constriction was defined as full-body extension with at least one hindpaw leaving the floor of the observation chamber.
INJ
SAL
MOR
SAL
MOR
SAL
MOR
24 24
21 22
29 27
23 13
9 8
9 7
4. Data analysis Hot-plate latencies, tail-withdrawal latencies, and total number of abdominal constrictions were the dependent
350
W.F. Sternberg et al. / Pain 113 (2005) 347–353
variables in three separate analyses. Each analysis consisted of a 2!3!2 factorial ANOVA, with sex, neonatal surgical condition, and neonatal drug condition as between-subjects variables. Neonatal USVs were also analyzed using 2!2 factorial ANOVA, with surgical and drug condition as between-subjects variables. Finally, maternal behavior over the 90-min observation period on days 0 and 7 was quantified as being pup-directed (PUP: nesting, grooming pups) or selfdirected (SELF: grooming self, eating/drinking) and analyzed using 2!2!2 mixed-factorial ANOVA, with neonatal surgical condition (SHAM or SURG), drug condition (SAL, MOR) as between groups variables and behavioral target (PUP vs. SELF) as a within-subjects variable. Significance level was set to aZ0.05; post-hoc comparisons were performed using Fisher’s post-hoc tests where warranted.
5. Results 5.1. Ultrasonic vocalizations Significant main effects were noted for surgery condition (F2,17Z12.43, P!0.001), drug (F1,17Z30.43, P!0.001), and a significant surgery!drug interaction (F2,17Z4.14, P!0.05) was present. Post-hoc analysis of the interaction indicates that in SAL-treated animals, the surgery manipulation increased USVs compared to INJ and SHAM controls (who did not differ). However, when MOR was administered following the surgery, USVs were reduced to that of unoperated controls (see Fig. 1). The main effect of drug (SALOMOR) suggests an overall reduction in USV induced by morphine. However, SHAM and INJ subjects given MOR did not differ from each other. 5.2. Maternal behavior The duration of time the dam spent attending to pups or to herself was quantified and expressed as a percentage of total time spent engaged in either type of behavior. Time spent wandering around the cage or lying alone was not
Fig. 2. Maternal behavior during a 90-min period following surgery (D0) and 1 week later (D7). Percentage of time spent attending to pups vs. self (proportion of each bar in the black region) did not differ based on condition (surgery vs. sham). Overall, dams spent a vast majority of the observation period engaged in pup-directed behavior.
included in the quantification. On both days 0 and 7, dams spent a far greater percentage of the 90-min observation period engaged in pup-directed as opposed to self-directed behaviors (F1,19Z137.47, P!0.001 on day 0; F1,17Z 321.97, P!0.001 on day 7). On neither day did the time spent engaged in pup vs. self-directed behavior vary according to surgical or drug condition (all two-and threeway interactions were non-significant). Thus, observable maternal behavior patterns were the same regardless of surgical/drug condition of the pups (see Fig. 2). 5.3. Tail-withdrawal Adult male mice displayed significantly longer tailwithdrawal latencies than females (significant main effect of sex F1,155Z8.19, P!0.001), regardless of neonatal condition. A significant surgical condition!drug interaction was present (F2,155Z8.01, P!0.001). Subjects undergoing the surgery without post-operative analgesic treatment (SURG-SAL) displayed longer tail-withdrawal latencies than injection-only subjects (INJ-SAL) and sham-operated controls (SHAM-SAL), as well as those undergoing surgery with post-operative morphine (SURG-MOR). However, subjects in the sham-operated control group who received morphine (SHAM-MOR) displayed elevated tail-withdrawal latencies compared to all other control subjects (see Fig. 3). 5.4. Hot-plate
Fig. 1. Elevation in stress vocalizations induced by neonatal surgical procedure. Surgery increased vocalizations compared to sham-operated control and injection-only subjects (who did not differ). Morphine reduced the surgery-induced vocalizations. Morphine also reduced vocalizations to a lesser extent in unoperated subjects.
Two significant main effects were present: surgical condition (F2,204Z10.83, P!0.001; SURGZSHAMO INJ) and drug condition (F1,204Z8.15, P!0.01; SALO MOR). There were no significant interactions, nor any sexdependent effects. Thus, male and female subjects exposed to both the neonatal surgical condition and the sham-operated
W.F. Sternberg et al. / Pain 113 (2005) 347–353
351
Fig. 3. Long-term changes in latency to tail-withdrawal from a noxious heat stimulus (50 8C water immersion) induced by untreated pain exposure on day of birth. *Indicates significant increase in withdrawal latency in salinetreated animals undergoing surgery treatment compared to sham-operated and injection-only controls receiving saline. #Indicates significant reduction in latency compared to surgery-saline. ^Indicates significant elevation compared to all other groups.
Fig. 5. Abdominal constrictions in response to .9% acetic acid (i.p.) in subjects exposed to treated or untreated surgical pain on the day of birth. Subjects receiving surgery without morphine analgesia exhibited significantly fewer constrictions than subjects receiving surgery with analgesic treatment, or unoperated controls (both sham-operated and injection-only) (*indicates significantly fewer constrictions than all other groups).
control procedure exhibited elevated hot-plate latencies compared to injection-only controls (see Fig. 4).
adult nociceptive sensitivity. The results are clearest in the AC assay, where neonatal surgery resulted in a dramatic reduction in visceral nociceptive behavior that was reversed by post-operative morphine administration immediately after the surgery. Subjects exposed to the stressful, but non-nociceptive, aspects of surgery (cold exposure, separation from the dam) were no different from injection-only control subjects, and morphine exposure alone did not elevate pain behavior on this test. Taken together, these results suggest that the long-term effect of the surgery on reducing visceral pain behavior in adulthood is due to the nociceptive aspects of the neonatal procedure. That the surgical procedure itself produced postoperative pain is supported by the results of the neonatal behavioral observations. Upon return to the dam, subjects undergoing surgery without post-operative analgesic treatment exhibited far more distress vocalizations than unoperated subjects (both sham-operated and injectiononly controls), or operated subjects exposed to morphine. In the hours following surgery, and 1 week later, maternal treatment among operated and unoperated groups did not differ, suggesting that the neonatal procedure, rather than maternal treatment during the pre-weanling period, was responsible for the adulthood alteration in visceral nociception. The results on the TW and HP tests lead to similar (but not identical) conclusions. As in the AC test, subjects in the SURG-SAL group displayed reduced pain behavior (longer withdrawal latencies) on the TW test compared to unoperated subjects, and those receiving morphine following surgery. However, our conclusions are limited by an apparent elevation in tail-withdrawal latency induced by neonatal morphine in the sham-operated control group. Since the elevation was not noted in SHAM-SAL subjects, this finding cannot simply be attributed to the stress of the neonatal procedure, and is unexplained. However, we note the similarity in pattern of the long-term reduction of pain
5.5. Abdominal constriction A significant surgical condition!drug interaction was present (F2,94Z7.35, PZ0.001), with subjects in the SURG-SAL condition displaying fewer constrictions than subjects in the SHAM-SAL and INJ-SAL groups, and fewer constrictions than subjects in the SURG-MOR group. Subjects in the three surgical conditions treated with MOR did not differ from one another (see Fig. 5). Neonatal morphine treatment alone did not alter adulthood constriction behavior (MORZSAL in INJ and SHAM groups). The effect was present for both male and female subjects (no significant sex-dependent interactions).
6. Discussion The results of this study provide strong support for the existence of long-lasting effects of neonatal surgical pain on
Fig. 4. The effects of neonatal stress on hot-plate latency (53 8C). Subjects in both the surgery and sham-operated control conditions displayed elevated hot-plate latencies compared to injection-only control subjects.
352
W.F. Sternberg et al. / Pain 113 (2005) 347–353
behavior resulting from neonatal surgery on the TW and AC tests. On the hot-plate test, the long-term alterations (again, elevations in latency) appear to be due to the stressful aspects of the procedure alone, with no additional alterations occurring due to the nociceptive aspects of the procedure. Subjects in the SURG and SHAM groups exhibited elevations in hot-plate latency compared to INJ subjects, regardless of drug treatment. Thus, the nociceptive circuitry underlying hot-plate responses is susceptible to early influences of stress (whether or not that stress results from a painful procedure). We are not entirely surprised that the findings differ on the HP test compared to the other two nociceptive assays. Both the TW and AC tests assess reflexive responses to noxious stimuli. The HP test measures a supraspinally organized behavior pattern that has been shown to be affected by non-painful perinatal stressors in previous studies (Sternberg and Ridgway, 2003). The results obtained here confirm and extend those of Bhutta et al. (2001) who demonstrated that neonatal formalin exposure in rats increased hot-plate and tail-flick latencies when tested in adulthood. As in the present study, morphine exposure following formalin treatment reversed the effects of neonatal surgery (Bhutta et al., 2001). Despite differing neonatal noxious stimuli (surgical pain on the day of birth; repeated formalin injections during the first week of life) and the cross-species comparison, the findings were quite similar, suggesting a particularly robust phenomenon. We do not know the mechanism underlying the decreased nociceptive sensitivity we observed in the present investigation. However, we believe the alteration to be a global nervous system phenomenon, given its generalizability across acute cutaneous and tonic, subcutaneous/ visceral algesiometric assays. It is evident that the subjects exposed to surgery without post-operative morphine on the day of birth displayed lower nociceptive sensitivity in adulthood in response to a variety of noxious stimuli. It is plausible that early pain exposure (perhaps mediated by neonatal activation of the hypothalamic–pituitary–adrenal stress-axis) upregulates basal endogenous opioid production or receptor density such that antinociceptive mechanisms are more readily evoked, perhaps even by the stress associated with nociceptive testing itself in adulthood. We have data to suggest that the neonatal surgical procedure elevates adulthood stress-induced analgesia levels compared to injection-only controls (unpublished observations), but we have not yet tested this procedure in sham-operated control subjects and therefore cannot conclude that the longterm alterations are due to surgical pain per se. Various perinatal manipulations, such as prenatal stress (Kinsley et al., 1988), prenatal ethanol exposure (Nelson et al., 1985, 1986) and post-natal handling (Pieretti et al., 1991) have been shown to affect opioid-mediated processes when tested in adulthood. We are currently testing the hypothesis that the alteration in pain behavior following neonatal pain
exposure is due to elevated endogenous opioid activity in adulthood. The results reported here are also consistent with a small number of clinical investigations on infant pain exposure. Children who were born at an extremely low birth weight (ELBW) and exposed to invasive procedures in the NICU are reported by their parents to display lowered sensitivity to everyday pains at 18 months (Grunau et al., 1994), and display lowered facial responsiveness to immunization at 4 and 8 months of age (Oberlander et al., 2000). Interestingly, pain behavior following immunization does not differ between toddlers who had major surgery as infants and control subjects when the surgeries were performed with pre-emptive morphine analgesia (Peters et al., 2003). There have also been reports of enhanced pain behavior during immunization in children that had undergone circumcision as neonates, an effect that is reversed by local anesthetic at the time of the procedure (Taddio et al., 1995, 1997). The difficulty in reconciling conflicting findings in the human literature is compounded by the barriers to conducting studies on human infants with appropriate placebo controls. Therefore, understanding the mechanisms of plasticity following neonatal pain exposure will require useful animal models. We believe the model of neonatal surgical pain in the mouse will be a useful one for future investigations. Abdominal surgery is a complex stimulus, involving nociception, skin wounding, local inflammation (indeed, the chromic gut sutures used here are disintegrated by an inflammatory response), cold exposure, and maternal separation. Thus in many ways, the experience is similar to that experienced by human pre-term neonates requiring intensive care. Local changes in nociceptive circuitry induced by the neonatal procedure are unlikely to account for the changes in pain behavior since we do not test adulthood nociceptive sensitivity at the site of the neonatal injury. However, we cannot rule out such a mechanism because the spinal segments responsible for transmitting visceral nociceptive stimuli and mediating the tail-withdrawal reflex may be close to that affected by the laparotomy. We did not test whether mechanical hyper- or hypoalgesia was present in adulthood at the site of cutaneous injury, which may provide useful information regarding the peripheral reorganization that may take place due to this procedure. Future investigations using the neonatal surgical model should confirm the extent of local changes induced by this noxious stimulus. Laboratory investigations of neonatal pain exposure are often aimed at improving clinical practice regarding pain management in the human neonate. Neonates (especially pre-term neonates) are often under-medicated for pain, despite undergoing frequent painful procedures (Johnston et al., 1997), due in large part to misconceptions regarding the inability of neonates to experience pain and the safety of analgesics or anesthetics during the neonatal period (Anand, 2001). However, it is clear that human neonates exhibit
W.F. Sternberg et al. / Pain 113 (2005) 347–353
behavioral and physiological responses during painful procedures (e.g. McCulloch et al., 1995), suggesting that infants are indeed capable of experiencing pain. Therefore, infants’ pain experience should be considered alongside of concerns regarding the safety of neonatal anesthesia or analgesia. The findings reported here suggest that in addition to considering ethical obligations to alleviate pain in the neonate, one should also consider the long-term (and potentially permanent) consequences of aberrant sensory stimulation during the neonatal period for the development of pain sensing pathways in the nervous system.
Acknowledgements The authors wish to thank James Steinemann and Hiro Takahashi for help with neonatal surgical procedures and Dr Jeffrey Mogil for proofreading an earlier version of this paper. This research was supported by a Research in Undergraduate Institutions grant and a Haverford College faculty research grant to WFS.
References Anand KJ. Effects of perinatal pain and stress. Prog Brain Res 2000;122: 117–29. Anand KJ. Consensus statement for the prevention and management of pain in the newborn. Arch Pediatr Adolesc Med 2001;155:173–80. Anand KJ, Coskun V, Thrivikraman KV, Nemeroff CB, Plotsky PM. Longterm behavioral effects of repetitive pain in neonatal rat pups. Physiol Behav 1999;66:627–37. Bernardi M, Genedani S, Bertolini A. Behavioral activity and active avoidance learning and retention in rats neonatally exposed to painful stimuli. Physiol Behav 1986;36:553–5. Bhutta AT, Rovnaghi C, Simpson PM, Gossett JM, Scalzo FM, Anand KJ. Interactions of inflammatory pain and morphine in infant rats: longterm behavioral effects. Physiol Behav 2001;73:51–8. Fitzgerald M, Beggs S. The neurobiology of pain: developmental aspects. Neuroscientist 2001;7:246–57. Fitzgerald M, Gibson S. The postnatal physiological and neurochemical development of peripheral sensory C fibres. Neuroscience 1984;13: 933–44. Garg S, Narsinghani U, Bhutta AT, Rovnaghi C, Anand KJS. Long-term effects of neonatal pain: the animal literature. In: McGrath PJ, Finley GA, editors. Pediatric pain: biological and social context, vol. 26. Seattle, WA: IASP Press; 2003. p. 1–22. Grunau RV, Johnston CC, Craig KD. Neonatal facial and cry responses to invasive and non-invasive procedures. Pain 1990;42:295–305. Grunau RV, Whitfield MF, Petrie JH. Pain sensitivity and temperament in extremely low-birth-weight premature toddlers and preterm and fullterm controls. Pain 1994;58:341–6. Johnston CC, Stevens B, Craig KD, Grunau RV. Developmental changes in pain expression in premature, full-term, two- and four-month-old infants. Pain 1993;52:201–8. Johnston CC, Collinge JM, Henderson SJ, Anand KJ. A cross-sectional survey of pain and pharmacological analgesia in Canadian neonatal intensive care units. Clin J Pain 1997;13:308–12.
353
Kinsley CH, Mann PE, Bridges RS. Prenatal stress alters morphine-and stress-induced analgesia in male and female rats. Pharmacol Biochem Behav 1988;30:123–8. Lidow MS. Long-term effects of neonatal pain on nociceptive systems. Pain 2002;99:377–83. Lidow MS, Song ZM, Ren K. Long-term effects of short-lasting early local inflammatory insult. Neuroreport 2001;12:399–403. McCulloch KM, Ji SA, Raju TN. Skin blood flow changes during routine nursery procedures. Early Hum Dev 1995;41:147–56. McLaughlin CR, Lichtman AH, Fanselow MS, Cramer CP. Tonic nociception in neonatal rats. Pharmacol Biochem Behav 1990;36: 859–62. Nelson LR, Taylor AN, Lewis JW, Branch BJ, Liebeskind JC. Opioid but not nonopioid stress-induced analgesia is enhanced following prenatal exposure to ethanol. Psychopharmacology 1985;85:92–6. Nelson LR, Taylor AN, Lewis JW, Branch BJ, Liebeskind JC. Morphine analgesia is potentiated in adult rats prenatally exposed to ethanol. Brain Res 1986;372:234–40. 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. Peng YB, Ling QD, Ruda MA, Kenshalo DR. Electrophysiological changes in adult rat dorsal horn neurons after neonatal peripheral inflammation. J Neurophysiol 2003;90:73–80. Peters JW, Koot HM, de Boer JB, Passchier J, Bueno-de-Mesquita JM, de Jong FH, Duivenvoorden HJ, Tibboel D. Major surgery within the first 3 months of life and subsequent biobehavioral pain responses to immunization at later age: a case comparison study. Pediatrics 2003; 111:129–35. Phifer CB, Terry LM. Use of hypothermia for general anesthesia in preweanling rodents. Physiol Behav 1986;38:887–90. Pieretti S, d’Amore A, Loizzo A. Long-term changes induced by developmental handling on pain threshold: effects of morphine and naloxone. Behav Neurosci 1991;105:215–8. Ren K, Anseloni V, Zou S-P, Wade EB, Novikova SI, Ennis M, Traub RJ, Gold MS, Dubner R, Lidow MS. Characterization of basal and reinflammation-associated long-term alteration in pain responsivity following short-lasting neonatal local inflammatory insult. Pain 2004; 110:588–96. Ruda MA, Ling QD, Hohmann AG, Peng YB, Tachibana T. Altered nociceptive neuronal circuits after neonatal peripheral inflammation. Science 2000;289:628–31. Shimada C, Kurumiya S, Noguchi Y, Umemoto M. The effect of neonatal exposure to chronic footshock on pain responsiveness and sensitivity to morphine after maturation in the rat. Behav Brain Res 1990;36:105–11. Sternberg WF, Ridgway CG. Effects of gestational stress and neonatal handling on pain, analgesia, and stress behavior of adult mice. Physiol Behav 2003;78:375–83. Sternberg WF, Smith LD, Scorr L. Nociception and antinociception on the day of birth in mice: sex differences and test dependence. J Pain 2004;5:420–6. Tachibana T, Ling QD, Ruda MA. Increased Fos induction in adult rats that experienced neonatal peripheral inflammation. Neuroreport 2001;12: 925–7. Taddio A, Goldbach M, Ipp M, Stevens B, Koren G. Effect of neonatal circumcision on pain responses during vaccination in boys. Lancet 1995;345:291–2. Taddio A, Katz J, Ilersich AL, Koren G. Effect of neonatal circumcision on pain response during subsequent routine vaccination. Lancet 1997;349: 599–603.
Long-term effects of neonatal surgery on adulthood pain behavior Wendy F. Sternberg*, Laura Scorr, Lauren D. Smith, Caroline G. Ridgway, Molly Stout Department of Psychology, Haverford College, 370 Lancaster Avenue, Haverford, PA 19041, USA Received 4 August 2004; received in revised form 3 November 2004; accepted 15 November 2004
Abstract The long-term consequences of neonatal noxious stimulation on adulthood pain behavior were investigated in male and female mice. On the day of birth, mouse pups were exposed to a laparotomy under cold anesthesia followed by an analgesic dose of morphine (10 mg/kg) postoperatively, or a saline control. An additional group of subjects was exposed to the non-noxious aspects of the surgical procedure (cold exposure, separation from the dam, injection) comprising a ‘sham’ surgery control group, whereas another group of control subjects was administered an injection of saline or morphine, but was otherwise undisturbed. Behavioral observations of the pups immediately following the procedure indicated that the laparotomy produced increased distress vocalizations in the ultrasonic range (40 kHz) compared to both groups of control subjects. During 90 min observations periods following the surgery and 1-week later, maternal care did not vary among treatment conditions. In adulthood, offspring were tested for nociceptive sensitivity on the hot-plate (HP; 53 8C), tail-withdrawal (TW; 50 8C) and acetic acid abdominal constriction test (AC). On both the TW and the AC tests, neonatal surgery decreased pain behavior relative to both groups of control subjects, an effect that was reversed by post-operative morphine treatment. On the HP test, both groups of subjects exposed to the stressful aspects of neonatal surgery (laparotomy or sham surgery) exhibited decreased pain behavior in adulthood. These findings suggest that early exposure to noxious and/or stressful stimuli may induce long-lasting changes in pain behavior, perhaps mediated by alterations in the stress-axis and antinociceptive circuitry. q 2004 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved. Keywords: Neonatal surgery; Tail-withdrawal; Hot-plate; Abdominal constriction
1. Introduction Pain pathways are formed and functional at birth in humans and in laboratory animals (Anand, 2000, 2001; Fitzgerald and Beggs, 2001; Fitzgerald and Gibson, 1984; Garg et al., 2003; Johnston et al., 1993; McLaughlin et al., 1990). Clinicians observe behavioral and physiological responses to routine noxious stimuli (e.g. needle sticks) in human infants (Grunau et al., 1990; McCulloch et al., 1995), and neonatal intensive care often involves painful invasive procedures. Thus, neonatal pain experience may contribute to variability in adult sensitivity to noxious stimuli. In rats, early injury leads to hyperinnervation of the injured area that persists into adulthood. Unilateral application of Complete Freund’s Adjuvant (CFA) to * Corresponding author. Tel.: C1 610 896 1237; fax: C1 610 896 4963. E-mail address: [email protected] (W.F. Sternberg).
the hindpaw causes long-lasting increases in adult pain behavior following re-inflammation, with increased density and lowered activation thresholds of ipsilateral superficial dorsal horn terminals (Peng et al., 2003; Ruda et al., 2000; Tachibana et al., 2001). In this model, basal thermal nociception is unaffected by the neonatal noxious stimulus (Ruda et al., 2000), although ipsilateral dorsal horn neurons display heightened responsiveness to noxious thermal stimuli applied to the affected paw in adulthood (Peng et al., 2003). In contrast, formalin injection applied to all four paws daily for the first post-natal week leads to longlasting thermal hyposensitivity of the hindpaw (Anand et al., 1999; Bhutta et al., 2001). Carrageenan inflammation of the hindpaw similarly leads to a long-lasting lowering of sensitivity to thermal noxious stimuli in the affected paw, but also causes hypersensitivity to thermal stimuli in the presence of ongoing inflammation following adulthood exposure (Lidow et al., 2001). These disparate findings
0304-3959/$20.00 q 2004 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.pain.2004.11.013
348
W.F. Sternberg et al. / Pain 113 (2005) 347–353
have led some to conclude that neonatal inflammation (of moderate intensity) leads to both basal hyposensitivity in the affected region, and hypersensitivity to re-inflammation or injury in adulthood (Lidow, 2002; Ren et al., 2004). These findings extend to a non-inflammatory noxious stimulus (neonatal footshock), which similarly leads to a diminished response to noxious thermal stimuli applied in adulthood (Bernardi et al., 1986; Shimada et al., 1990). Few investigations, however, have assessed overall nociceptive sensitivity (i.e. away from the site of injury) in adults exposed to locally applied neonatal noxious stimuli. It has been observed that some of the procedures that are used to model neonatal pain experience are unusually severe in their intensity and chronicity (Lidow, 2002). Although some clinically relevant neonatal procedures have been employed in animal studies (e.g. neonatal skin wounding, short-duration inflammation, repeated needle sticks), a more realistic animal model of neonatal pain experience could prove informative for understanding long-lasting alterations. The purpose of the current study was three-fold. First, we wished to assess the long-term alterations in pain behavior following neonatal abdominal surgery, a clinically relevant noxious stimulus. Second, we assessed adulthood alterations in nociceptive sensitivity at body loci other than the site of injury. Lastly, we attempted to extend the literature on long-term effects of early noxious stimulation from the rat (the species used in all previous studies) to the laboratory mouse, since we expect future investigations to be concerned with the interaction between early noxious stimulation and genetic profile.
2. Methods 2.1. Subjects The subjects in this study were the offspring of breeding pairs of outbred CD-1w mice, bred in our laboratory (breeders obtained from Harlan Sprague–Dawley, Indianapolis, IN). Animals were housed in a light- (12:12 h, lights on at 08:00) and temperature (20 8C)-controlled environment with free access to food (Harlan Teklad 8604) and tap water. Following weaning at 3 weeks of age, subjects were housed with same-sex littermates in groups of 2–6 in polypropylene cages until adult testing (7–12 weeks of age). All procedures were reviewed and approved by the Haverford College Animal Care and Use Committee. 2.2. Mating procedures Male and female CD-1w breeders were obtained commercially. Upon arrival, female mice were monitored for estrus cyclicity via daily vaginal smears for 8 days. Examination of cell cytology was used to predict which mice were sexually receptive (predominance of cornified epithelial cells). When determined to be receptive, the female was placed in a cage with two males to maximize likelihood
of conception. Conception was verified by observation of a mucus plug. Pregnant mice were housed singly until parturition. 2.3. Surgery The surgical procedure used as a noxious stimulus in the present study was a simple laparotomy. Pups were anesthetized using cryoanesthesia (Phifer and Terry, 1986)—each pup assigned to the surgical condition was covered in crushed ice for 2 min, then removed and placed atop a bed of ice for the duration of the surgical procedure. An abdominal incision was made extending from umbilicus to genitalia. A curved tip forceps was inserted into the incision and rotated. Sutures were placed using chromic gut (6–0 needle). Pups were then placed under a heat lamp on a warming bed to recover. 2.4. Behavioral observations Following the neonatal surgical procedure (or its control, described below) male and female pups undergoing the same procedure were returned to a single dam. In order to determine whether maternal behavior was responsible for any potential group differences in adulthood, maternal behavior was observed for a 90 min period immediately following the neonatal procedure, and again 1 week later. The dam was removed from the cage, the pups were scattered around the home cage and rolled in the bedding, and then the dam was re-introduced. During the first 15 min following reunion, the latency to gathering of the first pup and gathering 80% of the litter were recorded, and the number of 40 kHZ ultrasonic vocalizations (USVs) made by the litter as a whole was recorded using a bat detector (Ultrasonic Advice, London, UK). USVs were counted by an observer listening through headphones and using an electronic switch attached to a digital counting device. At the end of the 15-min period, pups who were not gathered were placed in the nest and a 90-min observation period ensued during which the duration of the following maternal behaviors were timed: nesting (crouching atop the litter of suckling pups); grooming self; grooming pups; eating/drinking; wandering/lying alone. The day of neonatal manipulation was identified as D0 (always performed on post-natal day 0 or 1); a follow up observation including the 15min gathering period and 90-min maternal behavior observation was conducted on D6 or D7 (6 or 7 days after the neonatal manipulation). These observation parameters were based on pilot studies that indicated the peak period of activity following disruption of the home cage. The 1 week follow-up was included to determine if effects of the procedure outlasted the period of discomfort induced by the surgery (as assessed in pilot studies). 2.5. Drugs Morphine sulfate (MOR; Sigma) was dissolved in saline and administered subcutaneously (10 mg/kg) under gathered loose skin on the back in an injection volume of 20 ml. Isovolumetric injections of saline (SAL) were used as a control for the injection procedure. This dose of morphine was chosen based on our previous observations of the analgesic effects of 10 mg/kg morphine in newborn mice on standard algesiometric assays (Sternberg et al., 2004).
W.F. Sternberg et al. / Pain 113 (2005) 347–353
2.6. Neonatal procedure All procedures took place between 09:00 and 14:00 h (during the light cycle). Two litters born within 24 h of each other were tested simultaneously. The pups in the two litters were removed from the dams and combined. The entire set of pups (combined across litter) was sexed, and half of the pups of each sex were placed in separate piles. Each pile was randomly assigned to one of the following two surgical conditions: surgery (laparotomy, as described above) or sham surgery (described below). All pups manipulated on a given day were exposed to the same drug condition (SAL, MOR), although the assignment of drug across days was random. Thus, each testing day produced two groups of subjects with an equal sex ratio: surgery/sham surgery, each of which was exposed to the same drug treatment. During the surgical procedure (lasting from 75 to 110 min), the mothers were habituated to the observation environment. The procedure took place as follows: one pup from the surgery group and a same-sex subject from the sham surgery group were selected simultaneously. Both members of a given pair of pups were placed in a container of crushed ice (for cryoanesthesia induction), and removed when both pups were immobile and failed to respond to tail-pinch. The pup assigned to the surgery group underwent the laparotomy procedure. The pup assigned to the sham surgery group was kept on the surface of a bed of crushed ice for the same duration as the matched pup in the surgery group. When the surgery was complete, both pups were removed from the ice and left to warm on a heating pad under a heat lamp. When both pups were warmed (at least 25 8C as determined by an infrared surface thermometer, Kant model 600C), had regained a pink color and were observed to move on their own, they were administered the assigned injectant. When both sets of pups (surgery and sham surgery) were fully recovered from the anesthesia, each group (pups undergoing the same treatment) was returned to a single dam. Maternal behavior was assessed in both dams simultaneously (by observers blind to the pup treatment group) for the 90 min period following reunion. Thus, for each set of ‘surgery’ pups, there was a matched group of ‘sham surgery’ pups, with length of exposure to the ice, duration of separation from the dam, sex composition, and number of pups returned to the dam matched between the groups. Furthermore, although each set of pups returned to a given dam underwent the same experimental procedure, each set consisted of male and female pups combined across two birth litters. Thus, genetic differences between the experimental groups were minimized. Separate litters were bred and were injected with either SAL or MOR (but not separated from the dam) on the day of birth, and were kept with the birth mothers to serve as injection-only controls. Thus, within each drug condition, subjects were either exposed to surgery (SURG), or served in one of two control groups: those exposed to anesthesia and separation from the dam for the same duration, and administered the same drug as a set of pups in the SURG group Table 1 Number of animals surviving to adulthood in each of the surgical/control conditions SURG
Male Female
SHAM
349
(SHAM), or those simply injected with the drug (INJ). The number of male and female subjects surviving to adulthood in each of the neonatal conditions is outlined in Table 1.
3. Adulthood pain measurement Male and female mice were subjected to three adulthood nociceptive tests: tail-withdrawal (TW), hot-plate (HP), abdominal constriction (AC), in that order, with at least 1 week intervening between testing sessions by experimenters blind to the neonatal treatment group. All of the animals surviving to adulthood were included in the HP/TW tests. Due to the increased amount of time required for the AC test compared to the TW and HP tests, only a subset of animals were tested in the AC test. No fewer than five subjects (randomly chosen across the 3–4 litters in each group) in any sex/surgical condition/drug condition group contributed data to the AC test. 3.1. HP test The hot-plate apparatus (Model 39C, IITC, Woodland Hills, CA) consists of a temperature-controlled metal surface (53 8C) with a cylindrical Plexiglas enclosure placed on the surface. Mice were placed inside the cylinder and a timer was started when all four paws were in contact with the surface. Latency to hind paw lick or shake was recorded, and the animal quickly removed from the surface. A 60-s cut-off latency was imposed on non-responsive animals. 3.2. TW test A hot-water circulator (Fisher Isotemp) with water maintained at a temperature of 53 8C was used for the TW test. Mice were lightly restrained by wrapping in a towel, and the distal end of the tail was immersed in the water. Latency to vigorously withdraw the tail from the hot water was recorded. A 15-s cut-off was employed. Two readings were taken for each test; latency was the average of the two readings. 3.3. AC test Subjects were habituated to Plexiglas cylinders for 20 min, after which they were given an intraperitoneal injection of 0.9% acetic acid (10 ml/kg injection volume). Abdominal constrictions were then observed for 30 min following injection. A constriction was defined as full-body extension with at least one hindpaw leaving the floor of the observation chamber.
INJ
SAL
MOR
SAL
MOR
SAL
MOR
24 24
21 22
29 27
23 13
9 8
9 7
4. Data analysis Hot-plate latencies, tail-withdrawal latencies, and total number of abdominal constrictions were the dependent
350
W.F. Sternberg et al. / Pain 113 (2005) 347–353
variables in three separate analyses. Each analysis consisted of a 2!3!2 factorial ANOVA, with sex, neonatal surgical condition, and neonatal drug condition as between-subjects variables. Neonatal USVs were also analyzed using 2!2 factorial ANOVA, with surgical and drug condition as between-subjects variables. Finally, maternal behavior over the 90-min observation period on days 0 and 7 was quantified as being pup-directed (PUP: nesting, grooming pups) or selfdirected (SELF: grooming self, eating/drinking) and analyzed using 2!2!2 mixed-factorial ANOVA, with neonatal surgical condition (SHAM or SURG), drug condition (SAL, MOR) as between groups variables and behavioral target (PUP vs. SELF) as a within-subjects variable. Significance level was set to aZ0.05; post-hoc comparisons were performed using Fisher’s post-hoc tests where warranted.
5. Results 5.1. Ultrasonic vocalizations Significant main effects were noted for surgery condition (F2,17Z12.43, P!0.001), drug (F1,17Z30.43, P!0.001), and a significant surgery!drug interaction (F2,17Z4.14, P!0.05) was present. Post-hoc analysis of the interaction indicates that in SAL-treated animals, the surgery manipulation increased USVs compared to INJ and SHAM controls (who did not differ). However, when MOR was administered following the surgery, USVs were reduced to that of unoperated controls (see Fig. 1). The main effect of drug (SALOMOR) suggests an overall reduction in USV induced by morphine. However, SHAM and INJ subjects given MOR did not differ from each other. 5.2. Maternal behavior The duration of time the dam spent attending to pups or to herself was quantified and expressed as a percentage of total time spent engaged in either type of behavior. Time spent wandering around the cage or lying alone was not
Fig. 2. Maternal behavior during a 90-min period following surgery (D0) and 1 week later (D7). Percentage of time spent attending to pups vs. self (proportion of each bar in the black region) did not differ based on condition (surgery vs. sham). Overall, dams spent a vast majority of the observation period engaged in pup-directed behavior.
included in the quantification. On both days 0 and 7, dams spent a far greater percentage of the 90-min observation period engaged in pup-directed as opposed to self-directed behaviors (F1,19Z137.47, P!0.001 on day 0; F1,17Z 321.97, P!0.001 on day 7). On neither day did the time spent engaged in pup vs. self-directed behavior vary according to surgical or drug condition (all two-and threeway interactions were non-significant). Thus, observable maternal behavior patterns were the same regardless of surgical/drug condition of the pups (see Fig. 2). 5.3. Tail-withdrawal Adult male mice displayed significantly longer tailwithdrawal latencies than females (significant main effect of sex F1,155Z8.19, P!0.001), regardless of neonatal condition. A significant surgical condition!drug interaction was present (F2,155Z8.01, P!0.001). Subjects undergoing the surgery without post-operative analgesic treatment (SURG-SAL) displayed longer tail-withdrawal latencies than injection-only subjects (INJ-SAL) and sham-operated controls (SHAM-SAL), as well as those undergoing surgery with post-operative morphine (SURG-MOR). However, subjects in the sham-operated control group who received morphine (SHAM-MOR) displayed elevated tail-withdrawal latencies compared to all other control subjects (see Fig. 3). 5.4. Hot-plate
Fig. 1. Elevation in stress vocalizations induced by neonatal surgical procedure. Surgery increased vocalizations compared to sham-operated control and injection-only subjects (who did not differ). Morphine reduced the surgery-induced vocalizations. Morphine also reduced vocalizations to a lesser extent in unoperated subjects.
Two significant main effects were present: surgical condition (F2,204Z10.83, P!0.001; SURGZSHAMO INJ) and drug condition (F1,204Z8.15, P!0.01; SALO MOR). There were no significant interactions, nor any sexdependent effects. Thus, male and female subjects exposed to both the neonatal surgical condition and the sham-operated
W.F. Sternberg et al. / Pain 113 (2005) 347–353
351
Fig. 3. Long-term changes in latency to tail-withdrawal from a noxious heat stimulus (50 8C water immersion) induced by untreated pain exposure on day of birth. *Indicates significant increase in withdrawal latency in salinetreated animals undergoing surgery treatment compared to sham-operated and injection-only controls receiving saline. #Indicates significant reduction in latency compared to surgery-saline. ^Indicates significant elevation compared to all other groups.
Fig. 5. Abdominal constrictions in response to .9% acetic acid (i.p.) in subjects exposed to treated or untreated surgical pain on the day of birth. Subjects receiving surgery without morphine analgesia exhibited significantly fewer constrictions than subjects receiving surgery with analgesic treatment, or unoperated controls (both sham-operated and injection-only) (*indicates significantly fewer constrictions than all other groups).
control procedure exhibited elevated hot-plate latencies compared to injection-only controls (see Fig. 4).
adult nociceptive sensitivity. The results are clearest in the AC assay, where neonatal surgery resulted in a dramatic reduction in visceral nociceptive behavior that was reversed by post-operative morphine administration immediately after the surgery. Subjects exposed to the stressful, but non-nociceptive, aspects of surgery (cold exposure, separation from the dam) were no different from injection-only control subjects, and morphine exposure alone did not elevate pain behavior on this test. Taken together, these results suggest that the long-term effect of the surgery on reducing visceral pain behavior in adulthood is due to the nociceptive aspects of the neonatal procedure. That the surgical procedure itself produced postoperative pain is supported by the results of the neonatal behavioral observations. Upon return to the dam, subjects undergoing surgery without post-operative analgesic treatment exhibited far more distress vocalizations than unoperated subjects (both sham-operated and injectiononly controls), or operated subjects exposed to morphine. In the hours following surgery, and 1 week later, maternal treatment among operated and unoperated groups did not differ, suggesting that the neonatal procedure, rather than maternal treatment during the pre-weanling period, was responsible for the adulthood alteration in visceral nociception. The results on the TW and HP tests lead to similar (but not identical) conclusions. As in the AC test, subjects in the SURG-SAL group displayed reduced pain behavior (longer withdrawal latencies) on the TW test compared to unoperated subjects, and those receiving morphine following surgery. However, our conclusions are limited by an apparent elevation in tail-withdrawal latency induced by neonatal morphine in the sham-operated control group. Since the elevation was not noted in SHAM-SAL subjects, this finding cannot simply be attributed to the stress of the neonatal procedure, and is unexplained. However, we note the similarity in pattern of the long-term reduction of pain
5.5. Abdominal constriction A significant surgical condition!drug interaction was present (F2,94Z7.35, PZ0.001), with subjects in the SURG-SAL condition displaying fewer constrictions than subjects in the SHAM-SAL and INJ-SAL groups, and fewer constrictions than subjects in the SURG-MOR group. Subjects in the three surgical conditions treated with MOR did not differ from one another (see Fig. 5). Neonatal morphine treatment alone did not alter adulthood constriction behavior (MORZSAL in INJ and SHAM groups). The effect was present for both male and female subjects (no significant sex-dependent interactions).
6. Discussion The results of this study provide strong support for the existence of long-lasting effects of neonatal surgical pain on
Fig. 4. The effects of neonatal stress on hot-plate latency (53 8C). Subjects in both the surgery and sham-operated control conditions displayed elevated hot-plate latencies compared to injection-only control subjects.
352
W.F. Sternberg et al. / Pain 113 (2005) 347–353
behavior resulting from neonatal surgery on the TW and AC tests. On the hot-plate test, the long-term alterations (again, elevations in latency) appear to be due to the stressful aspects of the procedure alone, with no additional alterations occurring due to the nociceptive aspects of the procedure. Subjects in the SURG and SHAM groups exhibited elevations in hot-plate latency compared to INJ subjects, regardless of drug treatment. Thus, the nociceptive circuitry underlying hot-plate responses is susceptible to early influences of stress (whether or not that stress results from a painful procedure). We are not entirely surprised that the findings differ on the HP test compared to the other two nociceptive assays. Both the TW and AC tests assess reflexive responses to noxious stimuli. The HP test measures a supraspinally organized behavior pattern that has been shown to be affected by non-painful perinatal stressors in previous studies (Sternberg and Ridgway, 2003). The results obtained here confirm and extend those of Bhutta et al. (2001) who demonstrated that neonatal formalin exposure in rats increased hot-plate and tail-flick latencies when tested in adulthood. As in the present study, morphine exposure following formalin treatment reversed the effects of neonatal surgery (Bhutta et al., 2001). Despite differing neonatal noxious stimuli (surgical pain on the day of birth; repeated formalin injections during the first week of life) and the cross-species comparison, the findings were quite similar, suggesting a particularly robust phenomenon. We do not know the mechanism underlying the decreased nociceptive sensitivity we observed in the present investigation. However, we believe the alteration to be a global nervous system phenomenon, given its generalizability across acute cutaneous and tonic, subcutaneous/ visceral algesiometric assays. It is evident that the subjects exposed to surgery without post-operative morphine on the day of birth displayed lower nociceptive sensitivity in adulthood in response to a variety of noxious stimuli. It is plausible that early pain exposure (perhaps mediated by neonatal activation of the hypothalamic–pituitary–adrenal stress-axis) upregulates basal endogenous opioid production or receptor density such that antinociceptive mechanisms are more readily evoked, perhaps even by the stress associated with nociceptive testing itself in adulthood. We have data to suggest that the neonatal surgical procedure elevates adulthood stress-induced analgesia levels compared to injection-only controls (unpublished observations), but we have not yet tested this procedure in sham-operated control subjects and therefore cannot conclude that the longterm alterations are due to surgical pain per se. Various perinatal manipulations, such as prenatal stress (Kinsley et al., 1988), prenatal ethanol exposure (Nelson et al., 1985, 1986) and post-natal handling (Pieretti et al., 1991) have been shown to affect opioid-mediated processes when tested in adulthood. We are currently testing the hypothesis that the alteration in pain behavior following neonatal pain
exposure is due to elevated endogenous opioid activity in adulthood. The results reported here are also consistent with a small number of clinical investigations on infant pain exposure. Children who were born at an extremely low birth weight (ELBW) and exposed to invasive procedures in the NICU are reported by their parents to display lowered sensitivity to everyday pains at 18 months (Grunau et al., 1994), and display lowered facial responsiveness to immunization at 4 and 8 months of age (Oberlander et al., 2000). Interestingly, pain behavior following immunization does not differ between toddlers who had major surgery as infants and control subjects when the surgeries were performed with pre-emptive morphine analgesia (Peters et al., 2003). There have also been reports of enhanced pain behavior during immunization in children that had undergone circumcision as neonates, an effect that is reversed by local anesthetic at the time of the procedure (Taddio et al., 1995, 1997). The difficulty in reconciling conflicting findings in the human literature is compounded by the barriers to conducting studies on human infants with appropriate placebo controls. Therefore, understanding the mechanisms of plasticity following neonatal pain exposure will require useful animal models. We believe the model of neonatal surgical pain in the mouse will be a useful one for future investigations. Abdominal surgery is a complex stimulus, involving nociception, skin wounding, local inflammation (indeed, the chromic gut sutures used here are disintegrated by an inflammatory response), cold exposure, and maternal separation. Thus in many ways, the experience is similar to that experienced by human pre-term neonates requiring intensive care. Local changes in nociceptive circuitry induced by the neonatal procedure are unlikely to account for the changes in pain behavior since we do not test adulthood nociceptive sensitivity at the site of the neonatal injury. However, we cannot rule out such a mechanism because the spinal segments responsible for transmitting visceral nociceptive stimuli and mediating the tail-withdrawal reflex may be close to that affected by the laparotomy. We did not test whether mechanical hyper- or hypoalgesia was present in adulthood at the site of cutaneous injury, which may provide useful information regarding the peripheral reorganization that may take place due to this procedure. Future investigations using the neonatal surgical model should confirm the extent of local changes induced by this noxious stimulus. Laboratory investigations of neonatal pain exposure are often aimed at improving clinical practice regarding pain management in the human neonate. Neonates (especially pre-term neonates) are often under-medicated for pain, despite undergoing frequent painful procedures (Johnston et al., 1997), due in large part to misconceptions regarding the inability of neonates to experience pain and the safety of analgesics or anesthetics during the neonatal period (Anand, 2001). However, it is clear that human neonates exhibit
W.F. Sternberg et al. / Pain 113 (2005) 347–353
behavioral and physiological responses during painful procedures (e.g. McCulloch et al., 1995), suggesting that infants are indeed capable of experiencing pain. Therefore, infants’ pain experience should be considered alongside of concerns regarding the safety of neonatal anesthesia or analgesia. The findings reported here suggest that in addition to considering ethical obligations to alleviate pain in the neonate, one should also consider the long-term (and potentially permanent) consequences of aberrant sensory stimulation during the neonatal period for the development of pain sensing pathways in the nervous system.
Acknowledgements The authors wish to thank James Steinemann and Hiro Takahashi for help with neonatal surgical procedures and Dr Jeffrey Mogil for proofreading an earlier version of this paper. This research was supported by a Research in Undergraduate Institutions grant and a Haverford College faculty research grant to WFS.
References Anand KJ. Effects of perinatal pain and stress. Prog Brain Res 2000;122: 117–29. Anand KJ. Consensus statement for the prevention and management of pain in the newborn. Arch Pediatr Adolesc Med 2001;155:173–80. Anand KJ, Coskun V, Thrivikraman KV, Nemeroff CB, Plotsky PM. Longterm behavioral effects of repetitive pain in neonatal rat pups. Physiol Behav 1999;66:627–37. Bernardi M, Genedani S, Bertolini A. Behavioral activity and active avoidance learning and retention in rats neonatally exposed to painful stimuli. Physiol Behav 1986;36:553–5. Bhutta AT, Rovnaghi C, Simpson PM, Gossett JM, Scalzo FM, Anand KJ. Interactions of inflammatory pain and morphine in infant rats: longterm behavioral effects. Physiol Behav 2001;73:51–8. Fitzgerald M, Beggs S. The neurobiology of pain: developmental aspects. Neuroscientist 2001;7:246–57. Fitzgerald M, Gibson S. The postnatal physiological and neurochemical development of peripheral sensory C fibres. Neuroscience 1984;13: 933–44. Garg S, Narsinghani U, Bhutta AT, Rovnaghi C, Anand KJS. Long-term effects of neonatal pain: the animal literature. In: McGrath PJ, Finley GA, editors. Pediatric pain: biological and social context, vol. 26. Seattle, WA: IASP Press; 2003. p. 1–22. Grunau RV, Johnston CC, Craig KD. Neonatal facial and cry responses to invasive and non-invasive procedures. Pain 1990;42:295–305. Grunau RV, Whitfield MF, Petrie JH. Pain sensitivity and temperament in extremely low-birth-weight premature toddlers and preterm and fullterm controls. Pain 1994;58:341–6. Johnston CC, Stevens B, Craig KD, Grunau RV. Developmental changes in pain expression in premature, full-term, two- and four-month-old infants. Pain 1993;52:201–8. Johnston CC, Collinge JM, Henderson SJ, Anand KJ. A cross-sectional survey of pain and pharmacological analgesia in Canadian neonatal intensive care units. Clin J Pain 1997;13:308–12.
353
Kinsley CH, Mann PE, Bridges RS. Prenatal stress alters morphine-and stress-induced analgesia in male and female rats. Pharmacol Biochem Behav 1988;30:123–8. Lidow MS. Long-term effects of neonatal pain on nociceptive systems. Pain 2002;99:377–83. Lidow MS, Song ZM, Ren K. Long-term effects of short-lasting early local inflammatory insult. Neuroreport 2001;12:399–403. McCulloch KM, Ji SA, Raju TN. Skin blood flow changes during routine nursery procedures. Early Hum Dev 1995;41:147–56. McLaughlin CR, Lichtman AH, Fanselow MS, Cramer CP. Tonic nociception in neonatal rats. Pharmacol Biochem Behav 1990;36: 859–62. Nelson LR, Taylor AN, Lewis JW, Branch BJ, Liebeskind JC. Opioid but not nonopioid stress-induced analgesia is enhanced following prenatal exposure to ethanol. Psychopharmacology 1985;85:92–6. Nelson LR, Taylor AN, Lewis JW, Branch BJ, Liebeskind JC. Morphine analgesia is potentiated in adult rats prenatally exposed to ethanol. Brain Res 1986;372:234–40. 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. Peng YB, Ling QD, Ruda MA, Kenshalo DR. Electrophysiological changes in adult rat dorsal horn neurons after neonatal peripheral inflammation. J Neurophysiol 2003;90:73–80. Peters JW, Koot HM, de Boer JB, Passchier J, Bueno-de-Mesquita JM, de Jong FH, Duivenvoorden HJ, Tibboel D. Major surgery within the first 3 months of life and subsequent biobehavioral pain responses to immunization at later age: a case comparison study. Pediatrics 2003; 111:129–35. Phifer CB, Terry LM. Use of hypothermia for general anesthesia in preweanling rodents. Physiol Behav 1986;38:887–90. Pieretti S, d’Amore A, Loizzo A. Long-term changes induced by developmental handling on pain threshold: effects of morphine and naloxone. Behav Neurosci 1991;105:215–8. Ren K, Anseloni V, Zou S-P, Wade EB, Novikova SI, Ennis M, Traub RJ, Gold MS, Dubner R, Lidow MS. Characterization of basal and reinflammation-associated long-term alteration in pain responsivity following short-lasting neonatal local inflammatory insult. Pain 2004; 110:588–96. Ruda MA, Ling QD, Hohmann AG, Peng YB, Tachibana T. Altered nociceptive neuronal circuits after neonatal peripheral inflammation. Science 2000;289:628–31. Shimada C, Kurumiya S, Noguchi Y, Umemoto M. The effect of neonatal exposure to chronic footshock on pain responsiveness and sensitivity to morphine after maturation in the rat. Behav Brain Res 1990;36:105–11. Sternberg WF, Ridgway CG. Effects of gestational stress and neonatal handling on pain, analgesia, and stress behavior of adult mice. Physiol Behav 2003;78:375–83. Sternberg WF, Smith LD, Scorr L. Nociception and antinociception on the day of birth in mice: sex differences and test dependence. J Pain 2004;5:420–6. Tachibana T, Ling QD, Ruda MA. Increased Fos induction in adult rats that experienced neonatal peripheral inflammation. Neuroreport 2001;12: 925–7. Taddio A, Goldbach M, Ipp M, Stevens B, Koren G. Effect of neonatal circumcision on pain responses during vaccination in boys. Lancet 1995;345:291–2. Taddio A, Katz J, Ilersich AL, Koren G. Effect of neonatal circumcision on pain response during subsequent routine vaccination. Lancet 1997;349: 599–603.