Spinal Anesthesia in Infant Rats: Development of a Model and Assessment of Neurologic Outcomes

• General anesthesia during infancy is associated in animals with enhanced apoptosis of neurons in brain, but spinal anesthesia has not been tested

• In 7-day-old infant rats, 6 h of exposure to isoflurane induced brain apoptosis whereas spinal bupivacaine did not have this effect in either the brain or spinal cord

THE potential for harmful effects of general anesthesia on the developing brain has been described as pediatric anesthesiology's greatest current challenge.1Anesthetic agents appear to have the greatest neurodegenerative effect when exposure occurs during the period of rapid synaptogenesis.2–15Among species, peak synaptogenesis occurs at different rates and at different parts of the life cycle. This critical period of synaptogenesis in the rat lasts approximately from the day of birth to day 10, peaking in cortex at day 7. In humans, this corresponds approximately to a period from the third trimester of gestation and the initial 3 yr of postnatal life. In rats and mice, brain growth is most rapid during the first 2 weeks of postnatal life. The brain's maturational state in postnatal day 7 mice is comparable to that of 7-day-old rats and corresponds to that of premature human neonates.16,17 

Studies in infant rat models have shown that exposure to a number of classes of sedative-hypnotic medications, including benzodiazepines, N -methyl-d-aspartate receptor antagonists, and inhalation anesthetics produce striking degrees of apoptotic neurodegeneration, particularly at 7 days postnatal age (P7) but not at postnatal day 21 (P21). Similar concerns have been raised by studies in other species, including infant mice and young primates in vivo , and in brain slice preparations in vitro .2–15Conversely, other investigators, including one of us (S.S.), have found increased brain apoptosis in sections of parietal and temporal cortex and thalamus in P7 rats with prolonged (6-h) ketamine exposure, but not with brief (1-h) exposure.4,18During programmed cell death, also termed apoptosis, a cascade of autodigestive enzymes known as caspases are activated. Cleavage of the final member of this cascade, caspase-3, will lead to activation of endonucleases, resulting in apoptotic cell death.19 

Case-control studies have shown that infants undergoing surgery during general anesthesia, especially with more than one procedure,20may have a higher frequency of neurodevelopmental difficulties than infants who did not require anesthesia in infancy.20Conversely, a recent twin study did not find evidence for harmful sequelae of anesthesia in infancy.21Case-control studies do not fully answer the question of whether the anesthesia per se  is associated with increased risk, due to the confounding effect of surgery, postsurgical stress, and other comorbidities that may be more common among infants requiring surgery.

To address these concerns, a large, international, randomized controlled trial is in progress examining neurodevelopmental outcomes at 2 and 5 years of age in a group of infants undergoing inguinal hernia repairs with spinal or general anesthesia.22The rationale for this trial is based in part on the assumption that spinal anesthesia would not be expected to produce harmful effects on neurodevelopment. Nevertheless, this assumption has not previously been examined in infant animal surrogate models.

The current study describes a technique for spinal anesthesia in the young rat. Our general hypothesis is that spinal anesthesia is neurologically benign in young rats. In particular, we hypothesized that: (1) a spinal bupivacaine dose could be identified that produced signs of lower body blockade without signs of high spinal blockade, severe physiologic instability, or high mortality; (2) these doses would not increase apoptotic neurodegeneration in the brain or spinal cord (as assayed by cleaved caspase-3 staining), as seen with prolonged exposure to a volatile anesthetic; (3) spinal anesthesia would not produce long-term deficits in motor performance in those rats when tested as adults; and (4) histologic effects of spinal anesthesia on lumbar spinal cord would appear benign.

With the approval of the Investigational Review Board and adherence to the Guide for the Care and Use of Laboratory Animals,23all experiments used Sprague-Dawley postnatal day 7 (P7), 14 (P14), and 21 (P21) male and female rat pups (Charles River Laboratories, Wilmington, MA). They were housed in a room on a 12-h light/dark cycle with free access to water. Rat pups were kept in cages with their littermates and mothers until weaning at postnatal age 21 and were housed with their littermates thereafter.

The experiments were conducted in a temperature-controlled acrylic container maintained at 36.7°C. In order to minimize handling of the pups, direct core temperature measurements were not obtained. However, similar conditions resulted in core body temperatures between 36.5 and 37.5°C.11 

The rat pups were handled repeatedly by investigators to diminish effects resulting from stress-induced analgesia. The P7 and P14 rats were separated from their mothers during spinal injection and behavioral testing, but were promptly returned for warmth and breast-feeding thereafter. The P21 rats were allowed free access to water and food at all times except during the spinal injection and behavioral testing. Rat pups for each age group were randomly assigned to spinal and general anesthesia as described in subsequent sections.

The P7, P14, and P21 rats underwent baseline testing of sensory and motor functions (see Behavioral Assessments for Sensory and Motor Blockade section below) immediately before spinal injection. The content of the spinal solutions were blinded to the investigator who performed the intrathecal injections (BY). With the animals in a prone position, the spinal solutions were injected intrathecally at the L4–L5 or L5–L6 level using a 100-μl syringe (26S-gauge needle, model 801 RN; Hamilton Company, Bonaduz, Switzerland). Intrathecal placement of the needle tip was confirmed by observation of a tail flick.

Preliminary experiments indicated that awake P7 rats could be held unanesthetized sufficiently still to perform these injections with high success rates, but that sufficient immobility was more difficult to achieve in awake P14 and P21 rats. For this reason, experiments in P14 and P21 rats were all performed after brief (less than 2 min) inhalation of isoflurane 2–4% (Baxter Healthcare Corporation, Deerfield, IL) vaporized in oxygen. The experiments in P7 rats were performed with separate groups of rats either awake or during brief isoflurane-anesthetized conditions, respectively. Doses were varied by maintaining a constant concentration of bupivacaine 7.5 mg/ml commercial solution (Hospira, Inc., Lake Forest, IL) and injecting varying volumes scaled to the rat pup's body weight. Injected doses were 3.75 mg/kg as a low dose and 7.5 mg/kg as a high dose in all rats. After spinal injections, animals were placed in a warm, flat cage and sensory and motor functions were measured every 10 min until full recovery and for a minimum of 40 min after injection. All rats receiving bupivacaine had successful blocks, as determined by sensory and motor functional impairments detailed in the Behavioral Assessments for Sensory and Motor Blockade section below.

Groups of rats were exposed to either 1 h or 6 h of 1% inspired concentration of isoflurane in oxygen in a temperature-controlled chamber. Isoflurane was delivered using calibrated Matrx vaporizer (Matrx Medical Inc, Orchard Park, NY). Isoflurane and oxygen concentration were monitored by Capnomac Ultima Anesthesia Monitor (Datex-Ohmeda, Instrumentarium Corp., Helsinki, Finland), and animals were visually inspected for respiratory effort and skin color. Although many previous publications on prolonged general anesthetic exposure in infant rats have not provided glucose supplementation during administration of the anesthetic, there remains a concern that the neurologic effects of prolonged general anesthesia could be exacerbated by lack of substrate availability.7In order to account for this possibility, separate groups of animals at each of the three ages receiving 6-h isoflurane exposures were treated with and without dextrose supplementation, respectively. Supplementation involved intraperitoneal injections of 5% dextrose, 0.05 ml, given hourly beginning after 1 h of anesthetic exposure.††

Pilot experiments were performed to evaluate the spread of injectates in the spinal subarachnoid space. Using the same technique and age-specific injection volumes previously described, rats received spinal injections of methylene blue 1 mg/ml solution. The extent of spread of the methylene blue was examined upon animal necropsy 10 min after intrathecal injections.

Blockade of thermal nociception was assessed using a modified hot plate test, similar to what we have used previously in studies of peripheral nerve blockade.24,25Hind paws were exposed in sequence (left then right) to a hot plate (model 39D hot plate analgesia meter; IITC Inc., Woodland Hills, CA) at 52°C for P7 and P14 pups and 56°C for P21 adolescents. The time (thermal withdrawal latency) until the rats lifted their paws was measured with a stopwatch. After 12 s, the tested paw was removed by the experimenter to avoid injury to the animal or the development of hyperalgesia. This test was repeated three times (with a 10-s pause between tests) for each rat at every time point. Thermal withdrawal latencies were measured every 10 min for at least 40 min after the intrathecal injection and until full recovery was observed.

Blockade of mechanical nociception was assessed by hind paw withdrawal to von Frey filaments. A series of filaments of increasing bending force was applied, and the force in grams of the filament that first evoked a withdrawal response was recorded. Mechanical withdrawal thresholds were recorded at baseline and every 10 min for at least 40 min after the intrathecal injection and until full recovery was observed. Of note, for both the thermal and mechanical withdrawal tests, animals were observed for the possibility of exhibiting motor blockade without sensory blockade, i.e. , by absence of lower limb movement accompanied by vocalization or signs of upper body escape responses. This was not observed for any animal.

Motor performance of the lower extremities was assessed by a qualitative score. For each leg, if there was no spontaneous or evoked movement, the contribution to the score was zero. If there was partial movement, the contribution was one, and if there was normal movement, the contribution to the score was two. Thus, in summing the values for both legs, the score could range from zero (complete blockade) to four (normal).

Motor impairment was assessed on the P30 rats that had undergone spinal bupivacaine or saline injections or 1 h or 6 h of 1% isoflurane general anesthesia interventions at postnatal day 7. These rats were introduced to a Dual Species Economex Rotarod (Columbus Instruments, Columbus, OH) using a spindle rotating at 10 rotations per min.11Each rat was tested three times with 10-min intervals between each assessment. The maximal latency for each trial was 300 s before removal from the spindle. The average of the three assessments was used for data analysis. The investigator was blinded to the treatment groups.

The animals were euthanized with intraperitoneal injection of sodium pentobarbital 100 mg/kg (Hospira, Inc.) and transcardiac perfusion performed with saline followed by 4% paraformaldehyde 6 h after the treatment periods ended. The spinal anesthesia groups (receiving bupivacaine or saline control), sham group, and 1-h isoflurane or 6-h isoflurane groups underwent perfusion 6 h after initiation of the anesthetic intervention. After gentle dissection of the brain and spinal cord, the organs were stored overnight at 4°C in 4% paraformaldehyde and then transferred to a 30% sucrose solution at 4°C until sectioning, as detailed in the subsequent section. All samples were coded to allow for blinded histologic evaluation.

A cohort of P7, P14, and P21 rats treated with spinal saline, spinal anesthesia at high and low dose, 1 h or 6 h (with or without dextrose as detailed previously in the section on General Anesthesia Exposure) of 1% isoflurane general anesthesia underwent cardiac puncture 10 min after either spinal injection or the end of general anesthesia for the blood gas analysis. For all groups receiving spinal injections, appropriate presence or absence of blockade was confirmed by neurobehavioral criteria as described previously. Rats were terminally anesthetized by intraperitoneal injection of sodium pentobarbital 100 mg/kg, the heart was exposed, and 0.3 ml of blood was taken by right ventricular cardiac puncture for immediate blood gas analysis using a blood gas analyzer (pHOx Plus L, Nova Biomedical, Mississauga, Ontario, Canada).

Fixed brains and spinal cords were placed in dry ice for 10–15 min and placed gently on a thin layer of Frozen Tissue Embedding Media (Fisher Diagnostic, Fairlawn, NJ) inside the cryostat box to create a block for sectioning. The brains and spinal cords were sectioned at 20-μm thickness using a cryostat (Leica CM3050S, Bannockburn, IL). Each slide contained four brain sections or six spinal cord sections. Coronal brain sections were obtained at a location that includes the lateral dorsal thalamus (ventrolateral division) and corresponds to an anteroposterior plane approximately 2 mm from the bregma in the adult rat, similar to figure 26 in the Paxinos and Watson atlas.26Lumbar sections of spinal cord were mounted onto Super Plus coated glass slides (Fisher Scientific, Pittsburgh, PA). The slides were allowed to dry at room temperature for 24 h. Initially, the blocking solution (bovine-serum-albumin phosphate-buffered-saline Triton × 100 -Sodium Azide) was applied on the sections for 30 min at room temperature, and then the blocking solution was removed from the sections to apply the primary rabbit anticleaved caspase-3 antiserum (1:2,500, Cell Signaling, Danvers, MA). Slides were then incubated overnight at room temperature. On the next day the sections were rinsed three times with phosphate-buffered saline, and then secondary antibody CY3 antirabbit (1:100, Jackson Immunoresearch, West Grove, PA) was applied. Slides were incubated at room temperature for 2 h. After 2 h, secondary antibody was removed, and the sections were washed three times in phosphate-buffered saline and then given a final wash with 0.0 5M phosphate-buffered saline. Finally, the slides were allowed to dry before they were mounted over three drops of 90% glycerol with a coverslip. Microscopic evaluation was done using a BX-60 light microscope (Olympus, Southall, United Kingdom) at 20× magnification. A Hamamatsu digital camera C10600 ORCA-R²(Photonic K.K., Hamamatsu City, Shizuoka, Japan) was used to obtain images of each section.

The investigator, blinded to the experimental cohort, counted the number of cleaved caspase-3 positive cells on one slide per animal either at the cortex in four brain sections or in six sections at the lumbar spinal cord. The number of cells counted reflected the sum of the cell counts per animal.

Animals from each age group were used for histopathologic evaluation of the lumbar spinal cord. For P7, 10 animals were used, with 5 of the animals treated with intrathecal injection of bupivacaine 3.75 mg/kg and 5 injected with the same volume of saline. For P14 and P21, four animals each received intrathecal bupivacaine 3.75 mg/kg or saline, respectively. Rats received spinal injections as detailed previously, and they were allowed to recover in their cages. All rats receiving bupivacaine (i.e. , not saline control subjects) had successful blocks, as determined by signs of sensory and motor blockade as previously described. Seven days after injections, they were terminally anesthetized with Tribromoethanol (Avertin®, Sigma Chemical Company, St. Louis, MO) 2.5% 20 ml/kg, and underwent cardiac puncture for perfusion with cold phosphate-buffered saline 30 ml, followed by 50 ml cold fixative, containing 1.25% paraformaldehyde, 2.5% glutaraldehyde, 0.03% picric acid in 0.1M cacodylate buffer. In brief, the histologic approach involved the following steps: fixation for 48 h, washing twice in 0.1M cacodylate buffer, postfixation incubation for 48 h in 1% OsO4 in 0.1M cacodylate buffer, washing twice for 1 h in 0.1M cacodylate buffer, dehydrating in graded ethanol solutions: 50, 75, 95, and 100% (two times), suspension in propylene oxide: twice for 1 h, propylene oxide: Epon (EMbed-812; EMS, Hatfield, PA) 1:1: overnight, embedding in Epon, and then sectioned in 1-μm sections. Sections were mounted on slides, counterstained with Epoxy Tissue Stain (EMS), and covered. Slides were coded and images have been obtained at 10× and 60× magnifications with a Hamamatsu digital camera. The images were then examined by an expert in nerve structure and pathology (G.C.) who was blinded to the treatments.

For P14 and P21 rats, where there is sufficient maturation of tracts in the white matter of the lumbar spinal cord by the time of sacrifice 1 week after injections, a semiquantitative rank ordered scoring system was used, as per Estebe and Myers27with the following six-point Likert scale: 0, no pathology in any portion of the field; 1, very few myelinated axons with any mild abnormality; 2, slightly more myelinated axons with any abnormality than 1, but less than 10% abnormal; 3, 10–20% abnormal myelinated axons; 4, signs of moderate axonal degeneration; 5, clear signs of axonal degeneration. Because P7 rats, when sacrificed 1 week later at P14, have much less developed myelination and immature laminar organization of the spinal cord, a more qualitative interpretation was used by the examiner, who rated slides only as normal or not normal for age.

Thermal withdrawal latencies and von Frey mechanical withdrawal thresholds before and after bupivacaine or saline intrathecal injection was compared using a two-way mixed-model factorial ANOVA with age group (P7, P14, P21) and treatment (saline, bupivacaine 3.75 mg/kg, bupivacaine 7.5 mg/kg) as factors, and F-tests were used to compare ages and treatments with a conservative post hoc  Sidak-Bonferroni correction for multiple comparisons to protect against type I errors. Blood gas parameters were compared between six treatment groups (spinal saline, spinal bupivacaine 3.75 mg/kg, spinal bupivacaine 7.5 mg/kg, isoflurane 1% 1h, isoflurane 1% 6 h, isoflurane 1% 6h + dextrose supplementation) using ANOVA with the Dunnett post hoc  procedure to assess mean differences relative to the spinal saline (control) reference group. Hind paw motor response data are presented as median and interquartile range and compared between treatment groups for each age using the two-way nonparametric Friedman test using time as a repeated measures factor. Motor performance as measured by the Rotarod apparatus was compared between four groups (saline, bupivacaine 3.75 mg/kg, isoflurane 1% 1h, isoflurane 1% 6 h) also by ANOVA with the F-test used to ascertain differences with 95% CIs constructed around the mean performance. Apoptosis as measured by caspase-3 activation using immunohistochemistry was compared between treatment groups in P7 rat pups by one-way ANOVA. Histologic injury scores were compared between spinal saline and spinal bupivacaine groups at ages P14 and P21 using Mann–Whitney U tests. Two-tailed values of P < 0.05 with correction as appropriate were considered statistically significant. Statistical analysis was performed with SPSS version 18.0 (SPSS Inc/IBM, Chicago, IL).

Among the rats used, the P7 rat pups weighed an average of 16.23 ± 1.9 g, the P14 rat pups weighed an average of 32.2 ± 2.4 g, and the P21 rat pups weighed an average of 48.8 ± 4.3 g.

Needle placement evoked a tail flick in all rat pups in one or two attempts. Spinal injections were performed only when a tail flick was observed. Pilot experiments using escalating doses of bupivacaine found that 10 mg/kg produced respiratory failure and death in four of five P7 rat pups. Preliminary experiments using bupivacaine 7.5 mg/kg dose injected over 30 s produced behavioral signs of high spinal block, including weakness and reduced response to pinch in the forepaws, labored respiration, and death in 4 of 17 animals at age P7 (23.5%), 5 of 19 at P14 (26.3%), and 5 of 20 at P21 (25%), respectively. Subsequent experiments showed lower death rates (less than 5%) with the 7.5 mg/kg dose when the dose was injected over 1 min and when the isoflurane was discontinued as soon as the tail flick was identified. Pilot experiments with bupivacaine 3.75 mg/kg showed preservation of motor and sensory function in the forepaws, and normal respiratory efforts. Additional animals received intrathecal injection of methylene blue 1 mg/ml in volumes age-matched to equal the 3.75 mg/kg bupivacaine dose (8 μl for P7, 16 μl for P14, and 25 μl for P21). Necropsy showed a subarachnoid spread of dye to thoracic levels.

In pilot experiments in P7 rats, needle placement either awake or under a 2-min isoflurane anesthetic (using an inhalation chamber followed by a nasal cone) gave equivalent technical success and time course of spinal anesthesia.

Hind paw thermal withdrawal latencies were determined for P7 (fig. 1A), P14 (fig. 1B), and P21 (fig. 1C) rats in groups receiving spinal bupivacaine at high dose (7.5 mg/kg) or low dose (3.75 mg/kg), or saline control. For all age groups, both the 3.75 mg/kg and 7.5 mg/kg bupivacaine doses produced dense thermal nocifensive blockade at the first measurement at 10 min following injection, and remained dense at 20 min. In P21 rats, thermal withdrawal latencies remained significantly greater than control values at 30 and 40 min for the high-dose bupivacaine treatment, and at 30 min for the low-dose bupivacaine treatment. In P14 rats, thermal withdrawal latencies remained significantly greater than control values at 30 and 40 min for the high dose, but not low-dose bupivacaine treatment. In the P7 rats, thermal withdrawal latencies remained significantly above baseline at 30 min for both high-dose and low-dose groups. Rats receiving intrathecal saline showed no impairment of responses to thermal stimuli in any age group (fig. 1).

Withdrawal responses to von Frey filaments applied to the hind paws are shown in figure 2for P7 (fig. 2A), P14 (fig. 2B), and P21 (fig. 2C) rats in groups receiving spinal bupivacaine at high dose (7.5 mg/kg) or low dose (3.75 mg/kg), or saline control.

Motor block scores are shown in table 1. For all age groups, no animal receiving spinal saline showed any signs of motor impairment. Bupivacaine at both low dose (3.75 mg/kg) and high dose (7.5 mg/kg) produced dense motor block in all animals that recovered completely by 40 min in all low-dose animals for all age groups, and by 50 min in all high-dose animals for all age groups. Nonparametric analysis using the Friedman test indicated longer motor block for each bupivacaine treatment group compared with saline for at least the first 20 min, with the high-dose bupivacaine of 7.5 mg/kg lasting longer than 3.75 mg/kg (table 1).

Animals receiving saline injections all responded to pinch with forceps on the skin of the back with a startle, escape behaviors, and turning of the head and movement of the forepaws. All animals receiving spinal injections with bupivacaine 3.75 mg/kg showed no behavioral response to pinch over the skin of the back at lumbar and lower thoracic levels, but they had identical brisk withdrawal behaviors to pinch at upper thoracic levels and on the forepaws. Animals receiving spinal injections with bupivacaine 7.5 mg/kg showed no behavioral response to pinch over the skin of the back at lumbar and lower thoracic levels. As noted previously, in pilot experiments, some animals died with this dose of bupivacaine, and others appeared sedated for 1–2 min, and then recovered. Although some sedated animals had sluggish responses to pinch at upper thoracic levels and on the forepaws for 1–2 min, all animals that survived showed brisk withdrawal behaviors to pinch at upper thoracic levels and on the forepaws by the 10-min testing point.

Separate groups of P7, P14, and P21 animals were given spinal anesthesia with bupivacaine 3.75 mg/kg, and then underwent full-thickness incision of the hind paws 10 min later. No animal responded to the incision with any vocalization, flinch, head turning, upper body movement, or escape response. Subsequent pinch of the area of the hind paw incision also failed to induce any behavioral response.

Mixed venous (right ventricular) blood samples were analyzed for blood gases, glucose, and lactate for groups of P7 rats that underwent spinal injections of saline, bupivacaine 3.75 or 7.5 mg/kg, or 1 and 6 h of 1% isoflurane general anesthesia are shown in table 2. Two 6-h isoflurane anesthesia groups were examined: one with and one without intraperitoneal supplementation of dextrose and fluid. General linear model ANOVA showed no overall effect of treatment on values of any parameter, except that the supplemental dextrose group had increased blood glucose concentration and slightly reduced bicarbonate concentrations compared with the other groups (table 2). Blood lactate concentrations were not statistically different among groups. In particular, they were not increased in the groups that received 1% isoflurane for 6 h, with or without dextrose supplementation (table 2).

Groups of rats at 30 days of age who had undergone exposure at P7 to spinal bupivacaine, spinal saline, isoflurane for 1 h, or isoflurane for 6 h, respectively, were tested for motor performance using the Rotarod apparatus. No group differences were found for this test. The times that rats remained on the Rotarod (mean + SD) for the saline, spinal bupivacaine 3.75 mg/kg, 1 h isoflurane, and 6 h isoflurane-treated groups were 272 ± 30, 255 ± 25, 278 ± 27, and 265 ± 27 s, respectively (table 3).

No animals were observed to have swelling at the injection sites, or signs of distress, hyperalgesia, or self-mutilation. Furthermore, none of the animals had any evident residual weakness or impairment of gait when reassessed at 24 h after spinal injections.

Blinded counts of cleaved caspase-3 staining cells showed no increase in apoptosis in the brain or spinal cord of rats receiving spinal anesthesia compared with control subjects or 1-h exposure to 1% isoflurane, but were significantly less than those with 6-h exposure to 1% isoflurane (fig. 3).

Our analysis revealed an overall effect of condition on average number of apoptotic cells in the brain (F = 34.98, P < 0.001, df = 44) and spinal cord (F = 8.48, P < 0.001, df = 35). Post hoc  Dunnett tests found that only the group with exposure to 1% isoflurane for 6 h showed a significant increase in the number of apoptotic cells in both the brain and spinal cord compared with control subjects (fig. 3).

To test the possibility that intrathecal bupivacaine injection has effects on spinal cord structure, we examined tissues (fig. 4) from animals injected with saline (fig. 4A, C, and E) or bupivacaine (3.75 mg/kg) (fig. 4B, 4D, 4F) 7 days after treatment. Examination of plastic embedded spinal cord sections of animals injected at P7 (fig. 4A and B), P14 (fig. 4C and D) and P21 (fig. 4Eand F) showed no differences between saline- and bupivacaine-injected animals at all ages. We found no alterations in the appearance of myelinated fibers, cell bodies, or perineural tissues. No signs of histopathology were noted at any age group. For the P14 animals, histologic injury scores in white matter, as outlined in the Methods section, were a median of 2 (range, 1–5) in the spinal saline group and a median of 1 (range, 0–2) in the spinal bupivacaine 3.75 mg/kg group. The single animal rated as “5” in the saline group was thought by the examiner to have had probable artifact due to processing or tissue trauma, rather than any sign of demyelination or neuronal injury. No other animal had a score above 2. For the P21 animals, median scores were 0 (range 0–0) in the saline group and 0 (range 0–2) in the bupivacaine group, respectively. For the P7 animals, all slides in both the spinal saline and spinal bupivacaine groups were graded as normal for age.

This study developed a model for spinal anesthesia with bupivacaine in infant rats, and identified a dose that produced anesthesia with apparent safety and tolerability. A comparison of the treatment groups demonstrated that there were no significant differences in the appearance of cleaved caspase-3 cells in sections of brain and lumbar spinal cord between the control, bupivacaine, and 1-h isoflurane groups. However, the 6-h isoflurane group showed significantly higher numbers of cleaved caspase-3 cells than the other treatment groups. The effect of prolonged isoflurane anesthesia on apoptosis could not solely be ascribed to lack of substrate from prolonged fasting, because rats who did not receive supplemental glucose did not show either hypoglycemia or lactic acidosis. These observations are consistent with previous studies that demonstrated isoflurane-induced neuroapoptosis without detectable changes in physiologic parameters as noted previously.7,8 

Despite the significant increase in cleaved caspase-3 cells in the 6-h isoflurane group, no motor deficits were detected between the groups in Rotorod testing at 30 days of age. This is consistent with the finding that prolonged isoflurane exposure in neonatal mice led to increased apoptosis but no decrements in spontaneous locomotion in adulthood.28Because the current investigation only tested motor performance, we cannot exclude effects on other more subtle neurologic outcomes based on these data.

Cleaved caspase-3 serves as a marker for apoptosis in the detection of anesthetic-induced developmental neurotoxicity. Exposure to 0.75% isoflurane and 75% N²O for 6 h resulted in increased cleaved caspase-3 cells in the spinal cord of P7 rat pups.11Despite the increase in neuroapoptosis after exposure to isoflurane and N²O, no functional deficits were detected in the tail flick latency or Rotorod tests. Our findings confirm an increase in apoptosis in the brain and spinal cord of neonatal rats exposed to 1% isoflurane for 6 h. However, cleaved caspase-3 counts in neonatal rats receiving intrathecal bupivacaine or 1% isoflurane for 1 h were not different from their control cohorts. A previous report in mice notes that a 1-h exposure to 2% isoflurane, double the concentration used in the current study, resulted in increased cleaved caspase-3 expression.7The discrepancy between our findings might be explained by the difference in concentration of isoflurane or the species used in these experiments.

The neurotoxic effect of local anesthetics on the spinal cord has been studied extensively in adult rats.29Bupivacaine appears not to be toxic to the spinal cord in older rats, although general anesthesia has been shown to induce apoptotic neurodegeneration in the neonatal rat spinal cord.11Our observations in neonatal rats show that intrathecal bupivacaine had a similar morphologic and apoptotic profile as the 1-h isoflurane and control groups. Furthermore, the motor function tested at 30 days showed no differences between the experimental groups.

The specter of developmental neurotoxicity of anesthetic agents in laboratory animals has incited the pediatric anesthesia community to examine their clinical practices. Several investigators have demonstrated that drugs commonly used for general anesthesia and sedation produce accelerated neurodegeneration and neurobehavioral deficits. Regional anesthetic techniques are routinely used as alternatives to general anesthesia in appropriate surgical procedures in pediatric patients. However, the effect of spinal anesthesia on the developing central nervous system has not been previously investigated. Infant animal studies have the theoretical potential to detect age-specific toxicities and thereby prevent harm to human infants. In the current study, we show that spinal anesthesia with bupivacaine is technically feasible in the infant rat, and that a dose of 3.75 mg/kg appears to produce thoracic-level sensory blockade, without evident signs of morbidity or distress. Rats that had undergone spinal or general anesthesia as infants showed no motor impairments as young adults on day 30 of life. Cleaved caspase-3 staining showed no increased frequency of neuronal cell death in spinal bupivacaine-treated animals, unlike those animals treated with isoflurane for 6 h. These findings demonstrate that spinal anesthesia does not produce neuroapoptosis in either brain or spinal cord, suggesting that it may be a feasible alternative to general anesthesia during the perinatal period.

The infant rat has been an important model for the study of the ontogeny of pain mechanisms30,31and for the study of age-related differences in actions of anesthetic and analgesic drugs.32–35Studies in infant rats and rhesus monkeys showing age-specific neurodegeneration after exposures to general anesthetics and sedative-hypnotics have raised serious concerns about how this might apply to human infants undergoing surgery.36,37Retrospective reports demonstrate a link between early and repeated exposure to general anesthetics to subsequent learning and behavioral deficits. As noted in the Introduction, some case-control human studies have reinforced these concerns. However, many factors complicate any attempt to generalize the results of these infant animal studies to infant humans. First, the experimental conditions under which data are collected from rodents differ from the conditions experienced by neonates undergoing surgery and general anesthesia.36,37Second, there are important species differences in the time course of neurologic development and in the critical periods for neurogenesis and synapse formation. An ongoing prospective randomized controlled trial comparing spinal and general anesthesia in infants may further clarify the short-term and longer-term risks of both of these anesthetic techniques.22 

An infant rat model was developed to study immediate and longer term behavioral responses and potential brain neuronal loss following spinal anesthesia. Spinal bupivacaine 3.75 mg/kg produced hind paw surgical anesthesia and did not induce apoptosis. As in previous studies, short-term exposure to subminimal alveolar concentrations of isoflurane appeared benign, while prolonged isoflurane exposure induced apoptosis. Within the limitations imposed by extrapolation from rats to humans, these experiments support the expectation that spinal anesthesia in infants should have benign neurodevelopmental effects.

The authors thank Marcin Liana, Ph.D., Research Associate, Neuroscience Program, Children's Hospital Boston, Boston, Massachusetts, from Dr. Corfas' laboratory for assistance with the histology portions of the research. The authors also acknowledge use of facilities and expertise in immunochemistry provided by Kathryn Commons, Ph.D., Assistant Professor, Department of Anesthesiology, Perioperative and Pain Medicine, Children's Hospital Boston.