Efficacy of limb cooling on the salvage of peripheral nerve from ischemic fiber degeneration

The effect of Angipars on diabetic neuropathy in STZ-induced diabetic male rats: a study on behavioral, electrophysiological, sciatic histological and ultrastructural indices

Diabetes mellitus is the most common metabolic disease with a high prevalence rate in human society that eventually leads to the peripheral nervous system complications in a great number of patients. In the present study, the effects of Angipars on nerve conduction velocity, histological alterations, and behavioral indices were investigated. Diabetes was induced in male rats by intraperitoneal injection of streptozotocin (STZ). Six weeks after STZ injection, animals were divided into five groups control, vehicle, and 3 experimental groups. The vehicle group received 1 mL distilled water daily for two weeks and three experimental groups received, respectively, intraperitoneal injection of 5, 10, and 20 mg/kg Angipars daily for two weeks. Intraperitoneal injection of Angipars, in some extent, could significantly improve behavioral indices of the experimental groups as compared to the vehicle group. Furthermore, mean nerve conduction velocity in the vehicle group showed significant difference with that in the control and the 2nd experimental groups; therefore, Angipars could increase nerve conduction velocity in neuropathic rats. Overall, Angipars exerted positive effects on the treatment and reduction of physiologic symptoms and improvement of sciatic morphological injuries in neuropathic rats.

The relationship between blood pressure and sciatic nerve blood flow velocity in rats with insulin-treated experimental diabetes

Peripheral nerve blood flow (NBF) does not autoregulate but, instead, responds passively to changes in mean arterial pressure (MAP). How this relationship is impacted by insulin-treated experimental diabetes (ITED) is unknown. We tested the hypothesis that ITED will reduce NBF across a range of MAP in Sprague Dawley rats. Following 10 weeks of control or ITED conditions, conscious MAP (tail-cuff) was measured, and under anaesthesia, the MAP (carotid artery catheter, pressure transducer) and NBF (Doppler ultrasound, 40 MHz) responses to sodium nitroprusside (60 µg/kg) and phenylephrine (12 µg/kg) infusion were recorded (regression equations for MAP vs NBF were created for each rodent). Thereafter, motor nerve conduction velocity (MNCV) and nerve vascularization (haematoxylin and eosin stain) were determined. Conscious MAP was higher and MNCV was lower in the ITED group (p < 0.01). In response to drug infusions, the ΔMAP and ΔNBF were similar between groups (p ≥ 0.18). Estimated conscious NBF (based on substituting conscious MAP values into each individual regression equation) was greater in the ITED group (p < 0.01). Sciatic nerve vascularization was similar between groups (p ≥ 0.50). In contrast to the hypothesis, NBF was not reduced across a range of MAP. In spite of increased estimated conscious NBF values, MNCV was reduced in rats with ITED.

Exercise training enhances insulin-stimulated nerve arterial vasodilation in rats with insulin-treated experimental diabetes

Insulin stimulates nerve arterial vasodilation through a nitric oxide (NO) synthase (NOS) mechanism. Experimental diabetes reduces vasa nervorum NO reactivity. Studies investigating hyperglycemia and nerve arterial vasodilation typically omit insulin treatment and use sedentary rats resulting in severe hyperglycemia. We tested the hypotheses that 1) insulin-treated experimental diabetes and inactivity (DS rats) will attenuate insulin-mediated nerve arterial vasodilation, and 2) deficits in vasodilation in DS rats will be overcome by concurrent exercise training (DX rats; 75–85% V̇o2 max, 1 h/day, 5 days/wk, for 10 wk). The baseline index of vascular conductance values (VCi = nerve blood flow velocity/mean arterial blood pressure) were similar ( P ≥ 0.68), but peak VCi and the area under the curve (AUCi) for the VCi during a euglycemic hyperinsulinemic clamp (EHC; 10 mU·kg−1·min−1) were lower in DS rats versus control sedentary (CS) rats and DX rats ( P ≤ 0.01). Motor nerve conduction velocity (MNCV) was lower in DS rats versus CS rats and DX rats ( P ≤ 0.01). When compared with DS rats, DX rats expressed greater nerve endothelial NOS (eNOS) protein content ( P = 0.04). In a separate analysis, we examined the impact of diabetes in exercise-trained rats alone. When compared with exercise-trained control rats (CX), DX rats had a lower AUCi during the EHC, lower MNCV values, and lower sciatic nerve eNOS protein content ( P ≤ 0.03). Therefore, vasa nervorum and motor nerve function are impaired in DS rats. Such deficits in rats with diabetes can be overcome by concurrent exercise training. However, in exercise-trained rats (CX and DX groups), moderate hyperglycemia lowers vasa nervorum and nerve function.

Glucose-stimulated insulin secretion causes an insulin-dependent nitric oxide-mediated vasodilation in the blood supply of the rat sciatic nerve

This study tested the hypothesis that acute hyperglycemia reduces sciatic nerve blood flow in Sprague-Dawley rats. Anesthetized rats underwent cannulation of their right jugular vein (for anesthetic/nutrient/drug infusion) and right carotid artery (for continuous blood pressure measurement via pressure transducer). The left sciatic nerve was exposed and nerve blood velocity (NBV) was assessed from an arterial segment lying superficially along the sciatic nerve (Doppler ultrasound, 40 MHz). NBV and mean arterial pressure (MAP) values were collected, and an index of nerve vascular conductance (NVC) was established (NBV/MAP) at baseline and at 5, 10, 20, and 30 min (and 80 min for insulin) following 1) low glucose infusion, 1 g/kg (50% solution); 2) high glucose infusion, 3 g/kg; 3) high glucose infusion in the absence of a functioning pancreas; 4) euglycemic hyperinsulinemic clamp-insulin infusion (10 mU·kg⁻¹·min⁻¹; 0.4 IU/ml); 5) high glucose infusion + NG-nitro-L-arginine methyl ester (L-NAME) infusion (30 mg/kg); and 6) L-NAME alone followed 20 min later by high glucose infusion. High glucose infusion increased NVC by ~120% relative to baseline (P < 0.001), and this dilation was attenuated in rats without a functioning pancreas (i.e., without insulin secretion) (P = 0.004) and following L-NAME infusion (P = 0.011). Therefore, the vasodilation in rat sciatic nerve during glucose infusion was dependent upon the insulin response and acted through a nitric oxide synthase pathway.

Comparison of the effects of subcutaneous anterior transposition and in situ decompression on the histologic and electrophysiologic properties of the ulnar nerve: an experimental study in a rabbit model

PURPOSE

Although anterior transposition of the ulnar nerve is a commonly performed procedure for treatment of cubital tunnel syndrome, there are concerns that surgical mobilization puts blood supply to the nerve at risk. The purpose of this study was to compare the effects of subcutaneous anterior transposition (SAT) and in situ decompression (ISD) on the ulnar nerve's histologic and electrophysiologic properties in a rabbit model.

METHODS

We assigned 30 male, young adult New Zealand white rabbits to 1 of 3 groups (each group, n = 10). The extensive dissection-SAT group underwent complete neurolysis of the right ulnar nerve from brachial plexus to wrist and a subsequent SAT, the ISD group underwent ISD with 4-cm-long exposure of the ulnar nerve, and the SAT group underwent SAT with 4-cm-long circumferential neurolysis of the ulnar nerve. We killed the animals 7 days after the operation. We assessed electrophysiologic changes by finding the percentage of compound motor action potential of the surgical limb compared with the contralateral limb, and histologic changes by ischemic fiber degeneration.

RESULTS

The extensive dissection-SAT group showed a significantly lower percentage of compound motor action potential and higher grade of ischemic fiber degeneration than the ISD and SAT groups. On the other hand, the percentage of compound motor action potential and the grade of ischemic fiber degeneration were similar between the latter 2 groups without a statistically significant difference.

CONCLUSIONS

This study demonstrated that surgical mobilization of the ulnar nerve during SAT has an effect similar to ISD on the nerve's histologic and electrophysiologic properties at the 7th postoperative day in a rabbit model.

CLINICAL RELEVANCE

In a rabbit model, ISD and SAT appear to affect histologic and electrophysiologic findings of the ulnar nerve equally at postoperative day 7, which suggests that they would have the same relative safety in a clinical setting.

Neurophysiologic responses of peripheral nerve to repeated episodes of anoxia

OBJECTIVE

Determine the effects of serial episodes of anoxia in an in vitro peripheral nerve preparation.

METHODS

The nerve action potential (NAP) from rat sciatic nerve was recorded during 5 cycles of anoxia and reperfusion. Multiple NAP parameters were analyzed as well as stimulus response curves.

RESULTS

The amplitude of the NAP declined to half baseline in 865 s on the first cycle of anoxia and recovered to half baseline during recovery in 470 s. These times increased with successive cycles of anoxia. The current required to produce a half maximal NAP showed a variable initial decrease before increasing with anoxia. The paired-pulse response showed a decline at 2-3 ms interstimulus interval during anoxia but was less dependent of interstimulus interval during recovery. NAP amplitude and velocity decrease over successive cycles of anoxia at a rate greater than in the absence of anoxia.

CONCLUSIONS

The NAP declines slowly when peripheral nerve is exposed to anoxia but returns at least twice as quickly when re-exposed to oxygen. Short periods of anoxia produce long lasting changes in the nerve suggesting greater resistance to anoxia. With serial episodes of anoxia there is gradual NAP amplitude reduction and increase in duration and latency.

SIGNIFICANCE

Anoxic-preconditioning appears in isolated peripheral nerve.

Ameliorative potential of montelukast on ischemia-reperfusion injury induced vasculitic neuropathic pain in rat

AIMS

Ischemia-reperfusion (I/R) event in vascular and nervous system has been documented to rising ischemic and vasculitic neuropathic pain, clinically resembles the complex regional pain syndrome (CRPS). The present study evaluated the effect of montelukast, a cysteinyl leukotriene receptor (Cys-LTC(4) and Cys-LTD(4)) antagonist on ischemia -reperfusion (I/R) induced vasculitic neuropathic pain in rats.

MAIN METHODS

Behavioral parameters were assessed at different time intervals (i.e. 0, 1, 7, 14 and 21st day) and biochemical analysis in sciatic nerve tissue samples were also performed along with histopathological studies.

KEY FINDINGS

Behavioral pain assessment has shown increase in paw and tail withdrawal threshold in montelukast treated groups against thermal and mechanical stimuli as compared to I/R control group. We observed a decrease in the total calcium, thiobarbituric acid reactive substance (TBARS) and myeloperoxidase (MPO) activity levels, whereas there is rise in reduced glutathione level in montelukast treated groups as compared to I/R control group. However, significant behavioral and biochemical results were observed only in medium and high dose of treated groups which were comparable to normal control group. Moreover, histopathological study has revealed the reduction of I/R induced neuronal edema and axonal degeneration due to montelukast.

SIGNIFICANCE

Montelukast has ameliorated I/R induced vasculitic neuropathic pain, these effects may be due to inhibition of lipid peroxidation, reduction of oxidative stress, release of inflammatory mediators and neuroprotective actions. Hence, it could be used as a novel therapeutic agent for the management of vasculitic inflammation related neuropathic pain.

Pharmacological evaluation of tacrolimus (FK-506) on ischemia reperfusion induced vasculatic neuropathic pain in rats

Background

Ischemia reperfusion (I/R) is common in various pathological conditions like diabetic complication, rheumatic arthritis, necrotizing vascular occlusive disease and trauma.

Methods

We have evaluated the effect of tacrolimus (1, 2 and 3 mg/kg, p.o. for 10 consecutive days) on femoral arterial ischemic reperfusion (I/R) induced neuropathic pain in rats. Behavioral parameters (i.e. hot plate, radiant heat, acetone drop, tail heat hyperalgesia, tail flick and tail cold allodynia tests) were assessed at different time intervals (i.e. 0, 1, 4, 7, 10, 13 and 16^th day) and biochemical analysis in serum and tissue samples were also performed along with histopathological studies.

Results

Behavioral pain assessment revealed increase in the paw and tail withdrawal threshold in tacrolimus treated groups against hyperalgesic and allodynic stimuli as compared to the sham control group. We observed a decrease in the serum nitrate and thiobarbituric acid reactive substance (TBARS) levels along with reduction in tissue myeloperoxidase (MPO) and total calcium levels, whereas, rise in tissue reduced glutathione levels in tacrolimus treated groups. However, significant results were obtained in medium and high dose treated group as compared to sham control group. Histopathological study had revealed the increase in the neuronal edema and axonal degeneration in the I/R group whereas, tacrolimus ameliorate these effects.

Conclusion

Our results indicate the anti-oxidative, anti-inflammatory and calcium modulatory actions of tacrolimus. Therefore, it can be used as a therapeutic agent for the treatment of vascular inflammatory related neuropathic pain.

Development of animal model for vasculatic neuropathy: Induction by ischemic-reperfusion in the rat femoral artery

Ischemic-reperfusion (I/R) is common in various pathological conditions like diabetic complication, complex regional pain syndrome type II (CRPS II), necrotizing vascular occlusive disease and trauma. We have developed an animal model of ischemic-reperfusion injury induced nociceptive sensory neuropathy in rats. The model was validated after 2, 4 and 6h of ischemia followed by prolonged reperfusion. The sensory behavioral assessment revealed thermal and mechanical hyperalgesia in paw and in tail which expressed the peripheral and central neuropathic pain respectively. We observed a decrease in the serum IL-10 and nerve conduction velocity and increase in the serum nitrate, malondialdehyde (MDA) and TNF-alpha levels in the 4 and 6h I/R groups in biochemical and electrophysiological evaluations. Histopathological study had revealed the decrease in nerve fiber density in the moderate and severe I/R groups. We selected the moderate (4h) ischemic-reperfusion injury as beneficial model because of the good correlation with clinical status for the development of neuropathy in human associated with severe pain disorders. This model can be used to explore pathophysiological mechanisms implied in the genesis of neuropathic pain and also to evaluate the new analgesic agents, peripheral neuro-vasoactive substances and neuroprotective drugs.

Ischemia–reperfusion injury of peripheral nerve in experimental diabetic neuropathy

The pathogenesis of human diabetic neuropathy likely involves the interplay of hyperglycemia, ischemia, and oxidative stress. Mild-moderate ischemia-reperfusion to streptozotocin (STZ)-induced diabetes results in florid fiber degeneration in diabetic but not in normal nerves. Uncertainty exists as to the influence of duration of diabetes on this susceptibility. We therefore studied diabetic tibial and sciatic nerves using a rat ischemia-reperfusion (IR) model after 1 month and 4 months of diabetes utilizing electrophysiological, behavioral, and neuropathological methods. Electrophysiological abnormalities were present in 1-month diabetic rats (D) and persisted over 4 months. Behavioral scores were decreased markedly at 4 months (p<0.05). Endoneurial edema and ischemia fiber degeneration (IFD) were observed at both the 1-month (p<0.01 and p<0.001) and 4-month (p<0.001) durations in diabetic nerves, whereas only mild or no damage was observed in age-matched control nerves. These findings demonstrate that STZ-induced diabetes exacerbates the morphological and electrophysiological pathology in peripheral nerve to IR injury both in the early timepoint of 1 month and late timepoint of 4 months, although there was a gradation of injury, which is more severe at the later timepoint. Reperfusion exaggerated morphological pathology in 1-month STZ-induced diabetic peripheral nerve.

Schwann cell is a target in ischemia-reperfusion injury to peripheral nerve

Ischemia-reperfusion (IR) causes oxidative injury and ischemic fiber degeneration due to injury of the neuron and axon. In this study, we explore the effect of oxidative stress on Schwann cells, as a specific peripheral nerve target, using our established rat model for IR injury. Fifty-six rats were used. Six groups (N = 8 each) underwent complete hindlimb ischemia for 4 h, followed by reperfusion durations of 0 h, 3 h, 7 days, 14 days, 28 days, and 42 days. One group underwent sham operation (N = 8). We evaluated immunohistochemical labeling for oxidative injury using anti-8-hydroxydeoxyguanosine (8-OHdG). To identify cells committed to apoptosis, we studied immunolabeling to caspase-3 and terminal deoxynucleotidyl transferase (TdT)-mediated dUTP-biotin nick end labeling (TUNEL) positivity. Only minimal positivity was seen in the sham, 0-h, and 3-h groups. Positivity to 8-OHdG, caspase-3, and TUNEL increased significantly in groups undergoing longer reperfusion (8-OHdG, 7-28 days; caspase-3, 14-42 days; TUNEL, 14-42 days). The positive cells surrounding axons were identified as being Schwann cells by their configuration and colabeling with S-100. We conclude that apoptosis of Schwann cells occurs during reperfusion and continues even when axons regenerate. Schwann cell apoptosis could contribute to impairment of axonal function and efficiency of fiber regeneration. Both these abnormalities are known to occur in experimental and human diabetic nerves.

Peripheral nerve ischemia: reperfusion injury and fiber regeneration

We continued our studies of ischemia-reperfusion (IR) injury, extending the reperfusion duration to 42 days to capture the fiber regeneration process. We used a rat model for IR injury produced by ligation and release of nooses around supplying vessels to the sciatic nerve. Fifty-six rats were used. One group (control N = 8) underwent sham ischemia; the other six groups (N = 8 each) underwent complete hind limb ischemia for 4 h followed by reperfusion durations of 0 h (ischemia alone), 3 h, 7 days, 14 days, 28 days, and 42 days. Behavioral and electrophysiological data were obtained immediately before euthanasia. Pathologically, three phases were identifiable: Phase 1 (0-3 h)-minimal pathological changes, minimal edema; phase 2 (7 days, 14 days)-prominent fiber degeneration, endoneurial edema; phase 3 (28 days, 42 days)-abundant small regenerating fiber clusters, minimal edema. Compound muscle action potential (CMAP) was the most sensitive index of neural deficits and recovery, showing progressive recovery beyond 14 days. Severe functional deficits developed immediately and persisted with a trend to recovery at the 42-day time-point. It was concluded that reperfusion, by oxidative injury, worsened nerve function and aggravated fiber degeneration, but in the longer time frame, permitted fiber regeneration to occur.

Nerve and ganglion blood flow in diabetes: an appraisal

Vasa nervorum, the vascular supply to peripheral nerve trunks, and their associated cell bodies in ganglia have unique anatomical and physiological characteristics. Several different experimental approaches toward understanding the changes in vase nervorum following injury and disease have been used. Quantative techniques most widely employed have been microelectrode hydrogen clearance palarography and [14C]iodoantipyrine autoradiographic distribution, whereas estimates of red blood cell flux using a fiber-optic laser Doppler probe offer real time data at different sites along the nerve trunk. There are important caveats about the use of these techniques, their advantages, and their limitations. Reports of nerve blood flow require careful documentation of physiological variables, including mean arterial pressure and nerve temperature during the recordings. Several ischemic models of the peripheral nerve trunk have addressed the ischemic threshold below which axonal degeneration ensues (< 5ml/100 g/min). Following injury, rises in local blood flow reflect acitons of vasoactive peptides, nitric oxide, and the development of angiogenesis. In experimental diabetes, a large number of studies have documented reductions in nerve blood flow and tandem corrections of nerve blood flow and conduction slowing. A significant proportions, however, of the work can be criticized on the basis of methodology and interpretation. Similarly, not all work has confirmed that reductions of nerve blood flow are an invariable feature of experimental or human diabetic polyneuropathy. Therefore, while there is disagreement as to whether early declines in nerve blood flow "account" for diabetic polyneuropathy, there is unquestioned eveidence of early microangiopathy. Abnormalities of vase nervorum and micorvessels supplying ganglia at the very least develop parallel to and together with changes in neurons, Schwann cells, and axons.

A small dose of the immunosuppressive agent FK506 (tacrolimus) protects peripheral nerve from ischemic fiber degeneration

The immunosuppressant agent FK506 (tacrolimus) has proven to be neuroprotective against brain ischemia, but there are no data on potential neuroprotective effects of FK506 in peripheral nerve ischemia. We examined the potential effects of two doses of FK506 in protecting peripheral nerve from ischemic fiber degeneration. Ischemia in the left sciatic nerve of the rat was produced by injecting 2 x 10(6) microspheres (14 microm) into the left femoral, hypogastric, and superior gluteal arteries in proportions of 47%, 37%, and 17%, respectively. After embolization, FK506 was injected into the left femoral, hypogastric, and superior gluteal arteries in doses of 9.4, 7.4, and 3.4 microg, respectively, for the high-dose group and 4.7, 3.7, and 1.7 microg, respectively, for the low-dose group. The control rats were injected with saline. FK506 treatment resulted in dramatic behavioral improvement in nerve function, in the number of functioning nerve fibers, and in the salvage of a majority of nerve fibers from ischemic fiber degeneration in a dose-dependent fashion. These results suggest that a small dose of FK506 protects peripheral nerve from ischemic fiber degeneration and that it may have potential in the treatment of ischemic neuropathy.

The expression of proinflammatory cytokine mRNA in the sciatic-tibial nerve of ischemia-reperfusion injury

We evaluated the proinflammatory cytokines, TNF-alpha and IL-1beta, mRNA expression in the rat sciatic and tibial nerves following ischemia-reperfusion (IR) injury, using competitive RT-PCR, to explore the role of cytokines in IR injury. The expressions of both TNF-alpha and IL-1beta mRNA were related to severity of ischemia and occurred with reperfusion rather than ischemia alone. TNF-alpha gene expression peaked at 24 h of reperfusion, while that of IL-1beta peaked at 12 h. These data support the notion that the proinflammatory cytokines TNF-alpha and IL-1beta are involved in the inflammatory response of IR injury to the peripheral nervous system and may be involved in the pathophysiology of ischemic fiber degeneration.

Hypothermic neuroprotection of peripheral nerve of rats from ischemia-reperfusion injury: intraischemic vs. reperfusion hypothermia

The pathophysiology of ischemic fiber degeneration (IFD) is not known, but mechanisms involved during nerve ischemia differ from those during reperfusion. We have previously demonstrated hypothermic neuroprotection of peripheral nerve from IFD. We now evaluate the efficacy of hypothermia in the intraischemic vs. the reperfusion period, using our established model of ischemia-reperfusion injury. Intraischemic hypothermia resulted in significant recovery of all indices (behavior score, electrophysiology and histology, P<0.01 or 0.05) while hypothermia during reperfusion period showed less improvement, significant only for the histological score compared to normothermia group (IFD index, P<0.05). Once hypothermia was applied in the ischemic period, the resultant neuroprotection continued into the reperfusion period, even if nerve temperature was then raised during the reperfusion period. These results indicate that hypothermic neuroprotection is more efficacious during the intraischemic period than during reperfusion, when a lesser degree of neuroprotection ensued.

Effects of hindlimb temperature on sciatic nerve laser Doppler vascular conductance in control and streptozotocin-diabetic rats

The effects of hindlimb temperature on sciatic nerve and skeletal muscle laser Doppler vascular conductance (LDVC) were assessed in anesthetized control and streptozotocin (STZ)-diabetic rats. With core temperature at 37 degrees C and exposed hindlimb temperature at 32 degrees C, nerve LDVC was significantly lower in rats after 8 weeks of STZ diabetes than in age-matched control rats. Subsequent warming of the exposed hindlimb of control rats from 32 degrees C to 37 degrees C significantly decreased nerve LDVC by 41% and increased muscle LDVC by 48%. Because nerve LDVC was unchanged by hindlimb warming in STZ-diabetic rats, there was no significant difference between control and diabetic nerve LDVC at 37 degrees C. In a second study, after 6 weeks of STZ diabetes, changes from control nerve LDVC were shown to depend on temperature rather than the duration of surgical exposure. These findings emphasize that information about hindlimb temperature is a prerequisite for interpreting the effects of experimental diabetes on hindlimb nerve blood flow.

Hypothermic neuroprotection of peripheral nerve of rats from ischaemia-reperfusion injury

Although there is much information on experimental ischaemic neuropathy, there are only scant data on neuroprotection. We evaluated the effectiveness of hypothermia in protecting peripheral nerve from ischaemia-reperfusion injury using the model of experimental nerve ischaemia. Forty-eight male Sprague-Dawley rats were divided into six groups. We used a ligation-reperfusion model of nerve ischaemia where each of the supplying arteries to the sciatic-tibial nerves of the right hind limb was ligated and the ligatures were released after a predetermined period of ischaemia. The right hind limbs of one group (24 rats) were made ischaemic for 5 h and those of the other group (24 rats) for 3 h. Each group was further divided into three and the limbs were maintained at 37 degrees C (36 degrees C for 5 h of ischaemia) in one, 32 degrees C in the second and 28 degrees C in the third of these groups for the final 2 h of the ischaemic period and an additional 2 h of the reperfusion period. A behavioural score was recorded and nerve electrophysiology of motor and sensory nerves was undertaken 1 week after surgical procedures. At that time, entire sciatic-tibial nerves were harvested and fixed in situ. Four portions of each nerve were examined: proximal sciatic nerve, distal sciatic nerve, mid-tibial nerve and distal tibial nerve. To determine the degree of fibre degeneration, each section was studied by light microscopy, and we estimated an oedema index and a fibre degeneration index. The groups treated at 36-37 degrees C underwent marked fibre degeneration, associated with a reduction in action potential and impairment in behavioural score. The groups treated at 28 degrees C (for both 3 and 5 h) showed significantly less (P < 0.01; ANOVA, Bonferoni post hoc test) reperfusion injury for all indices (behavioural score, electrophysiology and neuropathology), and the groups treated at 32 degrees C had scores intermediate between the groups treated at 36-37 degrees C and 28 degrees C. Our results showed that cooling the limbs dramatically protects the peripheral nerve from ischaemia-reperfusion injury.