Secondary Logo

Journal Logo

Intrathecal Administration of a Cylcooxygenase-1, but Not a Cyclooxygenase-2 Inhibitor, Reverses the Effects of Laparotomy on Exploratory Activity in Rats

Martin, Thomas J., PhD*; Buechler, Nancy L., BS*; Eisenach, James C., MD

Section Editor(s): Liu, Spencer S.

doi: 10.1213/01.ane.0000226093.46973.39
Analgesia: Research Report

Studies of hypersensitivity to mechanical stimuli after incisional surgery suggest that cyclooxygenase (COX)-1, but not COX-2, in the spinal cord participates in postoperative pain. In the current study, we sought to determine the role of COX isoenzymes after laparotomy, examining spontaneous exploratory behavior rather than withdrawal reflexes. Adult male Sprague-Dawley rats underwent subcostal laparotomy surgery under isoflurane anesthesia or received anesthesia without surgery. Exploratory locomotor activity was measured on the first postoperative day after intrathecal injection of dimethyl sulfoxide (vehicle) or COX-1 (SC-560) or COX-2 (NS-398) inhibitors. Laparotomy reduced ambulation, rearing, and rapid small movements (stereotypy) similarly in animals without intrathecal catheters and those receiving intrathecal vehicle control. SC-560 produced a dose-related return to normal exploratory behavior with complete return at doses of 20 μg and larger. In contrast, NS-398 in doses up to 50 μg failed to increase exploratory behavior. These data with exploratory behavior and laparotomy agree with studies with reflexive withdrawal responses after incisional surgery and indicate that COX-1 inhibition reduces pain responses after surgery. Spinal COX-1 inhibition completely restores exploratory activity, including rearing behavior that stretches the abdominal muscles. These data suggest that targeting COX-1 in the spinal cord may produce postoperative analgesia.

IMPLICATIONS: Using a model of abdominal surgery in rats, spinal administration of a cyclooxygenase 1 inhibitor was found to be more effective in restoring exploratory behavior in rats after surgery than a cyclooxygenase 2 inhibitor. These data suggest that selective cyclooxygenase 1 inhibitors may be useful for certain pain modalities.

From the Departments of *Physiology and Pharmacology and †Anesthesiology, the Center for the Study of Pharmacologic Plasticity in the Presence of Pain, Wake Forest University School of Medicine, Winston-Salem, North Carolina.

Accepted for publication April 27, 2006.

Supported, in part, by grants GM48085 (JCE), NS41386 (JCE), and NS38231 (TJM) from the National Institutes of Health, Bethesda, Maryland.

Address correspondence and reprint requests to Thomas J. Martin, PhD, Department of Physiology and Pharmacology, Wake Forest University School of Medicine, Medical Center Blvd, Winston-Salem, NC, 27157. Address e-mail to

Postoperative pain remains poorly treated in nearly half of a general surgical population (1) and severe postoperative pain is associated with major morbidity (2) and increased likelihood of developing chronic pain (3). Neural blockade, either peripherally or at the spinal level, with local anesthetics alone or with adjuvants, produces effective analgesia, but this approach is limited by technical complexity, cost, and side effects, and most patients receive systemic drug treatment for postoperative pain. Opioids are the mainstay for treatment of moderate to severe postoperative pain, but their utility is limited by frequent bothersome and less frequent dangerous side effects. Although research suggests that small doses of ketamine (4) or preoperative gabapentin (5) are effective adjuncts to opioid therapy, cyclooxygenase (COX) inhibitors are the most frequently used drugs for this purpose, and systemically administered COX-2 inhibitors are effective adjuncts to opioids for postoperative analgesia (6). Although COX-2 selective inhibitors reduce the risk of gastric ulceration and bleeding compared with nonselective drugs during chronic use, there is increasing concern regarding their safety, even with brief administration in the perioperative period (7). Some have even asserted that COX-2 inhibitors should not be used in the perioperative period until more safety information is available and have called for “black box” warnings for all COX-2 inhibitors (8).

There is no doubt that systemically administered nonsteroidal antiinflammatory drugs and COX-2 inhibitors provide postoperative analgesia, but they can act at multiple sites in the periphery and the central nervous system. The focus of the current study is to determine the role of COX-1 and COX-2 in the spinal cord on postoperative pain. COX inhibitors produce analgesia by blocking prostaglandin synthesis, both at the site of injury or inflammation and in the central nervous system, especially the spinal cord (9). In rats, peripheral nerve injury (10) or inflammation (11) is accompanied by increased expression of COX-2, but not COX-1, messenger RNA and protein in the spinal cord, as well as by a reduction in withdrawal threshold to mechanical and thermal stimuli, which is blocked by intrathecal (IT) injection of COX-2, but not COX-1, inhibitors. In contrast, we and others (12,13) have shown that hypersensitivity to mechanical stimuli after incisional surgery of the paw is accompanied by increased expression of COX-1 in spinal cord glia and reduction in hypersensitivity by IT injection of COX-1, but not COX-2, inhibitors. These data suggest that targeting COX-1 in the spinal cord might be appropriate for treating postoperative pain.

The purpose of the current study was to extend these previous studies, which used superficial incisional surgery of the extremity, to a more invasive procedure, subcostal laparotomy, and to examine spontaneous behavior rather than withdrawal reflexes as the outcome measure. A major goal of postoperative analgesic therapy is to return activity to normal, and we have previously observed qualitative and quantitative differences between reflex withdrawal and behavioral measures after surgery (13–15). For example, systemic morphine increases withdrawal threshold responses to normal in a dose-dependent fashion after incisional surgery (16) but only partially returns depressed rearing activity after subcostal laparotomy (14). We previously demonstrated that IT injection of the COX-1 preferring inhibitor ketorolac, currently in clinical trials for spinal analgesia (17), returned some exploratory behaviors to normal after subcostal laparotomy (15). In this report, we examine the efficacy of IT injection of selective COX-1 and COX-2 inhibitors in this model of postoperative pain and compare these results to those previously published with these same drugs administered IT in rats with incisional paw surgery (13).

Back to Top | Article Outline


Male Sprague-Dawley rats weighing 200–250 g (Charles River, Raleigh, NC; n = 112) were used for all studies. Rats were anesthetized and a 32-gauge polyethylene catheter (ReCathCo, Allison Park, PA) connected to a piece of Tygon external tubing (Saint-Gobain Performance Plastics, Akron, OH) inserted 6.0 cm through the cisterna magnum, as previously described (18), until the tip lay in the mid-thoracic area. Rats showing neurologic deficits were immediately killed with pentobarbital. After surgery, rats were housed individually in plastic cages in a climate-controlled room under a 12:12-h light-dark cycle with free access to food and water. Two additional groups of animals did not have IT catheters inserted and served as control subjects. All procedures were approved by the Animal Care and Use Committee of Wake Forest University and were performed according to the Guide for the Care and Use of Laboratory Animals, as adopted and promulgated by the National Institutes of Health.

One week after IT catheterization, laparotomy was performed, as previously described (14). Briefly, rats were anesthetized with isoflurane, the right upper quadrant of the abdomen shaved, and a 3-cm incision made diagonally 0.5 cm below and parallel to the lowest rib, penetrating the peritoneal cavity. The wound was vigorously dilated by inserting 5 cm of the index finger into the peritoneal cavity, and then the wound was closed in layers and dressed with antibiotic powder. Sham-treated animals were anesthetized and shaved only.

Exploratory behavior was assessed 24 h after laparotomy using commercially available equipment and software (Med Associates Inc., St. Albans, VT) as previously described (14). Briefly, animals were placed in activity chambers equipped with duplicate banks of 16 infrared transmitters spaced 2.5 cm apart with aligned detectors on the opposing sides of the chamber. A third bank of infrared transmitters and detectors, located 7 cm above the floor surface, allowed determination of rearing behavior. Data were collected in 6-min bins for 1 h and included total distance traveled in the X-Y plane, total beam breaks in the X-Y plane (ambulatory counts), repeated beam breaks within 3 cm of the animal in the absence of locomotion (stereotypy), and total beam breaks in the upper X direction (rearing).

Selective inhibitors of COX-1 (SC-560) and COX-2 (NS-398) were purchased from Cayman Pharmaceuticals Inc. (Westborough, MA). Each group consisted of 10–16 rats. Dose ranges administered, determined from pilot experiments, were 10, 20, and 50 μg after sham surgery or laparotomy. These drugs and this dose range have previously been shown to distinguish COX-1 from COX-2 activity in the spinal cord (12,13,19–23). All IT injections were administered in a 10-μL volume followed by a 10-μL saline flush, and dimethyl sulfoxide (DMSO) was used as vehicle. Exploratory behavior was assessed immediately after the IT injection. The investigator was blinded to drug treatment for all studies.

Behavioral data were analyzed using one-way analysis of variance (ANOVA) on the incision group and the sham-group separately. Post hoc analyses were performed using the Dunnett t-test for multiple comparisons to a control, with noncatheterized sham-animals serving as the control group. All data are presented as mean ± se. P < 0.05 was considered statistically significant.

Back to Top | Article Outline


Laparotomy decreased all variables of exploratory behavior, as previously reported (14) (Figs. 1–4). Comparing sham and incision control groups (no IT catheter), the incision decreased the total distance traveled by 42% ± 10% (F(1,15) = 10.9; P = 0.005). All other behavioral variables were significantly affected by the abdominal incision with ambulation, stereotypy, and rearing being decreased by 45% ± 11%, 25% ± 8%, and 32% ± 12% relative to the behavior recorded from sham-control subjects, respectively (ambulation F(1,15) = 10.1: P = 0.007; stereotypy F(1,15) = 7.3; P = 0.02; and rearing F(1,15) = 5.3; P = 0.04). IT catheterization and DMSO administration did not produce a significant effect on exploratory behaviors in either the sham-laparotomy or abdominal incision groups compared with their noncatheterized counterparts because there was no significant interaction between catheterization (IT catheter with DMSO administration versus no IT catheter) and surgical treatment (sham-abdominal surgery versus laparotomy; total distance F(1,49) = 1.7; P = 0.2; ambulation F(1,49) = 1.9; P = 0.18; stereotypy F(1,49) = 0.05; P = 0.8; and rearing F(1,49) = 0.5; P = 0.5). There were no significant differences between the behavior in noncatheterized subjects and IT-catheterized animals given DMSO after sham-laparotomy. Likewise, the behavior recorded from rats 24 h after laparotomy was not significantly different between catheterized animals administered DMSO and noncatheterized animals. Therefore, laparotomy decreased exploratory behaviors in rats in a similar manner and to a similar extent as our previous findings (14) in both IT-catheterized and noncatheterized animals.

Figure 1.

Figure 1.

Figure 2.

Figure 2.

Figure 3.

Figure 3.

Figure 4.

Figure 4.

SC-560 produced an increase in exploratory behavior after laparotomy for distance traveled (F(5,73) = 7.6; P < 0.0001) with doses of 20 or 50 μg producing an increase compared with DMSO administration and resulting in distance traveled that was not significantly different than the sham-control group (P ≤ 0.05; Fig. 1). SC-560 more than this range of doses did not affect total distance traveled in the sham-laparotomy group (Fig. 1). In contrast, administration of the selective COX-2 inhibitor NS-398 did not affect distance traveled at any of the doses up to 50 μg, and distance traveled was not significantly different after the administration of NS-398 or DMSO (Fig. 1).

The dose-response curve for SC-560's effect on ambulation was similar to that obtained for distance traveled after laparotomy (F(5,73) = 7.1; P < 0.0001), with 20 and 50 μg increasing behavior relative to DMSO administration, resulting in ambulatory counts that were not significantly different from the sham-control group (Fig. 2). As was found for total distance traveled, administration of NS-398 had no effect on ambulatory counts, and ambulation was not significantly different after DMSO or NS-398 administration at doses up to 50 μg (Fig. 2).

The dose-response curve for the effect of SC-560 on stereotypy after laparotomy was similar to that found with the above two variables (F(5,73) = 5.2; P = 0.0004), and doses of 20 and 50 μg resulted in behavior that was not significantly different from the sham-control group (Fig. 3). Administration of up to 50 μg of NS-398 failed to increase stereotypy to more than that obtained after DMSO administration, similar to the other behavioral measures (Fig. 3).

The dose-effect curve for SC-560's effect on rearing after laparotomy was similar to that found using the other three behavioral measures of activity (F(5,73) = 3.8; P = 0.0042), with 20- and 50-μg doses resulting in vertical counts after laparotomy that were not significantly different from the sham-control group (Fig. 4). Administration of up to 50 μg of NS-398 was without effect on rearing behavior, as was the case with the other measures (Fig. 4).

Back to Top | Article Outline


The partial efficacy of IT ketorolac alone and enhancement of IT morphine by ketorolac after subcostal laparotomy in rats (15), coupled with the current observations that this effect likely reflects inhibition of COX-1, suggest that targeting COX-1 in the spinal cord could yield a novel approach to the treatment of postoperative pain. COX-1 inhibition after systemic administration of existing drugs is complicated by gastrointestinal and renal side effects, but it may be feasible to develop selective COX-1 inhibitors for spinal administration. Thus, the results of the current study provide a clear rationale for the development of new drugs selective for COX-1 for postoperative analgesia.

Prostaglandins, synthesized in the spinal cord after nociceptive stimuli, result in sensitization to mechanical and thermal stimuli and spontaneous pain behavior (24). Both COX-1 and COX-2 isoenzymes are constitutively expressed in the spinal cord, yet their relative contribution to prostaglandin synthesis depends on the nature of the stimulus. For example, IT administration of substance P or glutamate and peripheral inflammation result in increased prostaglandin spillover in cerebrospinal fluid in animals, accompanied by hypersensitivity and spontaneous pain behaviors that are reduced by COX-2, but not COX-1, inhibitors (9,25). In contrast, IT injection of dynorphin results in hypersensitivity and spontaneous pain behavior that are reduced by both COX-1 and COX-2 inhibitors (26), and hypersensitivity after incisional surgery is reduced by COX-1, but not COX-2, inhibitors (13). The reasons for this differential activation are unclear but may reflect, in part, the different anatomic locations of these isoenzymes, with COX-2 present in dorsal horn neurons (27) and COX-1 present in microglia (13). After paw incision surgery, COX-1 expressing cells become hypertrophic, both in total cell volume and the number of cellular processes extending from the cell body (13). The number of cellular targets expressing COX-1 increases in dorsal spinal cord and gracile nucleus after paw incision surgery, and it is hypothesized that the increased sensitivity to IT administration of COX-1 relative to COX-2 inhibitors in reversing hypersensitivity after paw incision is related to this increase (13). Therefore, it is likely that the present data are also explained by an increase in microglia in the dorsal horn of the spinal cord expressing COX-1 after laparotomy.

The study of COX isoenzymes has been facilitated by development of inhibitors with several hundredfold selectivity. NS-398 is highly COX-2 selective, and IT injection of a 30-μg dose of this drug completely blocks nocifensive behaviors after IT N-methyl-d-aspartate acid or AMPA injection (12) and completely blocks hypersensitivity to thermal stimuli after carrageenan inflammation (23). In contrast, doses up to 50 μg of this COX-2 inhibitor failed to return exploratory behaviors to normal after laparotomy in the current study. SC-560, a selective COX-1 inhibitor, in an IT dose of 100 μg, fails to affect nocifensive behaviors from IT substance P injection (20), whereas a COX-2 inhibitor was effective in that model. A much smaller dose of SC-560 (20 μg) in the current study returned exploratory behaviors after laparotomy to normal, consistent with a specific effect on spinal COX-1. IT administration of NS398 at a dose smaller than that used in the present study also attenuates the development of tolerance to opioids, suggesting effective blockade of COX-2 at doses within the range used in the present study (28). In agreement with the present data, nociception was found to be blunted in COX-1, but not COX-2, knockout mice (29). Therefore, it seems unlikely that the lack of efficacy of NS398 in the present study was caused by insolubility or lack of adequate dosing, but rather a relative lack of involvement of COX-2 in mediating postoperative pain in the present model relative to COX-1.

We recognize that drugs were only studied for one hour at a single time period, one day after surgery, and these data do not provide guidance regarding timing or duration of drug effect. Of course, spinal anesthesia is used in a small minority of surgical patients, and, when used, it is nearly always provided as a single injection rather than a continuous method. Nonetheless, administration of IT ketorolac or SC-560 just before incisional surgery in rats reduces tactile hypersensitivity for as long as three days later (19), suggesting that inclusion of a COX-1 inhibitor with a spinal anesthetic may have long-lasting analgesic effects.

There are several strengths, but also weaknesses, to measuring spontaneous behavior rather than reflex withdrawal to tactile stimuli as an index of pain. Strengths include the face validity of spontaneous (exploratory) and motivated (self-administration of food) behaviors to goals in humans after surgery for rapid recovery of activity and eating and the ability of these measures to uncover unwanted effects of analgesics, such as sedation and ileus from opioids, that are less apparent using reflex withdrawal tests. Additionally, some measures that reduce tactile hypersensitivity after surgery in humans have no effect on opioid requirement or pain scores, questioning the relevance of allodynia per se to the overall postoperative pain experience (30). Other investigators (31,32) have used spontaneous behaviors in rats to study behavioral alterations produced by surgery and report sensitivity to systemic administration of COX inhibitors. The primary weakness of measuring spontaneous or motivated behaviors however, lies in the uncertainty of the determinants of changes in behavior. Whereas it is tempting to speculate that reduced activity after laparotomy compared with anesthesia alone reflects pain and that a drug treatment that restores this activity represents analgesia, other interpretations are possible. For example, exploratory activity is reduced by manipulations that produce an anxiety-like state in rodents (33). Therefore, anxiogenesis after surgery is one potential mechanism of reduced exploratory activity. Prostaglandins generated in the periphery also cross the blood-brain barrier, and one cannot exclude central behavioral actions of systemically generated prostaglandins as mediators of reduced exploratory activity after surgery (34). The ultimate utility of these models will be determined by their predictive value as compared with clinical trials.

In conclusion, IT administration of a selective COX-1 inhibitor, but not a selective COX-2 inhibitor, reversed the effects of abdominal surgery on exploratory behaviors in rats. These data are consistent with previous findings demonstrating an effect of SC-560, but not NS-398, against mechanical hypersensitivity after paw incision (13). These data support the study of spinal COX-1 inhibition to treat postoperative pain.

Back to Top | Article Outline


1. Apfelbaum JL, Chen C, Mehta SS, Gan TJ. Postoperative pain experience: results from a national survey suggest postoperative pain continues to be undermanaged. Anesth Analg 2003;97:534–40.
2. Beattie WS, Buckley DN, Forrest JB. Epidural morphine reduces the risk of postoperative myocardial ischaemia in patients with cardiac risk factors. Can J Anaesth 1993;40:532–41.
3. Perkins FM, Kehlet H. Chronic pain as an outcome of surgery: a review of predictive factors. Anesthesiology 2000;93:1123–33.
4. Schmid RL, Sandler AN, Katz J. Use and efficacy of low-dose ketamine in the management of acute postoperative pain: a review of current techniques and outcomes. Pain 1999;82:111–25.
5. Dirks J, Fredensborg BB, Christensen D, et al. A randomized study of the effects of single-dose gabapentin versus placebo on postoperative pain and morphine consumption after mastectomy. Anesthesiology 2002;97:560–4.
6. Gilron I, Milne B, Hong M. Cyclooxygenase-2 inbibitors in postoperative pain management: current evidence and future directions. Anesthesiology 2003;99:1198–208.
7. Ott E, Nussmeier NA, Duke PC, et al. Efficacy and safety of the cyclooxygenase 2 inhibitors parecoxib and valdecoxib in patients undergoing coronary artery bypass surgery. J Thorac Cardiovasc Surg 2003;125:1481–92.
8. Furberg CD, Psaty BM, FitzGerald GA. Parecoxib, valdecoxib, and cardiovascular risk. Circulation 2005;111:249.
9. Malmberg AB, Yaksh TL. Hyperalgesia mediated by spinal glutamate or substance P receptor blocked by spinal cyclooxygenase inhibition. Science 1992;257:1276–9.
10. Dirig DM, Yaksh TL. Spinal synthesis and release of prostanoids after peripheral injury and inflammation. Adv Exp Med Biol 1999;469:401–8.
11. Beiche F, Scheuerer S, Brune K, et al. Up-regulation of cyclooxygenase-2 mRNA in the rat spinal cord following peripheral inflammation. FEBS Lett 1996;390:165–69.
12. Yamamoto T, Sakashita Y. COX-2 inhibitor prevents the development of hyperalgesia induced by intrathecal NMDA or AMPA. Neuroreport 1998;9:3869–73.
13. Zhu XY, Conklin D, Eisenach JC. Cyclooxygenase-1 in the spinal cord plays an important role in postoperative pain. Pain 2003;104:15–23.
14. Martin TJ, Buechler NL, Kahn W, et al. Effects of laparotomy on spontaneous exploratory activity and conditioned of operant responding in the rat: a model for postoperative pain. Anesthesiology 2004;101:191–203.
15. Martin TJ, Zhang Y, Buechler N, et al. Intrathecal morphine and ketorolac analgesia after surgery: comparison of spontaneous and elicited responses in rats. Pain 2005;113:376–85.
16. Zahn PK, Gysbers D, Brennan TJ. Effect of systemic and intrathecal morphine in a rat model of postoperative pain. Anesthesiology 1997;86:1066–77.
17. Eisenach JC, Curry R, Hood DD, Yaksh TL. Phase I safety assessment of intrathecal ketorolac. Pain 2002;99:599–604.
18. Yaksh TL, Rudy TA. Chronic catheterization of the spinal subarachnoid space. Physiol Behav 1976;7:1032–6.
19. Zhu X, Conklin DR, Eisenach JC. Preoperative inhibition of cyclooxygenase-1 in the spinal cord reduces postoperative pain. Anesth Analg 2005;100:1390–3.
20. Yaksh TL, Dirig DM, Conway CM, et al. The acute antihyperalgesic action of nonsteroidal, anti-inflammatory drugs and release of spinal prostaglandin E2 is mediated by the inhibition of constitutive spinal cyclooxygenase-2 (COX-2) but not COX-1. J Neurosci 2001;21:5847–53.
21. Hefferan MP, Carter P, Haley M, Loomis CW. Spinal nerve injury activates prostaglandin synthesis in the spinal cord that contributes to early maintenance of tactile allodynia. Pain 2003;101:139–47.
22. Yamamoto T, Sakashita Y. The role of the spinal opioid receptor like1 receptor, the NK-1 receptor, and cyclooxygenase-2 in maintaining postoperative pain in the rat. Anesth Analg 1999;89:1203–8.
23. Yamamoto T, Nozaki-Taguchi N. Role of spinal cyclooxygenase (COX)-2 on thermal hyperalgesia evoked by carrageenan injection in the rat. Neuroreport 1997;8:2179–82.
24. Svensson CI, Yaksh TL. The spinal phospholipase-cyclooxygenase-prostanoid cascade in nociceptive processing. Annu Rev Pharmacol Toxicol 2002;42:553–83.
25. Samad TA, Moore KA, Sapirstein A, et al. Interleukin-1beta-mediated induction of Cox-2 in the CNS contributes to inflammatory pain hypersensitivity. Nature 2001;410:471–5.
26. Koetzner L, Hua XY, Lai J, et al. Nonopioid actions of intrathecal dynorphin evoke spinal excitatory amino acid and prostaglandin E2 release mediated by cyclooxygenase-1 and -2. J Neurosci 2004;24:1451–8.
27. Ghilardi JR, Svensson CI, Rogers SD, et al. Constitutive spinal cyclooxygenase-2 participates in the initiation of tissue injury-induced hyperalgesia. J Neurosci 2004;24:2727–32.
28. Wong CS, Hsu MM, Chou R, et al. Intrathecal cyclooxygenase inhibitor administration attenuates morphine antinociceptive tolerance in rats. Br J Anaesth 2000;85:747–51.
29. Ballou LR, Botting RM, Goorha S, et al. Nociception in cyclooxygenase isozyme-deficient mice. Proc Natl Acad Sci U S A 2000;97:10272–6.
30. Stubhaug A, Breivik H, Eide PK, et al. Mapping of punctuate hyperalgesia around a surgical incision demonstrates that ketamine is a powerful suppressor of central sensitization to pain following surgery. Acta Anaesthesiol Scand 1997;41:1124–32.
31. Roughan JV, Flecknell PA. Behavioural effects of laparotomy and analgesic effects of ketoprofen and carprofen in rats. Pain 2001;90:65–74.
32. Liles JH, Flecknell PA. A comparison of the effects of buprenorphine, carprofen and flunixin following laparotomy in rats. J Vet Pharmacol Ther 2004;17:284–90.
33. Kliethermes CL. Anxiety-like behaviors following chronic ethanol exposure. Neurosci Biobehav Rev 2005;28:837–50.
34. Dantzer R, Konsman JP, Bluthe RM, Kelley KW. Neural and humoral pathways of communication from the immune system to the brain: parallel or convergent? Auton Neurosci 2000;85:60–5.
© 2006 International Anesthesia Research Society