Alvimopan and COX-2 inhibition reverse opioid and inflammatory components of postoperative ileus


  • This work was supported by the National Institutes of Health (R01-GM58241, R01-DK068610, P50-GM53789) and Adolor Corporation.

Anthony J. Bauer PhD, Department of Medicine/Gastroenterology, University of Pittsburgh, S-849 Scaife Hall, 3550 Terrace Street, Pittsburgh, PA 15261, USA.
Tel: (412) 648 7204; fax: (412) 648 9731; e-mail:


Abstract  Our objective was to investigate the therapeutic potential of peripheral opioid antagonism with alvimopan and anti-inflammatory cyclooxygenase 2 (COX-2) inhibition in an animal model of postoperative ileus with pain management. Intestinal manipulation was conducted in mice and rats with or without postoperative morphine injection. Rodents were orally fed non-digestible fluorescein (FITC)-labelled dextran and transit measured after a period of 90 min. The immunomodulatory effects of morphine and alvimopan were determined on nitric oxide released from the organ cultured muscularis externa. Surgical manipulation of the intestine resulted in a delay in gastrointestinal transit after 24 h that worsened with exogenous morphine. Alvimopan did not significantly alter transit of control or manipulated animals, but significantly antagonized the transit delaying effects of morphine. However, when the inflammatory component was robust enough to obscure a further opioid induced delay in gastrointestinal transit, alvimopan ceased to be effective in improving postoperative intestinal function. Cyclooxygenase 2 inhibition significantly diminished the inflammatory component of postoperative ileus. Surgical manipulation resulted in an increased release of nitric oxide from the inflamed isolated muscularis externa in 24-h organ culture which was not altered by morphine or alvimopan. Two distinct mechanisms exist which participate in postoperative bowel dysfunction: a local inflammatory response which is antagonized by COX-2 inhibition, and a morphine-induced alteration in neural function which can be blocked with alvimopan.


geometric centre


surgical manipulation


light surgical manipulation


inducible nitric oxide synthase


nitric oxide


standard error of the mean


Evidence indicates that the clinical conundrum of postoperative ileus is caused by inflammatory and neural mechanisms. The inflammatory response within the immunologically active, postsurgical muscularis externa is beginning to be systematically delineated and it consists of the rapid activation of specific inflammatory transcription factors, the upregulation of cytokines and chemokines, and leucocyte recruitment.1–3 Numerous studies have also demonstrated the importance of neural mechanisms,2 including inhibitory adrenergic pathways,4 stress induced release of endogenous corticotropin releasing factor,5,6 and activation of the somato-sympathetic pathway.7 Additionally, an increased sensitization of primary afferents by the local muscularis inflammatory response also plays a significant role.2,8,9

In patients, postoperative ileus is often exacerbated by the standard use of opioids for analgesia. Although treatment with opioid analgesics is a common and integral constituent of postoperative care, it also directly contributes to postoperative bowel dysfunction and increases the morbidity of postoperative nausea, vomiting and aspiration.10 Four distinct opioid receptors modulate secretory and motility functions of the gastrointestinal tract.11,12 It has been shown that opioids decrease inhibitory neuromuscular transmission in the human small intestine and colon by acting on μ- and δ-opioid receptors located on postjunctional terminals of enteric motor neurons.13,14 Hence, this ‘dysinhibition’ disrupts the descending relaxation of peristalsis, thereby halting productive propulsive gastrointestinal motility. In animal models, the activation of μ, δ and κ-opioid receptors has also been shown to suppress excitatory pathways, resulting in delayed gastrointestinal transit.11,12 In contrast, nociceptin by acting on the opioid-receptor-like-1, or OP4 receptor, has been reported to accelerate colonic transit in rat.15

In addition to the untoward effects of opiates on gastrointestinal motility are their long recognized effects on immune functions.16,17 Morphine is generally thought to be immunosuppressive by decreasing phagocytosis and chemotaxis, as well as, suppressing the release of TNF-α, interleukins, oxygen intermediates and arachidonic acid products.18,19 Interestingly, the expression of opioid receptors has been shown to be increased in the inflamed intestine.19–21 Hence, the existence of a potential interaction between postoperative analgesia by morphine and the inflamed postoperative bowel could be hypothesized.

Currently, no pharmacological therapies for the management of postoperative ileus are approved by the Food and Drug Administration, although multimodal postoperative ‘fast track’ protocols that include patient education, spinal analgesia, avoidance of narcotics, early oral nutrition and ambulation have been used to accelerate recovery and hospital discharge after abdominal surgery.22,23 Alvimopan is reportedly a peripherally acting and selective μ-opioid receptor antagonist.24 Clinical studies have demonstrated that alvimopan reverses morphine-induced inhibition of gastrointestinal transit in humans without affecting analgesia,25 and that alvimopan significantly improves bowel function in patients who have undergone abdominal surgery.12,26–29

The present experiments sought to delineate the specific role of alvimopan in antagonizing the morphine-induced delay in gastrointestinal transit in the context of varying degrees of inflammatory postoperative ileus, and also to investigate the possible immune modulating potential of morphine and alvimopan on the surgically induced postoperative intestinal inflammatory response.


Animals and materials

The research protocol complied with the regulations regarding animal care as published by the National Institutes of Health and was approved by the Institutional Animal Use and Care Committee of the University of Pittsburgh. Male C57BL/6 mice weighing 20–25 g and male Sprague–Dawley rats weighing 300–350 g (Harlan, Indianapolis, IN, USA) were used in this study. Animals were not fasted prior to the experiments.

Drug injections

Alvimopan ((+)-[[2(S)-[[4(R)-(3-hydroxyphenyl)-3(R),4-dimethyl-1-piperidinyl]methyl]-1-oxo-3-phenylpropyl]-amino]-acetic acid dehydrate) was obtained from Adolor Corporation (Exton, PA, USA), DFU [5,5-dimethyl-3-(3-fluorophenyl)-4-(4-methylsulfonyl)phenyl-2(5H)-furanone] from Merck Frosst (Kirkland, QC, Canada) and morphine sulphate from Astra Pharmaceuticals (Westborough, MA, USA). Alvimopan was diluted in Dimethyl sulfoxide (DMSO)/0.9% NaCl (1 : 10, 10 mg mL−1) and was injected subcutaneously at 10 mg kg−1. DMSO/NaCl was used as a control vehicle for alvimopan. Morphine sulphate was intraperitoneally injected in concentrations of 0.1, 0.3, 1, 3, 5 and 10 mg kg−1. NaCl 0.9% was used as the control vehicle. Alvimopan was administered 30 min prior to each morphine injection. DFU (10 mg kg−1, s.c.) was given 30 min before surgery to block the COX-2 pathway selectively.

Animal model of surgery

Ileus was induced in mice and rats by opioid administration and/or by surgical manipulation (SM) of the small intestine, using the procedure described by Kalff et al.30 Briefly, under general anaesthesia through a midline laparotomy, the small intestine was carefully eventrated and spread out onto a gauze sponge moistened with 0.9% saline. The small intestine was manipulated along its length from the duodenal–jejunal junction to the ileocecal junction using a rolling motion of two moist cotton-tipped applicators. The light manipulation (LM) model was constructed by eventration of the bowel and light manipulation of the small bowel with the cotton-tipped applicators. The standard manipulation model was constructed by compression of the small bowel wall such that the lumenal contents was moved aborad without causing bleeding or visual structural damage to the bowel wall. To ensure even manipulation of all sections of the small intestine, this procedure was repeated three times. A similar procedure of standard manipulation was used for the rat, but is termed moderate manipulation to avoid inappropriate direct comparison of the procedures between rat and mouse. After completion of the standardized gut manipulation procedure, the intestine was returned to the peritoneal cavity and the incision was closed in two layers using running 4–0 silk suture. The time schedule of the various drug injections with surgery was as follows. In mice, alvimopan (10 mg kg−1, s.c.) or 10% DMSO/saline vehicle as a similar volume (1 μL g−1) was administered 30 min before surgery, 3.5 and 23.5 h after surgery to allow for a 30-min pretreatment before the injection of morphine analgesia (1 mg kg−1, i.p.). Hence, morphine or saline vehicle as a similar volume (1 μL g−1) was injected immediately, 4 and 24 h after surgery. In rats, DFU (10 mg kg−1) or a similar volume of DMSO vehicle (1 μL g−1) was used as a 30-min pretreatment to block the COX-2-dependent postoperative inflammatory response. Then, alvimopan (10 mg kg−1, s.c.) or 10% DMSO/saline vehicle as a similar volume (1 μL g−1) was administered 23.5 h after surgery to allow for a 30-min pretreatment before the injection of morphine (3 mg kg−1, i.p.) or saline vehicle as a similar volume (1 μL g−1).

Determination of gastrointestinal motility

To determine the effect of the opioid agents and SM of the small intestine on gut motility, we measured the aboral transit of a non-absorbable tracer, fluorescein isothiocyanate-labelled dextran with an average molecular mass of 70 kDa (FD70), as previously described.31,32 The transit of FD70 along the gastrointestinal tract was quantitated after 90 min by calculating the geometric centre (GC) for the distribution of the fluorescein-labelled dextran using the following formula: GC = ∑ (% of total fluorescent signal per segment × segment number)/100.33

Tissue culture preparation and NO measurement in the supernatant

The production of NO release was determined using the isolated intestinal muscularis externa. The small intestines of control and manipulated mice and rats were removed under sterile conditions. The small intestine was transferred to a sterile beaker containing DMEM culture medium with 200 U mL−1 penicillin G and 200 μg mL−1 streptomycin. The muscularis externa was isolated from the mucosa/lamina propria and aliquots of 70–100 mg were incubated in culture dishes containing supplemented DMEM-culture medium in a CO2-controlled incubator (Nu0078Aire, Plymouth, MN, USA). The muscularis from manipulated animals was incubated in the following solutions: (i) DMEM-culture medium alone; (ii) DMEM-culture medium with vehicles in vitro; (iii) alvimopan (10 μg mL−1) in vitro; (iv) morphine (10 μg mL−1) in vitro; and (v) alvimopan and morphine combined (10 μg mL−1) in vitro. After an incubation period of 24 h, the cultured tissue was inspected for contamination and the supernatant was frozen in liquid nitrogen and stored at −80 °C. The muscle tissue was blotted dry and weighed. Nitrite concentration in the culture media, a measurement of NO synthesis, was assayed by a standard Griess reaction adapted to microplates, as described previously.34 Nitrite was quantitated using NaNO2 as a standard and results were expressed as μmol L−1 nitrite per g tissue.

Statistical methods

Results are presented as mean values ± standard error of the mean (SEM). The data were analysed using Student’s t-test or analysis of variance (anova). EZAnalyze was used for F-test and Bonferroni post hoc group comparisons where appropriate. P-values <0.05 were considered significant.


Effect of alvimopan on morphine-induced bowel dysfunction

Opioids are known to delay gastrointestinal transit in humans and rats. We sought to determine the sensitivity of mouse gastrointestinal transit to morphine. C57BL/6 mice were injected with morphine at various doses (0.1, 0.3, 1, 3, 5 and 10 mg kg−1) and gastrointestinal transit was determined over the following period of 90 min. As shown in Fig. 1, morphine caused a dose-dependent delay in gastrointestinal transit. We used these data to determine the dose of morphine to use in the subsequent experiments. Morphine doses of 1 and 3 mg kg−1 caused approximately 75% of the maximal inhibition observed with 10 mg kg−1 in mouse and rat respectively. Therefore these doses were used to avoid excessive opioid induced bowel dysfunction and systemic central nervous system effects.

Figure 1.

 Dose–response curve for the morphine-induced delay in gastrointestinal transit time as calculated from the geometric centre of a 90-min gastrointestinal transit distribution histogram (mean ± SEM, n = 3). Geometric centres were calculated from each 90-min FITC-dextran transit distribution histogram using 15 gastrointestinal segments from C57BL/6 mice after the intraperitoneal injection of morphine (0.1, 0.3, 1.0, 3.0, 5.0 and 10.0 mg kg−1).

In control vehicle-treated mice, the non-absorbable fluorescently labelled dextran transited down the gastrointestinal tract over the 90-min measurement period and primarily localized to the terminal ileum (Fig. 2A). The administration of alvimopan did not alter gastrointestinal transit over the 90-min period. However, as expected morphine (1 mg kg−1) caused a significant delay in gastrointestinal transit as shown in the gastrointestinal transit distribution histogram in Fig. 2A and quantified by the calculation of the GC in Fig. 2B. When mice were pretreated with the opioid antagonist alvimopan and then injected with morphine i.p., the transit delay observed with morphine was significantly blocked (P < 0.05).

Figure 2.

 Panel 2A: Gastrointestinal transit distribution histograms from C57BL-6 mice injected with control-vehicles (NaCl for morphine and NaCl with 10% DMSO for alvimopan), morphine (1 mg kg−1, i.p.), alvimopan (10.0 mg kg−1, s.c.) or both drugs combined. The animals were pretreated with alvimopan or its respective vehicle 30 min prior to the intraperitoneal administration of morphine. Subsequent to the morphine/vehicle application, gastrointestinal transit time was analysed over the following 90 min. Morphine (*) treated animals showed a significant delay in gastrointestinal transit compared to the control group that was reversed by alvimopan (**) back to normal distribution pattern of the control animals. Panel 2B: Calculated geometric centers from the transit distribution histograms for the four groups of mice (anova, Bonferroni post hoc comparison * and **P-value <0.05, n = 5 each).

Similar to mice, FITC-dextran given orally to control Sprague–Dawley rats localized after a period of 90 min to the distal end of the small intestine and alvimopan alone did not alter this response (Fig. 3A,B). Morphine injected i.p. caused a 30% delay in gastrointestinal transit at a dose of 3 mg kg−1 (Fig. 3B), while in mice 1 mg kg−1 caused a 44% delay, as calculated from the change in GCs. As shown in Fig. 3, alvimopan significantly reversed the morphine-induced delay in rat gastrointestinal transit (P < 0.05).

Figure 3.

 Panel 3A: Gastrointestinal transit distribution histograms from Sprague–Dawley rats injected with control-vehicles (NaCl for morphine and NaCl with 10% DMSO for alvimopan), morphine (3 mg kg−1, i.p.), alvimopan (10.0 mg kg−1, s.c.) or both drugs combined. Similar to the mouse model alvimopan was given 30 min before the morphine injection and gastrointestinal transit was measured during the following 90-min period. The morphine (*) induced delay in gastrointestinal transit was significantly reversed by alvimopan (**) pretreatment. Panel 3B: Calculated geometric centres from the transit distribution histograms for the four groups of rats (anova, Bonferroni post hoc comparison * and **P-value <0.05, n = 5 each).

Using mice, we next investigated the dose-dependent blocking ability of alvimopan on the morphine-induced delay in gastrointestinal transit. Calculation of the GC for the individual mouse gastrointestinal transit distribution histograms showed that morphine significantly delay transit moving the GC orad from 11.1 ± 0.30 in controls to 6.9 ± 0.25 in morphine (1 mg kg−1) treated animals. Alvimopan at a low dose of 0.1 mg kg−1 did not significantly antagonize the morphine inhibition of gastrointestinal transit (GC = 9.4 ± 0.49), but the higher dose of 10 mg kg−1 caused a further and significant normalization of transit with less variability in the response (GC = 10.5 ± 0.11). Therefore, throughout this study we used the higher dose of alvimopan, as indicated above.

Interplay of surgical trauma and opioid-induced bowel dysfunction

Clinical postoperative ileus is often complicated by surgical opioid analgesia. To investigate a possible interaction of surgical trauma combined with morphine analgesia, we utilized our standard mouse model of intestinal manipulation to induce postoperative ileus.31 As seen in our surgical controls (Fig. 4) and reported above, a morphine-induced delay in gastrointestinal transit in control animals was blocked by alvimopan. Intestinal eventration and LM reduced the gastrointestinal transit non-significantly, which was further delayed by the addition of morphine, but to the same extent as that observed in control animals. As, in control animals, alvimopan significantly reversed the opioid-induced decrease in transit (P < 0.05). A further, more vigorous manipulation caused a more substantial degree of postoperative bowel dysfunction, decreasing the average GC by 58%. Interestingly, combining standard intestinal manipulation with morphine administration did not produce a further delay in gastrointestinal transit than that caused by manipulation alone (SM GC = 4.3 ± 0.3 vs SM + morphine GC = 3.9 ± 0.08, P > 0.05). In the standard model of manipulation, alvimopan did not significantly improve the transit delay in the manipulation alone group or in the combined manipulation with morphine group, although a slight change was noted in both groups.

Figure 4.

 Mean calculated geometric centres from individual gastrointestinal transit distribution histograms of three groups of mice with two degrees of surgical trauma (control, light surgical manipulation and standard surgical manipulation) with each group consisting of a subgroup of mice injected with vehicles, morphine (3 mg kg−1, i.p.), alvimopan (10.0 mg kg−1, s.c.) or both drugs combined (anova, Bonferroni post hoc comparison * and **P-value <0.05, n = 5 each).

Opioid modulation of immune function

μ-Opioids have been reported to have immunosuppressive properties and μ-opioid receptor deficient mice were observed to be more susceptible to colon inflammation by trinitrobenzene sulfonic acid (TNBS).19 Therefore, given the inflammatory genesis of this rodent model of postoperative ileus, alvimopan blockade of the endogenous effects of opioids at the hypothesized immunosuppressive μ-receptor could result in an adverse potentiation of the inflammatory response. However, alvimopan administered at three time points (30 min before surgery, 3.5 h after surgery and 23.5 h after surgery) did not significantly alter the 24-h gastrointestinal transit GC measurement in both the light and standard manipulation models compared to vehicle (Fig. 4).

Secondly, because of the importance of the inhibitory effect of nitric oxide (NO) on intestinal smooth muscle contractility,35 we designed experiments to measure the potential modulation of NO release from the postoperative inflamed muscularis by morphine and its antagonism by alvimopan. Nitric oxide released from the 24 h cultured mouse small intestinal muscularis was measured by the Griess reaction from control and manipulated animals 24 h after surgery in the presence or absence of the opioid agonist and antagonist. anova with adjusted Bonferroni individual group analysis statistically showed significant increases in NO release from all manipulated groups compared to the control muscularis (Fig. 5A). This analysis was also performed to determine if morphine, alvimopan or the drugs combined altered the release of NO from the surgically inflamed muscularis. The comparative analysis did not yield a significant difference among any of the manipulated groups in the mouse (P > 0.05). A similar set of experiments was also performed in the rat. anova with adjusted Bonferroni individual group analysis statistically showed significant increases in NO release from all manipulated groups compared to the control rat muscularis, as in the mouse (Fig. 5B). No statistical differences existed among any of the surgically inflamed rat muscularis groups that were pharmacologically treated.

Figure 5.

 Twenty-four hours after standard surgical manipulation the mouse or rat small intestinal muscularis externa was selectively harvested (n = 5) and organ cultured for 24 h in 2% PenStrep DMEM cell culture medium at 37 °C with 5% CO2 in the presence and absence of the vehicles and both opioid agents at concentrations of 10 μmol L−1 mL−1. NO released into the medium was measured by Griess reaction and normalized to the amount of tissue (μmol L−1 nitrite per g tissue). Small intestine muscularis harvested from untreated control animals released only trace amounts of NO, whereas the inflamed muscularis after surgical manipulation showed a significant increase of NO release for both mice (Panel 6A) and rats (Panel 6B). The addition of alvimopan, morphine or the drugs combined did not significantly alter the amount of NO released from the inflamed muscularis of either species (anovaF-value = 6.5 and 7.7, Bonferroni post hoc comparison P > 0.05, for mice and rats respectively).

Anti-inflammatory and opioid modulation of postoperative bowel dysfunction

Previously, we have demonstrated that the inhibition of the inflammatory COX-2 pathway can significantly lessen the severity of postoperative ileus in the rodent.36 As a pharmacological intervention, we used DFU, a potent and selective COX-2 inhibitor, to ameliorate rat postoperative ileus. To investigate the potential benefit of combined anti-inflammatory therapy and blockade of the untoward peripheral effects of morphine on postoperative bowel function, we used a moderate degree of SM in rats with morphine and treated them with DFU and alvimopan. For each of the eight groups of rats studied in this experimental series, we calculated GCs from the distribution histograms of gastrointestinal transit (Fig. 6, n = 5 each). Surgical induction of postoperative ileus decreased the GC by 27%. Consistent with the above results in mice, again alvimopan alone in the postoperative rats resulted in a non-significant improvement in transit, while the sole addition of morphine to the postoperative rat resulted in a further delay to an overall significant 47% decrease (P < 0.05). DFU was observed to quantitatively reverse the surgically induced ileus in the absence of morphine. As expected, morphine in the presence of the inflammatory inhibitor again caused a significant opioid-induced bowel dysfunction. This opioid component of the inhibition was again significantly antagonized by alvimopan, ‘normalizing’ the postoperative gastrointestinal transit in the presence of DFU. anova analysis of eight groups of animals yielded an F-value of 21.00 with a P-value <0.001. Multiple comparisons between each of the eight groups were performed using adjusted Bonferroni analysis (Table 1). These data demonstrate that combined DFU and opioid antagonism can significantly ameliorate postoperative bowel dysfunction caused by inflammation and morphine.

Figure 6.

 Geometric centre calculations from gastrointestinal transits of moderately surgically manipulated Sprague–Dawley rats injected with control-vehicles (NaCl for morphine, NaCl with 10% DMSO for alvimopan, 100% DMSO for DFU), morphine (3 mg kg−1, i.p.), alvimopan (10.0 mg kg−1, s.c.), DFU (10 mg kg−1, s.c.) and specific combinations of all drugs. See Table 1 for statistical analysis.

Table 1. anovapost hoc comparison test for the study of the effects of the combination of alvimopan with DFU on opioid-induced and postoperative ileus in rats
anovapost hoc comparison testsP– Bonferroni
  1. The table shows all groups included in the study presented in Fig. 6. All clusters were included in an anova with Bonferroni post hoc comparison test for significance. Displayed are the adjusted Bonferroni P-values for each group in comparison. The overall F-value was 21.003 and P-value was 0.0000000014. The significant group clusters are bolded (P < 0.05).

Control + vehicles vs SM + vehicles
Control + vehicles vs SM + alvimopan
Control + vehicles vs SM + morphine
Control + vehicles vs SM + morphine + alvimopan
Control + vehicles vs SM + DFU
Control + vehicles vs SM + morphine + DFU
Control + vehicles vs SM + morphine + alvimopan + DFU
SM + vehicles vs SM + alvimopan
SM + vehicles vs SM + morphine
SM + vehicles vs SM + morphine + alvimopan
SM + vehicles vs SM + DFU
SM + vehicles vs SM + morphine + DFU
SM + vehicles vs SM + morphine + alvimopan + DFU
SM + alvimopan vs SM + morphine
SM + alvimopan vs SM + morphine + alvimopan
SM + alvimopan vs SM + DFU
SM + alvimopan vs SM + morphine + DFU
SM + alvimopan vs SM + morphine + alvimopan + DFU
SM + morphine vsSM + morphine + Alvimopan
SM + morphine vsSM + DFU
SM + morphine vs SM + morphine + DFU
SM + morphine vsSM + morphine + alvimopan + DFU
SM + morphine + alvimopan vs SM + DFU
SM + morphine + alvimopan vs SM + morphine + DFU
SM + morphine + alvimopan vs SM + morphine + alvimopan + DFU
SM + DFU vsSM + morphine + DFU
SM + DFU vs SM + morphine + alvimopan + DFU
SM + morphine + DFU vs SM + morphine + alvimopan + DFU 0.173


For this study, we developed an animal model of postoperative ileus that mimics the clinical situation of abdominal surgery with postsurgical pain management. The results above clearly demonstrate the existence and interaction of two distinct mechanisms that participate in the cause of postoperative bowel dysfunction: a local inflammatory response and an opioid-induced alteration in neural function. In these experiments and as we have previously shown in rats,36 the inflammatory component of postoperative ileus can be significantly antagonized using the selective COX-2 inhibitor, DFU. Additionally, we show that the supplemental gastrointestinal effects of morphine can be successfully antagonized with alvimopan. However, when the inflammatory component is robust enough to obscure a further opioid-induced delay in gastrointestinal transit, alvimopan ceases to be effective in improving postoperative intestinal function.

In rodents, typical analgesic doses of morphine range from 5 to 30 mg kg−1.37 It would appear that the delay in gastrointestinal transit exhibits a similar dose–response relationship to this μ-opioid agonist, with a delay in the mouse GC beginning to occurring at 0.3 mg kg−1. The morphine doses used in this rodent study to investigate the antagonism of the opioid-induced delay in transit were considerably higher (1–10 mg kg−1) than the typical human dosage (0.25 mg kg−1), which might be used for clinical postoperative pain management. Hence, the probability of efficient antagonism by alvimopan of a much lower dose of morphine is likely in human. Indeed, recently in human studies, alvimopan has been observed to antagonize the opioid-induced delay in small intestine and colonic transit times.26,29,38 Our results using low dose morphine for a rodent, clearly demonstrates a mechanism of opioid antagonism by alvimopan in both mice and rats. However, we do not know what the result of alvimopan would be at high analgesic doses of morphine.

Interestingly, robust intestinal manipulation which induces an injury in which the GC is no longer altered significantly by morphine, also no longer responds to receptor antagonism with alvimopan (Fig. 5). A potential explanation for the loss of the morphine effect is that intense manipulation initiates an immunological response that disrupts enteric neuromuscular transmission, the specific neural elements which are modulated by morphine. Hence, neither morphine nor its antagonism by alvimopan can affect gastrointestinal transit. This observation further supports the selective interaction of alvimopan on the morphine-induced delay in gastrointestinal transit. Although, admittedly the fidelity of the gastrointestinal transit measurement is nearing the floor effect of this assay in this series of experiments, which could make a further delaying effect caused by morphine difficult to observe.

Fukuda et al.39 studied rat gastrointestinal transit over a 3-h time period commencing immediately after surgery. In their model, alvimopan significantly reversed the immediate postoperative delay in transit, in the absence of exogenous morphine. Our studies were not able to show that alvimopan, by itself, can accelerate transit using the 24-h postoperative inflammatory ileus model, so we cannot conclude that endogenous opioids play a significant postoperative role at this time point. Furthermore, alvimopan, by itself, did not alter control gastrointestinal transit indicating that the enteric μ-opioid system is only minimally active in the control state.

However, as Fukuda et al. suggested, the early beneficial effect of alvimopan may occur due to receptor blockade from the physical manipulation-induced release of endogenous opioids, which could happen due to the physical perturbation of the enteric nervous system.40 We both do show that alvimopan can significantly reverse the morphine-induced delay in transit following intestinal manipulation. The manipulation models used in these two separate studies differ in two important ways. Firstly, the time at which gastrointestinal transit was assessed was markedly different (immediate vs the 24-h inflammatory postoperative ileus model). In this study, we used the 24-h postoperative ileus model, because this time point would seem clinically relevant. And secondly, the manipulation procedure appears to be quite different as we have measured only very minimal effects of laparotomy alone on the transit GC, while a major decrease in transit was reported in the studies of Fukuda et al.7,39 Data from De Winter et al.41 on the effects of μ- and κ-opioids on the postoperative gut are difficult to compare to this study directly, because they also used animals shortly after surgery. However, like these results they also demonstrated a delaying effect of morphine on gastrointestinal transit.

In human patients with active inflammatory bowel disease, the μ-receptor has been reported to increase within the mucosal compartment, probably due to the increased presence of receptor-bearing inflammatory myeloid cells.42 Additionally, the same phenomenon has been observed in the inflamed mouse gastrointestinal tract.19–21 Functionally, selective agonist stimulation of the μ-receptor has been shown to reduce inflammation in experimental colitis and for μ-receptor deficient mice to be more susceptible to colitis.19 Given these observations along with the plethora of reported data on the immunomodulatory actions of opioids,16–18 we hypothesized an interaction between morphine and the inflamed postoperative small bowel. However, in our experiments morphine given twice during the onset of the postoperative inflammatory response and then at the time of transit measurement did not show an acceleration, through the hypothesized decrease in the inflammatory response, compared to a single dose of morphine given only at the time of measurement (data not shown). Likewise, alvimopan alone given at multiple times did not delay gastrointestinal transit, by a presumed increase in the inflammatory response due to the blockade of a potential beneficial immunosuppressive effect of endogenous opioids. We further evaluated the immunomodulatory potential of morphine and alvimopan in the setting of postoperative ileus by measuring NO output from the inflamed isolated muscularis in 24-h organ culture. In these experiments, measurements of NO output were not altered by the continual presence of either morphine or alvimopan in the culture media. This indicated that morphine did not alter NO release from the postoperative inflamed muscularis. Hence, we were not able to show an immunomodulatory effect of morphine in this model. Additionally, endogenous μ-agonist opioids that would have been blocked by alvimopan did not alter the postoperative immune response, as measured by the release of NO. Even though, we have shown that μ-agonists will decrease the release of neuronal NO at the neuromuscular junction, we have also shown that the Griess reaction in these experiments primarily measured inducible nitric oxide synthase (iNOS) derived NO from inflammatory cells.35

In this study, morphine appeared to be an additive component in further delaying transit of the postoperative inflamed gastrointestinal tract, most likely by classically interfering with neuromuscular transmission.14,43 Previously, Schwarz et al.36 showed that the postoperative inflammatory response within the muscularis externa can be efficiently prevented by pretreatment of the rat with the selective COX-2 inhibitor, DFU. In our last series of experiments, DFU consistently blocked the inflammation-induced ileus, further unmasking the morphine-induced postoperative ileus component, which could be significantly antagonized with alvimopan. Hence, a bimodal anti-inflammatory and peripheral acting opioid receptor blocker to allow for proper postoperative pain management may be the most beneficial approach to the surgical patient.