The effects of methylnaltrexone alone and in combination with acutely administered codeine on gastrointestinal and colonic transit in health


Dr M. Camilleri, Mayo Clinic, Charlton 8-110, 200 First St. S.W., Rochester, MN 55905, USA.


Aliment Pharmacol Ther 2010; 32: 884–893


Background  The short-term effects of methylnaltrexone (MNTX), a peripherally acting μ-opioid receptor antagonist, on gastrointestinal and colonic transit remain unclear.

Aim  To compare the effects of placebo, codeine, subcutaneous (s.c.) MNTX and codeine with s.c. MNTX on gastrointestinal and colonic transit of solids in healthy humans.

Methods  In a randomized, parallel-group, double-blind, placebo-controlled trial of 48 healthy volunteers, effects of 6 consecutive days of placebo [s.c. and p.o. (orally), n = 8], codeine (p.o. 30 mg q.d.s., n = 8), MNTX (s.c. 0.30 mg/kg, n = 16) and combined MNTX and codeine (same doses and routes, n = 16) on gastrointestinal and colonic transit were assessed. A validated scintigraphic method was used to measure transit during the last 48 h of treatment. Bowel function was estimated during treatment as well as 1 week preceding treatment using standard diaries. Analysis of covariance was used to assess treatment effects.

Results  Codeine delayed colonic transit [geometric centre at 24 h (P = 0.04) and ascending colon t1/2 (P = 0.02)] and reduced stool frequency (P = 0.002), but had no effect on stool form. MNTX did not affect transit, stool frequency or stool form, either alone or with codeine (P > 0.3). No drug interaction effects were detected (P > 0.15).

Conclusion  Methylnaltrexone does not alter gastrointestinal or colonic transit and does not reverse acute codeine-associated delayed gut transit in health.


Opioid medications are widely used to provide analgesia in painful conditions such as terminal cancer. Unfortunately, opioids commonly induce or exacerbate constipation, with symptoms severe enough to diminish quality of life and limit opioid use.1 The health-care burden of opioid-induced constipation (OIC) led to the development of peripherally acting μ-opioid receptor antagonists (PAMORAs), which relieve OIC without compromising analgesia or triggering opioid withdrawal as a result of their limited ability to cross the blood–brain barrier.2

Currently, two PAMORAs have approved indications in the USA, alvimopan and methylnaltrexone (MNTX). Alvimopan is approved for post-operative ileus following large or small bowel resection surgery with primary anastomosis, but contra-indicated in those who have used therapeutic doses of opioids in the preceding 7 or more consecutive days. MNTX is indicated for OIC in patients with advanced illness receiving palliative care after failing laxatives; the approved dose of MNTX to reverse such bowel dysfunction is 0.15 mg/kg on alternate days.3 The major difference between these approved indications is that MNTX is exclusively used in opioid-treated patients, whereas alvimopan is predominantly used in the opioid-naïve. The clinical utility of MNTX for reversing bowel dysfunction in the opioid-naive because of acute treatment with opioids, as used in primary care or following trama or minor surgery, remains unclear.

μ-Opioid receptors are present on the surface of myenteric neurons.4 These G protein-coupled receptors internalize on binding by their μ-opioid ligands, activating downstream intracellular signals to reduce neuronal excitability and neurotransmitter release.5 The physiological role of μ-opioid receptors in human small bowel and colonic motor function was illustrated by stimulation of colonic transit in healthy volunteers by alvimopan alone, suggesting that endogenous opioids such as β-endorphin, met-enkephalin and leu-enkephalin delay colonic transit via stimulation of μ-opioid receptors on myenteric neurons.6 It remains unclear whether the structurally different PAMORA MNTX reverses opioid-induced changes on gastrointestinal (GI) and colonic function and accelerates colonic transit when given alone in opioid-naïve, healthy volunteers.

Our hypothesis is that MNTX reverses the effects of codeine on GI and colonic transit in healthy volunteers. Our primary aim was to assess the ability of MNTX to reverse the effects of codeine sulphate administered for 5 days on gastric emptying, small bowel, ascending colon and overall colonic transit. Our secondary aims were to assess the effects of MNTX alone on regional GI transit, its safety and tolerability and its effects on bowel function.


Trial design and participants

We conducted a randomized, double-blind, placebo-controlled, parallel-group study of the effects of the PAMORA MNTX, given subcutaneously (s.c.), on GI and colonic transit, with and without co-administration of oral codeine sulphate, in 48 healthy volunteers (Figure 1). Codeine sulphate was the opioid chosen because of its widespread clinical use in the treatment of pain and its predictable effects on GI transit.7 Study participants included men and nonpregnant, nonbreastfeeding women aged from 18 to 55 years. All subjects had a minimum body mass index (BMI) of 22 kg/m2. The study was approved by Mayo Clinic Institutional Review Board and all participants signed informed consent. All data were collected from October 2009 to March 2010 at the Mayo Clinic Clinical Research Unit in Rochester, Minnesota. This study was registered at with the identifier NCT01055704.

Figure 1.

 Experimental design showing timing of study medications and transit measurements.

Exclusion criteria included the use of medications with potential effects on GI motility within 48 h of dose initiation, use of alcohol within 7 days of dose initiation, structural or metabolic diseases/conditions that affect the GI system, functional GI disorders, substance abuse, and enrolment in another clinical study within the past 30 days. For screening, the shortened Bowel Disease Questionnaire8 was used to exclude subjects with dyspepsia, irritable bowel syndrome or significant GI symptoms as previously described.9 We also excluded participants with previous abdominal surgery other than appendectomy, cholecystectomy, hysterectomy, or hernia repair and also patients with a history or current clinical evidence of significant cardiovascular, pulmonary, renal, hepatic, GI, haematological, neurological, psychiatric or any other diseases that interfere with the objectives of the study.

Participants received study medications on days 1–5. Scintigraphic transit was measured for 48 h starting on day 4. Subjects were contacted by telephone for follow-up between days 7 and 10.

Randomization and Treatment Allocation

Subjects were screened for eligibility within 28 days of day 1 of the study. They were stratified based on gender and BMI (<25, ≥25 kg/m2) and randomized into one of four treatment groups: placebo–placebo (n = 8), placebo–codeine (n = 8), MNTX–placebo (n = 16) and MNTX–codeine (n = 16). Treatment groups were randomly assigned in fixed block sizes according to a schedule provided by the study statistician (ARZ). Allocation sequence was concealed as follows: Mayo Clinic Research Pharmacy maintained the randomization schedule and only randomization personnel were allowed to access the randomization book to prevent the possibility of prediction by the investigators.

Study medications

Study medication dosages were codeine 30 mg orally four times daily and MNTX 0.30 mg/kg s.c. once daily for 5 consecutive days from days 1 to 5. The codeine dosage of 30 mg orally four times daily was chosen because of its short 3-h elimination half-life. This frequency of administration is usually necessary for adequate and uninterrupted pain relief in clinical situations following minor surgery or trauma, and the recommended dosage range for mild-to-moderate pain is 15–60 mg every 4–6 h.10 In addition, our previous study of alvimopan used the same codeine dosing schedule, which caused a significant delay in colonic transit.6 The codeine-induced delay in colonic transit was significantly reversed by alvimopan.6 Matching oral placebo for codeine and s.c. placebo for MNTX were used such that all groups received identical numbers and appearances of oral capsules and s.c. injections. Participants recorded the dosing times and any adverse events throughout the study period.

GI and colonic transit

An established, validated scintigraphic method was used to measure GI and colonic transit.11 Evidence supporting the use of this method as a biomarker of colonic dysmotility has been recently summarized.12111Indium chloride (111In) adsorbed on activated charcoal particles was delivered to the colon by means of a methacrylate-coated, delayed-release capsule. The capsule was ingested following an overnight fast. Colonic transit measurements were performed using the delayed release capsule administered at the same time as the dose of drug. After the capsule emptied from the stomach (documented by its position relative to radioisotopic markers placed on the anterior iliac crests) or 1 h, whichever occurred first, a 300 kcal radiolabeled meal was ingested. In this meal, technetium-99m sulphur colloid (99mTc) was used to label two scrambled eggs that were eaten with one slice of whole wheat bread and one glass of skim milk. This meal facilitated measurement of gastric and small bowel transit. Subjects ingested standardized meals for lunch and dinner at 4 and 8 h after the radio-labelled meal respectively. Relative to the time of breakfast meal ingestion, abdominal images were initially obtained at 30-min intervals, followed by hourly scans until 8 h. Standard meals were ingested at 4 h (chicken, potato, pudding, milk, 550 kcal) and at 8 h (roast beef sandwich, sugar cookie and milk, 750 kcal). Anterior and posterior gamma images were taken at each time point. The participant left the study centre at the end of the afternoon and returned briefly the following day (day 5) to obtain a camera image at 24 h after the first scan, and again the day after (day 6) to obtain a camera image at 48 h. The performance characteristics of this test are summarized elsewhere.11–14

Transit data analysis

A variable region of interest programme was used to quantify in anterior and posterior images the counts in the stomach and each of four colonic regions: ascending, transverse, descending and combined sigmoid and rectum. These counts were corrected for isotope decay, tissue attenuation and downscatter of 111In counts in the 99mTc window.

Data were analysed as in previous studies. Intragastric content was estimated in each image, and gastric emptying t1/2 was calculated by linear interpolation. The proportion of 99mTc reaching the colon at 6 h was estimated as a measure of orocecal (and a surrogate for small bowel) transit.

The primary variable of interest in overall colonic transit is the geometric centre at 24 h (GC24). A secondary endpoint is the geometric centre at 48 h (GC48). The geometric centre (GC) is the weighted average of counts in the different colonic regions: ascending (AC), transverse (TC), descending (DC), rectosigmoid (RS) and stool. At any time, the portion of colonic counts in each colonic region is multiplied by its weighting factors as follows: (%AC × 1 + %TC × 2 + %DC × 3 + %RS × 4 + %stool × 5)/100 = GC. Thus, a high GC implies faster colonic transit. A GC of 1 implies that all isotope is in the ascending colon, and a GC of 5 implies that all isotope is in the stool. A change in colonic GC24 of 0.65 units is associated with a one point change in stool consistency on the Bristol stool form scale.11

Ascending colon emptying t1/2 was estimated by power exponential analysis of the proportionate emptying over time of counts from the colon.

Assessment of stool frequency and consistency

All participants completed a Bowel Pattern Diary to record their bowel habits on the days that they were taking the study medications. The Bowel Pattern Diary was dispensed at the screening visit and completed diary was collected at the conclusion of the study. Stool frequency and stool consistency were tabulated for each treatment, but the study was not powered to assess statistically significant differences between treatment groups for these clinical endpoints.

Statistical analysis

The primary endpoint for analysis was colonic GC at 24 h. The secondary endpoints for analysis were t1/2 of ascending colon emptying, t1/2 of gastric emptying, colonic GC at 48 h, colonic filling at 6 h, stool frequency and stool consistency.

The overall effects of MNTX and codeine treatment on the primary and secondary endpoints were assessed using an analysis of covariance (ancova) that also included a cross-product term (MNTX × codeine) to examine potential drug interactions. The covariates included in the analyses were gender (gastric emptying and colonic filling at 6 h) and BMI (colonic transit). The specific comparisons of MNTX versus placebo and codeine versus codeine plus MNTX were of significant interest. As these comparisons were related to specific a priori hypotheses, the α-level used was 0.05.

Sample size assessment

Table 1 summarizes the a priori power analysis for the primary and secondary endpoints, using the % relative variation (coefficient of variation) to estimate the effect size detectable with 80% power based on a two-sample z-test at a two-sided α of 0.05. The effect size is the difference in group means as a percentage of the overall mean for each response. The ancova provided 80% power to detect similar (pairwise) differences using a pooled estimate of variation across all four groups and potentially even smaller effect sizes by adjusting for important covariates.

Table 1.   Effect sizes detectable with proposed sample sizes (number of subjects/study group)
Response typeMeans.d.CV (%)Effect size* detectable with 80% power (α = 0.05), %
16 vs. 1616 vs. 88 vs. 8
  1. s.d., standard deviation; CV, coefficient or variation; GE, gastric emptying; t1/2, halftime; CF, colonic filling; GC, geometric centre; AC, ascending colon.

  2. * Effect size is the difference between means as a percentage of listed mean.

GE t1/2 min (n = 63)1302922202528
CF at 6 h (n = 63)442966627683
GC at 24 h (n = 62)2.360.8536344145
AC emptying t1/2 h (n = 50)15.08.053506166

Handling of missing data

Data were collected as planned except in three of the 48 patients (Figure 2) who had some data points missing. To enable an analysis using all 48 randomized patients under the intention-to-treat paradigm, the imputed values for the missing data in these three subjects for each of the endpoints were obtained from the corresponding overall mean value of all nonmissing data. The residual error degrees of freedom in the ancova models were decreased by one for each missing data value (i.e. three) in the analysis of each endpoint to account for this imputation.

Figure 2.

 Trial flow diagram.



As shown in Figure 2, among 75 healthy volunteers who were screened, 19 did not meet study entry criteria and eight declined to participate; thus, 48 subjects (17 men, 31 women; mean age 34.4 ± 1.5 years; mean BMI 27.2 ± 0.6 kg/m2) were enrolled and randomized. None withdrew before the study, but three subjects dropped out after study initiation, two because of adverse events and one because of weather conditions that precluded travel for transit measurements. Missing transit data for these three subjects were imputed as described. All subjects were analysed within their originally assigned treatment groups. Table 2 shows the demographics for each treatment group.

Table 2.   Demographics of study participants
 Placebo p.o. q.d.s. + placebo s.c. (n = 8)Placebo p.o. q.d.s. + MNTX s.c. (n = 16)Codeine p.o. 30 mg q.d.s. + placebo s.c. (n = 8)Codeine p.o. 30 mg q.d.s. + MNTX s.c. (n = 16)
  1. p.o., oral; q.d.s., four times per day; s.c., subcutaneous; MNTX, methylnaltrexone; M, male; F, female.

  2. Values represent mean ± standard error of the mean (except gender).

Gender (M/F)3/55/113/56/10
Age (years)36.1 ± 3.633.9 ± 2.639.4 ± 3.831.6 ± 2.7
BMI (kg/m2)27.2 ± 1.227.0 ± 1.227.9 ± 1.826.9 ± 1.0

Effects of study medications on GI and colonic transit

Figure 3 shows representative examples of colonic transit scintiscans at 24 h from a patient in each of the four treatment groups. Table 3 summarizes the transit endpoints for each of the four treatment groups.

Figure 3.

 Colonic transit scintiscans at 24 h representative of each of the four treatment groups. Intensity of images reflects the concentration of counts in each region. Variable regions of interest are drawn around isotope in ascending, transverse and descending colonic regions. Note that codeine slows colonic transit with or without MNTX. MNTX alone has no effect on colonic transit relative to placebo, and similarly, MNTX with codeine has no effect on colonic transit compared with codeine alone.

Table 3.   Effects of treatment on scintigraphic transit and stool characteristics
 Placebo p.o. q.d.s. + placebo s.c. (n = 8)Placebo p.o. q.d.s. + MNTX s.c. (n = 16)Codeine p.o. 30 mg q.d.s. + placebo s.c. (n = 8)Codeine p.o. 30 mg q.d.s. + MNTX s.c. (n = 16)
  1. p.o., oral; q.d.s., four times per day; s.c., subcutaneous; MNTX, methylnaltrexone; GC, geometric centre; AC, ascending colon; t1/2, halftime; CF, colonic filling; GE, gastric emptying.

  2. Values represent mean ± standard error of the mean.

  3. P > 0.3 compared to same variable in Placebo p.o. + Placebo s.c. group.

  4. † P > 0.3 compared to same variable in Codeine p.o. + Placebo s.c. group.

Colon GC at 24 h2.3 ± 0.42.3 ± 0.2*1.8 ± 0.21.9 ± 0.3†
AC emptying t1/2 (h)14.1 ± 2.617.1 ± 2.2*24.0 ± 3.923.6 ± 3.3†
CF at 6 h (%)34.5 ± 14.121.9 ± 5.7*23.7 ± 11.228.0 ± 6.9†
GE t1/2 (min)101.1 ± 6.2102.7 ± 5.1*104.0 ± 14.1126.8 ± 11.9†
Colon GC at 48 h4.1 ± 0.23.9 ± 0.3*3.3 ± 0.42.8 ± 0.3†
Stool frequency (per day)1.5 ± 0.41.1 ± 0.1*0.63 ± 0.20.7 ± 0.1†
Stool form, Bristol3.7 ± 0.33.4 ± 0.3*3.1 ± 0.53.3 ± 0.3†

Overall, when compared with placebo, codeine (with and without MNTX) delayed colonic transit (P = 0.04 for GC24, P = 0.02 for AC t1/2, P = 0.02 for GC48), but had no effects on gastric emptying halftime (t1/2) or colonic filling at 6 h.

Methylnaltrexone had no effect on any of the transit endpoints, either compared with placebo, or in combination with codeine compared with codeine alone (P > 0.3; Table 3). No drug interaction effects between MNTX and codeine were detected (P > 0.15). At 48 h, the retardation of colonic transit by codeine was not inhibited by MNTX, as shown by the significant retardation of colonic transit by codeine and MNTX co-administration compared with placebo (P = 0.02).

Effect of study medications on bowel pattern

Overall, codeine reduced stool frequency (P = 0.002), but did not significantly affect stool form. MNTX had no effect on stool frequency or form, either when administered alone or in combination with codeine (P > 0.3) (Table 3). No drug interaction effects on bowel pattern were detected (P > 0.15).

Adverse events

Table 4 summarizes adverse events. No overall treatment group differences were detected for ‘Any adverse event’ (P = 0.76, Fisher’s exact test). Overall treatment group differences were detected only for nausea (P = 0.02, Fisher’s exact test).

Table 4.   Most common treatment-emergent adverse events
 Placebo p.o. q.d.s. + placebo s.c. (n = 8)Placebo p.o. q.d.s. + MNTX s.c. (n = 16)Codeine p.o. 30 mg q.d.s. + placebo s.c. (n = 8)Codeine p.o. 30 mg q.d.s. + MNTX s.c. (n = 16)
  1. p.o., oral; q.d.s., four times per day; s.c., subcutaneous; MNTX, methylnaltrexone.

  2. Values represent number of adverse events (%). No overall treatment group differences were detected for ‘Any adverse event’ (P = 0.76, Fisher’s exact test), whereas overall treatment group differences were detected only for Nausea (P = 0.02, Fisher’s exact test).

Any adverse event5 (63)12 (75)7 (88)11 (69)
Burn or sting with injection3 (38)7 (44)2 (25)2 (13)
Fatigue03 (19)3 (38)5 (31)
Flatulence2 (25)5 (31)2 (25)1 (6)
Headache03 (19)3 (38)2 (13)
Sleepiness01 (6)2 (25)5 (31)
Constipation02 (13)2 (25)2 (13)
Nausea003 (38)2 (13)
Vomiting1 (13)01 (13)3 (19)
Μusculoskeletal pain1 (13)1 (6)03 (19)
Abdominal bloating03 (19)00

Codeine was associated with more adverse events, including fatigue, sleepiness, constipation, nausea and vomiting. When MNTX was given with codeine compared with codeine alone, there was a trend towards decreased nausea, flatulence, and burn or sting with injection, but increased musculoskeletal pain. MNTX alone, compared with placebo, was associated with an increase of fatigue, headache, constipation and abdominal bloating.

Two female subjects discontinued intervention because of adverse events; after study blind was removed, it was identified that one who was receiving placebo had experienced vomiting, diarrhoea and flatulence. The second participant who withdrew had received codeine and experienced vomiting and flatulence. Nearly all of the adverse events were of mild-to-moderate intensity; none required further medical intervention and all resolved spontaneously during the period from days 7 to 10 after starting study medications observed at the post-trial follow-up.


In this study, we show that MNTX, a PAMORA, administered s.c. at a dose of 0.30 mg/kg daily, does not reverse the effects of 30 mg four times daily of codeine, a μ-opioid analgesic, on GI and colonic transit in opioid-naïve, healthy volunteers. In addition, MNTX alone at this dose does not accelerate colonic transit in these opioid-naïve subjects relative to placebo.

Methylnaltrexone is a quaternary amine derivative of naltrexone with an added N-methyl group providing an extra positive charge.15 This charge increases polarity and decreases lipid solubility, resulting in low permeability of MNTX through the blood–brain barrier and hence its restriction to the peripheral compartment.16 In previous trials, MNTX has been shown to be efficacious in treating OIC by facilitating laxation without attenuating analgesia or precipitating opioid withdrawal.17–20 In a decision-analysis model of advanced-illness patients with constipation, Earnshaw et al.21 demonstrated that treatment with MNTX plus standard care results in more days without constipation symptoms compared with standard care alone. Cost of MNTX was largely offset by reduction in other constipation-related costs, suggesting that MNTX treatment is cost-effective with an incremental cost per QALY of EUR €40 865 or approximately US$54 700.21

Results in this study with MNTX contrast with those of oral alvimopan 12 mg b.d. in a previous study with identical experimental design also in opioid-naïve subjects by our group, which accelerated colonic transit when given alone and reversed opioid-induced delay in small intestinal and colonic transit when given with codeine.6 There are potential explanations for the differences in outcomes between alvimopan and MNTX.

First, unlike alvimopan, a selective μ-opioid receptor antagonist, MNTX is less selective for the μ-opioid receptor, but antagonizes it about eightfold more potently than the κ-opioid receptor.22 MNTX does not interact with δ-opioid or non-opioid receptors.22 It remains unclear whether differences in selectivity account for some of the differences between alvimopan and MNTX in their effects on colonic transit in healthy humans.

Second, the efficacy of MNTX in treating OIC has typically been studied in chronic pain patients already treated for at least 2 weeks with high doses of opioid agonists prior to initiation of MNTX.17, 19, 20 Repeated exposure to opioids can induce drug tolerance, whereby opioid receptors become down-regulated or decoupled from their associated G proteins and their downstream signalling activity.23 This leads to the requirement for a higher dosage of the agonist to achieve the same drug effect, which conversely may allow a lower dosage of an antagonist or inverse agonist to reverse the agonist effect or, if centrally acting, to induce withdrawal. In fact the dose of s.c. MNTX used in the study of Thomas et al.17 was 0.15 mg/kg every other day for 2 weeks, and the median opioid dosages in morphine equivalent mg/day were 100 and 150 mg respectively, in the placebo and MNTX treatment arms. In such patients, the lowest effective MNTX dose to antagonize the μ-opioid agonist was sought to avoid induction of opioid withdrawal, so the dose used in this trial was approved for marketing.

As the participants in our study were started on codeine at the same time as MNTX, tolerance to opioids was not an issue. Our hypothesis is that a higher dose of MNTX may be required to induce the same degree of μ-opioid antagonist effect as that obtained with the dosages used in previous trials, in which subjects were pre-treated with opioid agonists. The MNTX dosage of 0.3 mg/kg/day in our study is double the strength and double the frequency (fourfold) of the dosing for the FDA-approved indication of opiate-induced constipation in adult patients with advanced illness receiving palliative care who failed laxatives. Patients satisfying this indication are highly likely to be tolerant to opioid agonists, and therefore to respond to relatively lower doses of MNTX. Yuan et al.20 showed in a pilot study that in four opioid-dependent subjects on methadone treatment, MNTX IV from 0.1 to 0.9 mg/kg/day induced immediate laxation and significantly increased stool frequency from an average of 1–2 BM/week to 1–2 BM/day. None of the participants had withdrawal symptoms, although the subject assigned 0.45 mg/kg/day had severe abdominal cramps.20 The authors decreased MNTX dosage for some of the subjects in the dose escalation study after noting high sensitivity to MNTX in the opioid-dependent subjects.20 In opioid-naïve, healthy subjects, however, our study suggests that a total MNTX dosage that is higher than fourfold the FDA-approved dosage may be necessary to reverse the retarding effects of acutely administered codeine on GI and colonic transit. A more comprehensive dose-response study would be necessary to identify the optimal MNTX dose for this indication.

In contrast to the MNTX dosage of this trial, which was only fourfold higher than the dose approved for treatment of opioid-dependent subjects, the dose of alvimopan used in our previous pharmacodynamic study6 was 12 mg b.d. (24 mg total daily). This dose was at least an order of magnitude higher than the total daily dosage of 0.5–2 mg of alvimopan demonstrated to be efficacious in managing opioid-induced bowel dysfunction in chronic noncancer pain patients requiring chronic opioid analgesia.24, 25 The 12-mg b.d. dosage of alvimopan is identical to the FDA-recommended dosage for treating post-operative ileus following bowel resection, confirming its efficacy in reversing short-term effects of opiates. In the early pivotal trial by Taguchi et al.26 of alvimopan’s effects on bowel function after bowel resection in opioid-naïve subjects, the dosage of 6 mg b.d. was efficacious in accelerating bowel recovery and shortening hospitalization. In contrast, the 1-mg b.d. dose showed no statistically significant effects on any of the study outcomes.26 Thus, the presence or absence of drug tolerance to opioid agonists appears responsible for the difference in alvimopan’s efficacy in opioid-dependent patients with OIC versus in opioid-naïve patients with post-operative ileus. Difference in pharmacokinetics, however, remains a possible contributing factor given that postsurgical patients may have impaired or delayed absorption and delivery of oral alvimopan to target sites in the bowel.

Our study was specifically designed to study motility endpoints measured objectively using validated scintigraphic methods. It was not powered to assess for statistically significant differences in the patient-reported endpoints of stool frequency and consistency. Nevertheless, codeine treatment significantly decreased stool frequency, but showed only a weak trend towards firmer stool consistency. We have previously shown that changes in colonic transit are moderately correlated with changes in stool consistency, stool frequency, as well as ease of stool passage.11 However, intestinal secretion and absorption also play an important role in the effects of opioids on bowel function in a manner that probably overlaps in part with the role of intestinal motility.27, 28 The exact relationships between intestinal motility and intestinal absorptive and secretory functions, as well as their relative contributions to bowel function, are complex and will require further investigation.

There is some evidence that MNTX may have reduced some of the adverse effects of treatment with opiates, as demonstrated by a trend towards decreased nausea and flatulence. On the other hand, it remains unclear why the MNTX dose tested was associated with an increase of fatigue, headache, constipation and abdominal bloating, compared with placebo. Nevertheless, these symptoms were not severe, and MNTX was generally well-tolerated.

The strengths of our study include its randomized, double-blind, placebo-controlled design, the well established and validated methods of measuring endpoints, the low numbers of drop outs after randomization, and the use of age, gender and BMI as covariates in statistical analysis.

There are limitations in this study. First, we did not evaluate dose-response. As discussed above, a higher dosage of MNTX may be needed to produce a detectable effect on GI and colonic transit as well as bowel pattern in healthy subjects who are naïve to opioid agonists during MNTX initiation. Dosages of MNTX up to 3 mg/kg/day orally18 and 1.2 mg/kg/day intravenously29 have been used short-term in opioid-dependent patients without serious adverse effects. It can be anticipated that dosages even higher than these can be tolerated by opioid-naïve subjects, as suggested by previous trials in healthy subjects.30, 31

Second, we explored the possibility of a type II error. However, the observed pooled coefficient or variation for GC24 (our primary endpoint) and AC t1/2 of 39% and 54% respectively, closely match those of 36% and 53% used in sample size calculations. Therefore, we believe a type II error is unlikely to explain the inability to demonstrate differences in colonic transit with MNTX.

In summary, s.c. MNTX at a dose of 0.3 mg/kg/day does not alter GI or colonic transit, or bowel pattern, relative to placebo. In addition, this dosage of MNTX does not reverse opioid-induced delay in gut transit or the changes in bowel pattern observed when co-administered with codeine 120 mg/day for 5 consecutive days. Further studies are needed to assess the role of μ-opioid receptors in gut motility and to identify the optimal dosage of MNTX in treating OIC associated with acute opioid use in opioid-naïve patients, as in primary care or in patients following trauma or surgery.


Declaration of personal interests: Dr. Camilleri has served as a consultant and an advisory board member for Albireo Pharma, SK Bio, Takeda, Theravance, ARYx Therapeutics, BioKier, Alkermes, Ikaria, Ocera, Synergy, Procter & Gamble, Johnson & Johnson, McNeil, Dynogen, IPSEN, Tioga, AstraZeneca, Helsinn Therapeutics, Axcan Pharma, Tranzyme, GlaxoSmithKline, Adolor, Domain, Ironwood and BioMedical Insights. He has received research grants from Wyeth, Tsumura, Rose, Albireo and NIH. Declaration of funding interests: This study was funded in part by a grant from Wyeth (now a part of Pfizer).