Customize your JAMA Network experience by selecting one or more topics from the list below.
Yuan C, Foss JF, O'Connor M, et al. Methylnaltrexone for Reversal of Constipation Due to Chronic Methadone Use: A Randomized Controlled Trial. JAMA. 2000;283(3):367–372. doi:10.1001/jama.283.3.367
Author Affiliations: Committee on Clinical Pharmacology (Drs Yuan, Foss, and Roizen) and Departments of Anesthesia and Critical Care (Drs Yuan, Foss, O'Connor, Moss, Roizen, and Mr Osinski) and Health Studies (Dr Karrison), Pritzker School of Medicine, University of Chicago, Chicago, Ill.
Context Constipation is the most common chronic adverse effect of opioid pain
medications in patients who require long-term opioid administration, such
as patients with advanced cancer, but conventional measures for ameliorating
constipation often are insufficient.
Objective To evaluate the efficacy of methylnaltrexone, the first peripheral opioid
receptor antagonist, in treating chronic methadone-induced constipation.
Design Double-blind, randomized, placebo-controlled trial conducted between
May 1997 and December 1998.
Setting Clinical research center of a university hospital.
Participants Twenty-two subjects (9 men and 13 women; mean [SD] age, 43.2 [5.5] years)
enrolled in a methadone maintenance program and having methadone-induced constipation.
Main Outcome Measures Laxation response, oral-cecal transit time, and central opioid withdrawal
symptoms were compared between the 2 groups.
Results The 11 subjects in the placebo group showed no laxation response, and
all 11 subjects in the intervention group had laxation response after intravenous
methylnaltrexone administration (P<.001). The
oral-cecal transit times at baseline for subjects in the methylnaltrexone
and placebo groups averaged 132.3 and 126.8 minutes, respectively. The average
(SD) change in the methylnaltrexone-treated group was −77.7 (37.2) minutes,
significantly greater than the average change in the placebo group (−1.4
[12.0] minutes; P<.001). No opioid withdrawal
was observed in any subject, and no significant adverse effects were reported
by the subjects during the study.
Conclusions Our data demonstrate that intravenous methylnaltrexone can induce laxation
and reverse slowing of oral cecal-transit time in subjects taking high opioid
dosages. Low-dosage methylnaltrexone may have clinical utility in managing
Opioids are widely used analgesics in patients with advanced cancer.
Constipation is the most common long-term adverse effect of opioid pain medications
in patients with metastatic malignancy1,2
and can be severe enough to limit opioid use or dose.3-5
This significant negative impact on the quality of life of these patients
has received insufficient attention. Thus, there is a need to enhance palliative
care in terminal cancer patients.6
Methylnaltrexone (N-methylnaltrexone bromide;
Mallinckrodt Specialty Chemicals, St Louis, Mo) is the first quaternary ammonium
opioid receptor antagonist that does not cross the blood-brain barrier in
humans.7,8 It offers the therapeutic
potential to reverse adverse effects of opioid pain medications mediated by
receptors peripherally located (eg, in the gastrointestinal tract), while
sparing opioid effects mediated at receptors in the central nervous system,
most importantly, analgesia. In healthy volunteers, intravenous methylnaltrexone,
0.45 mg/kg, effectively blocked acute morphine-induced delay in oral-cecal
transit time without affecting analgesia.9
Recently, in a pilot study of 4 subjects with long-term methadone-induced
constipation, we observed immediate laxation and a dramatic reduction in oral-cecal
transit time after a low intravenous dose of methylnaltrexone, suggesting
a clinical utility for the compound.10 The
present study is a double-blind, randomized, placebo-controlled trial, evaluating
the effects of methylnaltrexone in treating long-term opioid-induced constipation.
We conducted this trial with subjects in a methadone maintenance program,
in which approximately 60% of the long-term methadone users have constipation.11 These subjects served as a proxy group for patients
with advanced cancer to evaluate the efficacy of methylnaltrexone on long-term
With approval from the University of Chicago Institutional Review Board,
9 men and 13 nonpregnant, nonbreastfeeding women were enrolled in the study
(Figure 1). Mean (SD, range) age
was 43.2 (5.5, 25-52) years.
Subjects met the following inclusion criteria: (1) enrollment in a methadone
maintenance program for 1 month or longer; (2) methadone-induced constipation,
ie, no or 1 bowel movement in the previous 3 days, or 2 or fewer bowel movements
in the previous week12,13 and
(3) no laxative use 2 days before the study or during the study. Exclusion
criteria were (1) history or current evidence of significant cardiovascular,
respiratory, endocrine, renal, hepatic, hematological, or psychiatric disease;
(2) any laboratory findings indicating hepatic or renal impairment, or abnormal
physical examination findings; (3) current use of other medications, including
street drugs; (4) known hypersensitivity to lactose or lactulose; and (5)
participation in any investigational new drug study in the previous 30 days.
An investigator explained the study procedures and obtained written,
informed consent from 22 paid subjects. These subjects, who continued to receive
their usual dosage of methadone during the study, were admitted to the Clinical
Research Center at the University of Chicago Medical Center for 2 days. An
intravenous catheter was placed in each arm, one for test drug administration
(placebo or methylnaltrexone) and the other for blood drawing.
On day 1, at 9 AM, after a restricted supper with no fiber the night
before (required for the oral-cecal transit time measurement) and overnight
fasting, subjects were instructed to ingest 10 g of lactulose (Solvay Pharmaceuticals,
Marietta, Ga) in 100 mL of tap water. Subjects were also given placebo (normal
saline) in 4 syringes (35 mL each) for intravenous injection (single-blinded
to the subject).
At 5 PM, subjects were given placebo or methylnaltrexone up to 0.365
mg/kg in 4 syringes. Each syringe contained placebo or methylnaltrexone in
35 mL of normal saline and was administered intravenously over 9 minutes.
For the methylnaltrexone group, syringes 1, 2, 3, and 4 contained 0.015, 0.05,
0.1, and 0.2 mg/kg of the study drug, respectively. The interval between administration
of each syringe in both groups was 1 minute. The continued administration
of each syringe depended on the absence of a clinical laxation response (ie,
elimination of any stool) and/or potential adverse effects. Immediate laxation
was defined as defecation either during or within 1 minute after cessation
of the infusion. The injection was discontinued if the subject had a bowel
After a fiber-free supper and overnight fasting, on day 2 at 9 AM, subjects
were again given the test drug intravenously. Subjects were also given 10
g of lactulose at this time. Day 2 studies were performed to test the constancy
of effect and measure the oral-cecal transit time; this study did not have
a crossover design.
Injection assignment was prepared using a table of random numbers from
which sealed envelopes were prepared and opened sequentially as subjects were
enrolled in the study. No stratification or blocking factors were used, except
to ensure that equal numbers of subjects were assigned to each treatment group
after enrollment of the last (22nd) subject. Randomization and test drug preparation
was done by a biostatistician and a physician, respectively, who did not participate
in data acquisition and evaluation.
Vital signs were obtained 0, 5, 10, 30, 60, 90, and 120 minutes after
each test drug administration. Illicit drug use was monitored by random urine
Seven 5-mL blood samples were obtained 0, 5, and 30 minutes and 1, 2,
4, and 8 hours after each test drug administration. Three urine samples were
collected 0, 2, and 4 hours after drug administration. Plasma and urine methylnaltrexone
levels were determined by high-performance liquid chromatography,9,14 with a detection limit of 2 ng/mL.
Subjects were asked to record frequency and consistency of stools during
the study period. Subjects' bowel movements were confirmed and recorded by
a research nurse blinded to the group assignment. At the end of the study,
the subjective opinion of the participants was gathered to rate subjects'
satisfaction with respect to bowel movement.
Oral-cecal transit time was assessed by measuring pulmonary hydrogen
produced when unabsorbed lactulose is fermented by colonic bacteria and excreted
in the breath. The time between ingestion and the earliest detectable and
sustained rise in pulmonary hydrogen excretion, ie, a sudden rise to the peak
(>25 ppm) or an increase of at least 2 ppm above the baseline, maintained
and increased in 3 consecutive samples, indicates that lactulose has reached
Hydrogen breath tests were conducted every 15 minutes until oral-cecal transit
time was determined.
To evaluate possible opioid withdrawal with methylnaltrexone, before
and 10 minutes after the completion of test drug administration, subjects
were asked to complete an adjective checklist of withdrawal symptoms modified
from Fraser et al18 and Jasinski.19
Items rated (none, mild, moderate, severe) were yawning, lacrimation (excessive
tearing), rhinorrhea, perspiration, tremor, piloerection (goosebumps), and
restlessness. The ratings for individual items were translated to a 0-to-3
scale and summed to give a total symptom score. The total scores before and
after test drug administration were compared between groups. Potential opioid
withdrawal symptoms were also monitored by an investigator throughout the
Laxation responses were compared between groups with the Fisher exact
test. The Mann-Whitney U test was used to compare
change from baseline in oral-cecal transit time between the 2 groups and evaluate
statistical differences between sexes in oral-cecal transit times with P<.05 considered statistically significant. Changes
in opioid withdrawal symptoms were analyzed similarly.
Mean (SD) stool frequency per week of the 22 subjects before the study
was 1.5 (0.7). All 22 subjects showed no response to placebo on the morning
of day 1.
Eleven subjects were randomly allocated to each treatment group. Those
in the placebo group received all 4 syringes in the day 1 afternoon and day
2 morning sessions. As shown in Table 1, subjects showed no laxation response after placebo and reported
no abdominal cramping. At the end of the trial, 7 of the subjects who received
placebo reported that they were disappointed when asked about bowel movement
satisfaction. There were no significant frequency changes in bowel movement
before and during the study, no opioid withdrawal, and no significant adverse
effects in these subjects.
Ten subjects in the methylnaltrexone group had immediate laxation response
in the day 1 afternoon session, and all 11 subjects had immediate laxation
in the day 2 morning session (Fisher exact P<.001
vs placebo group response for both days 1 and 2). The stool of most subjects
(>90%) was soft to loose and in large quantity. The mean (SD, range) methylnaltrexone
dose received was 0.09 (0.10, 0.01-0.37) mg/kg and 0.10 (0.10, 0.01-0.37)
mg/kg for day 1 and day 2, respectively. Figure 2 shows the relationship between effective methylnaltrexone
dose and peak plasma concentration.
During and immediately after each study drug injection, all subjects
reported mild to moderate abdominal cramping, which they described as being
similar to a defecation sensation, without discomfort. No opioid withdrawal
symptoms were observed in any of these subjects during the study. No significant
adverse effects were reported by the subjects. Subject 13 reported mild light-headedness
that resolved spontaneously. No subject showed any clinically significant
change in blood pressure or heart rate from baseline with either the placebo
or study drug infusions. Subjects did not have additional bowel movements
after drug-induced immediate laxation, except subject number 15, who reported
mild diarrhea. At the end of the study, all 11 subjects who received methylnaltrexone
were satisfied with their bowel movement activity (Table 1).
Oral-cecal transit time data are presented in Figure 3. The mean (SD, range) transit times for subjects in the
placebo group (n = 11) at baseline and after placebo injection were 126.8
(48.3, 60-195) minutes and 125.3 (45.0, 60-180) minutes, respectively. The
transit times for subjects in the methylnaltrexone group (n = 11) showed that
the study drug reduced the mean (SD, range) transit time from the baseline
level of 132.3 (36.0, 60-180) minutes to 54.5 (19.3, 30-105) minutes. The
average (SD) change in the methylnaltrexone group (−77.7 [37.2] minutes)
was significantly greater than the average change in the placebo group (−1.4
[12.0] minutes) (P<.001). There were no statistical
differences in oral-cecal transit times between sexes.
Mean (SD, range) peak plasma levels of the 11 subjects in the methylnaltrexone
group for day 1 and day 2 were 162 (237, 30-774) ng/mL and 166 (177, 33-658)
ng/mL, respectively. The percentage (SD, range) of the intravenous dose excreted
unchanged in urine from 0 to 4 hours for day 1 and day 2 was 23.7% (10.5%,
9.6%-39.9%) and 37.6% (17.8%, 13.2%-73.6%), respectively.
The effect of opioids on gastrointestinal motility and transit is well
appreciated as a clinical phenomenon. Opioids inhibit gastric emptying and
propulsive motor activity of the intestine, thereby decreasing the rate of
intestinal transit and producing constipation. Opioid receptors and endorphins
are widely distributed in the central nervous system and throughout the gastrointestinal
tract.20 Based on data obtained from previous
animal experiments, the site of opioid action (central vs peripheral) of exogenous
opioid-induced gut motility change or constipation is still controversial.21-25
Since the translation of animal experiment data in the literature to humans
may be problematic because of differences in the physiology of the opioid
systems, the action site for opioid-induced constipation in humans remains
a matter of investigation. Methylnaltrexone, a peripheral opioid receptor
antagonist, very effectively reversed chronic opioid constipation in this
clinical trial. Our data provide the first strong evidence that the methadone
constipating effect in humans is predominantly mediated by receptors located
in the peripheral gastrointestinal tract.
All 11 subjects who received intravenous methylnaltrexone had a laxation
response immediately after administration of methylnaltrexone on day 1 or
day 2, and all reported some degree (mild to moderate) of abdominal cramping
prior to their bowel movement. We interpret their abdominal cramping as a
physiological desire to defecate, because the cramping disappeared after bowel
movement. Because the half-life of methylnaltrexone is approximately 2 hours,9,26 one would expect the cramping caused
by hyperactivity of the gut to be much more prolonged.
We used the lactulose hydrogen breath test in this study, the method
most commonly used for gut transit time measurement.27
Subjects received placebo the morning of day 1 to establish an oral-cecal
transit time baseline. We observed a reduction in gut transit time in all
subjects after methylnaltrexone treatment compared with baseline levels. This
result is consistent with the methylnaltrexone-induced clinical laxation response
in these individuals. Lactulose, a nonabsorbable osmotic agent that acts in
the colon by increasing water content of the stool without directly stimulating
gut peristaltic activity, may have laxative effects itself and could affect
interpretation of our results. However, the dose used in this study (10 g)
is one half to one third of a single dose and one sixth to one twelfth the
daily dose recommended to produce soft stools. That this small dose of lactulose
had no effect in our study is indicated by the absence of a laxation response
and no change in oral-cecal transit time in the placebo group.
A relatively wide dose range of intravenous methylnaltrexone was used
to achieve clinical laxation. However, the laxation doses for day 1 and day
2 for individuals were very similar, and no tachyphylaxis was noticed. The
peak plasma concentrations of the compound and the percentage of the dose
excreted unchanged in urine in our subjects were comparable to those reported
in healthy volunteers who received similar doses.9,26
Tertiary opioid receptor antagonists, such as naloxone, naltrexone,
and nalmephene, cross the blood-brain barrier and block both the beneficial
pain-relieving effect and the adverse effects of morphine. Although oral naloxone
may reverse opioid-induced constipation, the therapeutic index is very narrow;
ie, reversal of the gut effects with naloxone occurred at doses near the reversal
of analgesia.28 Naloxone may also induce opioid
As a novel quaternary peripheral opioid receptor antagonist, methylnaltrexone,
even at high doses, has not shown reversal of analgesic effect of morphine
in rats32 or humans.9
In this study, no opioid withdrawal symptoms were observed in subjects receiving
long-term methadone treatment, which further indicates that methylnaltrexone
does not penetrate into the brain in humans.
Data from our previous studies in animals33
and healthy human volunteers34 suggest that
oral methylnaltrexone is efficacious in preventing morphine-induced gut motility
changes. The effects of the oral compound in long-term opioid users need to
be investigated in future clinical trials. In this study, none of the 11 subjects
in our methylnaltrexone group experienced significant adverse effects. However,
the upper 95% confidence limit for the adverse event rate based on the rule
of 3 is 25%.35 Thus, the absence of significant
adverse effects in this small group does not preclude the discovery of adverse
effects in larger populations, especially those with terminal illnesses and
In the United States, approximately 500,000 patients die of cancer annually.
Opioid pain medication is used in the terminal phase of care for more than
50% of these patients, and constipation, a significant clinical problem, affects
40% to 50% (approximately 125,000) of patients with metastatic malignancy
who receive opioid pain medications.36,37
A significant number of hospice patients receiving long-term opioid treatment
for pain would rather endure their pain than face the severe, incapacitating
constipation that opioids cause. Results from this clinical trial demonstrate
that individuals receiving long-term methadone treatment are very sensitive
to intravenous methylnaltrexone compared with healthy, opioid-naive subjects
in our previous trial, who received 0.45 mg/kg of methylnaltrexone without
any laxation response.9 Our data suggest that
cancer patients receiving long-term opioid treatment also may have increased
sensitivity to methylnaltrexone and that low-dose methylnaltrexone administration
may have clinical utility in managing opioid-induced constipation, thus potentially
improving the quality of life in these patients.
Create a personal account or sign in to: