OUP user menu

Efficacy of Subcutaneous Methylnaltrexone in the Treatment of Opioid-Induced Constipation: A Responder Post Hoc Analysis

Edward Michna MD, Arnold J. Weil MD, Marc Duerden MD, Seth Schulman MD, Wenjin Wang PhD, Evan Tzanis BA, Haiying Zhang PhD, Dahong Yu PhD, Amy Manley BS, Bruce Randazzo MD
DOI: http://dx.doi.org/10.1111/j.1526-4637.2011.01189.x 1223-1230 First published online: 1 August 2011

Abstract

Objective. Methylnaltrexone, a selective peripherally acting mu-opioid receptor antagonist, effectively treats opioid-induced constipation (OIC) in patients with advanced illness and shows efficacy in patients with chronic nonmalignant pain. The objective was to identify patients who achieved maximal treatment effect based on response to initial four methylnaltrexone doses.

Design. A post hoc analysis of a randomized, double-blind, placebo-controlled study evaluating patients with OIC and chronic nonmalignant pain who received 12 mg subcutaneous methylnaltrexone daily for 4 weeks was performed to determine if response to the first four methylnaltrexone doses predicted overall response during the study. Patients receiving ≥8 doses were included.

Outcome Measures. Patients having ≥3 rescue-free bowel movements (RFBMs)/week; change from baseline in RFBMs/week; percentage of doses with RFBMs within 4 hours after dosing.

Results. Of 137 patients, 58 patients (42.3%) had RFBMs after ≥2 of four doses. Among those with response to ≥2 of four doses, 81% had ≥3 RFBMs/week vs 43% for those with response to <2 of four (P < 0.0001). Those with RFBMs after ≥2 of first four doses averaged 4.8 RFBMs/week vs 2.0 RFBMs/week for those with <2 of four (P < 0.0001). Percentage of subsequent injections resulting in RFBMs within 4 hours was 45.9 ± 27.6 for those with response to ≥2 of four doses vs 17.1 ± 19.1 for those with response to <2 of four (P < 0.0001). Abdominal pain was the most frequently reported adverse event.

Conclusion. Early response to ≥2 of first four doses of methylnaltrexone identified patients who demonstrated a particularly robust effect of treatment over the duration of use.

  • Methylnaltrexone
  • Constipation
  • Opioids
  • Pain
  • Opioid-Related Disorders

Introduction

Opioids are increasingly used for the management of chronic severe pain of nonmalignant origin [1]. Although clinical trials have demonstrated that opioids are effective in decreasing pain and improving sleep, functioning, and quality of life in patients with chronic nonmalignant pain, chronic opioid use is also associated with multiple adverse effects that can lead to discontinuation of treatment [2–7]. Of these adverse effects, constipation is perhaps the most common, debilitating, and persistent and may be difficult to manage [4,8,9]. Standard measures to prevent or treat opioid-induced constipation (OIC), including increased fiber intake and physical activity, laxatives, and enemas, are frequently ineffective [9].

The effects of opioids on the gut are primarily mediated by mu-opioid receptors in the gastrointestinal tract. Opioid binding to these receptors decreases enteric nerve activity and gastrointestinal propulsive motor activity, inhibits ion and fluid secretion, and increases resorption of water, leading to constipation [10,11]. Methylnaltrexone (Relistor; Pfizer Inc., Philadelphia, PA, USA and Progenics Pharmaceuticals, Tarrytown, NY, USA), a selective peripherally acting mu-opioid receptor antagonist, reverses the peripheral effects of opioids without affecting central analgesia [12] and is approved for the treatment of OIC in patients with advanced illness receiving palliative care when response to laxative therapy is insufficient [13]. In placebo-controlled studies of patients with advanced illness and OIC, methylnaltrexone induced laxation within 4 hours of administration and improved stool frequency, reduced laxative use, and decreased constipation distress [14–16]. Methylnaltrexone has a favorable safety profile, with abdominal pain, nausea, diarrhea, and flatulence being the most commonly reported adverse events.

The efficacy and safety of methylnaltrexone have also been investigated in patients with chronic nonmalignant pain [17]. In a double-blind study, 460 patients on stable-dose opioids with <3 rescue-free bowel movements (RFBMs) per week were randomly assigned 1:1:1 to receive methylnaltrexone 12 mg daily (QD), every other day (QOD) with placebo QOD, or placebo QD for 4 weeks [17]. Response rates within 4 hours after the first dose were 34.2% in the combined methylnaltrexone groups and 9.9% in the placebo group (P < 0.001). Mean response rates within 4 hours of all doses were 28.9% for methylnaltrexone QD compared with 9.4% of placebo QD, and 30.2% for methylnaltrexone QOD compared with 9.3% for placebo QOD (P < 0.001, both comparisons).

A post hoc analysis of data from this study was conducted to identify which patients were likely to have a particularly good overall response to methylnaltrexone treatment. The analysis used the three major efficacy variables of the study as criteria for response: whether a patient had ≥3 RFBMs per week, change from baseline in the number of RFBMs per week, and percentage of doses resulting in RFBMs within 4 hours. None of the baseline patient characteristics explored (sex, age [<65 vs ≥65], body mass index [<25 vs ≥25], or baseline mean daily oral morphine use) were effective predictors of superior response. However, analysis of responses to the initial four doses of methylnaltrexone in the QD group demonstrated that it was a good predictor of a superior response to subsequent doses of methylnaltrexone. The full results of this analysis are reported.

Methods

Study Design

This double-blind, randomized, placebo-controlled phase 3 study was conducted at 91 centers in the United States and Canada. Complete details of the design of this study have been reported previously [17].

All patients provided written informed consent prior to study enrollment. The study was approved by independent ethics committees at each participating institution and was conducted in accordance with the International Conference on Harmonisation Guideline for Good Clinical Practice and the Declaration of Helsinki.

Patients

Eligible patients were adults aged 18 years and older with OIC and chronic pain caused by a nonmalignant condition. Eligibility was determined during a 14-day screening period and was defined as having <3 RFBMs per week and either hard or lumpy stools, straining during bowel movements, or a sensation of incomplete evacuation after a bowel movement. Eligible patients were required to have a history of chronic pain lasting for ≥2 months prior to study enrollment and treatment with opioid medications for ≥1 month, with a mean daily dose of ≥50 mg oral morphine equivalents for ≥2 weeks.

The major exclusion criteria were: 1) inflammatory bowel disease within the previous 6 months; 2) evidence of bowel obstruction or impaction; 3) history of rectal bleeding not due to hemorrhoids or fissures; 4) malignancy within the previous 5 years; or 5) a history of chronic constipation before initiation of opioid therapy.

Treatment

Eligible patients were randomly assigned in a 1:1:1 ratio to receive subcutaneous methylnaltrexone 12 mg QD, methylnaltrexone 12 mg QOD, or placebo for 4 weeks. Patients in the methylnaltrexone QOD group also received placebo injection on days on which they did not receive methylnaltrexone. All laxatives were discontinued at study entry. During the study, the administration of rescue laxatives was permitted if the patient had no bowel movement for three consecutive days. Rescue laxatives could not be administered within 4 hours of study drug.

Assessments

Efficacy was assessed using a patient-reported diary system in which bowel movement timing, frequency, laxative use, and stool quality parameters were recorded. Safety was assessed through a physical examination, vital signs, laboratory assessments, and an assessment for adverse events. Patients were followed for 14 days after the last dose of methylnaltrexone.

The coprimary efficacy end points of the study were the proportion of patients having an RFBM within 4 hours of the first dose and the percentage of active injections resulting in an RFBM within 4 hours. Primary and secondary end point results were reported previously [17].

In this post hoc analysis, we examined the percentage of patients with ≥3 RFBMs/week, the mean number of RFBMs/week, and the percentage of doses resulting in a RFBM within 4 hours during the study. Patients were stratified by their responses to the first four doses of methylnaltrexone. The analyses included only the outcomes from doses of methylnaltrexone administered after the first four doses.

Statistical Analysis

Only patients who received ≥8 doses of methylnaltrexone QD during the double-blind treatment period were included in the analysis. The major efficacy variables were derived based on the responses from dose 5 to the last dose of the double-blind treatment period.

Summary results (n, mean or proportion, standard deviation) were presented by the number of responses to the initial four doses. Response to a dose was defined as having an RFBM within 4 hours of the methylnaltrexone injection. A logistic model was fitted for the end point of whether a patient had ≥3 RFBMs per week, and the odds ratios were reported. Regression models were fitted for the change from baseline in RFBMs per week and the percentage of doses with RFBMs within 4 hours, respectively. The changes in the subsequent response per one increase in the number of responses to the initial four doses were reported.

Additional comparisons were made between patients with a lower number (<2) of initial responses (Group 1) and patients with higher number (≥2) of initial responses (Group 2). Chi-square test was used to compare the proportions of patients who had ≥3 RFBMs per week between the two groups. Two-sided t-tests were used to compare the change from baseline in the RFBMs per week and the percentage of doses with RFBMs within 4 hours. Summary statistics and P values were reported.

Results

Patients

A total of 150 patients received QD methylnaltrexone, of whom 137 received ≥8 doses of methylnaltrexone and were included in the analysis (Table 1). Of these, 58 patients had RFBM responses to ≥2 of the initial four doses, and 79 patients had responses to <2 of the initial four doses or no response to any of the initial four doses. Baseline characteristics were similar between these response groups, with the exception of median baseline morphine equivalent daily dose, which was slightly higher in patients with RFBM responses to ≥2 of the first four doses (180.0 and 131.4, respectively; P = 0.046) (Table 2). The majority of patients were female, and the most common pain condition was back pain. A numerically greater percentage of females (40/87, 46%) had a response to ≥2 of the initial four doses compared with males (18/50, 36%), but the difference was not significant.

View this table:
Table 1

Disposition of patients who received ≥8 doses of methylnaltrexone administered once daily

Total
N (%)
Received ≥8 doses137
Completed study122 (89.3)
Reasons for withdrawal
  Adverse event5 (3.6)
  Protocol violation6 (4.4)
  Failed to return3 (2.2)
  Other1 (0.7)
View this table:
Table 2

Baseline demographics and clinical characteristics among patients who received ≥8 doses of methylnaltrexone administered once daily

Patients Who Responded to <2 of First Four Doses (Group 1) N = 79Patients Who Responded to ≥2 of First Four Doses (Group 2) N = 58Total N = 137
Mean age, years (SD)46.7 (11.2)50.0 (9.9)48.1 (10.8)
Sex, N (%)
  Male32 (40.5)18 (31.0)50 (36.5)
  Female47 (59.5)40 (69.0)87 (63.5)
Mean body mass index (SD)29.9 (7.7)30.3 (8.2)30.1 (7.9)
Median baseline morphine equivalent dose, mg/d (min, max)131.4* (45, 738)180.0 (54, 733)151.1 (45, 738)
Pain condition, N (%)
  Back pain51 (64.6)38 (65.5)89 (65.0)
  Fibromyalgia6 (7.6)6 (10.3)12 (8.8)
  Osteoarthritis6 (7.6)5 (8.6)11 (8.0)
  Cervical/neck pain4 (5.1)2 (3.4)6 (4.4)
  Neuropathic pain2 (2.5)2 (3.4)4 (2.9)
  Lower extremity/hip pain3 (3.8)03 (2.2)
  Complex regional pain syndrome1 (1.3)2 (3.4)3 (2.2)
  Migraine/headaches2 (2.5)02 (1.5)
  Upper extremity/shoulder pain2 (2.5)3 (5.2)5 (3.6)
  • * P = 0.046 between groups.

Efficacy

Response by Number of Responses to the First Four Doses

A response to a greater number of the first four doses of QD methylnaltrexone was associated with a superior response to subsequent doses during the rest of the study. The observed means and percentages are presented in Figures 1A,B and 2. For doses subsequent to the first four, the percentages of patients who had ≥3 RBFMs per week within 4 hours after dosing were 39.0%, 47.4%, 75.9%, 90%, and 77.8% for those who responded to 0, 1, 2, 3, or all of the first four doses, respectively. The coefficient estimate of the initial response from the logistic regression model was 0.705 (P < 0.0001). The odds ratios of having ≥3 RBFMs per week for those who responded to 1–4 doses compared with those who responded to none of the first four doses ranged from 2.02 to 16.77 (Table 3).

Figure 1

Response to subsequent doses of methylnaltrexone based on initial response within 4 hours after the first four doses (observed patient data). (A) Percentage of patients with >3 RFBMs/week, and (B) Percentage of injections resulting in a RFBM. % responders = percentage of patients who had ≥3 RFBMs/week during the rest of the treatment period. RFBM = rescue-free bowel movement.

Figure 2

Percentage of patients with ≥3 rescue-free bowel movements/week in patients who responded to ≥2 of first four doses and patients who responded to <2 of first four doses of methylnaltrexone. Error bars indicate ± 1 standard error. RFBM = rescue-free bowel movement.

View this table:
Table 3

Response to subsequent doses by initial rescue-free bowel movement responses within 4 hours after the first four doses

Number of Initial Four Doses with a Response*01234P value
N = 41N = 38N = 29N = 20N = 9
Mean percentage of patients with ≥3 RFBMs/week during the rest of the treatment period39.047.475.990.077.8<0.0001
Odds ratio for % responders vs patients with response* to 0 doses (95% CI)2.02 (1.44, 2.85)4.10 (2.07, 8.10)8.29 (2.98, 23.06)16.77 (4.28, 65.64)
Change from baseline in RFBM/week, mean ± SD2.0 ± 2.61.9 ± 1.94.4 ± 2.74.9 ± 2.25.8 ± 3.8<0.0001
Percentage of subsequent injections with a response,* mean ± SD14.5 ± 18.519.9 ± 19.536.2 ± 27.553.6 ± 24.660.1 ± 24.5<0.0001
  • * Rescue-free bowel movement within 4 hours after dosing.

  • RFBM = rescue-free bowel movement; CI = confidence interval; SD = standard deviation.

Similarly, the mean change from baseline in the number of RFBMs per week was 2.0, 1.9, 4.4, 4.9, and 5.8, respectively, for those who responded to 0, 1, 2, 3, or 4 of the first four doses (Table 3). The coefficient estimate of the initial response from the regression model was 1.06 (P < 0.0001); therefore, one additional response to the initial four doses was associated with 1.06 additional increase from baseline in the number of RFBMs per week.

The mean percentage of subsequent injections resulting in an RFBM within 4 hours after dosing ranged from 14.5 to 60.1 (Table 3). The coefficient estimate of the initial response from the regression model was 12.5 (P < 0.0001); therefore, on average, one additional response to the initial four doses was associated with a 12.5% increase in the percentage of subsequent injections resulting in an RFBM within 4 hours after dosing.

Response by Initial Response To ≥2 or <2 of the First Four Doses

Patients who had a response within 4 hours after 2 of the first four doses of QD methylnaltrexone (Group 2) also had a superior response to subsequent doses compared with patients who had a response to <2 of the first four doses (Group 1). Eighty-one percent (47/58) of Group 2 patients had ≥3 RFBMs per week during subsequent treatment compared with 43% (34/79) of Group 1 patients (P < 0.0001) (Figure 2); the odds ratio was 5.7 in favor of Group 2 patients. Similarly, the mean (95% confidence interval [CI]) changes from baseline in the number of RFBMs per week were significantly higher in Group 2 than in Group 1, 4.8 (4.1, 5.5) and 2.0 (1.4, 2.5), respectively (P < 0.0001) (Figure 3). The mean percentages (95% CIs) of subsequent injections resulting in an RFBM within 4 hours were 46% (39%, 53%) and 17% (13%, 21%) for Groups 2 and 1, respectively (P < 0.0001) (Figure 4).

Figure 3

Mean change from baseline in number of RFBMs/week among patients who responded to ≥2 of the first four doses and patients who responded to <2 of the first four doses of methylnaltrexone. Error bars indicate ± 1 standard error. RFBM = rescue-free bowel movements.

Figure 4

Percentage of injections followed by RFBM within 4 hours among patients who responded to ≥2 of the first four doses and patients who responded to <2 of the first four doses of methylnaltrexone. Error bars indicate ± 1 standard error. RFBM = rescue-free bowel movement.

As a reference, 158 patients received ≥8 doses of placebo during the double-blind period. In this placebo group, only 5.7% (9 subjects) had RFBM responses within 4 hours after at least two of the first four doses.

Safety

Detailed safety results for the entire population of this study are reported elsewhere [17]. The most commonly reported adverse events among patients who received ≥8 doses of QD methylnaltrexone were abdominal pain, nausea, diarrhea, and hyperhidrosis (Table 4), which is similar to the safety profile reported in other studies of methylnaltrexone. Patients in Group 2 had slightly higher rates of diarrhea, hyperhidrosis, headache, and hot flushes compared with those in Group 1 (Table 4).

View this table:
Table 4

Adverse events in >2% of patients who received ≥8 doses of methylnaltrexone administered daily

Adverse Event, N (%)Patients Who Responded to <2 of First Four Doses (Group 1)Patients Who Responded to ≥2 of First Four Doses (Group 2)Total
N = 79N = 58N = 137
Abdominal pain14 (17.7)11 (19.0)25 (18.2)
Nausea 5 (6.3) 6 (10.3)11 (8.0)
Diarrhea 3 (3.8) 5 (8.6) 8 (5.8)
Hyperhidrosis 3 (3.8) 5 (8.6) 8 (5.8)
Flatulence 4 (5.1) 3 (5.2) 7 (5.1)
Headache 1 (1.3) 4 (6.9) 5 (3.6)
Dizziness 2 (2.5) 1 (1.7) 3 (2.2)
Anxiety 2 (2.5) 1 (1.7) 3 (2.2)
Hot flush 1 (1.3) 3 (5.2) 4 (2.9)
Orthostatic hypotension 3 (3.8) 0 3 (2.2)

Serious adverse events occurred in three patients (pancreatitis, car accident, diuretic-induced dehydration, and myoclonic jerks in one patient each) in Group 1 and in one patient in Group 2 (pneumonia). None of these events were considered related to methylnaltrexone. Withdrawal due to adverse events occurred in four patients (5.1%) in Group 1 (due to gastrointestinal bleeding, car accident, increased seizure activity, and dizziness in one patient each) and one patient (1.7%) in Group 2 due to diarrhea.

Discussion

This study demonstrated that patients with chronic nonmalignant pain and OIC who had greater responses to the initial QD doses of methylnaltrexone also had more robust responses to subsequent doses. Patients who had a greater response to the first four doses had a higher likelihood of having ≥3 RFBMs/week during subsequent dosing with methylnaltrexone. These patients also had a greater number of RFBMs/week and a higher percentage of subsequent injections likely to result in an RBFM. Baseline demographics and baseline opioid dosage, on the other hand, were generally not useful indicators of response to methylnaltrexone. While a slightly greater percentage of females had a response to ≥2 of the initial four doses compared with males, there was no difference between males and females in the responses to subsequent treatment.

The findings of this study showing that the initial response to methylnaltrexone is predictive of subsequent response are consistent with those of a double-blind, placebo-controlled study of patients with advanced illness and OIC [14,16]. In that study, patients on stable doses of opioids were randomly assigned to receive methylnaltrexone 0.15 mg/kg (N = 62) or placebo (N = 71) subcutaneously every other day for 2 weeks. Analysis based on RFBM responses to all previous doses indicated that patients responding within 4 hours to all of their previous doses had response rates to each subsequent dose (doses 2–7) of 57–100%, e.g., response to dose 1 was associated with a 57% response to dose 2, whereas response to both doses 1 and 2 was associated with an 88% response rate to dose 3. Conversely, those who did not respond to all of the previous doses had response rates of 34% at dose 2 diminishing to 0% at dose 7. This overall pattern of increasing likelihood of response with increasing response to initial doses is consistent across the two studies in different patient populations, although the differences in the way the data were analyzed makes more specific comparisons between the two studies impossible.

The impact of OIC on the lives of patients can lead to reduction or discontinuation of opioid treatment with suboptimal pain control [18,19]. A 2010 study of opioid-treated patients with either advanced or non-advanced illness demonstrated significantly lower scores on the EuroQol 5-dimension quality of life questionnaire among patients taking opioids who had OIC than those without; the impact of OIC on quality of life was similar between advanced illness and non-advanced illness patients [20]. A subsequent cost-benefit analysis of the published data from a methylnaltrexone trial in advanced illness patients using these quality of life data concluded that methylnaltrexone is cost-effective in treating patients with OIC, primarily owing to the offsetting of drug acquisition costs by a decrease in additional constipation-related costs [21].

The ability to determine whether patients with a given medical condition are likely to respond to a particular drug or treatment is the cornerstone of evidence-based medicine. However, clinical studies evaluate the likelihood of response over a large general population of patients and can be less useful in predicting which subgroups of patients are likely or not likely to benefit from a specific therapy, especially when overall response rates are not high. Ineffective therapies add cost, can cause unnecessary adverse effects, and can divert attention from other therapies that may be effective for that patient. When specific individual patient or disease characteristics that can predict a response to therapy are unknown, clinicians must rely on the initial response of the patient to determine whether or how long to continue therapy. In patients with nonmalignant pain and OIC, the frequent presence of other medical conditions and their associated treatments may further impact bowel function and change bowel movement behavior in addition to that related to opioid administration. It is difficult and costly in clinical practice to differentiate patients who may have a superior response to methylnaltrexone from those who will not. This study indicates that clinicians and patients can determine who is likely to derive maximum benefit from methylnaltrexone by the response to as few as four doses. However, it should be noted that even among patients with no response to any of the first four doses, although the likelihood of response to subsequent doses is decreased proportionally, methylnaltrexone may still be beneficial to those whose lives are highly negatively impacted by OIC or in whom other treatments may be ineffective.

Safety findings in this analysis were consistent with the overall favorable safety profile of methylnaltrexone as reported in published clinical trials [15,16]. The reason for the observed higher rate of hyperhidrosis, headache, and hot flushes among patients who responded to ≥2 than <2 of the first four doses in the current analysis is versus. The higher rate of diarrhea observed in the responder population may be consistent with reversal of constipation in patients who responded to methylnaltrexone.

In conclusion, findings of this study suggest that the degree of response to the first four doses of methylnaltrexone identifies patients who demonstrated a particularly robust effect of treatment over the duration of its use. Methylnaltrexone is generally safe and well tolerated.

Acknowledgments

The authors thank Naomi Pliskow, MD, of On Assignment for medical writing assistance, and John H. Simmons, MD, of Peloton Advantage, LLC, for editorial assistance, which was funded by Pfizer Inc.

Footnotes

  • Funding Disclosure: This research was funded by Wyeth Research, which was acquired by Pfizer Inc. in October 2009. Wyeth also provided funding for writing support, provided by On Assignment, and editorial support, provided by Peloton Advantage. No author received an honorarium or other form of financial support related to the development of this manuscript.

  • ClinicalTrials.gov Registry No.: NCT00529087.

  • TOC Summary: A post hoc analysis of data from a double-blind study of methylnaltrexone in chronic nonmalignant pain patients aimed to identify patients most likely to have a particularly good treatment response. Response to the initial four methylnaltrexone doses appeared to be a good predictor of a superior response to subsequent doses.

References

View Abstract