To determine what evidence exists for the efficacy of milnacipran in treating FMS, we performed a search of Medline and Cochrane Library databases through September 2009 for randomized-control trials using the key words milnacipran and fibromyalgia limiting results to those in the English language. There were four published studies identified by the search,33
the results of which are summarized below and in . It should be noted that none of the milnacipran trials utilized a placebo run-in phase to exclude placebo responders, as was done in trials of pregabalin, since these run-in phases can have important implications for external validity.
Overview of published milnacipran fibromyalgia treatment trials
Studies 1 and 2: The first two studies will be discussed together since they are separate reports of the same study.34
We will limit our discussion to results from the later report by Gendreau et al since a more conservative analysis of the data was used than that in the earlier report.35
The study was a 12 week double-blind, randomized, placebo-controlled, dose-ranging, multi-center (14 sites “experienced in treating” FMS) trial performed in the US comparing milnacipran to placebo in patients with FMS. Inclusion criteria required adult patients to meet the 1990 ACR classification criteria for FMS,5
have a pain score of ≥10 on the 20-point Gracely logarithmic pain scale, and be willing to discontinue all CNS-active therapies (although stable doses of aspirin, acetaminophen and NSAIDs were allowed during the study). Exclusion criteria included psychosis; active suicidality; alcohol or substance abuse; concurrent auto-immune, inflammatory, infectious, or malignant disease; known sleep apnea; or abnormal baseline liver or kidney function tests.
The study included 125 subjects composed almost entirely of female Caucasians. The subjects recorded pain during a 2-week baseline phase and were then randomized in a 3:3:2 ratio to receive milnacipran 25 mg once daily, milnacipran 12.5 mg twice daily, or placebo with all patients receiving capsules twice daily. Milnacipran dose was uptitrated as tolerated in weekly intervals up to a maximum of 200 mg/day over 4 weeks with the ability to decrease the dose if dose-limiting toxicities (DLT) occurred. The patients then continued at the maximum tolerable dose for an additional 8 weeks. The primary endpoint was change in average daily patient self-reported pain scores over the final 2 weeks of the trial compared to the 2-week baseline period. To eliminate bias involved in asking individuals to recall symptoms and to improve compliance,37
an innovative system using Palm®
-based electronic diaries (e-diaries) was employed to measured pain level through Gracely’s anchored logarithmic scale (0 to 20).38
Patients were asked to rate pain level in the e-diaries each morning, retrospectively once weekly, and in response to 4 to 6 random prompts. Patients also completed standard paper questionnaires to assess pain in monthly intervals using the McGill pain questionnaire and Gracely pain intensity scale. At the end of the study, patients completed the patient global impression of change (PGIC) scale and numerous secondary endpoints were assessed to determine the impact of once-daily and twice-daily milnacipran on fatigue, mood, physical function, sleep quality and global function as determined by the fibromyalgia impact questionnaire,39
the Medical Outcomes Study Short Form 36 (SF-36),40
the Beck depression inventory,41
the Jenkins sleep scale,42
and the Arizona sexual experience scale.43
An intention to treat analysis with last observation carried forward was used for data analysis.
Subjects receiving milnacipran dosed twice daily reported significantly better pain relief than those receiving milnacipran once daily, but in other parameters the once-daily and twice-daily dosing regimens were similar. Thirty-seven percent of milnacipran subjects dosed twice daily reported at least a 50% reduction in pain intensity compared with only 22% of subjects receiving milnacipran once daily and 14% of those receiving placebo. There was a statistically significant difference seen when comparing the percent of subjects who rated some degree of improvement in endpoint PGIC ratings between milnacipran and placebo treatment (73% for twice daily and 77% for once-daily milnacipran treatment versus 38% for placebo). While no differences were seen for changes in FIQ global scores, milnacipran-treated subjects reported significant increases in the number of days they “felt good” and the FIQ VASs for pain, fatigue and stiffness were significantly improved by milnacipran treatment along with improvements in physical function scores that neared statistical significance (P
= 0.074). This is in contrast to results from the SF-36, Jenkins and ASEX scores which showed no significant improvements in self-reported physical function, sleep quality, or sexual function, respectively. However, it must be noted that randomization of patients with baseline comorbid depression as determined by the Mini International Neuropsychiatric Interview (MINI) was not consistent across treatment groups (placebo 32%, milnacipran twice daily 16% and once daily 7%). This may have significantly affected the results as, similar to numerous other studies,44
the authors found depressed subjects had significantly higher placebo response rates and it is likely the differences would have been larger if depressed patients had been equally distributed.
The main importance of this study was that it proved the efficacy of milnacipran for managing FMS and that twice-daily dosing was more efficacious than once-daily dosing. It also showed that milnacipran was well tolerated, with no significant difference in dropout rates between active drug and placebo. Nearly all patients completing the trial (81% of once daily and 92% of twice-daily participants) achieved dose escalation to the maximum of 200 mg daily. These results set the stage for future phase 3 trials that ultimately led to the FDA approval of milnacipran for FMS management.
Study 3: This 15-week, multicenter (86 centers in US), double-blind, placebo-controlled, fixed-dose trial compared milnacipran 100 mg and 200 mg divided twice daily.36
As in the Gendreau et al study,35
subjects were required to meet the ACR 1990 criteria for FMS;5
in contrast, a linear rather than logarithmic VAS pain scale was used to ensure included subjects had significant pain at baseline (≥40 on a linear 0 to 100 VAS pain scale), making the inclusion criteria similar to that used in FMS trials of the other FDA-approved agents. Also contrary to the Gendreau et al study,35
subjects with a current major depressive episode as determined by the MINI were excluded from participation, but similar exclusions included severe psychiatric illness; active suicidality; abuse of alcohol, benzodiazepines or other substances; a history of non-compliant behavior; active cardiovascular, pulmonary, hepatic, renal, gastrointestinal, or autoimmune disease; current systemic infection; active cancer; unstable endocrine disease; severe sleep apnea; prostate enlargement or other male genitourinary disorder; pregnancy or breastfeeding. As in the Gendreau et al study, patients were required to discontinue any centrally-acting medications that might interfere with the evaluation of pain or other symptoms associated with FMS. However, contrary to the Gendreau et al study and all other FMS trials of indicated agents, patients were allowed up to 60 mg/day of hydrocodone as a rescue analgesic therapy provided it was not used during the 2-week data-collection period preceding the primary efficacy evaluation or during the 48-hour period immediately before study visits. This provision for opioid use during the study is surprising, as treatment trials used to obtain FDA approval of the other indicated FMS medications only allowed for acetaminophen as a rescue analgesic. Also, narcotic use in treating FMS is actively discouraged to prevent the development of rebound pain that can lead to dependence.
After a 2-week baseline period, subjects were randomized to receive milnacipran or placebo tablets twice daily with food. Doses were escalated over the course of 2 to 3 weeks to 50 mg twice daily (100 mg/day) or 100 mg twice daily (200 mg/day) and continued through 15 weeks total. The primary endpoints in this study were different from those used in prior FMS studies and evaluated composite response rates to investigate the ability of milnacipran to manage global FMS symptoms as well as FMS pain. FMS global composite responders were defined as subjects who simultaneously experienced clinically meaningful improvement in pain (30% or greater improvement in average morning VAS pain e-diary scores for final 2 weeks), global status (much or very much improved on the endpoint PGIC scale), and physical function (improvement of ≥6 points on the SF-36 physical component summary scale from baseline to week 15). FMS pain composite responders were defined as those who had improvement in pain and global status as defined above. It is important to realize these composite response outcomes are more difficult to achieve than the primary outcomes that were studied in the trials used to obtain FDA approval of the other indicated FMS medications which looked solely at reduction in pain. Secondary end points included time-weighted averages of the individual components of the composite responder analyses as well as improvement in sleep quality, cognitive function and fatigue. For the primary endpoints, missing data were handled using baseline-observation-carried forward (BOCF) methodology. BOCF is a much more conservative methodology than the more commonly used last-observation-carried forward (LOCF) method, and for this reason BOCF is now being required instead of LOCF by the FDA for approval trials. However, the study authors also reported LOCF and observed cases (OC) analyses. OC analysis only includes data from trial subjects who fully completed the trial with no missing values and is thought to better reflect the treatment response that can be expected clinically in patients who tolerate the drug.
Primary analysis included 1196 subjects with approximately equal numbers randomized to each treatment group. Both primary endpoints were met, with a significantly greater proportion of subjects in the milnacipran treatment groups achieving FMS global (15% 100 mg/day, 14% 200 mg/day versus 9% placebo) and pain (23% 100 mg/day, 25% 200 mg/day versus 17% placebo) composite response outcome rates compared to those receiving placebo using BOCF. While the responder percentages from BOCF analysis were small, the more clinically relevant OC analyses showed about one-quarter of milnacipran treated subjects achieved FMS global composite response rates (25% 100 mg/day, 26% 200 mg/day versus 13% for placebo) and nearly half had a FMS pain composite response (39% 100 mg/day, 46% 200 mg/day versus 25% for placebo), rates similar to our clinical experience. The authors also noted that analgesic response to milnacipran began as early as one week after starting treatment. This time was during the dose-escalation phase, and this observation is consistent with our clinical experience of many patients responding to lower than indicated milnacipran doses. A published extension study indicates the efficacy of milnacipran in improving FMS symptoms may be sustainable for up to 1 year.45
In analysis of secondary endpoints, both milnacipran doses were associated with significant improvements in fatigue but only the 200 mg/day dose improved cognitive dysfunction. There was no significant effect of milnacipran on sleep quality. This is somewhat surprising as other dual reuptake inhibitors including duloxetine have been associated with insomnia, presumably due to their noradrenergic activity. Since the majority of FMS patients have nonrestorative sleep, it is reassuring that milnacipran treatment does not appear to worsen sleep quality.
Study 4: The most recently published milnacipran study was a fixed-dose, multicenter (59 centers), randomized, placebo-control trial. Study 4 was very similar to Study 3 but lasted twice as long (27 weeks) to show the benefits of milnacipran treatment are maintained over time.33
The strategy of performing a 3-month trial to demonstrate efficacy followed by a six month trial to show durability has now been used to gain FDA approval of all three indicated FMS medications. Inclusion and exclusion criteria were similar to those used in Study 3, except that patients were required to have baseline VAS pain scale scores ≥50 on a linear 0–100 scale (slightly higher than the typical ≥40 scores required in Study 3 and other pain studies). While milnacipran dosing was identical to that used in Study 3, Study 4 randomized fewer subjects (888) and distributed them unequally between treatment groups using a 2:1:1 ratio between the 200 mg/day, 100 mg/day, and placebo, respectively. Primary outcomes were the same as those in Study 3, with composite responder rates used to demonstrate improvement in FMS global symptoms (simultaneous improvement in pain [30% or greater improvement in average morning VAS pain e-diary scores for final 2 weeks], global status (much or very much improved on the endpoint PGIC scale), and physical function (improvement of ≥6 points on the SF-36 physical component summary scale from baseline to week 27)] and improvement in FMS pain (simultaneous improvement in VAS pain e-diary and PGIC scores as defined above). Secondary end points were also identical to those used in Study 3 and included time-weighted averages of the individual components of the composite responder analyses as well as improvement in sleep quality, cognitive function and fatigue. In contrast to study 3 where strict BOCF analysis was used, a modified BOCF methodology was utilized for imputing missing data in Study 4, with BOCF used for patients who discontinued the trial before week 15 and LOCF used for patients who discontinued after week 15. However, as before, the study authors also reported LOCF and OC data analyses. Secondary outcomes were identical to those used in Study 3 and included changes in fatigue, sleep quality and cognitive function.
In contrast to Study 3, Study 4 did not meet its primary endpoints at 27 weeks. While a higher percentage of milnacipran treated FMS subjects achieved the predetermined FMS global composite response outcome compared to placebo (18.3% for 100 mg/day and 18.1% for 200 mg/day versus 13% for placebo), the differences were not statistically significant. The FMS pain composite outcome was met for the 200 mg/day milnacipran group, as significantly more of these subjects reached this endpoint compared with placebo (25.6% versus 18.4%, P = 0.034). While the 100 mg/day group had a responder rate that was higher than the rate in the 200 mg/day group (26.8% versus 25.9%, respectively), this difference failed to reach statistical significance in comparison to placebo, likely caused by lower statistical power due to the lower number of subjects in the 100 mg/day versus 200 mg/day group. Using the more clinically relevant OC comparison that limits analysis to patients who completed the trial, the 200 mg/day milnacipran treatment group met both primary outcomes, but the 100 mg/day group only met the FMS pain composite response outcome. In secondary endpoint analyses using LOCF analysis, milnacipran at 200 mg/day significantly reduced fatigue compared to placebo at both 15- and 27-week time points. While the 100 mg/day dose improved fatigue scores at 15 weeks, statistical significance compared to placebo was lost at 27 weeks likely due to the lower number of patients in the 100 mg/day versus 200 mg/day group and a higher placebo response rate observed at the last visit. Patients treated with milnacipran at 200 mg/day had significant improvement in cognition compared to placebo at both 15 and 27 week endpoints. While the 100 mg/day dose also improved cognition, the differences were not statistically significant. As in Study 3, there were no statistically significant effects on sleep quality for the milnacipran treatment groups compared to placebo.
Forty-two percent of randomized patients prematurely discontinued the study, which is a rate higher than that seen in the other milnacipran studies. However, this was a longer study and similar high discontinuation rates have been seen in other 6 month FMS studies.46
Also, the rates of serious TEAEs did not differ across treatment groups. Despite these caveats, higher discontinuation rates due to TEAEs were seen in the milnacipran versus placebo arms (27% for 200 mg/day and 19.6% for 100 mg/day versus 10.3% for placebo). Gastrointestinal-related complaints were the most common TEAEs, with nausea being particularly common in all three groups (40.1% 200 mg/day, 32.6% 100 mg/day and 21.1% placebo). Higher rates of constipation and vomiting were also seen in active treatment groups, as were increased rates of headache, hyperhidrosis, hot flush, heart rate increase and palpitation. These TEAEs are typical for patients taking norepinephrine active medications, and, on average, there were no clinically relevant changes in laboratory values, heart rate or blood pressure seen in the trials. We have found that starting with a low dose of medication, stopping at the dose patients find effective, and taking milnacipran with food can limit the occurrence of these side effects in the majority of FMS patients.