The Effects of Milnacipran on Sleep Disturbance in Fibromyalgia: A Randomized, Double-Blind, Placebo-Controlled, Two-Way Crossover Study

Abstract

Objective:

This study examined the effects of milnacipran on polysomnographic (PSG) measures of sleep and subjective complaints in patients with fibromyalgia and disturbed sleep.

Methods:

This was a single-site, double-blind, placebo-controlled, two-period crossover PSG study. Eligible subjects (aged 28–72 y) were randomized (1:1) to milnacipran (100 mg/d) or placebo for crossover period 1, and vice versa for period 2. Each crossover period comprised a dose-escalation and dose-maintenance phase, with a 2-w taper/washout between periods. In-laboratory PSGs were collected at baseline, and at the end of each treatment period. The primary endpoints were the difference in PSG-recorded wake after sleep onset (WASO), number of awakenings after sleep onset (NAASO), and sleep efficiency (SE) between 4 w of maintenance treatment with milnacipran and placebo. Other PSG measures, subject-rated sleep, fatigue, physical functioning, and pain were assessed. Post hoc analysis was performed in subjects showing at least 25% reduction in pain from baseline in the Brief Pain Inventory Score (responders).

Results:

Of 19 subjects randomized, 15 completed both periods. Subjects treated with milnacipran showed no significant improvements in WASO and NAASO, but showed reduced SE (p = 0.049). Milnacipran did not show significant improvement in other PSG parameters or subjective endpoints. Two thirds of completers met responder criteria and additionally showed a significant improvement in daily effect of pain (p = 0.043) and subjective sleep quality (p = 0.040).

Conclusion:

The data suggest that milnacipran is not sedating in most patients with fibromyalgia and improvements in sleep are likely a result of pain improvement.

Clinical Trial Registration:

ClinicalTrials.gov, identifier: NCT01234675

Citation:

Ahmed M, Aamir R, Jishi Z, Scharf MB. The effects of milnacipran on sleep disturbance in fibromyalgia: a randomized, double-blind, placebo-controlled, two-way crossover study. J Clin Sleep Med 2016;12(1):79–86.

INTRODUCTION

Fibromyalgia (FM) is a chronic pain disorder considered to affect an estimated five to six million Americans, with most complainants being female.¹ FM is characterized by widespread pain and tenderness, pervasive fatigue, cognitive impairment, and complaints of nonrestorative sleep.^1–3 Sleep deprivation or fragmentation was proven to increase pain sensitivity in both healthy individuals^4,5 and individuals with FM.^2,7 There is a bidirectional relationship between sleep and pain, such that pain increases with disrupted sleep, and disrupted sleep increases with pain.^2,6,8,9 Given this, improvements in both pain and sleep have been a key focus of FM therapeutic interventions. In recognition of the high prevalence and important nature of sleep disturbance as a clinical problem in FM, and its effect on pain, the American College of Rheumatology suggested that the presence of sleep disturbances be added to the diagnostic criteria for FM.¹⁰

Several sleep abnormalities have been reported in FM, including a prolonged sleep latency, lower sleep efficiency (SE), increased stage 1 sleep, decreased stage 2 and 3 along with more frequent arousals, increased wakefulness, and sleep disordered breathing compared with healthy controls.^1,11 Further, sleep disturbance seems to increase in patients with FM as symptoms worsen.^12,13 Although most studies in FM support these findings, a number of studies report conflicting results. For example, one study³ failed to establish differences in sleep disturbance between women with FM and healthy control patients using actigraphy and another reported fewer differences in standard polysomnography (PSG) parameters.¹⁴

BRIEF SUMMARY

Current Knowledge/Study Rationale: Pain and disturbed sleep are hallmark features of fibromyalgia. The effects of pharmacotherapy on sleep physiology may provide insight into this relationship.

Study Impact: Milnacipran seems to improve sleep by reducing pain. The study suggests the need to evaluate fibromyalgia in a more homogeneous patient population.

Milnacipran, a selective serotonin norepinephrine inhibitor with weak binding affinity for the N-methyl-D-aspartate (NMDA) receptor received the US Food and Drug Administration (FDA) approval for the management of FM. The compound was shown to improve the core symptoms of FM, including pain, fatigue, and physical function.^15–17 Additionally, milnacipran was shown to improve subjective sleep parameters in patients with FM.^18,19 To date, no studies have evaluated the effects of milnacipran on PSG measures of sleep in patients with FM.

The current study was undertaken to evaluate the effects of milnacipran on PSG determined measures of sleep in patients with FM. The secondary objective of the study was to evaluate subjective measures of sleep and FM symptoms.

METHODS

Study Design

A single-site, randomized, double-blind, placebo-controlled, two-way crossover PSG design was used to explore the effects of milnacipran on sleep disturbance. Subjects participated from February 2011 to November 2013. The study was conducted in compliance with the declaration of Helsinki and International Conference on Harmonization guidelines. The protocol and informed consent were reviewed and approved by an independent institutional review board (IRB). Written informed consent was obtained from all the subjects before any study related procedures were initiated.

The study comprised two double-blind crossover treatment periods, each having a 7-d dose escalation, 28-d treatment, and 7-d taper period. A 7-d washout period between completion of period 1 and the start of period 2 was included (Figure 1). Given that the half-life of milnacipran is reported to be 6–8 h, essentially all milnacipran was assumed to be completely washed out by the end of 7 d. Subjects were required to satisfy subjective and objective sleep disturbance criteria to be eligible for randomization.

Figure 1. Schematic overview of study design.

Randomization

Computer-generated random numbers were used for enrollment and allocation to sequence (1:1): milnacipran → placebo or placebo → milnacipran. The investigator, clinical staff, subjects, and the study sponsor were blinded to sequence allocation. A noninvolved staff member generated the random allocation sequence and kept an electronic copy in a secure location. Study drug was supplied as masked tablets of milnacipran (12.5 mg, 25 mg, and 50 mg), and matching placebos. To minimize nausea, subjects were advised to take the study drug with food. Baseline PSG was obtained prior to randomization, during the screening period and used for inclusion/exclusion. Baseline subjective assessments were carried out at the randomization visit (V3) of the first treatment period. Treatment was initiated with dose escalation for 7 d, with morning and evening doses of 12.5 mg milnacipran or matching placebo for 3 d, followed by 25 mg twice a day, on days 4 to 7. The maintenance dose of 50 mg given morning and evening was continued for the next 4 w. Following the PSG and subjective assessments at the end of each 4-w maintenance treatment period, the drug or matching placebo was tapered to 25 mg morning and evening for 4 d, and 12.5 mg morning and evening for the final 3 d. Drug taper was included to avoid any drug-placebo carryover and possible withdrawal effects.

Concomitant Medications

Subjects were allowed to continue on stable doses/regimens of medications that in the opinion of the investigator would not interfere with the conduct of the study. Medications for sleep difficulty were excluded, but some subjects were allowed to remain on narcotics/opioids that were being used as treatment for conditions other than FM. Low-dose escitalopram, dopamine agonists, pain killers such as acetaminophen, aspirin and other nonsteroidal anti-inflammatory medications were allowed, provided they did not interfere with the efficacy analysis of the study drug. However, the subjects were not allowed to take other serotonin-norepinephrine reuptake inhibitors. The prohibited medications were washed out of the subjects over 14 d prior to the baseline PSG visit.

Study Population and Procedures

After each subject signed the IRB approved informed consent form, a detailed medical and FM history was obtained as well as a tender point evaluation and history of concomitant medications. Other assessments included a physical and neurological examination, an electrocardiogram and depression levels (Beck Depression Inventory/BDI-II). Subjects were given a sleep diary with instructions to record their daily sleep times, as well as a numeric rating for subjective sleep quality. After the review of the clinical laboratory results, the subjects were instructed to wash out any medications necessary for enrollment. Sleep diaries were reviewed to ensure that subjective sleep difficulty measurements met the entry criteria to establish regular bedtimes during the 28-d screening period. Subjects were instructed to report to the sleep laboratory 1 hour before their usual bedtime for their baseline PSG visit. The results of the screening visit assessments, sleep diary, baseline PSG, and inclusion and exclusion criteria were thoroughly reviewed before the subjects were randomized into the study.

Inclusion Criteria

Male or female subjects ages ≥ 18 y meeting the American College of Rheumatology (1991) criteria for FM at screening along with clinically significant sleep disturbance, defined as subjective complaint of maintaining sleep at least three times per week for at least 1 mo and the subjective sleep diary demonstrating sleep disturbance for at least 2 w prior to randomization. The subjects understood and were willing to cooperate with the study procedures, before they signed the informed consent form. They were instructed to maintain a normal daytime awake and nighttime sleep schedule, and with a customary bedtime between 21:00 and midnight, and rise time 05:00 and 09:00.

Exclusion Criteria

Subjects presenting with unstable uncontrolled medical conditions were excluded from further participation. Subjects showing obstructive sleep apnea with an apnea-hypopnea index of 15 or more episodes per hour of sleep, and/or periodic limb movements associated with arousal (PLMAI) of 15 or greater episodes per hour during the baseline PSG were also excluded. However, subjects with a history of obstructive sleep apnea controlled with nasal continuous positive airway pressure (CPAP) with demonstrated nightly compliance were allowed to participate in the study. Subjects with psychiatric illnesses were accepted, but excluded if they were severely depressed or deemed to be at significant risk for suicide. Other exclusion criteria included subjects with uncontrolled glaucoma, subjects unable to discontinue prohibited medications, and females who were lactating or pregnant. Subjects with a history of alcohol, narcotic, benzodiazepines, or other substance abuse within 1 y prior to the study; excessive caffeine use, defined as a consumption of more than 500 mg of caffeine or other xanthines; smoking more than one-half pack/day or alcohol use > 14 units/w; and history of allergy to milnacipran were also excluded.

Efficacy Assessments

Polysomnography Assessment

The primary endpoints were the difference in sleep maintenance defined by PSG-recorded wake after sleep onset (WASO), number of awakenings after sleep onset (NAASO), and SE, between 4 w of maintenance treatment with milnacipran and placebo. Secondary PSG endpoints included: Latency to persistent sleep onset (LPS), defined as time from lights out to the first consecutive 2 min of uninterrupted sleep; total sleep time (TST); arousal index (AI), number of arousals per hour from sleep onset to lights on; and SWS as percent of TST. The total PSG recording time was 8 h and PSGs were obtained at baseline (visit 2) and at the end of each 4-w treatment maintenance period. All PSGs were carried out to match the subjects' usual bedtime. The PSG recording and scoring were performed on-site in accordance with the 1968 Rechtschaffen and Kales manual modified by Iber et al.^20,21

Secondary Endpoints

Secondary efficacy endpoints included the following subject reported assessments: the medical outcomes study sleep scale (MOS-SS); the Fibromyalgia Impact Questionnaire (FIQ) the Brief Pain Inventory (BPI) short form; the Fatigue Severity Scale (FSS); and the numeric rating scale (NRS) of sleep quality as part of subjective sleep questionnaire (sleep diary) administered throughout the study. The MOS-SS, FIQ, BPI, and FSS were administered at baseline (visit 3) and at the end of each treatment period.

MOS-SS: This self-report instrument consists of 12 items that assess perceived initiation and maintenance of sleep, respiratory problems during sleep, sleep duration, perceived adequacy of sleep and daytime somnolence.²² For 10 of the items, subjects respond to questions on how often each symptom or problem applied to them on a 6-point categorical scale ranging from “all of the time” to “none of the time”. An item on sleep latency, i.e., the time required to fall asleep, was answered on a 5-point categorical response scale ranging from “0 to 15 min” to “more than 60 min”, and an item on quantity of sleep was reported as the average number of hours slept per night. Subjects' responses to the MOS-SS scale were aggregated into a 9-item Sleep Problem Index II score and scored on a 0–100 possible range with higher scores indicating more severe sleep dysfunction. Answers were based on retrospective assessment over the past 4 w. The MOS-SS scale was scored according to the Spritzer and Hays scoring manual, version 1.0.²²

SLEEP QUALITY SCALE: In the daily diary assessment, subjects reported the quality of their sleep over the past 24 h on an 11-point NRS ranging from 0 (“very poor') to 10 (“excellent”). Subjects were instructed to complete the scale each morning upon awakening. The scores were computed as the average rating over the 7 d ending with each polysomnography recording.

FIBROMYALGIA IMPACT QUESTIONNAIRE: The scale is composed of 10 items relating to FM symptoms experienced in the past week. Item 1 is based on 11 questions related to the ability to perform large muscle tasks rated on a 4-point Likert type scale. Items 2 and 3 ask the subject to mark the number of days they felt well and the number of days they were unable to work (including housework) because of FM symptoms. Items 4 through 10 are visual analog scales (VAS) ranging from 0 to 10 on which the subject rates work difficulty, pain, fatigue, tiredness, stiffness, anxiety, and depression. The FIQ is scored in such a way that a higher score indicates a greater effect of FM on a person's life. Each of the 10 items has a maximum possible score of 10, thus the maximum possible total FIQ score is 100.²³

BRIEF PAIN INVENTORY: The scale measures the pain intensity (severity) and the effect of pain on functioning (interference) in the past 24 h on a 0 to 10 numerical rating scale. The pain severity scale was derived as the average score of 4 pain items assessing pain at its “worst”, “least”, “average” and “now”. The BPI interference scale is the average score of 7 items interfering with general activity, mood, walking ability, normal work, relations with other people, sleep and enjoyment of life. Higher scores reflect greater pain.²⁴

FATIGUE SEVERITY SCALE: This 9-item self-report of fatigue in the past week was scored on a 7-point scale with 1 = strongly disagree and 7 = strongly agree. The scores range from 9 to 63, with higher scores indicating higher fatigue severity. A total score ≥ 36 suggests fatigue.²⁵

Safety Assessments

Any abnormal findings such as abnormal laboratory results, clinically significant symptoms, physical/neurological examination findings, hypersensitivity to milnacipran or any other medications, worsening of the underlying disease, emergence and worsening of depression, drug overdose/dependency/misuse, and/or pregnancy were considered as adverse events (AEs).

Vital signs were obtained at every visit to monitor changes in heart rate and blood pressure. Electrocardiograms and physical examinations were performed at Screening and at End of Study visit. Urinalysis was performed on site by urine dipstick at Screening and Baseline/randomization visits. Pregnancy test using human chorionic gonadotrophin (HCG)-dipstick was performed at Screening, Randomization, and End of Study visits. Laboratory assessments (hematology and clinical chemistry) were done at Screening and at the end of each treatment period and include the following parameters: red blood cell count, hemoglobin, hematocrit, white blood cell count with differential, platelets, alkaline phosphatase, aspartate amino-transferase, alanine transaminase, bilirubin, protein, albumin, urea nitrogen, creatinine, creatinine kinase, urea, gamma-glutamyl transferase, cholesterol, glucose, potassium, sodium, and calcium creatinine clearance. The laboratory analyses were performed at Quest Diagnostics, Twinsburg, OH.

The BDI-II²⁶ was administered at the screening visit, baseline, and at the end of each treatment period to assess severity of current symptoms of depressive disorders.

All subjects receiving at least one dose of study medication or matching placebo were monitored for safety. The treatment-emergent adverse events, subject-reported or observed by the investigator, were assessed for severity and causality. Any serious adverse events were reported to the sponsor and the institutional review board within 24 h of the site notification.

Statistical Analysis

The primary efficacy endpoints were changes from placebo in WASO, NAASO, and SE. A sample size of 20 subjects was estimated by a priori power analysis to achieve 92% power for a two-sided 5% paired t-test with an effect size of 0.80. However, only 19 subjects were enrolled with 15 completers, reducing the power of the study to 80%. The primary efficacy analysis was performed on the per-protocol population who completed the study taking 100 mg/day of the study drug. For missing data, the last-observation-carried-forward approach (LOCF) while receiving the study medication or placebo was used for the sleep diary quality of sleep scale. The LOCF approach was used for each crossover period separately.

All data were analyzed using IBM SPSS Statistics 20 manufactured by IBM Watson Analytics. Descriptive statistics are reported for the primary and secondary end points by baseline visit (Table 1).

Table 1.

Polysomnography parameters and symptoms of patients with fibromyalgia at baseline and after 4-w maintenance treatment with milnacipran and placebo.

Polysomnography and secondary subjective end points were each analyzed using a paired t-test. Only the 95% confidence intervals (95% CIs) for the true mean treatment difference are reported for primary and secondary efficacy endpoints (Table 1).

A post hoc responder analysis was also implemented, defining responders to milnacipran as showing at least a 25% reduction in BPI pain interference from baseline. This definition is more stringent than the Initiative on Methods, Measurement and Pain Assessment in Clinical Trials (IMMPACT) provisional benchmark for interpreting minimal clinically important differences (MCIDs), which recommends a one-point decrease in the BPI interference score being seen as a minimally important improvement. Changes in objective and secondary subjective endpoints for the identified responder subgroup were each evaluated using the paired t-test.²⁷

All randomized subjects who received one or more doses of the study and placebo treatment were included in safety assessments. Safety and tolerability analyses focused primarily on frequency of adverse events and the changes in vital signs and clinical laboratory assessments over time.

RESULTS

Subject Disposition and Characteristics

A total of 45 subjects were screened, of whom 19 (42.2%) were randomized to receive study medication (9 milnacipran → placebo sequence and 10 placebo → milnacipran sequence) (Figure 2). During the study, treatment was discontinued in four subjects, with 2 discontinued because (AEs occurred during milnacipran treatment in two, a serious AE (gallstones) unrelated to study drug occurred during placebo treatment in one, and consent following milnacipran treatment was withdrawn in one patient. Two of the patients in whom treatment was discontinued had a study drug-related AE, petechial rash and pruritus, respectively.

Figure 2. Flow of subjects through the study.

*Baseline subjective assessments are carried out at V3 (end of screening period/randomization visit). AE, adverse event.

In total, 15 subjects (78.9%) subjects received drug and placebo treatment and were included in the per-protocol analysis for the assessment of primary polysomnography endpoints and secondary subjective endpoints.

Subjects were predominantly women and white (17 of 19 [89.5%]), with a mean age of 49.2 y (range, 28–72 y) and a mean weight of 196.7 lb (n = 19; ± 54.0). The mean duration of FM was 9.2 y (n = 18; ± 6.9) and the mean y since diagnosis of FM was 4.2 (n = 17; ± 5.1).

Objective sleep characteristics and secondary subjective measures recorded at baseline are shown in Table 1. Study subjects showed sleep disturbance at baseline as shown by PSG parameters for sleep continuity and architecture. They also showed clinically relevant sleep disturbance in the MOSSS sleep problem index 2 subscale and sleep quality scale. The average total FIQ score was 50, a value consistent with reported averages for moderate FM severity.²³ Mean FSS score was 50.4, with values above 36 reflecting fatigue.

Efficacy Analyses

Polysomnography Assessments

Subjects treated with milnacipran showed no significant reduction in PSG-determined WASO and NAASO versus treatment with placebo: (end of treatment paired difference: 22.6 minutes; t₁₄ [p value] = 2.086 [0.056] and 4.6; t₁₄ [p value] = 2.086 [0.056]). SE improved with placebo treatment versus milnacipran (end of treatment paired difference: −6.1%; t₁₄ [p value] = −2.159 [0.049]). Milnacipran did not show significant improvement in other PSG parameters (LPS, TST, AI, and SWS) (Table 1).

Secondary Subjective Efficacy Endpoints

Treatment with milnacipran did not result in statistically significant differences from placebo in any of the subjective scales (MOS-SS sleep problem index 2, subjective sleep quality, FSS, FIQ, and BPI) (Table 1).

Post hoc Responder Analysis

Ten of the 15 completers met the criteria for responder analysis and showed a significant reduction in the daily effect of pain while on milnacipran as compared to placebo (end of treatment paired difference: −1.44; t₉ [p value] = −2.350 [0.043]). The reduction in this pain parameter was accompanied by significant improvement from placebo in subjective sleep quality (sleep dairy) (end of treatment paired difference: 0.64; t₉ [p value] = 2.396 [0.040]) (Table 2).

Table 2.

Brief Pain Inventory interference scores and sleep quality for responders at the end of the treatment period with milnacipran and placebo (n = 10)

Safety Assessments

No drug-related SAEs were observed in the study. The incidence of treatment emergent adverse events was 64.2% in subjects treated with milnacipran compared to 35.7% in subjects treated with placebo. The AEs were mostly mild to moderate with nausea/vomiting and headache being the most commonly reported (Table 3). Nausea and vomiting resolved when the subjects followed the instructions to take the drug with their morning and evening meals.

Table 3.

Summary of reported treatment-emergent adverse events in each treatment group.

The two subjects with adverse events of pruritus and pete-chial rash were immediately withdrawn from the study, with both cases related to the study drug. There was only one non-drug-related SAE (gallstones) for which one subject was withdrawn while on placebo. The subject experiencing excessive urination reported concomitant initiation of hydrochlorothiazide. No clinically significant laboratory abnormalities were reported.

DISCUSSION

Although the most frequent complaint in FM is pain, poor sleep quality and daytime fatigue are almost as common. All currently FDA-approved medications for FM treatment (duloxetine, milnacipran, and pregabalin) have shown improvements in pain. However, the effects of these therapeutic agents on sleep are variable.²⁸

In the current study, milnacipran administration showed no statistical improvements in the primary sleep maintenance parameters (WASO, NAASO) or secondary PSG-determined measures of sleep in FM subjects. On the contrary, subjects treated with milnacipran showed a statistically significant decrease in SE (6.1%). Also no significant improvements were noted in subjects' subjective assessments of sleep quality (sleep diaries, MOS-SS) or the multiple symptom domains of FM. Milnacipran improved pain significantly in up to two thirds of study completers based on the BPI interference sub-scale. This was consistent with previous reports of the impact of milnacipran on pain.^15,16,29 When analysis of the data was limited to pain responders, there was a significant improvement in subjective sleep quality based on daily sleep diaries. Thus, the reduction of pain seems to be due to the analgesic effect of milnacipran in these subjects and any sleep benefit is an indirect consequence. Interestingly, in another study, milnacipran administration in depressed subjects improved sleep only when depression was also improved, further supporting the finding that milnacipran does not affect sleep continuity directly.³⁰ The lack of agreement between sleep diaries and PSG is consistent with reports from the literature that good sleepers show concordance between PSG and sleep diaries while poor sleepers do not.^31,32

Our results extend the findings of Hindmarch et al.,³³ who first reported a lack of sedative effects with milnacipran administration in young and elderly normal volunteers. As an SNRI, milnacipran results in increased synaptic levels of norephinephrine and 5-hydroxytryptamine. The affinity of milnacipran for norepinephrine receptors has been shown to be three times that for serotonin.³⁴ Whereas, both monoamines contribute to pain and sleep modulation, the higher norephinephrine reuptake activity may account for the lack of sedation seen in these subjects.^29,35 In addition, milnacipran has been shown to increase orexinergic transmission in the hypothalamus and histaminic transmission in the frontal cortex. Both of these neurotransmitters also contribute to wakefulness.³⁵

The current study had several limitations that hindered our ability to draw any firm conclusions. Given the small sample size and the resultant expanded variability, virtually all or most of the subjects would have had to show improvements in pain in order to demonstrate a significant effect on sleep (see endnote). Also, the design of the study included only single PSG recordings for baseline and end of treatment periods. The well-established effects of adaptation and readaptation to the laboratory may have compromised our results.^36,37 Laboratory adaptation effects include decreases in TST and increases in sleep fragmentation.³⁶ Further, it was noted in this study that percent time spent in SWS was considerably lower at baseline than at the end of treatment with milnacipran and placebo (4.2%, 8.4%, and 9.4%, respectively). Because reduced SWS is also a part of laboratory adaptation, distinguishing between the contribution of low SWS values inherent in FM and laboratory adaptation is difficult. We had hoped that our study would allow for a comparison between genders but only two males completed the study.

Given the complexity of sleep disturbance in FM, future studies are needed to precisely measure the microstructural abnormalities of sleep in this condition (e.g., alpha intrusions, cyclic alternating patterns, microarousals, and power spectral analyses of the electroencephalogram). This will allow for a deeper understanding of the mechanisms of sleep disturbance in FM and shed light on treatment effects. In a previous study, decreases in alpha intrusions in subjects with FM were accompanied by improvements in TST, SWS, subjective sleep quality, tender point index, pain, fatigue, and cognitive clarity.³⁶

Future PSG study designs for evaluating pharmacotherapies for FM should include entry criteria requiring the presence of alpha intrusion in order to ensure a more homogeneous study population. Future studies should also incorporate multiple nights for objective data collection and should focus on a larger sample of pain responders and responders based on ratings of the Patient Global Impression of Change (PGIC). Studies with durations of treatment with milnacipran beyond 4 w might reveal improvement in sleep quality as evidenced by the long-term therapeutic response of milnacipran.³⁸

In the current study, milnacipran was generally well tolerated in the majority of the study population. One subject developed rash while on active drug and another complained of itching; both were withdrawn from the study. Sexual side effects reported by a male subject were resolved upon discontinuation of milnacipran. The remainder of subjects showed only nausea, headache, and probably hot flushing as more common AEs related to milnacipran administration. These AEs were consistent with reports from the literature.³⁹ All AEs were resolved.

CONCLUSION

In short, our data suggest that milnacipran 100 mg/day is not sedating in most subjects with FM and is not more effective than placebo in terms of improvements in objective and subjective measures of sleep. Pain responder analysis suggests that when the effect of pain on daily living is reduced, subjective sleep quality improves.

DISCLOSURE STATEMENT

Funding source for this study was Forest Research Institute, Jersey City, NJ. The authors have indicated no other financial conflicts of interest. This was an investigational use of the study drug; however, there was no off label use.

ABBREVIATIONS

AE

adverse event

AI

arousal index

BDI-II

beck depression inventory

BPI

brief pain inventory

CPAP

continuous positive airway pressure

FIQ

fibromyalgia impact questionnaire

FM

fibromyalgia

FSS

fatigue severity scale

HCG

human chorionic gonadotrophin

IMMPACT

initiative on methods, measurement and pain assessment in clinical trials

IRB

independent institutional review board

LOCF

last-observation-carried-forward

MCIDs

minimal clinical important differences

MOS-SS

medical outcomes study sleep scale

NAASO

number of awakenings after sleep onset

NMDA

N-methyl-D-aspartate

NRS

numeric rating scale

PGIC

patient global impression of change

PLMAI

periodic limb movements associated with arousal

PSG

polysomnography

SE

sleep efficiency

SWS

slow wave sleep

TST

total sleep time

VAS

visual analog scale

WASO

wake after sleep onset

ENDNOTE

After the fact, a power analysis would suggest that if the effects on pain reduction are large (Cohen d = 0.80), the power of the paired t-tests with n = 10 is only 0.62, whereas if the effect is medium in size (d = 0.50), the power is only 0.29. Thus, the probability of obtaining significant effects with a sample size of 10 is quite low even if the hypothesis is true.