Jabfm.com

ORIGINAL RESEARCH
Bupropion and Restless Legs Syndrome:A Randomized Controlled Trial Max Bayard, MD, Beth Bailey, PhD, Deep Acharya, MD, Farhana Ambreen, MD,Sonia Duggal, MD, Taran Kaur, MD, Zia Ur Rahman, MD, Kim Roller, MD,and Fred Tudiver, MD Introduction: Restless legs syndrome (RLS) is a common neurological disorder affecting 10% of the
population. Most antidepressants exacerbate symptoms; however, correlational studies have noted
symptom improvement with bupropion. The purpose of the current study was to examine whether, in a
controlled study, bupropion would improve the symptoms of RLS, or at least not exacerbate them.
Methods: This was a double-blinded, randomized controlled trial. Twenty-nine participants with
moderate to severe RLS received 150 mg sustained-release bupropion once daily, and 31 control partic-
ipants received a placebo. Participants were followed for 6 weeks and completed standardized tools,
including the International Restless Legs Syndrome Study Group (IRLSSG) severity scale.
Results: The primary outcome was change from baseline in IRLSSG severity score; lower scores were
associated with improved symptoms. At 3 weeks, IRLSSG scores were 10.8 points lower in the bupro-
pion group and 6.0 points lower in the placebo group (P ؍ .016). At 6 weeks, IRLSSG scores were10.4
points lower in the bupropion group and 7.6 points lower in the placebo group (P ؍ .108). Bupropion
was more effective than placebo in the treatment of RLS at 3 weeks; however, this difference was not
statistically significant at 6 weeks.
Conclusions: The data from our study suggest that bupropion does not exacerbate the symptoms of
RLS and may be a reasonable choice if an antidepressant is needed in individuals with RLS. Larger stud-
ies that include titration of bupropion should be considered to determine if bupropion is appropriate
for primary treatment of RLS, particularly considering the lower cost and favorable side effect profile
compared with currently recommended first-line dopamine agonists. (J Am Board Fam Med 2011;24:
422– 428.)
Keywords: Bupropion, Neurology, Restless Legs Syndrome
Restless legs syndrome (RLS) is a common neuro- In addition to discomfort, individuals with RLS logical movement disorder. It affects approximately have overall worse quality of life than the general 10% of the adult population of the United States population,3 and depression and anxiety are more and is more common with advancing age.1 RLS is common among those with RLS.4 Sleep distur- diagnosed clinically based on 4 criteria (Table 1).2 bances are also very common and are often thepresenting complaint. Recent studies also showhigher incidence of cardiovascular disease in pa-tients with RLS.5 This article was externally peer reviewed.
Submitted 23 July 2010; revised 9 February 2011; accepted RLS may be a primary or a secondary condition.
Secondary causes of RLS include iron deficiency, From the Department of Family Medicine (MB, BB, FT), East Tennessee State University, Johnson City (DA, FA, chronic kidney disease, pregnancy, and various medications. Treatment of the secondary causes Funding: This study was funded in part by a grant from the East Tennessee State University Research Development may improve or resolve the symptoms of RLS.
Committee (MB) and in part by the Department of Family Antidepressants, particularly serotonin-specific re- Medicine, East Tennessee State University.
Conflict of interest: none declared.
uptake inhibitors (SSRIs), exacerbate RLS and peri- Corresponding author: Max Bayard, East Tennessee State odic limb movement disorder, which is closely corre- University, Johnson City Family Practice, 917 W. Walnut,Johnson City, TN 37604 (E-mail: mbayard@nmcinc.org).
lated with RLS. In one study, venlafaxine, citalopram, 422 JABFM July–August 2011 Vol. 24 No. 4
Table 1. Diagnostic Criteria for Restless Legs Syndrome
caused by uncomfortable or unpleasantsensations in the legs (urge to move may notbe accompanied by uncomfortable sensations,and arms or other body parts may beinvolved) Urge to move or unpleasant sensation begins or worsens during periods of rest or inactivity Urge to move or unpleasant sensation partially or totally relieved by movement (such aswalking or stretching) as long as activitycontinues Urge to move or unpleasant sensation worse in the evening or at night than during the day,or only occurs during the evening or at night;in very severe cases, worsening at night maynot be noticeable, but must have beenpreviously present fluoxetine, paroxetine, and sertraline all caused in- 150-mg, sustained-release dose of bupropion for 6 creasing frequency of periodic limb movements, weeks. Control group participants received identi- whereas bupropion decreased periodic limb move- cal placebo capsules. Outcome measures were as- ments compared with controls.6 In another study, sessed at baseline and weekly for 6 weeks.
RLS was noted as a possible side effect of the use offluoxetine, paroxetine, citalopram, sertraline, escitalo-pram, venlafaxine, duloxetine, and mirtazapine.7 Fi- Participants and Recruitment
nally, in one small case series, patients taking bupro- Participant recruitment was accomplished via local pion had improvement of their RLS symptoms.8 To radio advertisements, notices in local newspapers, and date, no controlled clinical trials have studied the flyers placed in patient areas of academic medicine effect of bupropion on the symptoms of RLS.
primary care clinics. To maximize recruitment, po- Because depression is common in RLS and most tential participants were screened for study eligibility commonly prescribed antidepressants exacerbate using the RLS diagnosis questions presented in Table the symptoms of RLS, it would be beneficial to 1. Screening over the phone included positive re- identify an antidepressant that did not exacerbate sponses to the 4 criteria in general terms. Criterion symptoms, particularly if it improved symptoms. If one (the urge to move legs usually accompanied or bupropion was found to improve, or at least not caused by uncomfortable or unpleasant sensations in exacerbate, the symptoms of RLS, it would be a the legs) was addressed over the phone with the ques- good choice for treating depression in individuals tion, Do you get an uncomfortable feeling in your with RLS. We conducted a double-blind, random- legs associated with an urge to move? The form ized controlled trial comparing the effect of 150 mg completed by the physician screener during the initial sustained-release bupropion daily with placebo on face-to-face evaluation listed each of the diagnostic the symptoms of RLS. We hypothesized that bu- criteria, with a choice of “Y” (yes) or “N” (no), which propion would improve the symptoms of RLS or at would be marked by the physician. For example, to the least would not exacerbate RLS symptoms.
determine whether criterion one was present, thescreener would ask, “Do you have an urge to moveyour legs usually accompanied by uncomfortable or unpleasant sensations in the legs?” A potential partic- Overview of Study Design
ipant had to answer affirmatively to all 4 questions The study was a randomized controlled trial of a and had to score 15 or higher on the International bupropion regimen for adults with RLS. Active RLS Study Group (IRLSSG) scale to be eligible for drug group participants were given a once-daily, the study (see below). In addition, participants had to be available for phone assessment weekly and be will- severity was assessed using the IRLSSG scale. This ing to travel to the study sites for in-person assess- validated 10-item questionnaire assesses symptom se- ment at 3 and 6 weeks. Furthermore, potential par- verity, frequency, and impact on daily life.9 The max- ticipants had to be willing to accept randomization imum score is 40, and a higher score indicates more into either the active drug or placebo group.
severe RLS. Participants completed the IRLSSG dur- Participants were excluded from study participa- ing the in-clinic assessments at baseline, 3 weeks, and tion for the following reasons: history of seizure 6 weeks. The scale was also completed via phone at 1, disorder, alcoholism, suicidal history or ideation, inability to return for 3- and 6-week assessment, no Depression was assessed via the Beck Depres- telephone access, eating disorders, age younger sion Inventory (BDI-II). This 21-item scale is well- than 18, pregnancy, and unwillingness or inability validated and commonly used to screen for depres- to discontinue current medications for the treat- sive symptoms in both research and clinical ment of RLS. Participants currently taking medi- settings. Scores can potentially range from 0 to 63, cations for the treatment of RLS were not auto- and higher scores indicate higher levels of depres- matically excluded from the study but had to sion. Study participants completed the BDI-II at complete a 2-week washout period off of the med- baseline and at the 3- and 6-week assessments.
ication before becoming eligible. Information Finally, as part of every in-person and phone about exclusion criteria were obtained from poten- assessment after baseline, participants were asked tial participants by a study physician during a how many days during the last week they had failed screening physical. The study physician was then to take their study medication and whether they responsible for determining participant eligibility had experienced any problems with or had any for the study. The study was approved by the uni- concerns about the medication they were taking.
versity institutional review board and informedconsent was obtained consistent with federal guide- Data Analysis
lines. The study was registered with the Clinical Examination of potential attrition bias involved ␹2 Trials Registry before participant recruitment.
analysis and t tests using P Ͻ .05 as criteria forsignificance. Evaluation of baseline and study out- Procedures
come differences by treatment group also used ␹2 After baseline data collection, eligible and consent- analysis and t tests where appropriate. All analyses ing participants were placed in a pool and random- of outcomes were intention to treat. Thus, all eli- ized in waves of 10 using a computer generated gible participants who were initially randomized to random list. Screeners were blinded to the alloca- a treatment group were included in all analyses, tion. Participants were randomized, in equal num- regardless of study completion status. For those bers within each block of 10, to either the active lost to follow-up and missing a score at a particular drug or control groups. Both groups received a time point, the most recent previous score for that bottle of 42 capsules with instructions to take one participant on that measure was used in analysis.
per day 2 hours before bedtime. The capsules givento the active drug group participants contained 150mg sustained-release bupropion, and those given to the control group contained no active ingredients.
Study Completion
Both groups received phone calls from study phy- One hundred fifty-one adults were screened for the sicians at 1, 2, 4, and 5 weeks after baseline. Study study during the period of January 2008 through physicians were blinded to randomizations. All par- February 2009. Seventy-two were determined to be ticipants also were scheduled to meet in the clinic eligible for the study, but 12 were excluded from with a study physician for more detailed assessment the study and considered screening failures before at 3 and 6 weeks after baseline assessment.
obtaining any data from them. Reasons for consid-ering participants as screening failures included no Measures
telephone or inability to contact at week 1 (n ϭ 4), Basic demographic information was obtained from failure to complete washout (n ϭ 2), inability to participants at baseline. Basic medical history was return to the clinic at 3 and 6 weeks (n ϭ 3), or they obtained as part of the screening physical. Restless leg chose not to participate before starting the medi- 424 JABFM July–August 2011 Vol. 24 No. 4
Figure 1. Participant flow diagram.
Did not meet inclusion criteria (n=75)Refused parƟcipaƟon cation (n ϭ 3). The remaining 60 participants be- the study in group assignment or by sex. In addition, came the study sample, of which 29 were random- no significant attrition bias with respect to study mea- ized to the active drug and 31 to the control group sures was noted; those lost to follow-up did not differ from remaining participants on baseline RLS or de- By the 3-week assessment, 11 of the 60 participants pression scores. However, those who remained in (18%) had either withdrawn from the study, could the study were significantly older than those who not be contacted, or failed to keep the 3-week clinic appointment. By week 6, an additional 3 participantswere lost to follow-up, for a total participant comple- Study Group Differences
tion rate of 77%. As shown in Table 2, those who Descriptive characteristics of study participants are failed to complete assessments at either time period presented in Table 3. The average age of partici- did not differ significantly (P Ͼ .05) from those still in pants was approximately 50 years (range, 18 –72 Table 2. Comparison of Those Who Completed the Study With Those Lost to Follow-up
*P Ͻ .05 for t test of significant difference between the completion groups.
IRLSSG, International Restless Legs Syndrome Study Group; BDI, Beck Depression Inventory.
Table 3. Participant Descriptives by Study Group Assignment
*Values presented as mean (SD).
IRLSSG, International Restless Legs Syndrome Study Group; BDI, Beck Depression Inventory.
years). Three quarters of the sample were women.
scores and IRLSS differences from baseline that Because of the study inclusion criteria, all partici- were evident at 3 weeks. However, the placebo pants had elevated IRLSS scores at baseline, with a group reported some additional improvement in total score range of 15 to 37. The 2 study groups RLS symptoms at 6 weeks. As a result, the study did not differ significantly (P Ͼ .05) on demo- group differences in RLS symptoms found at 3 graphic characteristics or baseline scores on any weeks, while trending in the expected direction, were no longer significant at 6 weeks.
Study group differences at 3 weeks are presented Follow-up analyses were performed to deter- in Table 4. As described earlier, all analyses were mine if lack of medication compliance or elevated intention to treat, and participants lost to follow-up levels of depression may at least partially explain were assigned their most recent previous score on why the treatment effect seen at 3 weeks was no each measure for analysis. Compared with those longer statistically significant at the 6-week assess- who received placebo, participants who took bu- ment. No specific instructions had been given re- propion had significantly lower IRLSSG scores at 3 garding compliance, and participants did not dou- weeks, indicating symptom improvement. In addi- ble their dosage in response to a missed dose. No tion, the average difference between baseline and verification of compliance other than participant 3-week IRLSS scores differed significantly between self-reporting was obtained. Those who received the 2 study groups, with those taking bupropion bupropion were no more likely to skip pills than experiencing nearly an 11 point drop in total score.
those who received the placebo (t(58) ϭ 0.23; P ϭ The 2 groups did not have significantly different .760). In addition, on average, active drug group scores on the depression measure at 3 weeks.
participants reported missing fewer than one pill Differences at 6 weeks also are presented in per week (mean, 0.7). Finally, as shown in Table 4, Table 4. At 6 weeks after baseline, the active drug placebo group participants reported a decrease in group had maintained their decrease in total IRLSS depression symptoms at 6 weeks compared with Table 4. Three- and Six-Week Outcomes by Study Group Assignment*
*All 60 participants were included in the above intention to treat analyses. For those who dropped out by 3 weeks (n ϭ 11) and by6 weeks (n ϭ 14), the most recent previous score for that participant on that measure was used.
†Values presented as mean (SD).
IRLSSG, International Restless Legs Syndrome Study Group; BDI, Beck Depression Inventory.
426 JABFM July–August 2011 Vol. 24 No. 4
depression scores at 3 weeks. However, though the ever, all those with whom we were able to make placebo group had somewhat lower depression contact and who actually began taking the medica- scores at 6 weeks compared with the active drug tion were included in the analysis. Thus, the anal- group, because of small sample size these differ- ysis did include all participants who chose to dis- ences were not statistically significant, and further continue the medication for any reason. It also follow up analyses could not be performed.
included all participants we were unable to reach orwho failed to show up for a research appointmentafter the first week. We believe this is appropriate Discussion
because, although intention to treat analysis may be Bupropion improved the symptoms of RLS com- defined to include all who were initially determined pared with placebo at 3 weeks. The difference at 6 to be eligible, this is not essential. Rather, studies weeks was not statistically significant because of should be transparent and identify their definition improvements in the placebo group; however, the of intention to treat as well as any effects the miss- improvements in the bupropion group persisted ing responses may have on the validity of the data.12 through 6 weeks. The degree of improvement with We feel the effect is minimal for the reasons men- bupropion was similar to the improvement seen tioned above. Another weakness of our study was with dopamine agonists currently approved for the our relatively high drop-out rate. Reasons for with- treatment of moderate to severe RLS.10,11 drawing from the study are listed in Table 5. How- In our study, bupropion did not exacerbate the ever, all who dropped out were included in the symptoms of RLS. This contrasts with other anti- intention to treat analysis. An additional weakness depressants, which have been shown to exacerbate is that we did not titrate the dose of bupropion to symptoms.7 As such, bupropion should be consid- symptom improvement. The dose for all partici- ered for depressed patients with RLS.
pants throughout the study was the starting dose of Our study did not definitively answer the ques- the sustained-release bupropion; it is possible that tion of whether bupropion might be an effective increasing the dose would have resulted in further primary treatment for the symptoms of RLS.
improvement. It is also possible that it would have There was improvement at 3 weeks and a trend to resulted in increased side effects. Our primary ob- continued improvement at 6 weeks. The improve- stacle to titrating this dose was lack of funding.
ment in symptoms was independent of the effect ofbupropion on depression.
Conclusions
Strengths of our study included evaluation of a Our study demonstrates that bupropion does not generic, relatively inexpensive medication with exacerbate the symptoms of RLS. Furthermore, which family physicians have experience. Our en- bupropion improved the symptoms of RLS at 3 tire minimal funding of $2,000 was provided by our weeks compared with placebo. The improvement university, and thus we received no industry sup- persisted at 6 weeks, but further improvement port. In addition, randomization was appropriate, among those who received the placebo resulted in yielding similar groups at baseline. Finally, the de-gree of improvement in RLS symptoms amongthose who received bupropion in the current study Table 5. Reasons for Withdrawing from Study
is similar to improvements seen in studies of med-ications indicated for treatment of RLS.
Weaknesses of our study included the following.
First, because of our limited resources, we were unable to recruit our target of 100 participants, which may have contributed to type II error at 6 weeks; that is, there may have been an effect that was not evident because of insufficient statistical power due to a small number of participants. Sec-ond, though analysis was by intention to treat, *Participant had denied pregnancy at screening but had miscar-riage during study. After this event, we began to require preg- individuals were excluded from this analysis if they nancy testing of all women younger than 55 years of age unless were inappropriately screened (see Results). How- lack of statistical significance, though there was a treatment of restless legs syndrome (RLS) in a primary care population: the REST (RLS Epidemiology, Further studies may clarify the role of bupropion Symptoms, And Treatment) Primary Care Study.
Sleep Med 2004;5(3):237–46.
in the treatment of RLS, and larger studies may 4. Sevim S, Dogu O, Kaleagasi H, Aral M, Metin O, determine if bupropion is effective as a primary Camdeviren H. Correlation of anxiety and depres- treatment for the symptoms of moderate to severe sion symptoms in patients with restless legs syn- RLS. Another important line of study would be to drome: a population based survey. J Neurol Neuro- determine if adding bupropion to an SSRI would surg Psychiatry 2004;75(2):226 –30.
improve SSRI-induced symptoms of RLS. We did 5. Winkelman JW, Shahar E, Sharief I, Gottlieb DJ.
not exclude from our study those taking SSRIs, but Association of restless legs syndrome and cardiovas-cular disease in the Sleep Heart Health Study. Neu- because of small sample size, we were unable to analyze that group separately. Future studies of bu- 6. Yang C, White DP, Winkelman JW. Antidepres- propion and RLS should include titration of dose to sants and periodic leg movements of sleep. Biol Psy- improvement. This is common in studies of dopa- mine agonists, but our limited resources and small 7. Rottach KG, Schaner BM, Kirch MH, et al. Restless sample size precluded this. Because our dose of legs syndrome as side effect of second generation sustained-release bupropion was the starting dose, antidepressants. J Psychiatr Res 2008;43(1):70 –5.
it is possible that increasing the dose could result in 8. Kim SW, Shin IS, Kim JM, Yang SJ, Shin HY, Yoon greater improvement. With bupropion’s reason- JS. Bupropion may improve restless legs syndrome: areport of three cases. Clin Neuropharmacol 2005; able cost and favorable side effect profile, it is our hope that some of these questions will be addressed 9. Walters AS, LeBrocq C, Dhar A, et al. Validation of the International Restless Legs Syndrome StudyGroup rating scale for restless legs syndrome. Sleep References
1. Phillips B, Young T, Finn L, Asher K, Hening WA, 10. Walters AS, Ondo WG, Dreykluft T, et al. Ropini- Purvis C. Epidemiology of restless legs symptoms in role is effective in the treatment of restless legs adults. Arch Intern Med 2000;160(14):2137– 41.
syndrome. TREAT RLS 2: a 12-week, double-blind, 2. Allen RP, Picchietti D, Hening WA, et al. Restless randomized, parallel-group, placebo- controlled study.
legs syndrome: diagnostic criteria, special consider- ations, and epidemiology. A report from the restless 11. Winkelman JW, Sethi KD, Kushida CA, et al. Effi- legs syndrome diagnosis and epidemiology workshop cacy and safety of pramipexole in restless legs syn- at the National Institutes of Health. Sleep Med 2003; drome. Neurology 2006;67(6):1034 –9.
12. Hollis S, Campbell F. What is meant by intention to 3. Hening W, Walters AS, Allen RP, Montplaisir J, treat analysis? Survey of published randomized con- Myers A, Ferini-Strambi L. Impact, diagnosis and trolled trials. BMJ 1999;319(7211):670 – 4.
428 JABFM July–August 2011 Vol. 24 No. 4

Source: http://www.jabfm.com/content/24/4/422.full.pdf