|
A correction has been published: N Engl J Med 2005;353(10):1078.
|
Background Patients with shift-work sleep disorder chronically have excessive sleepiness during night work and insomnia when attempting to sleep during the day. We evaluated the use of modafinil for treating sleepiness in patients with this disorder.
Methods In a three-month, double-blind trial, we randomly assigned 209 patients with shift-work sleep disorder to receive either 200 mg of modafinil or placebo before the start of each shift. Assessments were performed with the use of the nighttime Multiple Sleep Latency Test, the Clinical Global Impression of Change, the Psychomotor Vigilance Test, diaries of patients, and daytime polysomnography. After randomization, we conducted monthly assessments.
Results Treatment with modafinil, as compared with placebo, resulted in a modest improvement from baseline in mean (±SEM) nighttime sleep latency (the interval between the time a person attempts to fall asleep and the onset of sleep) (1.7±0.4 vs. 0.3±0.3 minutes, respectively; P=0.002), and more patients had improvement in their clinical symptoms (74 percent vs. 36 percent, respectively; P<0.001). Patients who were receiving modafinil also had a reduction in the frequency and duration of lapses of attention during nighttime testing of their performance on the Psychomotor Vigilance Test (change from baseline, a reduction in lapse frequency of 2.6 vs. an increase of 3.8, respectively; P<0.001), and proportionally fewer patients reported having had accidents or near accidents while commuting home (29 percent vs. 54 percent, respectively; P<0.001). Despite these benefits, patients treated with modafinil continued to have excessive sleepiness and impaired performance at night. Modafinil did not adversely affect daytime sleep as compared with placebo. Headache was the most common adverse event.
Conclusions Treatment with 200 mg of modafinil reduced the extreme sleepiness that we observed in patients with shift-work sleep disorder and resulted in a small but significant improvement in performance as compared with placebo. However, the residual sleepiness that was observed in the treated patients underscores the need for the development of interventions that are even more effective.
Methods
Patients
Adults between the ages of 18 and 60 years were eligible if they worked each month at least five night shifts for 12 hours or less, with 6 hours or more worked between 10 p.m. and 8 a.m. and at least three shifts occurring consecutively. Patients were diagnosed with shift-work sleep disorder in accordance with criteria stipulated in the International Classification of Sleep Disorders.5 Our diagnostic criteria included a primary symptom of excessive sleepiness on the night shift and insomnia during opportunities for daytime sleep and the absence of other primary sleep disorders, other medical conditions, and medications that might cause sleepiness. Patients had to have reported chronic excessive sleepiness (
3 months) during night shifts; a Clinical Global Impression of Severity12 rating of moderately ill or worse for sleepiness on work nights, including the commute home from work; an average latency to sleep onset of 6 minutes or less during 20-minute nap opportunities at 2-hour intervals during the night, as measured by the Multiple Sleep Latency Test13,14; and a sleep efficiency of 87.5 percent or less as determined by daytime polysomnography. Participants provided written informed consent.
Study Design and Conduct
The three-month randomized, double-blind, placebo-controlled study was conducted in the United States between December 2001 and September 2002. At each of 28 centers, an institutional review board approved the informed-consent statement and protocol. The study included a screening visit to assess eligibility; a baseline visit on the night after the patient had worked three or more consecutive night shifts in order to establish pretreatment levels of alertness and performance, the severity of sleepiness, and results of daytime polysomnography; and a randomization visit for the administration of the study drug. Thereafter, patients were evaluated monthly during an overnight laboratory shift after having worked for three or more consecutive nights.
Patients were randomly assigned (in a 1:1 ratio) to receive 200 mg of modafinil (Provigil, Cephalon), formulated as 100-mg tablets, or an identical-appearing placebo, taken 30 to 60 minutes before the start of each night shift. Assessment of treatment adherence was performed at visits after the initial baseline visit.
Measures of Efficacy
Sleep latency during laboratory night shifts was measured by polysomnography at two-hour intervals, starting at 2 a.m., with the use of the Multiple Sleep Latency Test (see the Supplementary Appendix, available with the full text of this article at www.nejm.org).13,14 To assess alertness with the use of a performance measure, a 20-minute Psychomotor Vigilance Test15 was administered every two hours, starting at 1 a.m. The level of sleepiness as reported by patients was assessed hourly using the Karolinska Sleepiness Scale,16 which ranges from 1 (very alert) to 9 (very sleepy). The investigator-rated Clinical Global Impression of Change,12 the scoring of which ranges from 1 (very much improved) to 7 (very much worse), was used to assess changes from baseline in the severity of sleepiness during night shifts, including the commute to and from work. Patients also completed electronic diaries containing questions about sleepiness, sleep, and caffeine use during the night shift and the commute home.
There were two prespecified primary efficacy variables. The first was the rating on the Clinical Global Impression of Change test for sleepiness during the night shift, including the commute to and from work, at the final visit. This test assessed the extent to which treatment effects could be recognized by patients and physicians. The second prespecified primary efficacy variable was the change between baseline and the final visit (i.e., at the third month or at withdrawal from the study) in overall mean sleep latency on the basis of results of the nighttime Multiple Sleep Latency Test. This test has been validated as a measure of sleepiness during the day but not at night, so questions remain regarding which results on this test indicate pathological sleepiness during the night and what level of improvement at night is clinically meaningful. Therefore, as recommended,17 one of our secondary outcome measures — the frequency and duration of lapses of attention during performance on the Psychomotor Vigilance Test15 — served as a validated and objective measure of alertness at night.
Safety Assessments
Adverse events were monitored throughout the study. Blood pressure and heart rate were monitored, and clinical laboratory tests (including chemical and hematologic studies) were conducted at each visit. Physical examination and electrocardiography were performed at the screening visit and the final visit.
Other Assessments
Polysomnography was conducted for eight hours, starting at 10 a.m. after baseline and after final laboratory night shifts. Melatonin concentrations were measured from saliva samples.18
Statistical Analysis
A total of 204 patients were selected for enrollment in the study (see the Supplementary Appendix for calculations regarding the sample size). Randomization was performed with the use of a central randomization process and stratified by center with the use of permuted blocks of two. On the basis of the sample that was available for efficacy analysis, there was at least 80 percent power (with the use of post hoc power calculations) for the statistical inference on both prespecified primary end points (assuming an alpha level of 0.05).
Comparisons of continuous demographic variables between groups were conducted with the use of analysis of variance, with treatment as a factor. Discrete categorical demographic variables were compared with the use of the chi-square test or Fisher's exact test. Included in efficacy analyses were patients who had been randomly assigned to treatment and received at least one dose of study drug and who had had a baseline assessment and at least one assessment after baseline on either the Multiple Sleep Latency Test or the Clinical Global Impression of Change. For the final-visit efficacy analysis, data from the patients' last visit on or before the third month were used.
Comparisons between the groups were conducted on the change between the baseline visit and the final visit with regard to variables on the Multiple Sleep Latency Test, scores on the Karolinska Sleepiness Scale, polysomnographic measures, and data from patients' diaries related to sleepiness ratings, unintentional and intentional sleep episodes during the night shift, the consumption of caffeinated drinks, and sleep efficiencies. All of these variables were analyzed with the use of analysis of variance, with treatment and site as factors. Data from the Clinical Global Impression of Change test were evaluated with the use of a Cochran–Mantel–Haenszel chi-square test, with adjustment made for site and modified ridit scores used to account for ordered categories.
Comparisons were made with the use of a chi-square test or Fisher's exact test on data from diaries that were related to the percentage of patients reporting mistakes, near accidents, or accidents during the night shift; unintentional sleep episodes, accidents, or near accidents during the commute home; and rates of adverse events. Comparisons of performance on the Psychomotor Vigilance Test and of melatonin phase were performed with the use of the Wilcoxon nonparametric rank test. There were no interim analyses of the data. All reported P values are two-sided and not adjusted for multiple testing.
Six academic investigators and four representatives of the sponsor designed the study and analyzed the data. Drs. Czeisler, Walsh, Roth, and Dinges conceived of and designed the study in collaboration with representatives of the corporate sponsors, Drs. Hughes, Niebler, Arora, and Kingsbury. Data were fully accessible to all group members, with the study sponsor placing no limits on interpretation or publication. The study designers vouch for the completeness and accuracy of the analyses. All authors were involved with the preparation of the manuscript.
Results
Disposition and Baseline Characteristics of Patients
A total of 4533 shift workers were prescreened through telephone calls placed to a central agency; 2765 workers were referred to study sites. Of 609 patients undergoing laboratory testing, 400 were judged ineligible or withdrew (Figure 1 and Supplementary Appendix). Most commonly, ineligible patients did not meet inclusion criteria for polysomnography (107 patients) or sleep latency (53) or either withdrew their consent or were lost to follow-up (118). Of 209 patients who were randomly assigned to receive the study drug, 204 patients received the drug, and 153 patients completed the study. At baseline, there were no significant differences in demographic variables, shift-work type, sleepiness, performance, and results on polysomnography between the group that received modafinil and the one that received placebo (Table 1). The patients who completed the trial and those who did not had similar baseline values for the primary outcome variables (as measured by the Multiple Sleep Latency Test and the Clinical Global Impression of Severity test) and similar results on polysomnography. Patients were severely sleepy at baseline, with overall mean (±SD) sleep latencies of 2.0±1.8 minutes and 2.1±1.5 minutes for the placebo and modafinil groups, respectively.
|
|
Seventy-four percent of patients in the modafinil group were rated as at least minimally improved on the Clinical Global Impression of Change test at the final visit, as compared with 36 percent in the placebo group (P<0.001) (Figure 2A, and Table 2 in the Supplementary Appendix). Overall mean (±SEM) sleep latency, as measured by the Multiple Sleep Latency Test, increased from 2.1 minutes at baseline to 3.8 minutes at the final visit with modafinil (change, 1.7±0.4 minutes; P<0.001) but not with placebo (2.04 at baseline vs. 2.37 at the final visit; change, 0.3±0.3; P=0.24) (Figure 2B). Sleep latency was significantly greater in the modafinil group than in the placebo group (P=0.002). This improvement in sleep latency with modafinil versus placebo was found at 2 a.m. (P=0.02) and 4 a.m. (P<0.001) (Figure 2C), but not at 6 a.m. (P=0.45) or 8 a.m. (P=0.17) (Figure 2D). A higher proportion of patients receiving modafinil had a positive change in the sleep-latency score from pretreatment to the final visit (Figure 1 in the Supplementary Appendix). Notwithstanding this improvement, sleep latencies during the night shift averaged less than six minutes, which is below the level considered normal during the daytime.
|
The duration of lapses showed a similar result, decreasing from baseline (780 msec) to the final visit (669 msec) for patients receiving modafinil and increasing from baseline (852 msec) to the final visit (1235 msec) for those receiving placebo. This resulted in a significant difference at the final visit (P=0.004) and in the change from baseline to the final visit in favor of modafinil versus placebo (P=0.019). Sleepiness levels on the Karolinska Sleepiness Scale were also significantly reduced for patients receiving modafinil (baseline mean, 7.3; final visit mean, 5.8; change, –1.5±0.2), as compared with placebo (baseline, 7.1; final visit, 6.7; change, –0.4±0.2) (P<0.001) (Figure 2F). In general, the results of efficacy measures at the final visit were observed at the first visit after baseline and sustained throughout subsequent visits (see the Supplementary Appendix).
Data Derived from Electronic Diaries
There were significant effects for three of the seven efficacy variables in the patients' diaries (Table 2). As compared with placebo, 200 mg of modafinil reduced the maximum level of sleepiness during night-shift work (P<0.001 for the change from baseline vs. placebo) and the level of sleepiness during the commute home (P=0.01), and 25 percent fewer patients receiving modafinil reported having had accidents or near accidents during the commute home (P<0.001). Modafinil treatment during night shifts had no statistically significant effects on unintentional or intentional sleep episodes, mistakes, accidents or near accidents, or caffeine consumption (Table 2). During days following nights off, there were no significant differences in caffeine use and sleep efficiency between the modafinil group and the placebo group (Table 2). The use of sleeping pills was not specifically monitored, although concomitant use of medications was queried at each visit. One of 96 patients in the modafinil group reported the use of a prescription hypnotic agent, whereas none of the 108 patients in the placebo group did. Five of the 96 patients in the modafinil group reported the use of over-the-counter sleep aids versus 1 of the 108 patients in the placebo group (P=0.102).
|
Headache was the most common adverse event associated with treatment in both groups (Table 3). No serious adverse events were reported for patients in the modafinil group. More patients in the modafinil group than in the placebo group had insomnia (6 percent vs. 0 percent, respectively; P=0.01). Adverse events that were not serious but resulted in the inability to carry out usual activities were defined as severe. Eleven patients reported such events (six in the modafinil group and five in the placebo group) (see Table 3 of the Supplementary Appendix). No clinically meaningful differences in vital signs, clinical laboratory measures, physical examinations, or electrocardiographic findings were observed between treatment groups.
|
There were no significant differences between modafinil and placebo with respect to any measurement of daytime sleep, including sleep duration, latency, and efficiency and the proportion and distribution of sleep stages (Table 4). Patients receiving modafinil did not differ significantly from those receiving placebo in the mean change in melatonin phase from baseline to the final visit (0.4 hour and –0.1 hour, respectively).
|
Circadian-rhythm sleep disorders have long been recognized as important disruptions of sleep–wake behaviors in a subgroup of people who are substantially more impaired than others with similar schedules.19,20,21 Such differential vulnerability regarding cognitive impairment that is induced by extended wakefulness at night is a stable characteristic of these persons.22,23 Estimates of the proportion of night-shift workers who meet the clinical criteria of both excessive sleepiness and daytime insomnia that we used to diagnose shift-work sleep disorder range from 5 to 10 percent.4,5,6,24 The burden of illness in persons with shift-work sleep disorder is substantial, as compared with shift workers without the disorder.6,7,24,25 Thus, shift-work sleep disorder is more than simply being tired on the night shift.
In this randomized, placebo-controlled study — the first such trial in the investigation of shift-work sleep disorder — improvements in alertness and performance were found with 200 mg of modafinil in measures of sleep latency, clinical-impression rating, sustained-attention performance, and patient-estimated sleepiness. Consistent with this profile were reductions in patient-estimated sleepiness on work nights and during the morning commute home. Despite these benefits, patients treated with modafinil continued to have high levels of sleepiness and impaired performance at night.
Although patients receiving 200 mg of modafinil continued to have lapses in performance on the Psychomotor Vigilance Test,26,27,28,29 there were twice as many lapses per night at the final visit in the placebo group as there were in the modafinil group, and the mean duration of these lapses in the modafinil group was nearly twice as long as that in the placebo group. It is likely that the effects of modafinil on sustained-attention performance derive, at least in part, from its effects on reducing the instability of wakefulness caused by brief episodes of sleep intruding into waking performance.15,30,31 Although lapses of attention were reduced, they remained at a high level in the treatment group. This suggests that although modafinil improves the measured levels of performance, it is far from what is needed for these patients to function at a normal level.
The results of this study also suggest that 200 mg of modafinil does not affect circadian adaptation to night-work schedules. Thus, the ability of modafinil to treat symptoms of excessive sleepiness in patients diagnosed with shift-work sleep disorder is a result of an improvement in wakefulness during the nocturnal work shift, similar to the improved alertness shown in other disorders of sleep and wakefulness,8,9,10,11 and not an improvement in the alignment between internal circadian rhythms and the work–sleep schedule.
Several considerations limit the interpretation and applicability of the findings. There remains a need for validated criteria and clinical instruments for assessing excessive sleepiness in shift-work sleep disorder. Although the Multiple Sleep Latency Test is sensitive to changes in sleepiness during nighttime hours32,33 and is recommended for assessing sleepiness at night in this population,5 it has not been specifically validated as a clinical instrument for measuring nighttime sleepiness, particularly in the absence of objectively monitored sleep in the laboratory on the day before testing. As recommended in the literature,17 we therefore used a validated performance measure — the Psychomotor Vigilance Test — to assess alertness at night, the results of which were consistent with the nighttime data from the Multiple Sleep Latency Test. Because patients worked in a variety of industries, actual work performance was not evaluated. We do not know how the laboratory sleep and performance variables that were used in the study may apply to actual on-the-job performance, although we do show concordance of results for measures of alertness, performance on the Psychomotor Vigilance Test, and diary data that collectively suggest a positive effect on personal and public safety. Although the study was open to both permanent and rotating night-shift workers with shift-work sleep disorder, the vast majority of study participants (90 percent) were permanent night-shift workers. Thus, it is not appropriate to generalize the findings of the study to patients who work on other types of shifts that include nighttime hours. The patients who met the criteria of having shift-work sleep disorder are only a subgroup of shift workers,7 a fact that limits the applicability of the findings to the broader shift-work population in whom the safety and efficacy of modafinil have not been evaluated. Our study was 12 weeks in duration; the effects of long-term modafinil use in this population are unknown.
In summary, we found that patients with shift-work sleep disorder had excessive sleepiness during night work, similar to that seen during the day in patients with narcolepsy. Even after treatment with modafinil, these patients still showed evidence of excessive sleepiness during the night shift. Although modafinil did not restore sleepiness to normal daytime levels, treatment was associated with improvements in symptoms of sleepiness, as well as objective measures of sleep propensity and performance. Modafinil is of some value in the clinical management of sleepiness associated with shift-work sleep disorder. Concern remains that even with treatment with 200 mg of modafinil, the excessive sleepiness observed in this underrecognized population requires the development of yet more effective therapies.
Supported by Cephalon, Frazer, Pa.
Drs. Czeisler and Dinges report having received consulting fees and speaker's fees from Cephalon and having received clinical trial research contracts, investigator-initiated research grants, and unrestricted research and education funds from Cephalon. Dr. Czeisler serves as the incumbent of an endowed professorship provided to Harvard University by Cephalon. Drs. Walsh, Roth, and Wright report having received consulting fees and speaker's fees from Cephalon and having received clinical trial research contracts and investigator-initiated research grants from Cephalon. Drs. Hughes, Kingsbury, Arora, and Niebler are employees of and report having an equity interest in Cephalon. Dr. Schwartz reports having received consulting fees and clinical trial research contracts from Cephalon and serving on the speakers bureau for Cephalon.
We are indebted to Mark Riotto, Angela Kaya, and Richard Malone for assistance in the preparation of the manuscript.
Source Information
From the Division of Sleep Medicine, Harvard Medical School, and the Division of Sleep Medicine, Department of Medicine, Brigham and Women's Hospital — both in Boston (C.A.C., K.P.W.); the Sleep Medicine Research Center, St. Luke's Hospital, Chesterfield, Mo. (J.K.W.); the Sleep Disorders and Research Center, Henry Ford Hospital, Detroit (T.R.); Cephalon, Frazer, Pa. (R.JH., L.K., S.A., G.E.N.); the Sleep and Chronobiology Laboratory, Department of Integrative Physiology, University of Colorado, Boulder (K.P.W.); Integris Southwest and Baptist Medical Centers, Oklahoma City (J.R.L.S.); and the Division of Sleep and Chronobiology, Department of Psychiatry and Center for Sleep and Respiratory Neurobiology, University of Pennsylvania School of Medicine, Philadelphia (D.F.D.).
Drs. Walsh, Roth, and Hughes contributed equally to this article.
Address reprint requests to Dr. Czeisler at Brigham and Women's Hospital, Harvard Medical School, 221 Longwood Ave., Suite 438, Boston, MA 02115, or at caczeisler{at}hms.harvard.edu.
References
In addition to the authors, members of the U.S. Modafinil in Shift Work Sleep Disorder Study Group are as follows: J.E. Black, Stanford, Calif.; R.K. Bogan, Columbia, S.C.; M.H. Bonnet, Dayton, Ohio; M.A. Carskadon, Providence, R.I.; J.S. Cook, Danville, Ind.; B.C. Corser, Cincinnati; M.K. Erman, La Jolla, Calif.; N.T. Feldman, St. Petersburg, Fla.; J.M. Ferguson, Salt Lake City; Y. Furman, Los Angeles; S.C. Hardy, Charlotte, N.C.; J.R. Harsh, Hattiesburg, Miss.; M. Hirshkowitz, Houston; S.G. Hull, Overland Park, Kans.; V.K. Mahajan, Toledo, Ohio; G.V. Pegram, Jr., Birmingham, Ala.; J. Pinto, Las Vegas; G.S. Richardson, Detroit; R. Rosenberg, Atlanta; M.H. Rosenthal, San Diego, Calif.; M.H. Schmidt, Dublin, Ohio; P.K. Schweitzer, Chesterfield, Mo.; D. Seiden, Pembroke Pines, Fla.; D.R. Wagner, White Plains, N.Y.; C.C. Wells, Jr., Macon, Ga.; J.K. Wyatt, Chicago; and G.K. Zammit, New York.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
