Introduction 1 2 3 4 Patients and methods Thirteen outpatients (5 males) participated in a randomized double-blind crossover placebo controlled study. Their mean age was 43.5 years (SD: 13.9 years). Age of onset of symptoms was before 12 years of age in 3 patients; all three lived independently at the time of study, although they needed professional social support with respect to their household or daily activities. Three patients were employed in highly responsible jobs. The remaining patients had been considered unfit for normal paid employment for reasons related to their disorder. They were mainly involved in housekeeping. With the exception of two elderly men with considerable weakness of the legs, weakness had little impact on activities of daily living (ADL) functions. Even these two men were ambulant, although they used a wheelchair regularly. Twelve patients had a partner or housemate. Medication was given during two periods of 14 days, separated by a one-week washout period. The study was preceded by a two-week period free of all drugs with an exception being made for contraception. Patients were randomized for either placebo first or modafinil first. The modafinil dose was 200 mg per day for the first week. The patients were instructed to double the dose during the second week of each period if they perceived an insufficient effect. novel have you, during the past two weeks, been more active than during the baseline period? no (0 points) to some degree (1 point) definitely (2 points) can you give one (1 point), two (2 points) or three (3 points) substantial and observable examples of activities/specific actions you undertook that you would otherwise not have done? 5 6 After completion of the trial the remaining capsules in the medication boxes were counted to assess compliance. The institutional Committee of Medical Ethics had approved the study. Patients gave written informed consent after study information was provided orally and in writing at the patients’ home. Results All patients completed the trial. Medication compliance was good: only three patients had omitted one dose each. The only reported side effect was slight headache in one patient using modafinil. Ten patients doubled the dose of both modafinil and placebo after the first week, meaning that results largely concern a daily dose of 400 mg modafinil. More often than not both patients (67%) and their partners/housemates (77%) correctly guessed when they had been taking either modafinil or Placebo, usually on the basis of a ‘decreased sleepiness’ and/or ‘increased activity’. The structured interview regarding activity and actions did not show significant differences between modafinil and placebo (p = 0.2 for patients and p = 0.5 for partners/housemates). The RAND-36 questionnaires revealed a poor perception of general health for the whole group with a mean value of 29 points out of 100 (range 0–50) on the General Health rating. The ratings were virtually identical for each patient over the four assessments (p = 1,Wilcoxon test). The perception of Role Limitations varied widely: mean 66 out of 100 (range 0–100). A medication related change was not observed (p = 0.7, Wilcoxon test). This also held for the perception of Social Functioning (p = 0.6), Vitality (p = 0.2) and Mental Health (p = 0.5). The ESS revealed a significant improvement with modafinil, in that the mean score decreased from 10.5 (range: 3–18) to 6.8 points (range: 1–15); for placebo the corresponding values were 10.5 (range: 3–18) and 10.7 (range: 2–17) points (p = 0.015, Wilcoxon test). There was no suggestion of a difference in outcome between patients with high and those with low scores. There was no significant relationship between the increase in activity/actions as perceived by the patient/ partner and indicated by the structured interview, and changes in perceived hypersomnolence as measured by the ESS (p = 0.38, Spearman’s test). Discussion The present study confirmed the beneficial effect of modafinil on excessive sleepiness in MD, but did not detect a concomitant effect on spontaneous activity as measured by a structured interview of the patients and their partners. This interview, not formally validated, was designed to reflect a clinically relevant and observable increase in daily activity by asking for specific actions. Examples might be that patients went to the theatre after a busy day, when they would otherwise have postponed such a visit, or cleaning up the shed. By asking for specific actions we hoped to distinguish actions from the mere feeling of being active or the intention to become so. The study was small, leaving open the possibility that minor changes have been missed. The study was also focused on short-term effects and thus it is not able to detect changes of behaviour that take more time to become manifest, but we believe that a fortnight is long enough to detect relevant improvements in activity as defined above. A further consideration is the unblinding we have observed, which was most probably due to a correct perception of an effect on somnolence. This might have confounding effects on the interpretation of intended double-blind studies of modafinil on symptoms other than hypersomnolence. In the present study this does not seem to have happened, as the effect on hypersomnolence was neither related to perceived improved activity, nor to perceived aspects of general health. That many patients and partners reported more activity in addition to less sleepiness when asked why they thought that modafinil or placebo had been used, might be the result of the expectations implied in the aim of the study as discussed with the participants. The structured interview did not detect this increased activity, which we feel speaks in favour of its validity. 2 3 4 7