Chronic fatigue syndrome (CFS) is a medically unexplained condition characterized by persistent or relapsing fatigue lasting at least 6 months, which substantially interferes with normal activities. CFS is primarily a problem in women's health in that it occurs in women nearly twice as often as in men (5). A disabling and characteristic feature of CFS is that even minimal exertion produces a dramatic worsening of symptoms (6). In an earlier work, we showed that activity levels fell several days after patients performed a standard cardiac-type stress test (13). We recently replicated and extended this finding using real-time assessment techniques and demonstrated that CFS symptoms do worsen several days after maximal exercise but that neither mood nor cognitive function was effected (17). Thus, it is unclear to what extent exercise influences the symptom complex of CFS. Because exercise training is recognized as an important and efficacious treatment option for many patients (1), it is important to understand the degree to which acute exercise influences other debilitating aspects of CFS.
One aspect of CFS and acute exercise that has not been systematically examined is its influence on sleep. We have reported that for several days after exercise, the circadian activity rhythm of the patients but not of the controls lengthened significantly, suggesting that the exercise had weakened entrainment to the usual 24-h zeitgeber (10). We hypothesized that alterations in sleep duration and/or quality might have produced these exercise-related changes. We have recently shown that CFS patients as a group have abnormalities in sleep morphology indicative of sleep disruption (14). Because CFS is identified using a clinical case definition, we reasoned that patients' sleep data would be heterogeneous with some patients sleeping normally and others having disrupted sleep. In our earlier work, we found that we could substantially reduce this heterogeneity by stratifying the patients on the basis of whether they reported increased or decreased sleepiness after a night in the sleep laboratory. Those who reported being sleepier after a night of sleep showed evidence for disrupted sleep, whereas those who reported some reduction in sleepiness after a night's sleep had relatively normal sleep morphology.
In healthy people, sleep duration increases after exercise (18). We hypothesized that the reverse would occur in CFS patients-especially in those reporting increased sleepiness after a night of sleep. To our knowledge, no one has evaluated sleep morphology in CFS patients before and after exercise. The purpose of this study was to fill this gap using a relatively homogeneous CFS patient sample, namely, women without comorbid major depressive disorder and studied while they were in the same menstrual phase, to determine the influence of an acute bout of exercise on polysomnographic (PSG) and self-reported measures of sleep.
The subjects were women, 17 with CFS and 16 healthy controls, ranging in age from 25 to 50 yr. Subjects with CFS were either physician-referred or self-referred in response to media reports about our research. Healthy controls were acquaintances of patients or responded to recruitment flyers. All of these subjects had been screened for sleep disorders by a previous night of diagnostic polysomnography (14), and all were negative. Patients fulfilled the 1994 case definition for CFS (4) and thus had no medical explanation for their symptoms on the basis of history, physical examination or laboratory tests, and no serious psychiatric diagnoses including schizophrenia, eating disorders, and substance abuse. Psychiatric diagnosis according to the criteria of the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, was made using the computerized version of the Diagnostic Interview Schedule (12a). This diagnostic interview allowed us to identify subjects with major depressive disorder (n = 5) who were then not further studied because depression itself is known to relate with sleep morphology (3). The existence of comorbid fibromyalgia was diagnosed on the basis of the American College of Rheumatology's 1990 criteria (16). All subjects provided informed consent, approved by the medical school's institutional review board to participate in this research.
After instructions to refrain from alcohol and caffeine ingestion and to avoid engaging in prolonged and/or strenuous exercise in the daytime of study nights, subjects underwent PSG recording in a quiet, shaded hospital room for an initial habituation night and were excluded if any sleep pathology was identified. Patients enrolled in the study subsequently underwent two additional nighttime PSG studies for the results reported here. Subjects went to bed at their usual bedtime and slept until 7:15 to 8:00 a.m. the next morning. Exercise was performed in the afternoon before the last study night. Subjects were all studied during the follicular phase of their menstrual cycles.
Within 6 months of their habituation session in the sleep laboratory, subjects returned to the sleep laboratory for two more nights of study during which they were instrumented to allow recording of electroencephalogram (EEG; C3/A2, O1/A2, and FZ/A2), electrooculogram, submental EMG, and a lead II ECG. The first of these nights was used as a control for the final night before which subjects performed the maximal exercise test detailed below. On these two nights, subjects had indwelling venous catheters from which blood was sampled remotely without disturbing the subject three times during the course of the night. Sleep was scored by a single scorer according to the standard criteria of Rechtstchaffen and Kales (12) every 30 s. Sleep onset was defined as the first three consecutive epochs of sleep stage 1 or the first epoch of other stages of sleep. An arousal was defined according to standard criteria of the American Academy of Sleep Medicine (2) as a return to alpha or fast-frequency EEG activity, well differentiated from the background, lasting at least 3 s but no more than 15 s.
A maximal exercise test was performed on an electronically braked cycle ergometer (Lode Corival, Groningen, the Netherlands) during the late afternoon. The seat height and toe clips were adjusted to the desired fit of the subject. The exercise test began with a 3-min unloaded warm-up. After warm-up, exercise began at 20 W. Exercise intensity was then increased by 5 W every 20 s until volitional exhaustion or the point where the subject could no longer maintain the prescribed minimum pedal rate (60 rpm). During the exercise test, measurements of oxygen consumption (V˙O2), carbon dioxide production (V˙CO2), and expired ventilatory volume (V˙E) were obtained breath-by-breath using a MedGraphics Gas Exchange and Pulmonary Function System (Medical Graphics Corporation, St. Paul, MN). HR was monitored during exercise by ECG using a Quinton Q4000 (Quinton Instruments, Seattle, WA). Because of technical difficulties, HR data were not acquired on three CFS participants and two control participants. Participants were encouraged during the test to continue as long as possible and to give their best possible effort. Acceptable effort was determined as achieving 80% of age-predicted maximum HR and/or a RER ≥ 1.1. All subjects were able to achieve at least one of the two criteria, and the majority (13/17 CFS and 13/16 controls) met both criteria. CFS and control groups did not differ significantly in percent peak HR (CFS = 0.87 ± 0.06, controls = 0.91 ± 0.07), peak RER (RER; CFS = 1.26 ± 0.17, controls = 1.30 ± 0.09), or peak V˙O2 (CFS = 20.1 ± 5.4 mL·kg−1·min−1, controls = 24.5 ± 5.1 mL·kg−1·min−1).
Sleep continuity was evaluated by generating a nonparametric survival curve calculated from the combined data within each group (11,15) of the varying durations of sequential sleep runs (i.e., continuous epochs of sleep separated from one another by epochs of wakefulness) and was expressed as the median duration of all continuous epochs scored as sleep in each subject. A run of sleep was defined using the sequence of epoch-based sleep stages represented in the hypnogram. A run began with a change from wake to any stage of sleep. A sleep run continued until there was a change from any stages of sleep to wakefulness. To compare sleep continuity between groups, all data from all subjects in each group were pooled, and a group survival curve was generated using standard statistical techniques, which take into account the multiple runs of sleep in each subject (11,15); this method was derived from an earlier one (9). We expected to replicate our earlier finding of sleep disruption in the subgroup of CFS patients reporting increased sleepiness after a night's sleep.
A visual analog scale (0-15.5 cm) was used to estimate perceived sleepiness, fatigue, pain, and feeling blue before and after each PSG recording. Visual analog scales have consistently been shown to provide valid measures of subjective feelings (8,14).
The Centers for Epidemiological Study - Depression scale was used as an indicator of depressed mood. This 20-item scale required respondents to rate how often certain symptoms occurred during the past week on a scale from rarely or none (score = 0) to most of the time (score = 3). Items were summed to yield a total score. High values indicated more depressed mood.
We dichotomized data on the basis of subjects' self-reported sleepiness before and after the baseline PSG night. We labeled those with more sleepiness in the morning than on the night before as "AM sleepier" and those with less sleepiness in the morning than on the night before as "AM less sleepy." Using these criteria, we were able to divide healthy controls and CFS patients into four groups with almost equal numbers of subjects. Changes of sleepiness before and after sleep as well as changes in the other variables captured via the visual analog scale were assessed using paired t-test. Differences in measured variables between groups were assessed using nonpaired t-test or ANOVA. Post hoc analyses used Tukey Student range tests. Changes in duration of median sleep run between baseline and postexercise nights for patients with CFS in the AM sleepier and AM less sleepy groups were assessed using paired t-test. Statistical significance was accepted when P < 0.05.
Table 1 depicts sleep structures for healthy controls and CFS patients on their baseline and postexercise nights in the sleep laboratory. Baseline total sleep time was similar for controls and patients, but patients had poorer quality sleep than controls as manifested by their having longer durations of awake plus stage 1 (P = 0.049) as well as reduced rapid eye movement (REM) sleep (P = 0.013). Exercise improved sleep compared with the baseline night for both patients and controls: The CFS group showed decreased number of arousals (P = 0.02) and shorter duration of stage 1 sleep (P = 0.005), but sleep efficiency remained low and duration of awake plus stage 1 sleep remained longer than in controls on the postexercise night. Healthy subjects showed a shortened latency to fall asleep after exercise (P = 0.025), whereas the durations of their individual sleep stages did not change; nonetheless, their sleep efficiency did improve enough to be significantly higher than that of patients (P = 0.016).
Subjective sleepiness, fatigue, and pain before and after the baseline and postexercise nights were significantly (P < 0.01) higher in patients than healthy controls as would be expected (Table 1, bottom). On the morning after exercise, healthy controls reported significantly less sleepiness (P = 0.008) and fatigue (P = 0.048) than after the postexercise night. In contrast, on that morning, patients reported no change in their sleepiness and an actual small but significant increase in fatigue than on their baseline night (P = 0.014).
Because we had found substantial differences in sleep morphology after stratifying the data on the basis of changes of sleepiness before and after the baseline night in our earlier work (8,14), we repeated the analysis using similarly stratified data. The number of subjects and range in changes of sleepiness for healthy subjects in the AM less sleepy and AM sleepier groups were 9 and −9.2 to −2.3 and 7 and 0.0 to 11.3, respectively, and those for CFS patients in the AM less sleepy and AM sleepier groups were 8 and −11.3 to −0.7 and 9 and 0.0 to 14.1, respectively. Probably because of our having a much smaller sample size here, we found little difference between the sleep of either patient group and that of either control group (Table 2). Stratifying by changes of sleepiness did provide further information as to the effects of exercise on sleep, however.
Healthy subjects in the AM sleepier group showed the most dramatic change: Their sleep efficiency increased significantly (P = 0.033) from that determined on the baseline night and averaged even higher than for controls in the AM less sleepy group. Commensurate with this improvement, awake time (with and without stage 1) decreased (P = 0.49 and P = 0.48, respectively) and REM sleep increased (P = 0.49) relative to values on the baseline night.
Whereas healthy subjects in the AM less sleepy group showed only a significant decrease in sleep latency (P = 0.032) from their baseline night, patients in the AM less sleepy group showed further improvement in their sleep structure on the night after exercise. Specifically, total sleep time (P = 0.011) and REM sleep increased (P = 0.048) relative to the baseline night; in addition, sleep efficiency went up. Duration of the median sleep run (14 ± 7 min) increased (P = 0.018) relative to the baseline night (8 ± 5 min). In contrast, although patients in the AM sleepier group did show fewer arousals (P = 0.007) and less time in stage 1 sleep (P = 0.006), their sleep efficiency remained low and was in fact significantly lower than any of the other groups studied (P < 0.05). Duration of the median sleep run (6 ± 4 min) after exercise did not change from that on the baseline night (6 ± 2 min).
The survival curve of all sleep runs depicted in Figure 1 shows that patients in the AM less sleepy group had a higher percentage of long runs of sleep on the post exercise night compared with their baseline night (i.e., less continuous sleep). For example, the proportion of runs lasting more than 10 min was 45.2% and 54.5% on the baseline and postexercise nights, respectively. This directional change was not seen in AM sleepier patients.
When the entire data set was dichotomized on the basis of the CFS group's median sleep efficiency on the baseline night (i.e., 80%), four healthy subjects had poor sleep efficiency too, and this group showed improved sleep efficiencies (P = 0.049) after exercise, whereas CFS patients in the group with the poor sleep efficiency (<80%) did not.
Both groups of CFS patients had higher depression scores than both groups of controls on both PSG nights, but means did not exceed the cutoff for mild depression in otherwise well people. As expected, fatigue and pain scores before and after both study nights were significantly higher in both groups of patients than in both groups of healthy controls (P < 0.01). Patients in the AM less sleepy group had increased fatigue (P = 0.039) on the morning after exercise relative to their baseline morning. Healthy women in the AM sleepier group reported more sleepiness (P = 0.043) on the night after exercise than on the baseline night; no other groups showed this increase in sleepiness. In contrast, patients in the AM sleepier group showed the converse: reduced sleepiness on the night after exercise compared with those in the AM less sleepy group but the highest sleepiness scores (P < 0.05) of all the groups on the morning after exercise. Both groups of controls as well as patients in the AM less sleepy group reported reduced sleepiness and reduced fatigue (P < 0.05) on the morning after exercise compared with the evening before, but this was not the case for those in the AM sleepier group.
As we showed in our previous article using data from the initial habituation and diagnostic night, these subjects spent in the sleep laboratory (14), CFS patients as a group have disrupted sleep characterized by significantly poorer quality sleep than controls. However, in contrast to our expectation, the patients as a group showed evidence of improved sleep after exercise.
The results were clearer after we used the same stratification strategy that we had used in our earlier work, that is, splitting subjects into those who were either sleepier or less sleepy after a night's sleep. Here, as expected, exercise improved the sleep quality of healthy controls who had reported increased morning sleepiness after the baseline sleep night. Contrary to expectation, it had the same result in CFS patients with decreased morning sleepiness. However, patients who reported increased morning sleepiness showed no improvement in sleep disruption, but exercise did not exacerbate their sleep pathology. These patients also had the lowest average sleep efficiency of any of the groups studied.
Because exercise did not produce a significant worsening of sleep morphology in CFS, the complaints of symptom worsening, which are reported to occur the next day after exertion, cannot be explained by disruption in sleep. After exercise, approximately half the patients actually sleep better than on their baseline study night, whereas the rest simply did not improve.
Data from this report and our previous report indicate that a subset of patients with CFS have an underlying sleep disorder characterized by sleep disruption leading to the patients' feeling sleepier after sleep than before it. Our success in being able to stratify patients on the basis of self-report data suggests that collecting data on sleepiness before and after a night of sleep is a useful way of stratifying patients without the need for formal polysomnography. The fact that patients showed no worsening in sleep architecture after performing a maximal exercise test is useful information for patient education. Patients are often afraid to exert themselves because of the fear of symptom worsening after such exertion. Although symptom exacerbation may in fact occur, it cannot be due to further disturbances in sleep in that at least half the patients studied here showed actual improvement in their sleep after a rather strenuous exercise. And importantly, the remaining patients showed no further deterioration in their sleep patterns. Although the healthy controls demonstrated greater improvements in sleep, CFS patients demonstrated improvements in time spent awake and arousals during stage 1 sleep that are similar to that generally seen in healthy good sleepers (18). These data strongly suggest that gentle physical conditioning, which has been suggested as a treatment of CFS, will also not disrupt sleep and may in fact lead to symptom improvement. This suggestion is consistent with meta-analytic data demonstrating that exercise does not need to be strenuous to improve sleep. Youngstedt (18) reported similar effect sizes on positive sleep outcomes for light, moderate, and vigorous exercise in healthy good sleepers.
Unfortunately, there is a paucity of research aimed at determining the influence of exercise on sleep in populations with disturbed sleep. Limited evidence suggests that the sleep-promoting influences of exercise are more pronounced when sleep is disrupted such as in elderly populations, patients with depression, or patients with restless leg syndrome (18). Further, exercise has been shown to reduce feelings of depression and anxiety in patients with major depressive and anxiety disorders (7) as well as in CFS patients (1). However, these studies have relied almost exclusively on self-report measures of sleep quality. The present investigation extends this research by objectively assessing sleep morphology in CFS and demonstrating both the lack of deleterious effects of exercise in patients with the most disturbed sleep and the benefit of sleep-promoting effects of exercise in patients with decreased morning sleepiness.
In conclusion, exercise improved sleep structure for healthy volunteers as well as for CFS patients reporting less sleepiness after a night's sleep than before it. In contrast, sleep structure did not change for patients reporting more sleepiness after a night's sleep than before it. In general, these patients were the ones with the lowest sleep efficiencies on baseline sleep polysomnography. The finding that exercise did not worsen sleep leads to two conclusions: first, that any ill effects of exercise are not due to altered sleep; and second, that exercise does not have a deleterious effect on sleep and in fact helps it in some CFS patients. This latter finding should prove important in helping patients deal with increasing their activity without worrying about negative health consequences.
The work reported here was supported by National Institutes of Health no. AI-54478.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Afari N, Buchwald D. Chronic fatigue syndrome: a review. Am J Psychiatry
2. American Sleep Disorders Association. EEG arousals: scoring rules and examples. Sleep
3. Ford DE, Cooper-Patrick L. Sleep disturbances and mood disorders: an epidemiologic perspective. Depress Anxiety
4. Fukuda K, Straus SE, Hickie I, Sharpe MC, Dobbins JG, Komaroff A. The chronic fatigue syndrome: a comprehensive approach to its definition and study. International Chronic Fatigue Syndrome Study Group. Ann Intern Med
5. Jason LA, Richman JA, Rademaker AW, et al. A community-based study of chronic fatigue syndrome. Arch Intern Med
6. King C, Jason LA. Improving the diagnostic criteria and procedures for chronic fatigue syndrome. Biol Psychol
7. Martinsen EW. Physical activity in the prevention and treatment of anxiety and depression. Nord J Psychiatry
. 2008;62(suppl 47):25-9.
8. McCormack HM, Horne DJ, Sheather S. Clinical
applications of visual analogue scales: a critical review. Psychol Med
9. Norman RG, Scott MA, Ayappa I, Walsleben JA, Rapoport DM. Sleep continuity measured by survival curve analysis. Sleep
10. Ohashi K, Yamamoto Y, Natelson BH. Activity rhythm degrades after strenuous exercise in chronic fatigue syndrome. Physiol Behav
11. Pena EA, Strawderman R, Hollander M. Nonparametric estimation with recurrent event data. J Am Stat Assoc
12. Rechtstchaffen A, Kales A. A Manual of Standardized Terminology, Techniques and Scoring System for Sleep States of Human Subjects
. Washington (DC): US Government Printing Office, National Institutes of Health Publication; 1968. 1-62 p.
12a. Robbins LN, Cottler LB, Bucholz KK, Compton WM, North CS, Rourke KM. Diagnostic interview schedule for the DSM-IV (DISIV)
. St. Louis, Washington University School of Medicine; 2000.
13. Sisto SA, Tapp WN, LaManca JJ, et al. Physical activity before and after exercise in women with chronic fatigue syndrome. Q J Med
14. Togo F, Natelson BH, Cherniack NS, Fitzgibbons J, Garcon C, Rapoport DM. Sleep structure and sleepiness in chronic fatigue syndrome with or without coexisting fibromyalgia. Arthritis Res Ther
15. Wang M-C, Chang S-H. Nonparametric estimation of a recurrent survival function. J Am Stat Assoc
16. Wolfe F, Smythe HA, Yunus MB, et al. The American College of Rheumatology 1990 criteria for the classification of fibromyalgia. Report of the Multicenter Criteria Committee. Arthritis Rheum
17. Yoshiuchi K, Cook DB, Ohashi K, et al. A real-time assessment of the effect of exercise in chronic fatigue syndrome. Physiol Behav
18. Youngstedt SD. Effects of exercise on sleep. Clin Sports Med