| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Department of Psychiatry and Psychotherapy, University of Marburg, Marburg, Germany (B.K., J.S., M.T.H., J-C.K.); and Physiological Psychology, University of Bamberg, Bamberg, Germany (S.L.).
Address correspondence and reprint requests to Bernd Kundermann, Department of Psychiatry and Psychotherapy, Philipps University Marburg, Rudolf-Bultmann-Str. 8, Marburg D-35033, Germany. E-mail: kunderma{at}staff.uni-marburg.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
METHODS: Twenty healthy volunteers were randomly assigned either to two nights of total sleep deprivation or to two nights of undisturbed night sleep. Sleep deprivation nights were separated by two days with normal night sleep. Heat and cold pain thresholds as well as warmth and cold detection thresholds were measured by use of a peltier thermode in the evening before and the morning after each deprivation or control night. Pain complaints were examined by use of a questionnaire in parallel.
RESULTS: During treatment nights, sleep deprivation produced a significant overnight decrease in heat pain thresholds. Cold pain thresholds tended to decrease also during sleep deprivation, whereas the warmth and cold detection thresholds remained unaffected. Accordingly, no substantial contributions of the changes in thermal detection thresholds to the changes in thermal pain thresholds were determined by regression analyses. Pain complaints were not induced by sleep deprivation.
CONCLUSIONS: The present findings suggest that sleep deprivation produces hyperalgesic changes that cannot be explained by nonspecific alterations in somatosensory functions.
Key Words: pain perception, • somatosensory sensitivity, • sleep deprivation.
Abbreviations: REM = rapid eye movement;; SEM = standard error of the mean.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
A few studies in humans support the notion that sleep deprivation produces hyperalgesic changes (for review, see Kundermann et al. (6)). The pioneering study on the effect of sleep deprivation on pain was performed by Cooperman et al. (7), showing that total sleep deprivation of 60 hours produced an increased sensitivity to pain, whereas the tactile sensitivity remained unchanged. The scientific value of this study is limited by the fact that the presentation of the data was merely descriptive, and no statistical tests were applied. In considering the possible role of specific sleep stages in the pathogenesis of fibromyalgia, Moldofsky et al. investigated the effect of selective non-rapid eye movement (REM) sleep deprivation over three consecutive nights (8). They observed an increased pressure pain sensitivity as well as more musculoskeletal pain after selective deprivation of stage 4 sleep. In a subsequent study (9), the marked changes found after stage 4 sleep deprivation could not be replicated by use of REM sleep deprivation. In a similarly designed study, Lentz et al. (10) also demonstrated hyperalgesic changes after sleep deprivation. Interestingly, musculoskeletal pain developed not earlier than after the third night of sleep deprivation. This finding suggests that spontaneous pain is less likely to occur before an altered responsiveness to noxious stimuli after a disruption of sleep. Although the finding of a hyperalgesic effect of sleep deprivation was recently replicated by Onen et al. (11), some other studies failed to demonstrate an effect of total (12) or non-REM (13, 14) sleep deprivation on pain.
Taken together, the findings regarding the effects of sleep deprivation on pain have not always been consistent and were also limited by the fact that only two studies (11,13) used a controlled experimental design. Nevertheless, the results suggest that sleep deprivation produces hyperalgesic states. However, it remains yet unclear whether sleep deprivation produces in parallel alterations in responsiveness to other types of stimuli, such as auditory, visual, or nonnoxious somatosensory stimuli. This leads to the question of whether the effects of sleep deprivation on pain are truly specific or due to more general changes in perception. To our knowledge, no study thus far has investigated the effect of sleep deprivation in a comparative manner including perceptual parameters from other sensory modalities. In an attempt to do so, we used quantitative sensory tests (QST) for the responsiveness to noxious and nonnoxious thermal stimuli applied to the skin. The comparisons of heat pain sensitivity to heat sensitivity and of cold pain sensitivity to cold sensitivity, which constitute the truly innovative part of our study, should answer best the question of the specificity of the effects of sleep deprivation because of the perceptual, physical, and methodological vicinity of the sensory modalities studied.
Therefore, the primary aim of the present study was to investigate in a controlled experimental design the effect of (total) sleep deprivation on both thermal pain sensitivity and thermal sensitivity in healthy volunteers. By that, changes in thermal pain sensitivity produced by sleep deprivation can be studied under the perspective of a more generalized alteration in somatosensory sensitivity. In addition, we attempted to determine the role of sleep deprivation in the development of clinical pain complaints.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
Four subjects were excluded because of noncompliance to the sleep deprivation procedure (N = 2), no indication of heat pain below the upper temperature cutoff of 52°C (N = 1), or misunderstanding of instructions (N = 1) during the thermal threshold assessment. The mean age of the remaining 20 participants was 35.8 years (standard error of the mean (SEM) 1.5), 34.9 (N = 10; SEM 2.3) years for the sleep deprivation group and 36.6 (N = 10; SEM 2.0) years for the control group. A t test for independent samples revealed no significant between-group difference in age (t = –0.563; df = 18; p = .580). The gender ratio within the groups (6 males and 4 females in the sleep deprivation group, 5 males and 5 females in the control group) was comparable (
2 = 0.202; df = 1; p = .653).
The protocol was approved by the ethics committee of the Medical School of the Philipps-University Marburg; all subjects gave written informed consent and were paid for participation.
Experimental Design
The subjects were randomly assigned to one of two groups, either to the experimental group with sleep deprivation or to the control group without sleep deprivation. During the week preceding the study period, all subjects were instructed to maintain a regular sleep-wake schedule. Thereafter, the experiment was performed during a period of 5 consecutive days.
Treatment (two nights of either total sleep deprivation in the experimental group or undisturbed night sleep in the control group) and measurements were conducted in the following order: evening testing session (7.00 PM) on day 1 with subsequent first treatment night and a morning testing session (8:00 AM) on day 2. The same experimental protocol as just described was used on day 4 to day 5. Accordingly, treatment nights were separated by intervals of two days with normal (recovery) night sleep.
Treatment
Sleep deprivation was performed according to a standard protocol on a specialized ward of the Department of Psychiatry and Psychotherapy. A staff member monitored the subject and ensured that from 8:00 PM to 7:00 AM, the participant stayed awake and was engaged in standardized activities (including conversation, watching television, going for a walk, and playing games). She/he was told to protocol the subject’s behavior every hour. The subjects received a standard breakfast before the morning testing session at 8.00 AM
The protocol for the control group with undisturbed night sleep started also immediately after the evening testing at 8:00 PM. The subject was welcomed on the same ward and assigned to a single-bed room by a staff member. Lights were turned off between 10:00 PM and 11:00 PM to enable sleep. Each subject of the control group was woken up at 7:00 AM, and a standard breakfast was served before the morning measurements (8:00 AM) started.
Thermal Threshold Assessment
The thresholds for cold sensation (CS), warmth sensation (WS), cold pain (CP), and heat pain (HP) were determined in that order using a Medoc TSA-2001 (Medoc Ltd, Ramat Yishai, Israel) with a contact thermode (stimulation surface of 3.2 x 3.2 cm2) attached to the skin of the center of the volar forearm. Beginning at a baseline temperature of 32.0°C, five stimuli were applied for each of the four thermal thresholds. The rate of temperature change was ±1°C/second for the two detection thresholds and ±1.5°C/second for the two pain thresholds (safety limits at 0°C and 52.0°C), respectively. Subjects were instructed to press a mouse button as soon as they felt a change in temperature (CS and WS) or the onset of pain sensation (CP and HP). Each time they pressed the button, the temperature returned to the baseline temperature, which was held constant until the next trial. The measure of each threshold was the mean of the 5 trials (relative to baseline in the case of the thermal detection thresholds, absolute in the case of the thermal pain thresholds). The retest reliabilities for thermal detection thresholds (CS, WS) as well as for heat pain thresholds were reported to be above an r of 0.7 (16). Thus, the reliability of these measures is sufficient for assessing changes in sensitivity to innocuous and noxious thermal stimuli over time. In comparison to these measures, lower levels of reliability were demonstrated for the measurement of cold pain thresholds (17).
Pain Complaints
Pain complaints regarding the number of painful sites as well as the intensity and unpleasantness of pain at these sites were assessed by means of a pain questionnaire, which was designed by the senior author (S.L.) and used in former studies (18). For the purpose of the present study, the instruction of the questionnaire was slightly adapted to evaluate current instead of chronic pain complaints. According to Jensen and Karoly (19), self-rating scales of pain have shown adequate reliability and validity, especially in the case of healthy subjects.
Sleep Quality
In order to control for sufficient night sleep in the control condition, a sleep questionnaire was administered on both morning testing sessions to evaluate sleep quality during the preceding night. This questionnaire was designed by Hemmeter et al. (20) and consists of questions concerning sleep duration, sleep latency and frequency, and duration of awakenings. Furthermore, the questionnaire evaluates several subjective aspects of night sleep (calmness, depth, restfulness) on a bipolar five-point Likert scale (eg, depth of sleep: +2 = "very deep"; +1 = "deep"; 0 = "balanced"; –1 = "superficial"; –2 = "very superficial").
Statistical Analysis
Data were statistically analyzed using SPSS version 11.0 for Windows. Results are presented as mean and SEM. Changes in thermal thresholds and pain complaints between the groups and over time were analyzed with an analysis of variance (ANOVA) for repeated measurements with one between-subject factor (treatment) and two within-subject factors, which evaluate short-term (from evening to morning) and long-term (from first treatment to second treatment night) effects. A verification of the hypothesis that sleep deprivation leads to a hyperalgesic state would be indicated by a significant interaction between treatment and the short-term time effect. If significant effects in ANOVA were observed, post hoc tests (two-tailed) were used for further analysis. Because for post hoc analysis of very small samples (post hoc within comparisons include not more than 10 subjects in our study) nonparametric tests are recommended, we applied Wilcoxon matched pairs signed-rank tests (exact test procedure) for that purpose. In addition, effect sizes were computed to determine the relative magnitude of a statistical significant treatment effect. According to Cohen’s (21) guidelines, effect sizes were calculated by dividing the mean difference between pretreatment and posttreatment scores by the standard deviation of the difference scores. In order to determine whether changes in thermal detection thresholds account for the changes observed in thermal pain thresholds, the measures of thermal detection thresholds were used as covariates. For this purpose, linear regression analyses for each of the four time points of measurement were conducted with cold detection threshold as the independent variable to predict cold pain threshold and warmth detection threshold as the independent variable to predict heat pain threshold. These scores, which were now statistically freed from the influence of somatosensory thresholds, were finally entered in an ANOVA with the same factorial design as for the raw values. This approach was preferred to a regular covariance analysis because it allowed entering the thermal detection thresholds four times as covariates (instead of only once) and to control, by that, for the statistical contribution of the thermal detection threshold to the thermal pain thresholds at each individual time point of measurement. A substantial change in the results of the ANOVAs for the raw and the corrected values would indicate a contribution of the covariates. The significance level was set to 
0.05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
|
Figure 2 shows the effect of sleep deprivation on thermal detection thresholds. For the cold detection thresholds, ANOVA with repeated measurements revealed no significant main effects or interactions but a tendency of increasing threshold values during each night (short-term: F(1,18) = 4.374; p = .051). For the warmth detection threshold, a significant interaction was detected between the two within-subject factors (short-term x long-term: F(1,18) = 8.003; p = .011). In order to specify this interaction, paired Wilcoxon tests were performed by using pooled data from both experimental groups to examine differences between the thresholds in the evening and the morning. The results indicate an insignificant decrease in warmth detection thresholds during the first night (T = 52.5; n = 20; p = .278) and a tendency to an increase during the second night (T = 42; n = 20; p = .052).
|
|
The data obtained from the sleep questionnaire revealed that the sleep quality of both control nights was sufficient. The mean duration of sleep was 7.1 hour (SEM 0.3) during the first night and 7.2 hours (SEM 0.3) during the second night. Two subjects reported at least three awakenings during the first night, whereas all (N = 10) subjects experienced a frequency <3 of awakenings during the second night. The average sleep latency was 24.2 minutes (SEM 7.8) on night 1 and 17.7 minutes (SEM 5.3) on night 2, respectively. The results of the Likert scale ratings (concerning calmness, depth, and restfulness of sleep) showed that not more than 2 subjects per night expressed estimation below the neutral category.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
Our finding of hyperalgesic changes produced by sleep deprivation is in accordance with previous human studies in which sleep deprivation in its total or selective form decreases pain thresholds (7–11). Furthermore, the results of several animal studies also support the view that sleep deprivation enhances nociception. Although these animal studies are limited by the exclusive focus on REM sleep deprivation, they demonstrated hyperalgesic effects of sleep deprivation when using electrical (22, 23), mechanical (24–27), and thermal (27) stimuli for triggering nociception. Interestingly, up to now, human studies have demonstrated hyperalgesic effects of sleep deprivation only in mechanical (pressure) pain paradigms (7–11). In contrast, the only two studies that made use of thermal pain (11,12) failed to reveal a hyperalgesic effect of sleep deprivation. The relatively small sample size of these two studies may be a reason for the negative finding. To our knowledge, this is the first human study that demonstrates a hyperalgesic action of sleep deprivation on thermal pain thresholds.
This action was short-term and abolished after two nights of recovery sleep. Besides its significant effect on heat pain thresholds, sleep deprivation also tended to decrease cold pain thresholds. The failure of finding a significant effect for cold pain thresholds was probably due to the lowered statistical power (caused by missing data). Another reason for the only tentative finding in the case of cold pain may be a lower reliability of this measure in comparison to heat pain (17).
Although several human and animal studies showed hyperalgesic effects produced by sleep deprivation, no study has addressed so far the question of whether these effects are truly specific to pain or part of a more general change of somatosensory sensitivity. Our results indicated that detection thresholds for (innocuous) warmth and cold were not affected by sleep deprivation. Furthermore, regression analysis revealed no significant correlations between thermal detection thresholds and thermal pain thresholds. This lack of a substantial relationship is in accordance with a previous study demonstrating moderate and insignificant correlations between thermal somatosensory thresholds and pain thresholds in healthy volunteers (28). Taking into statistical consideration the thermal detection thresholds as covariates, the effect of sleep deprivation on both pain thresholds remained nearly unchanged. Accordingly, we can exclude that sleep deprivation causes an unspecific alteration in the perception of thermal stimuli, which underlies the observed overnight decreases in thermal pain thresholds.
The present study failed to identify an effect of sleep deprivation on the development of pain complaints. This finding seems to contrast with observations of earlier studies, in which an increase of musculoskeletal symptoms (as measured by self-rating questionnaires) after stage 4 sleep deprivation was found (8). The discrepancy to our results is probably caused by differences in the study design, especially in conditions of sleep deprivation. Because we investigated the effect of two nights of sleep deprivation, which were separated by recovery sleep, and not of consecutive nights of sleep deprivation, one might speculate that the observed increase of musculoskeletal symptoms is due to cumulative effects. This idea is also supported by the finding that selective (stage 4) deprivation of three consecutive nights induced a significant increase in musculoskeletal discomfort (compared with baseline) only after the third night of sleep deprivation (10). Thus, it may be that the development of clinical pain complaints depends on a longer lasting disruption of sleep, whereas the responsiveness to noxious stimuli is already vulnerable to short-term manipulations of sleep. A second potential explanation for the finding, that we did not observe the development of clinical pain complaints as others, is the use of our pain questionnaire. Our questionnaire was exclusively designed for the assessment of pain symptoms and therefore does not include –as in the other studies (8,10) –the self-report of other somatic symptoms or unpleasant sensations (eg, stiffness, tight muscles). Sleep deprivation, however, is known to exert broad and nonspecific effects on subjective states like sleepiness, increased fatigue, negative mood, and cognitive alterations.
The underlying neurobiological mechanisms by which sleep deprivation decreases thermal pain thresholds remain unclear. This is –at least in part –due to the fact that most of the human studies dealing with the effect of sleep deprivation on pain were conducted without the assessment of neurobiological covariates or the use of psychopharmacological agonist/antagonist strategies. There are a few animal studies that have focused on the activity of the opioid and monoaminergic systems and their association with the changes in nociception produced by sleep deprivation. The study of Ukponmwan et al. (24) revealed that the analgesic action of endogenous and exogenous opioids is dependent on an undisturbed sleep architecture/continuity, whereas on the other hand, selective REM sleep deprivation prevents opioid analgesia. Furthermore, sleep deprivation has been shown to affect the serotoninergic system (29,30), which also plays a key role in the descending pain inhibitory control system (31). These findings suggest that sleep deprivation produces a transient disturbance of the descending pain inhibitory control system.
Methodological limitations of the present study mainly concern the control group, which lacked an adaptation night and a polysomnographic control of the nocturnal night sleep. However, the subjective ratings of sleep quality led us to believe that our experimental control was successful because the members of the control group reported sufficient night sleep during the two control nights. Although subjective measures of sleep quality are not as reliable as objective polysomnographic recordings, at least a few subjective sleep variables (sleep latency, number of prolonged awakenings, restfulness of sleep) have been reported to correspond with recorded objective sleep characteristics (32).
In summary, the present study has demonstrated hyperalgesic effects of sleep deprivation by use of thermal pain thresholds without an alteration of thermal detection thresholds, ie, without alteration of somatosensation in general. Because this suggests an effect truly specific for pain, research may now focus on its underlying mechanisms of action.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
Received for publication January 20, 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  B. Kundermann, J. Hemmeter-Spernal, M. T. Huber, J.-C. Krieg, and S. Lautenbacher Effects of Total Sleep Deprivation in Major Depression: Overnight Improvement of Mood is Accompanied by Increased Pain Sensitivity and Augmented Pain Complaints Psychosom Med, January 1, 2008; 70(1): 92 - 101. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
