Abstract

Long-term exposure to stress is usually seen as a major cause of insomnia. Cross-sectional studies consistently support this view and demonstrate close relations between reported stress and impaired sleep. Prospective studies are rare, however, but the few that are available indicate that new cases of insomnia follow increases in prior stress at work. With respect to polysomnography, minor effects, such as a slightly increased sleep latency or sleep efficiency, follow periods of increased stress. The level of stress in these studies is usually rather modest, however. One key factor seems to be worries at bedtime, often related to upcoming difficulties during the subsequent day: anticipatory stress seems to be important and often involves bedtime rumination. One interesting link between sleep loss and stress is their similarity of endocrine/metabolic effects. Both involve increased levels of cortisol, lipids, and insulin resistance, for example. Furthermore, typical “stress diseases” such as cardiovascular disease or type 2 diabetes are linked also to prior sleep disturbance. Burnout, a characteristic outcome of long-time exposure to stress involves extreme fatigue, cognitive impairment, and lowered mood. Patients with burnout also show pronounced sleep fragmentation, increased sleep latency, reduced sleep efficiency, and reduced slow-wave sleep. Sleep impairment seems to be a prerequisite for developing burnout. It is concluded that there is supporting evidence that stress affects sleep negatively but that we know too little about the details of the effects. The details include the critical levels and durations of stress necessary to cause insomnia and the process of the development of stress-related insomnia across time.

Stress research often refers to the observation that repeated stress with insufficient restitution in between episodes of stress will lead to a gradual increase in physiologic activation and eventually an allostatic upregulation of wear and tear.¹ Although the duration or type of rest in-between bouts of stress may be important factors, it is obvious that sleep is one of the major physiologic means of restitution.² This involves the elimination of subjective or behavioral fatigue, the physiologic substrates of this elimination, and the peripheral physiologic changes that maintain proper metabolic balance and a well-functioning immune system. The amount of available data on stress and sleep is relatively modest, however.³^,⁴ Some of the findings are summarized here.

From a general perspective, psychosocial stress refers to the “the rate of wear and tear in the organism,” and the biological definition of stress refers to the nonspecific response to any demand ⁵ to increase the chances of successful handling of a threatening situation. Contemporary physiologic stress models derive from the pioneering work of Cannon⁶ and of Selye.⁵ Cannon⁶ developed the concept of the “fight–flight” response, which linked the emotional perception of a “threat” to physiologic changes in the periphery. Selye⁷ proposed a model of stress, the general adaptation syndrome, which was composed of three stages—alarm, resistance, and exhaustion—and reflected the physiologic nonspecific response to a challenge. The resistance stage of the general adaptation syndrome has extreme energy requirements, which, if persistent over time, deplete the person’s capacity and leads to exhaustion.

Markers of the fight–flight response are the catecholamines epinephrine and norepinephrine and other physiologic indicators associated with the autonomic nervous system.⁸^,⁹ The hypothalamic-pituitary-adrenocortical (HPA) axis is also fundamental in the stress reaction. When the sympathetic-adrenal-medullary system is activated and neuropeptides such as corticotropin-releasing hormone (CRH) and vasopressin are released, they, in turn, stimulate release of adrenocorticotropic hormone (ACTH) into the general blood circulation within the pituitary.⁹^,¹⁰

Long-term effects of stress are described by the term allostasis, which refers to the ability of the body to increase or decrease the activation level of vital functions to new steady states dependent on the characteristics of the challenge and the person’s emotions and appraisal of the event.¹ The resulting allostatic load represents the cumulative cost to the body when the systems start to malfunction after a stressful event. It is suggested that serious pathophysiology can occur if overload is not relieved in some way.¹ One of the outcomes may be insomnia.

With respect to working life there are systematic theories, however. One of the leading ones in work stress research is the so-called demand/control model.¹¹^,¹² According to this model, high demand and low decision latitude have been found predictive of cardiovascular and other types of stress diseases.¹³^–¹⁵ Another approach is that of the “effort/reward model,” which focuses on the contrast between demands and resources.¹⁶ Resources include salary, career opportunity, and other rewards. In this model, “immersion” has an important role, representing major commitment of effort. Immersion (or commitment) involves experiences of “not being able to stop thinking of work in the evening” or “starting to think of work immediately on awakening,” for example. Another important work-related factor may be the amount of social support received at work. Several studies have indicated the impact of lack of such support on cardiovascular disease, depression, and other outcomes.¹²^,¹⁷

The Cross-Sectional Connection Between Stress and Sleep

Considering the physiologic activation involved in the stress response it seems logical to expect a connection with sleep disturbances. In fact, stress is considered the primary cause of persistent psychophysiologic (primary) insomnia.¹⁸ The evidence is, however, surprisingly modest, at least in terms of systematic studies of causal relations in a longitudinal perspective. Cross-sectional epidemiologic studies are readily available, however. They indicate a strong link between questionnaires on stress and sleep.¹⁹^–²⁶

There are far fewer studies on stress and sleep using polysomnography (PSG). However, Shaver and associates ²⁷ studied women with insomnia and did not find that stress ratings differed from those of good sleepers. Differences were found, however, in neuroendocrine activation and similar measures, suggesting increased stress.

With respect to the particular character of the stress involved, Ribet and colleagues ²¹ studied more than 21,000 subjects in France, using a sleep disturbance index and logistic regression analysis. It was found that shift work, a long work week, exposure to vibration, and “having to hurry” appeared to be the main risk factors for reports of impaired sleep, when controlled for age and gender.

The demand component of the demand/control model of stressful work conditions has also been shown to have a link to reported sleep impairment.¹⁹^–²³ In Åkerstedt and associates,²² it was found that the strongest item of the demand index was “having to exert a lot of effort at work,” not simply “having too much to do,” for example. The distinction may be interpreted as a lot of work per se may not be the key factor but that the response to the work situation is more important. It was also found that when the item “not being able to stop thinking about work in the evening” was added to the regression this variable became the most important predictor of disturbed sleep. Again, it is the response to the work load that predicts sleep impairment. The effort-reward imbalance has also been related to disturbed sleep.²⁸

With regard to socioeconomic group, sleep complaints are more frequently found in blue-collar workers. Thus, Partinen and coworkers ²⁹ investigated several occupational groups and found disturbed sleep to be most common among manual workers and much less so among physicians or managing directors. Geroldi and associates³⁰ found in a retrospective study of older individuals (older than 75 years) that former white-collar workers reported better sleep than blue-collar workers. Kupperman and colleagues³¹ reported fewer sleep problems in subjects who were satisfied with their work. On the whole, blue-collar workers exhibit higher levels of stress than white-collar workers but the origins seem more related to living conditions in general (e.g., economy, social situation, neighborhood) than work stress.

Social support is usually seen as a countervailing force in demand/control models and seems to counteract the effects of high demands.¹⁷ Lack of social support may thus function as a risk factor for disturbed sleep.³² Nordin and associates³³ demonstrated an interactive effect on sleep between social support, high demands at work, and lack of control. In a second study the same group showed that disturbed sleep may be a mediator between poor social support and coronary heart disease.³³ Poor social support has been associated for instance with sleep complaints in Vietnam War veterans.³⁴

The Prospective Connection Between Stress and Sleep

The question if stress actually causes impaired sleep can only be resolved in longitudinal studies. One such study is that of Dahlgren and colleagues,³⁵ who followed white-collar workers before and during a period of intense work stress. Reported sleep duration and sleep quality (by sleep diary) was reduced, as was evening alertness. The latter was probably due to the intense work pressure. Finally, the cortisol pattern was flattened. In another study, Sadeh and associates³⁶ found reduced sleep (by actigraphy and diary ratings) during a period of examination stress. Still, with both studies it is not clear if some of the reduced sleep could be partly due to simply trading sleep time for more study time rather than stress affecting the ability to sleep.

Linton and coworkers ³⁷ studied employees with no reported initial sleeping problems and found that 14.3% developed a sleeping problem during the ensuing year. Even when controlling for possible confounders, stress in the form of a “poor” psychosocial work environment doubled the risk of developing a sleep problem. In a similar vein, Jansson and associates³⁸ showed the effects of present stress on later complaints of disturbed sleep. The results showed that among individuals with no insomnia at baseline, high work demands increased the risk of developing insomnia 1 year later. Among participants with insomnia at baseline, high leader support decreased the risk of still reporting insomnia at follow-up. Finally, low influence over decisions and high work demands were related to the maintenance of insomnia.

In another longitudinal study, Vahtera and colleagues showed that the effect of a self-rated tendency to respond strongly to stress predicted later sleep disturbances as a response to negative life events.³⁹ This points to a consistent pattern of disturbed sleep in response to stress. Similarly, Drake and coworkers showed that those who reported a higher habitual sleep vulnerability to stress also showed longer sleep latency and lower sleep efficiency on the first night in the sleep laboratory.⁴⁰ This suggests that sleep vulnerability to stress is somewhat of a trait. The scale Ford Insomnia Response to Stress (FIRST, from the Ford hospital) has also been shown to predict long sleep latency in response to caffeine.⁴⁰ In the earlier mentioned study by Sadeh and associates,³⁶ it was demonstrated that a stronger sleep duration response to examination stress was seen in those individuals who showed a high emotion-focused coping.

Åkerstedt and colleagues ⁴¹ recorded sleep in the home of 50 participants on four occasions across several weeks and showed that nighttime ratings of stress/worries at bedtime were related to reduced sleep efficiency, increased wake time after sleep onset and increased latency to slow-wave sleep. Every-3-hour ratings of “stress” in the sleep diary were also increased both on the day before and the day after this sleep. Also in this study, differential vulnerability was seen. Those individuals who showed an increase in stress ratings showed a significantly higher level of depression on the Hospital Depression and Anxiety Scale.⁴²

In other PSG studies, sleep on the night before an important examination ⁴³ and also before a day of skydiving⁴⁴ was investigated. The results indicate a slightly negative effect on sleep efficiency and the amount of deep sleep. In addition, a number of early laboratory studies of stress and sleep have been performed but the stressors have been rather artificial (e.g., an unpleasant movie) and the results are unclear.⁴⁵ It is probably the case that the stressor needs to be of some significance to the individual to have any effect on PSG measures of sleep.

There are also studies of the effects of major life events, including stressors at the national/societal level. Cernovsky and coworkers ⁴⁶ demonstrated a clear increase in population-wide negative life events preceding an outbreak of insomnia. Haynes has shown similar results.⁴⁷ The economic recession in Finland in 1993 was apparently related to reported reductions in sleep quality.⁴⁸ It has also been shown that sleep is disturbed in response to threats to national security, for example, after the nuclear accident at the facility at Three Mile Island and during the scud missile attacks on Israel during the Gulf War.⁴⁹^,⁵⁰ The effect of losing a life partner has in one study been shown to have surprisingly modest effects, and then mainly an increase in rapid-eye-movement (REM) intensity.⁵¹

Rumination and Anticipation

Indirectly, many of the studies just described suggest that it is not the stress experience itself that impairs sleep but rather the anticipation of untoward events and worries about the immediate future. Examinations, high work demands, and external threats are likely to cause this type of worried anticipation. Cropley and colleagues ⁵² coupled the relation between strain and poor sleep in teachers to rumination. The previous cross-sectional study by Åkerstedt and associates²² found that not being able to turn off thoughts of work in the evening was strongly related to subjectively disturbed sleep. This may manifest itself as a form of rumination, and rumination is seen as a major cause of disturbed sleep.⁵³ Hall and coworkers²⁶ have demonstrated in a cross-sectional study that intrusive thoughts at bedtime are related to increased alpha and beta power in the subsequent sleep EEG. Similarly, increased cognitive arousal at bedtime is related to increased sleep latency.⁴⁷^,⁵⁴

Closely related to rumination is the worrying and the tension before sleep owing to a very early (and unpleasant) time of awakening before an unusually early morning shift.⁵⁵^,⁵⁶ The resulting sleep contained less slow-wave sleep, which supports the notion that it is the anticipation of difficulties that is important in the stress reaction. A similar study was carried out with machine officers on container ships.⁵⁷ Sleep recorded during a night on call (but without any call occurring) showed a reduction of slow-wave sleep and a corresponding increase in stage 2 sleep. Also, heart rate was increased. This was interpreted as an activation in the face of a possibility that the alarm would sound, which happened on average every second night on call. In the previously mentioned study on stress/worries at bedtime,⁴¹ sleep recordings preceded by moderately increased subjective “stress/worries” at bedtime showed a moderate impairment, which may be interpreted as due to rumination.

Interestingly, Meerlo and colleagues ⁵⁸ have shown that intense stress increases deep sleep in rats. Presumably this is due to the increase in the metabolic rate of the central nervous system due to the stress, without any anticipatory component. Whether stress has a similar effect on humans is unknown, but several studies in humans have shown that brain use (increased mental activity) leads to more intense sleep⁵⁹ and that nonuse leads to reduced sleep intensity.⁶⁰ The observations on the effects of brain use could mean that acute stress will improve subsequent sleep as long as there is no remaining negative anticipation. This question has not been addressed, however.

Posttraumatic Stress

Posttraumatic stress disorder (PTSD) is another well-established cause of disturbed sleep, even if many of the more common indicators of sleep quality (e.g., sleep latency, efficiency of sleep, total length of sleep, and amount of slow-wave sleep) are only moderately affected ⁶¹^–⁶⁴ and sometimes not affected at all.⁶⁵ Instead, these studies suggest that the major effect of PTSD is to disturb REM sleep by either increasing or decreasing its duration and by increasing its intensity. It also increases the number of awakenings. The unpleasant dreams associated with traumatic memories also tend to produce conditioned avoidance responses in affected individuals, resulting in postponements on a daily basis of retiring or of even entering the sleeping area. The effects on sleep seem to mediate long-term health effects.⁶⁶

Only gold members can continue reading. Log In or Register to continue