David A. Sack

Seasonal Affective Disorder and Phototherapy

Recognition that the timing of depressive episodes is related to the seasons goes back at least to Hippocrates (circa 400 B.C.),’ who noted that “melancholia occurs in the spring.” In the last 150 years many researchers have noted seasonal variations in affective episodes and in suicide, an indirect reflection of the incidence of depression. The suicide data, well reviewed by Kevan2 and A ~ c h o f f , ~ show a peak incidence in late spring and early summer. The incidence of depressive episodes is highest in spring and fall, whereas mania seems to occur most frequently in late summer.‘ Several authors have correkted the incidence of these occurrences with climatic variables such as sunshine and temperature and have shown statistically significant relationship^.^*^-^ Such survey studies, however, cannot establish a causal link between environmental influences and psychopathology. In order to probe the mechanisms by means of which these seasonal rhythms might be mediated, we have adopted a different strategy. We have recruited and studied a population of patients who have a history of extreme seasonal vulnerability to changes in mood and behavior. In this paper we summarize our experience with these patients and their condition, and discuss our interventions in their seasonal rhythms, and the extensive animal literature on which these interventions were patterned.

Longitudinal sleep EEG, temperature, and activity measurements across the menstrual cycle in patients with premenstrual depression and in age-matched controls

After a 2-month evaluation period, eight women with moderate to severe premenstrual depression and eight age- and sex-matched controls underwent sleep electroencephalographic (EEG) and temperature recordings 2 nights a week over the course of one menstrual cycle. Overall, patients had more Stage 2 (%) sleep and less rapid eye movement (REM) sleep (% and minutes) than normal controls. Stage 3 sleep and number of intermittent awakenings varied with phases of the menstrual cycle. Temperature minima were earlier in patients compared with controls, but this difference was not statistically significant, and there was no significant effect of menstrual cycle phase on the timing of temperature minima. Wrist motor activity did not change during the menstrual cycle in patients or controls. Thus, in this sample of women with premenstrual depression, we did not find sleep EEG alterations similar to those reported in some patients with major depressive disorder. In light of the small number of subjects and the large individual variability, the absence of marked changes with the menstrual cycle may be a function of a Type II error.

Phototherapy for Seasonal Affective Disorder

is not a new one. Indeed, books were written on the subject in the early decades of this century (Humphris, 1924; Kovacs, 1932). However, there are important distinctions between the treatments outlined in these earlier texts and the phototherapy currently in use. These earlier clinicians regarded phototherapy as a broad-spectrum treatment for a wide variety of ills. Ultraviolet light was considered an important component of the treatments, and the entire body was exposed to the light. The emphasis was thus on exposing the skin to light, and special &dquo;light baths&dquo; were designed to maximize skin exposure and to shield the eyes from toxic ultraviolet rays. In contrast, phototherapy is currently regarded as helpful only with certain specific types of patients; the light sources used have contained relatively little ultraviolet light; and emphasis has been placed on the eyes rather than the skin as the portal of entry of the therapeutic effect. The majority of phototherapy studies conducted in the past few years have involved patients with SAD, a condition characterized by regular winter depressions alternating with nondepressed periods in the spring and summer (Rosenthal et al., 1984). There are a few recorded historical examples of manipulations of lighting and climate specifically aimed at reversing certain winter difficulties such as depression and lethargy. Esquirol (1845) reported a case of a man with symptoms remarkably similar to those of SAD, whom he treated successfully by advising him to move from

Neurobiology of Seasonal Affective Disorder and Phototherapy

I. To whom all correspondence should be addressed, at the Clinical Psychobiology Branch, NIMH. Building 1.0/Room 4S-239, 9000 Rockville Pike, Bethesda. Maryland 20892. The investigation of the biochemical basis of seasonal affective disorder (SAD) and light therapy is a new field. As with any body of knowledge undergoing rapid expansion, the data, though interesting, are incomplete and at times even conflicting. In this paper, we attempt to organize and synthesize the information currently available into a coherent and understandable form.

Frequency analysis of the sleep EEG in depression

Eight patients with major depressive disorder (seven bipolar and one unipolar) and matched controls had sleep studies, on which frequency analysis of the electroencephalogram (EEG) was performed. Total sleep and sleep efficiency were decreased in the patients, but there was no significant difference in rapid eye movement (REM) latency between the two groups. Frequency analysis revealed no group differences in power in the delta band (0.23-2.5 Hz) or the whole EEG spectrum (0.23-25 Hz). These findings suggest that mean REM latencies are not always shorter in major depression. The results are discussed in light of a previous report of decreased delta energy in the sleep EEG of unipolar patients.

Predicting suicidal risk in schizophrenic and schizoaffective patients in a prospective two-year trial

BACKGROUND Enhanced ability to reliably identify risk factors for suicidal behavior permits more focused decisions concerning treatment interventions and support services, with potential reduction in lives lost to suicide. METHODS This study followed 980 patients at high risk for suicide in a multicenter prospective study for 2 years after randomization to clozapine or olanzapine. A priori predictors related to diagnosis, treatment resistance, and clinical constructs of disease symptoms were evaluated as possible predictors of subsequent suicide-related events. RESULTS Ten baseline univariate predictors were identified. Historical predictors were diagnosis of schizoaffective disorder, history or current use at baseline of alcohol or substance abuse, cigarette smoking, number of lifetime suicide attempts, and the number of hospitalizations in the previous 36 months to prevent suicide. Predictive clinical features included greater baseline scores on the InterSePT scale for suicidal thinking, the Covi Anxiety Scale, the Calgary Depression Scale (CDS), and severity of Parkinsonism. Subsequent multivariate analysis revealed the number of hospitalizations in the previous 36 months, baseline CDS, severity of Parkinsons, history of substance abuse, and lifetime suicide attempts. Clozapine, in general, was more effective than olanzapine in decreasing the risk of suicidality, regardless of risk factors present. CONCLUSIONS This is the first prospective analysis of predictors of suicide risk in a large schizophrenic and schizoaffective population judged to be at high risk for suicide. Assessment of these risk factors may aid clinicians in evaluating risk for suicidal behaviors so that appropriate interventions can be made.

Sleep and circadian rhythms in affective patients isolated from external time cues

Sleep electroencephalographic activity, circadian rhythms in motor activity and rectal temperature, and clinical state were monitored longitudinally in four affectively ill patients (two depressed, one manic, and one rapidly cycling between depression and mania) who lived in isolation from external time cues (zeitgebers) for 3 to 4 weeks. In these conditions it was possible to observe the intrinsic or free-running behavior of circadian pacemakers and thereby to test several hypotheses about the role of sleep and circadian rhythms in the pathogenesis of depression. No hypothesis was consistently supported by the results. We found that the intrinsic rhythm of a circadian pacemaker appeared to free-run with an abnormally fast frequency in one patient. No patient remained stably depressed during temporal isolation. Our experience suggests that this type of study can be carried out safely with appropriate precautions. Temporal isolation is a means to test decisively predictions of several chronobiological hypotheses about affective illness and should be applied to additional patients.

Nocturnal TSH and prolactin secretion during sleep deprivation and prediction of antidepressant response in patients with major depression

In order to test the hypothesis that changes in the hypothalamic-pituitary axis during sleep deprivation are related to the antidepressant effects of this procedure, we measured thyroid-stimulating hormone (TSH) and prolactin levels in 32 depressed patients at 2:00 AM during a night before, during, and after total sleep deprivation (TSD). TSH levels increased significantly (p less than 0.05) during TSD, and prolactin levels decreased significantly (p less than 0.0001). When we divided the patients into responder and nonresponder groups based on a 30% reduction in the Hamilton Rating Scale, there was no difference between the two groups in their hormone levels on the baseline, TSD, or recovery nights. Changes in prolactin or TSH were not correlated with clinical improvement when the two groups were considered together or in the responder/nonresponder groups separately. Baseline values of both hormones were significantly (p less than 0.01) correlated with their respective levels during TSD and recovery sleep. These findings indicate that the relative levels of nocturnal TSH and prolactin are stable even within acutely depressed individuals and that changes in their levels are not related to the clinical response to sleep deprivation.

Deficient nocturnal surge of TSH secretion during sleep and sleep deprivation in rapid-cycling bipolar illness

Rapid-cycling bipolar patients have a high prevalence of hypothyroidism, and this disturbance in their hypothalamic-pituitary-thyroid (HPT) function may provide a model for understanding the less severe thyroid dysfunction present in other forms of affective disorder. For these reasons, we investigated HPT function in eight rapid-cycling bipolar patients and eight normal controls by measuring plasma levels of thyroid-stimulating hormone (TSH) and cortisol every 30 min during a baseline 24-h period and during an additional night of sleep deprivation. Thyrotropin-releasing hormone (TRH) (500 micrograms) challenge tests were also performed in the patients. Controls exhibited a significant circadian variation in TSH with a nocturnal rise that was augmented by sleep deprivation. In the rapid cyclers, the nocturnal rise in TSH was absent, and sleep deprivation failed to raise their TSH levels significantly compared with baseline. Low nocturnal TSH levels were associated with blunted TSH responses to TRH infusions; due to the relatively brief sampling interval used in the TRH challenge tests, however, these results do not reliably discriminate between hypothalamic and pituitary dysfunction as an etiology for low nocturnal TSH levels. Additional studies are needed to determine the precise nature of the HPT disturbance in rapid-cycling patients.

Measuring cholinergic sensitivity: I. Arecoline effects in bipolar patients

In a previous study, bipolar patients were found to be more sensitive than controls to induction of rapid eye movement (REM) sleep by the direct muscarinic agonist arecoline. We have replicated this observation in a new group of patients and controls. However, much overlap between groups was seen. Muscarinic supersensitivity appears to be present in some bipolar affective patients, but the REM induction procedure may not be the optimal method for measuring it.