Treatment of seasonal affective disorder: a review

OBJECTIVE

To review the status of current treatment of seasonal affective disorder (SAD).

METHOD

Treatment studies of SAD published between January 1989 and March 1995 were identified using a computerized MEDLINE literature search. Additional citations were obtained from the reference sections of these articles. Studies included in this review were selected using operational methodologic criteria.

RESULTS

Many studies support the efficacy of bright light therapy using a fluorescent light box. The best studied protocol is > 2500 lux white light for 2 hours per day, but newer protocols using 10,000 lux for 30 minutes have comparable response rates. Studies of light visors and other head-mounted devices also report similar response rates, but have not yet shown superiority over putative control conditions. There are fewer medication studies in SAD, but controlled studies suggest that fluoxetine, d-fenfluramine and propranolol are effective. Other treatments such as dawn simulation require further study. No studies of psychological treatments for SAD were found. Many studies had methodologic limitations, including brief treatment periods, small sample sizes, and lack of replication, that limit the generalizability of findings.

CONCLUSION

There are several well-studied, effective treatments for SAD, including light therapy and medications. However, further research must be done to demonstrate sustained treatment response over time, to clarify the intensity-response relationship of light therapy, to clarify the role of light therapy and medications, and to assess combination treatments.

A comparison of Tridimensional Personality Questionnaire dimensions in bipolar disorder and unipolar depression

The harm avoidance (HA) personality dimension has been hypothesized to be a vulnerability factor for unipolar depression (UD) but not for bipolar disorder (BD). The reported difference on HA scores between these diagnostic groups may have been compromised by the assessment of BD patients who had not fully recovered. To test the diagnostic specificity of elevated HA scores and to elucidate whether assumptions about differences between patients with UD or BD might be attributed to the lingering effects of mood state, the Tridimensional Personality Questionnaire (TPQ) was administered to recovered patients with either BD or UD and a nonpatient comparison group. Both patient groups scored higher on the HA dimension than the nonpatient comparison group, but the patient groups did not differ from one another on this dimension. Moreover, novelty seeking (NS) scores were elevated in subjects with BD compared with both UD patients and nonpatient subjects. These results suggest that high HA scores may be associated with a mood disorder diagnosis, whereas high NS scores may be associated with the BD subtype.

Seasonal fluctuations in mood and eating behavior in bulimia nervosa

Thirty-five bulimia nervosa (BN) patients with high seasonality scores on a modified version of the Seasonal Pattern Assessment Questionnaire (SPAQ) were administered semistructured clinical interviews to further assess their seasonal symptom patterns. A diversity of patterns were identified, consistent with a multidimensional model of seasonality in the BN population. Overall, the results suggest that seasonality may be an important dimension to consider in patients with BN.

Change in folate status with antidepressant treatment

Ninety-nine consecutive unmedicated outpatients with a major depressive illness had blood drawn for measurement of serum folate (SF), red cell folate (RCF), and vitamin B12 within 24 hours of completion of ratings of severity of depression at the beginning and ending of a 5-week trial of desmethylimipramine (mean dose = 149.2 mg/day, range = 75-225 mg). As compared with nonresponders, responders had a significantly higher mean SF at baseline (nonresponders = 13.8 nmol/l; responders = 17.7 nmol/l) and RCF showed a significant inverse correlation with severity of depression and a significant positive correlation with age of onset of illness. At week 5, change in severity of depression was significantly correlated with change in RCF, and significantly more responders than nonresponders had an increase in RCF. The possible role of folate status in the regulation of mood and response to treatment is discussed.

Dim versus bright red (light-emitting diode) light in the treatment of seasonal affective disorder

Forty-three subjects with seasonal affective disorder were randomly assigned to receive 2 weeks of treatment with either bright- (mean 4106 lx) or dim-light (mean 96 lx) therapy, using red light-emitting diode light sources, in a head-mounted unit. Defining response as a 50% reduction in the 21-item Hamilton Depression Rating Scale score to a post-treatment score of less than 8, there was no significant difference in response rate between patients receiving bright light (67%) as compared with patients receiving dim light (68%). Possible explanations for the similar response rate to 2 very different illuminances of red light are discussed.

Is hypersomnolence a feature of seasonal affective disorder?

We examined hypersomnolence as experienced among individuals meeting standardized diagnostic criteria for Seasonal Affective Disorder (SAD). Data were available from 115 individuals attending a mood disorders clinic specializing in treatment of this disorder. Three modes of assessment were employed: retrospective self-reports (Seasonal Patterns Assessment Questionnaire), cross-sectional interviews (Standardized Interview Guide for the Hamilton Depression Scale, SAD Version), and prospective sleep diaries. Results indicated that self-reported total hours of sleep varied significantly across the seasons, with longest sleep occurring in winter and shortest sleep in summer. Seasonal sleep changes, as indicated by the SPAQ did not correlate significantly, however, with severity of depressive symptoms as indicated by the Hamilton scale. Multiple regression analyses indicated that only social activity levels (one of seven SPAQ items) was significantly and uniquely related to the severity of depression. When data obtained by the three instruments were compared, self-reported hours of sleep (whether measured by SPAQ or Hamilton interview) were significantly higher than indicated by prospective sleep diaries. We conclude that hypersomnolence may not be a central feature of SAD and that the validity of the SPAQ as an index of this disorder requires further investigation.

Predictors of response to lithium and triiodothyronine augmentation of antidepressants in tricyclic non-responders

Although both lithium and triiodothyronine (T3) augmentation strategies may be rapid, safe, and effective ways of treating patients who fail to respond to tricyclics, little is known about the clinical or biochemical correlates of response. We examined clinical and endocrine variables which distinguished T3 and lithium responders from each other and from non-responders in a post-hoc analysis of two studies involving 51 subjects who had received these augmentation strategies under double-blind conditions. Lithium non-responders were more severely depressed and had more insomnia and weight loss than responders. T3 responders could be distinguished from lithium responders by greater weight loss. There were no reliable clinical differences between T3 responders and non-responders. Thyroid function tests did not differ between any of the treatment or response groups. Our findings suggest that clinical but not endocrine variables may distinguish responders to these two augmentation strategies.

A comparison of personality characteristics of seasonal and nonseasonal major depression

During the acute depressive episode, seasonal affective disorder (SAD) patients (N = 24) differ significantly from non-SAD major depressives (N = 17) on five of 13 personality variables measured, although severity of depression appears to be similar. SAD patients score significantly lower on the self-criticism and dependency dimensions of the Depressive Experiences Questionnaire (DEQ) and significantly higher on three personality trait scales (including schizotypal, narcissistic, and avoidant) of the Millon Clinical Multiaxial Inventory (MCMI). Our data suggest that those with seasonal depression may represent a psychologically distinct subgroup of depressives.

Anxious and non-anxious bipolar disorder

Eighty-one outpatients with bipolar disorder (BD) were grouped by SADS anxiety symptom scores (high vs. low) or diagnosis of generalized anxiety disorder, and/or panic disorder. BD patients with high anxiety scores were more likely to have suicidal behaviour (44% vs. 19%), alcohol abuse (28% vs. 6%), cyclothymia (44% vs. 21%) and an anxiety disorder (56% vs. 25%) with a trend toward lithium non-responsiveness. Diagnosis of an anxiety disorder was related only to high anxiety and lower GAS scores. Thus, anxiety may have similar clinical relevance in BD as it does in unipolar patients.

The Tridimensional Personality Questionnaire in major depression

The authors examined Tridimensional Personality Questionnaire scores in 40 patients with unipolar nonpsychotic major depression before and after antidepressant treatment. They found that scores on the novelty seeking and reward dependence dimensions were not affected by depressed state or by treatment response status. However, scores on the harm avoidance dimension were significantly lower in antidepressant responders and were altered by depressed state.

A placebo-controlled comparison of lithium and triiodothyronine augmentation of tricyclic antidepressants in unipolar refractory depression

OBJECTIVE

To directly compare the efficacy of lithium carbonate and liothyronine sodium (triiodothyronine) in the augmentation of therapeutic response in antidepressant nonresponders.

DESIGN

A randomized, double-blind, placebo-controlled study of 2 weeks' duration.

SETTING

The Mood Disorders Program, Clarke Institute of Psychiatry and the University of Toronto, Ontario.

PATIENTS

Fifty outpatients, males and females, with unipolar, nonpsychotic major depression who had failed to respond to treatment with desipramine hydrochloride or imipramine hydrochloride.

RESULTS

Both liothyronine and lithium were more effective than placebo in reducing scores on the Hamilton Rating Scale for Depression. However, the antidepressant augmenting effect of these two compounds did not differ from each other. When response was defined as a 50% or more reduction in the Hamilton Rating Scale for Depression scores and a final score less than 10, we found that 10 of 17 subjects responded to liothyronine, nine of 17 responded to lithium and three of 16 responded to placebo.

CONCLUSIONS

Our study suggests that both lithium and liothyronine may be considered as alternatives in augmenting antidepressant response in patients who do not respond to treatment with a tricyclic antidepressant.

Anxiety disorders and anxiety symptoms in a clinic sample of seasonal and non-seasonal depressives

Thirty-eight patients with seasonal affective disorder (SAD) were compared with 33 non-seasonal recurrent major depressives (non-SAD) who presented during the winter months for differences in the prevalence of concurrent anxiety disorders and the impact of anxiety on treatment response. SAD patients received light therapy, whereas non-SAD patients received antidepressant medications. There was no differences in the prevalence of any anxiety disorder, or on scores of anxiety on the Hamilton Rating Scale for Depression between the SAD and non-SAD groups. The presence of any anxiety disorder was associated with a better response rate in SAD patients, and an inferior response rate in non-SAD patients. The findings refute previous suggestions that anxiety is more common in SAD than in non-SAD, but suggest that the presence of anxiety may be associated with differential treatment response rates.

Side effects of light therapy in seasonal affective disorder

The authors report the frequency of side effects of light therapy in 105 patients with seasonal affective disorder treated with three intensities of light. Common symptoms to emerge during treatment were headache (19%), eyestrain (17%), and feeling "wired" (14%). There was no relationship between side effects and intensity of light used.

Light visor treatment for seasonal affective disorder: a multicenter study

The effectiveness of light therapy in seasonal affective disorder (SAD) was evaluated in 105 subjects across five centers. Three intensities of light (60 lux, 600 lux, and 3500 lux) were used in a 2-week randomized, parallel design. There was no significant difference in antidepressant efficacy of the three intensities of light. All three intensities produced a similar frequency of antidepressant response to each other and to that reported in previous studies. There were site differences in the severity of depression during light treatment, but diagnosis and medication status did not affect antidepressant response. These findings suggest that light therapy has an antidepressant action by a nonspecific effect or that light is biologically active in the treatment of SAD across a wide range of intensities.

Clinical features of situational and nonsituational major depression

Clinical lore has long supported the subtype of situational or reactive depression. To date, however, there has been limited empirical research support for this subtype of major depression. We examined demographic, clinical and personality features of situational and nonsituational depression in 89 outpatients with unipolar nonpsychotic major depressive disorder. Situational depressives had a less recurrent course of illness and appeared to respond more completely to the antidepressant used for their current episode. Demographic and personality measures did not distinguish situational and nonsituational depression.

Folate, vitamin B12 and cognitive impairment in patients with Alzheimer's disease

This study examines the relationship between folate, vitamin B12 and severity of cognitive impairment in patients with Alzheimer's disease (AD) as compared with other disorders associated with cognitive impairment. The patients were 97 consecutive referrals to an AD clinic. Forty patients had either possible or probable AD, 31 had other dementias (OD) and 26 had mild cognitive impairment (cognitively impaired, not demented; CIND). Patients had blood drawn for serum, red cell folate and B12, as well as other biochemical indicators of nutrition, within 24 h of the Mini-Mental State Examination (MMSE). In the AD group, only B12 was significantly correlated with MMSE. Using regression analysis, B12 contributed significantly to variance in MMSE. There was no correlation between MMSE and serum, red cell folate or B12 in the OD or CIND group and no significant correlation between MMSE and other nutritional indices in any group. These findings suggest the possibility of a specific relationship between B12 levels and severity of cognitive impairment in patients with AD.

Diagnostic specificity of the dependent and self-critical personality dimensions in major depression

Validational studies of self-critical and dependent personality dimensions as vulnerability factors for depression have been tested primarily with depressed samples, employing research designs devised to address state vs. trait and trait-situational congruity issues. In this study we examined the diagnostic specificity to depression of these two personality dimensions, comparing Self-Criticism and Dependency scores as measured by the Depressive Experiences Questionnaire (DEQ) in two samples of outpatients: (1) panic disorder with agoraphobia; and (2) non-psychotic, unipolar major depression. As hypothesized, the two groups differed on Self-Criticism, with the depressed group scoring higher, but no differences were found for Dependency. These findings were similar even when depressed mood was partialed out. These results complement a growing body of research associating Self-Criticism, as specifically measured by the DEQ, with depression.

Major depression and subclinical (grade 2) hypothyroidism

Subclinical hypothyroidism (SCH) has been reported to occur in patients with a variety of affective syndromes. However, the clinical correlates of SCH in patients with major depression have received limited attention. We therefore examined demographic, clinical and treatment response variables in a cohort of patients with unipolar, nonpsychotic major depression with and without SCH. Of 139 subjects, 19 had SCH defined as an elevated basal TSH with normal circulating levels of T3 and T4. Major depression with SCH differed from that without SCH by the presence of a concurrent panic disorder and a poorer antidepressant response.

The thyroid and melancholia

Several studies that have examined heterogeneous groups of patients suggest that altered thyroid function may distinguish melancholic from nonmelancholic depression. We therefore measured basal thyroid hormone levels in 90 unipolar depressed patients who were divided into melancholic and nonmelancholic subgroups according to three definitions. Levels of thyroxine, triiodothyronine, and thyrotropin, obtained using an ultrasensitive assay, did not distinguish the subtypes of depression. However, severity of depression contributed significantly to the difference between these subtypes according to DSM-III and Research Diagnostic Criteria.

Coping styles, psychopathology and intellectual performance in patients with familial adenomatous polyposis

Thirty-eight patients with familial adenomatous polyposis (FAP) were compared with 19 patients with ulcerative colitis (UC) for differences in illness-related variables, coping styles, psychiatric symptomatology, and intellectual performance. Patients with FAP had significantly less education, longer time since recent surgery, less psychiatric illness, and evidence of less preoccupation with their illness, as compared with UC patients. FAP patients with a positive family history (N = 28) scored significantly lower on both verbal and performance intellectual tests, even when taking education into account, compared with FAP patients without a family history (N = 9). The relevance of these findings to the ongoing monitoring and surveillance of patients with FAP is discussed.