Depressive Symptoms and Family History in Seasonal and Nonseasonal Mood Disorders

Transcription

1 Depressive Symptoms and Family History in Seasonal and Nonseasonal Mood Disorders Judith M. Allen, M.D., Raymond W. Lam, M.D., Ronald A. Remick, M.D., and Adele D. Sadovnick, Ph.D. Objective: The authors goal was to compare the symptoms and family history of seasonal affective disorder with those of nonseasonal mood disorders. Method: From a subspecialty mood disorders clinic, 34 patients with major depression, seasonal pattern (seasonal affective disorder), diagnosed with DSM-III-R criteria, were matched in age, sex, and diagnostic subtype (recurrent unipolar, bipolar I, or bipolar II) to 34 patients with nonseasonal mood disorders. Data on symptoms during the most recent depressive episode were obtained by chart review and compared by using chi-square tests. Family history data for first-degree relatives ofpatients with seasonal and nonseasonal mood disorders were gathered by using the family history method, and diagnoses were based on Family History Research Diagnostic Criteria. Results: Patients with seasonal affective disorder reported significantly more hypersomnia, hyperphagia, and weight gain and reported less suicidal ideation and morning worsening of mood than the patients with nonseasonal mood disorders. No differences were found in family histories ofmood disorders, other psychiatric disorders, and any psychiatric disorder between the groups with seasonal versus nonseasonal mood disorders. Alcoholism was found more frequently in the relatives of the patients with seasonal affective disorder. Conclusions: Diff erences in symptoms between seasonal and nonseasonal mood disorders provide some support for seasonal affective disorder as a diagnostic subtype of mood disorders. However, the genetic loading for mood disorders (of unspecified seasonality), as determined by the family history method, is similar for seasonal and nonseasonal mood disorders. (Am J Psychiatry 1993; 150: ) S easonal changes in the course of psychiatric disordens have been recognized since the early nineteenth century with the writings of Pinel and Esquirol (as cited by Wehr and Rosenthal [1]). However, there has been renewed interest in the seasonality of mood disorders since the report by Rosenthal et al. (2) describing 29 patients with recurrent fall-winter depression (seasonal affective disorder) that responded to exposure to bright light (phototherapy, or light therapy). Presented in part at the 9th annual meeting of the West Coast College of Biological Psychiatry, San Diego, April 4-6, 1991, and at the 42nd annual meeting of the Canadian Psychiatric Association, Saskatoon, Sask., Oct. 2-4, Received Nov. 13, 1991; revision received June 22, 1992; accepted July 10, From the Departments of Psychiatry and Medical Genetics, University of British Columbia, and the University Hospital, UBC Site. Address reprint requests to Dr. Allen, Department of Psychiatry, University Hospital, UBC Site, 2255 Wesbrook Mall, Vancouver, B.C., Canada V6T 2A1. Supported in part by the Canadian Psychiatric Research Foundation, the University Hospital Foundation, and the B.C. Medical Services Foundation. The authors thank Arlene Tompkins, Ph.D., and Shauna Woolley, M.S.W., for their help in data collection. Copyright I 993 American Psychiatric Association. In DSM-III-R, seasonal affective disorder is not separately coded as a diagnosis but, instead, is included as a subtype, termed seasonal pattern, of recurrent major depression. Diagnostic criteria for a seasonal pattern include regular seasonal onset and remission of depnessive episodes in the absence of recurrent seasonal psychosocial stressors, at least three such episodes in three separate years with at least two of the years being consecutive, and seasonal episodes outnumbering any nonseasonal episodes of the disturbance by at least three to one. These criteria do not include any descniptive symptoms for a seasonal depressive episode, so that seasonal affective disorder is distinguished from nonseasonal mood disorders solely by the seasonal pattern of the mood episodes. There is still controversy about the existence of seasonal affective disorder as a separate diagnostic category (3, 4), but some sources have suggested that there is stronger evidence for the diagnostic validity of seasonal affective disorder than for diagnoses like dysthymia or melancholia (5). Several studies indicate that seasonal affective disonden is associated with a symptom profile different from that of nonseasonal mood disorders. Clear differences AmJ Psychiatry 150:3, March

2 SEASONAL AND NONSEASONAL MOOD DISORDERS in symptoms would provide additional support for seasonal affective disorder as a distinct clinical subtype of mood disorders. The depressions in seasonal affective disorder are reportedly characterized by atypical or reversed vegetative features, including increased appetite with carbohydrate craving and weight gain, reverse diurnal mood variation (afternoon and evening mood worsening), and increased sleep duration. Atypical depressive symptoms occur in 65% to 85% of study subjects with seasonal affective disorder (2, 6-9), compared with 15% to 30% of a depressed outpatient study group unselected for seasonality ( I 0). However, atypical features are more commonly found in anergic bipolar depression (10), which may be overrepresented in seasonal affective disorder; initial reports suggested that up to 90% of patients with seasonal affective disorder met diagnostic criteria for bipolar disorder (2, 6, 9). Few published studies have compared the symptoms of seasonal affective disorder and nonseasonal mood disordens. Garvey et al. (1 1 ) compared 1 8 patients with seasonal depression (defined as at least two episodes of DSM-III major depression that occurred in the same season over an unspecified number ofyears) with 13 patients who had similarly diagnosed major depression occurring in different seasons. Increased sleep duration, carbohydrate craving, seasonal sleep changes, and cloudy-day dysphoria significantly differentiated the patients with seasonal depression from those with nonseasonal depression. Reduced libido, increased appetite, irritability, fatigue, anxiety, and reduced concentration were not significantly different between the groups. The two groups, however, were not equivalent in diagnosis: 55% of the patients with seasonal depression had a history of mania or hypomania, compared with only 38% of the nonseasonal group. The two groups also differed significantly in age, although the difference was not stated. In addition, it is not known whether the patients defined as having seasonal depression would meet DSM-III-R criteria for a seasonal pattern. It remains unclear whether there are true differences in symptoms between seasonal and nonseasonal depression. Family psychiatric history may also help differentiate between seasonal and nonseasonal mood disorders; there is increasing evidence that genetic factors are involved in the etiology of psychiatric illnesses. Studies of seasonal affective disorder have reported a wide range (25%-67%) of positive family histories for mood disorders in first-degree relatives (6-9, 12). Similarly, the findings for a positive family history of alcoholism range from 8% to 38% (6-9, 12). The rates of a family history of seasonal depression in first-degree relatives of patients with seasonal affective disorder have been reported as 13% to 17% (6, 7, 9); the rate increases to 26% when second-degree relatives are included (12). To date, family studies of seasonal affective disorder have obtained family history data only from a psychiatnic interview of patients, without any attempt to systematically document the nature of the reported psychiatric disturbances in family members. In a clinical interview, depressed patients may underestimate the prevalence of mood disorder in relatives (13) or, conversely, may overestimate these disorders because they are sensitized to the symptoms of depression. A more reliable method of obtaining family history data is the family history method, in which knowledgeable family informants are also interviewed. This method has been shown to be more accurate than clinician assessment in identifying psychiatric illness in family members (13, 14) and is an acceptable alternative to the more comprehensive-but also more time-and resource-consuming-family study method, in which all available relatives are interviewed (15). To our knowledge, neither the family study method nor the family history method has yet been used in studies of seasonal affective disorder. In the present study we compared the symptom profiles during depressive episodes of patients with seasonal affective disorder and patients with nonseasonal mood disorders attending specialized clinics. We were specifically interested in whether matching for unipolar and bipolar diagnoses would affect the symptom profiles of the seasonal and nonseasonal groups. We also used the family history method to compare the family psychiatric histories of patients with seasonal affective disorder and matched patients with nonseasonal mood disorders. METHOD Thirty-four consecutive patients with a diagnosis of seasonal affective disorder attending our seasonal mood disorders clinic were cornpared with 34 patients with nonseasonal mood disorders attending an outpatient mood disorders service. This mood disorders service is the only tertiary care psychiatric program for the province of British Columbia that specializes in the assessment and treatment of mood disorders. All patients are referred to the service by family physicians and psychiatrists, and all medical services are covered under the provincial medical plan. Patients are assessed by experienced psychiatrists using unstructured clinical interviews. Diagnoses assigned by one psychiatrist using DSM-III-R criteria are routinely verified by another psychiatrist (R.W.L., R.A.R.) on chart review. Diagnostic interrater and intrarater reliability are regularly monitored. Patients with seasonal affective disorder met DSM-III-R criteria for recurrent major depression with a seasonal pattern. Comparison patients for each of these subjects were identified from the mood disorders service database and matched for age, sex, and diagnosis (recurrent unipolar, bipolar I, or bipolar II). The comparison patients did not meet DSM- III-R criteria for seasonal pattern. Demographic information and symptoms of the most recent depressive episode were obtained by chart review. Symptom items of interest were established a priori and included the atypical depressive symptoms reported to occur more frequently in seasonal affective disorder (2, 6-10). The family history method was used to obtain detailed family history data for patients with seasonal affective disorder and patients with nonseasonal mood disorders. The identified patient and a knowledgeable family informant were interviewed by a master s level genetic counselor who recorded detailed family pedigrees. When psychiatric symptoms were reported in family members, a family history research diagnosis questionnaire asking about specific symptoms ( 15) was administered and relevant psychiatric records were obtained, whenever possible, with appropriate consent. Using all available information on each family member, one of us (R.A.R.) used the Family History Research Diagnostic Criteria (FH-RDC) to assign diagnoses to relatives of both groups of subjects. The FH-RDC diagnoses were operationally defined as definite, probable, or possible. An FH-RDC 444 Am J Psychiatry I 50:3, March 1993

3 ALLEN, LAM, REMICK, ET AL. diagnosis was considered definite if supported by descriptive data and corroborating medical records. The diagnosis was defined as probable if medical records could not be obtained but sufficient descriptive data were available to clearly fulfill the FH-RDC and medical attention was received at the time of the illness or substantial impairment of function occurred. A possible FH-RDC diagnosis was assigned if only descriptive data were present. To increase diagnostic reliability, only definite and probable FH-RDC diagnoses for first-degree relatives were included. Intrarater and interrater reliability were regularly monitored and considered adequate. For example, for the diagnosis of depression, intrarater reliability was 73% (kappa=0.46) and interrater reliability was 80% (kappa=0.58). Family psychiatric history in first-degree relatives was classified according to FH-RDC into mood disorders, alcoholism, other psychiatric disorders, and any psychiatric disorder. Other psychiatric disorders included schizophrenia, anxiety disorders, and antisocial personality disorder. Any psychiatric disorder included mood disordens, alcoholism, or other disorders. Parametric data were analyzed with Student s t tests, and nominal data were compared with chi-square tests. Symptom items were included in the data analysis only if information was available on at least 80% (27 of 34) ofeach group of patients. Due to the number of comparisons in the symptom items, statistical significance was established a priori at the p<0.01 level for these items. Significance was defined at the p<0.0s level for other comparisons. All data analyses were performed on a microcomputer using the SPSS/PC+ 4.0 statistical software package (16). TABLE 1. Clinical Characteristics of Patients With Seasonal Affective Disorder and Patients With Nonseasonal Mood Disorders Patients With Seasonal Affective Disorder Patients With Nonseasonal Mood Disorders Characteristic N % N % Diagnosis Unipolar Bipolar! Bipolarll Alcoholism S IS 2 6 Clinical history Previous psychiatric carea Previous use of antidepressantsb is Previous use of lithiumc Previous use of anxiolytics ax2s.9, df=1, p<o.o2. b128, df=1, p< CXLO.3, df=1, p<0.6. X21.4, df=1, p<o.2. RESULTS There were 12 men and 22 women in each group. The mean age of the patients with seasonal affective disorden (38.4 years, SD=10.4, range=22-63) did not differ significantly from the mean age of the patients with nonseasonal mood disorders (39.2, SD=9.9, range=23-63) (t=0.34, df=66, p<o.8). The mean number of previous mood episodes of the patients with seasonal affective disorder (mean=10.2, SD=S0) did not differ from the mean number of previous mood episodes of the patients with nonseasonal mood disorders (mean= 10.8, SD=8.2) (t=0.36, df=66, p<o.7). Thus, the matching of the groups was verified. Only three patients in each group had experienced manic episodes, but 1 1 in each group had hypomania (table 1 ). Twenty patients in each group had recurrent unipolar depression. Significantly more patients with nonseasonal mood disorders than patients with seasonal affective disorder had previously used antidepressant medications, but previous use of anxiolytics and lithium did not differ between groups (table 1 ). A significantly greater proportion of patients with nonseasonal mood disorders had received previous psychiatric care (table 1). Three (9%) of the patients with seasonal affective disorder and seven (21 %) of the patients with nonseasonal mood disorders had past hospitalizations. Symptom item comparisons are presented in table 2. Items were compared only if there was a response rate of at least 80% (27 of 34 patients). Therefore, the percentages given are based on the total number of patients reporting for each item. Hypensomnia, hyperphagia, and weight gain occurred significantly more frequently in the patients with seasonal affective disorder (table 2). Significantly more patients with nonseasonal mood dis- orders reported suicidal ideation and morning worsening of mood when depressed (table 2). Energy changes, social activity, and job impairment did not differ between patients with seasonal affective disorder and those with nonseasonal mood disorders. Insufficient item response precluded a comparison of carbohydrate craving or irritability. The two diagnostic groups did not differ in the mean number of relatives assessed per patient by the family history method (6.4 relatives/patient with seasonal affective disorder versus 6.0 relatives/patient with nonseasonal mood disorders) (t=1.3, df=65, p<o.s). The FH-RDC diagnoses for each group are shown in table 3. No significant differences were found between the patients with seasonal affective disorder and those with nonseasonal mood disorders for family history of mood disorders, other psychiatric disorders, or any psychiatnc disorder. Alcoholism, however, occurred significandy more often in the family history of patients with seasonal affective disorder (table 3). In each group, only two patients (6%) had relatives with bipolar disorder. In the patients with a family history of mood disordens, the nine patients with seasonal affective disorder (seven unipolar, one bipolar I, and one bipolar II) had eight relatives with unipolar disorder and two relatives with bipolar disorder. The I 3 patients with nonseasonal mood disorders (seven unipolar, five bipolar I, and one bipolar II) had 12 relatives with unipolar and two with bipolar disorder. DISCUSSION This study, based on 34 patients with seasonal affective disorder and 34 matched comparison patients with nonseasonal mood disorders, confirms that there are Am J Psychiatry 1 50:3, March

4 SEASONAL AND NONSEASONAL MOOD DISORDERS TABLE 2. Symptoms of Patients With Seasonal Affective Diso rder and Patients With Nonseasonal M ood Disorders Patients With Seasonal Patients With Nonseasonal Affective Disorder Mood Disorders Number Number of Valid of Valid Symptom Cases N % Cases N % x2 df p Sleep < Increaset Decrease Nochange Appetite < Increasec Decrease Nochange Weight < Increase Decrease Nochange 34 S Diurnal variation <0.01 Morning wonsee Evening worse Nochange Energy <0.3 Increase Decrease Nochange Suicidalthoughts <0.01 Social withdrawal <0.7 Job impairment <0.6 apercentage of valid cases. b206, df=1, p<0.000s. cxl=235 df=1, p<0.000s. y=16.2, df=1, p<0.000s. exh9.4, df=1, p<0.01. TABLE 3. Positive Family History by Family History Method for Patients With Seasonal Affective Disorder and Patients With Nonseasonal Mood Disorders Patients With Patients With Seasonal Nonseasonal Affective Mood Disorder Disorders Family History Diagnosis N % N % Mood disorders Alcoholisma Other psychiatric disordert Any psychiatric disonderc ax24.s, df=1, p<0.03. bschizophrenia or anxiety disorders. CMood disorder, alcoholism, or other psychiatric disorder. significant differences in symptom patterns of sleep, appetite, weight, and diurnal variation between these diagnostic groups. The relatively small size of the study group and the case-control design of this study make the subject ascertainment methods important because of the possibility of a systematic selection bias between groups. Patient ascertainment, however, was similar for both seasonal and nonseasonal groups. All patients required a physician referral and were assessed in a subspecialty setting by the same group of clinicians. Universal medical coverage ensures that clinic participation does not select for patients who cannot obtain cane otherwise. Although selection bias cannot be ruled out, these factors may minimize its effect. The atypical depressive symptom profile found in our patients with seasonal affective disorder replicates the findings from several series of patients with seasonal affective disorder from research groups in different countries (2, 6-9). The predominance of atypical features in seasonal depression cannot be explained by an overrepresentation of bipolar patients because we carefully matched for unipolar and bipolar diagnoses. In addition, studies of seasonal affective disorder using more stringent DSM-III-R criteria reported that only 8%-21 % of patients have bipolar disorders (7, 12). There has been interest in diurnal mood variation associated with seasonal affective disorder because it may be a marker for hypothesized circadian rhythm disturbances. Although a significantly greater proportion of our patients with nonseasonal mood disorders showed early morning worsening of their mood (96%), many patients with seasonal affective disorder also reported morning worsening (59%) or no diurnal mood vanation (25%). Only a minority (16%) had afternoon or evening worsening of mood, which might be consistent with a hypothesis of delayed circadian rhythms in seasonal affective disorder. 446 Am J Psychiatry 1 50:3, March 1993

5 ALLEN, LAM, REMICK, ET AL. No significant differences were found in energy changes between seasonal affective disorder and nonseasonal mood disorders. Although low energy is considered a prominent feature in winter depression, it is also a common symptom of nonseasonal depression. Fewer patients with seasonal affective disorder than of patients with nonseasonal mood disorders had suicidal ideation. The patients with seasonal affective disorder were also less likely to have had previous psychiatric care and antidepressant treatment. This does not necessarily imply that seasonal depression is less severe than the nonseasonal mood disorders. Indeed, there were no significant differences in social withdrawal or job impairment between the two groups. Anecdotal reports from patients with seasonal affective disorder suggest that the predictable springtime remissions of past episodes reduces the hopelessness experienced during a winter depression, and this may explain the lower rate of suicidal ideation. The lower rate of previous treatment may reflect differences in help-seeking behavior or differential attribution of symptoms to a physical rather than a psychological disturbance. Further studies are needed to clarify these differences. The percentage of positive family histories of mood disorder as determined by the family history method was not significantly different between patients with seasonal affective disorder (27%) and patients with nonseasonal mood disorders (38%). In each group, only 6% of patients had a family history of bipolar disorder, despite the large percentage of patients with bipolar disorder (41%). This may be due, in part, to a higher threshold for the diagnosis of hypomania by the FH-RDC. Few other studies have subtyped mood disorders in the relatives of patients with seasonal affective disorder. White et al. (12) reported that 7% of patients with seasonal affective disorder had first-degree or 5ccond-degree relatives with a bipolar mood disorder; 7% also had first-degree or second-degree relatives with schizophrenia. Our rates of family history of mood disorders in patients with seasonal affective disorder are substantially lower than those reported in several other studies (55% to 67%) that used clinician assessment rather than the family history method (8, 9, 12). The family history method as applied in this study is a rigorous evaluation that is likely more reliable than an interview by a clinician. Other studies have shown that there is relatively poor agreement between the family history method and clinician assessments in diagnosing mood disorders and that a clinician assessment may be inadequate for family history studies (14). This study provides evidence that the genetic loading for mood disorders in seasonal affective disorder, as detenmined by family history, is similar to the genetic loading for nonseasonal mood disorders. It does not, however, help distinguish seasonal affective disorder from nonseasonal mood disorders on the basis of a genetic etiology. One possible explanation is that the precise genetic factors associated with transmission of mood disorders may not specifically influence the expression of seasonal or nonseasonal mood episodes. An important question that remains unanswered in our study is whether family members of patients with seasonal affective disorder are more likely to have seasonal affective disorder than nonseasonal mood disorders. We did not include seasonal pattern as a subtype of mood disorders for relatives in this study because that diagnosis was felt to be unreliable without a direct interview with the relative. The DSM-III-R diagnosis of seasonal affective disorder requires careful recognition of recurrent depression with specific seasonal onsets and remissions of several episodes. Given that a clinician assessment does not accurately diagnose specific mood disorders in family members, it is unlikely that seasonal affective disorder can be reliably diagnosed in relatives by patient interview alone, or even by the family history method. Therefore, the validity of reported rates of seasonal affective disorder in family members in previous studies must be questioned. Only the family study method, in which all relatives are directly interviewed, will likely supply sufficient information to reliably establish a seasonal affective disorder diagnosis for family members. The higher rate of alcoholism in the relatives of the patients with seasonal affe#{232}tivedisorder remains unexplained and open to speculation. Five patients with seasonal affective disorder had a lifetime diagnosis of alcoholism, compared with two patients with nonseasonal mood disorders (x2=l.4 df=1, p<0.23). Therefore, alcoholism was not significantly overrepresented in the patients with seasonal affective disorder. The number, age, and sex distribution of relatives assessed also did not differ between groups. It is possible that there is some genetic link between alcoholism and seasonality. Cases of seasonal alcoholism and seasonal substance abuse have been reported (17, 18), and there may also be seasonal changes in melatonin secretion in alcoholics (19). These preliminary findings suggest an intriguing connection between seasonality of mood and alcoholism that invites further study. In summary, the differences in symptoms between seasonal and nonseasonal depression found in this study support the hypothesis that seasonal affective disorder is clinically distinct from nonseasonal mood disorders. Etiologic differences between seasonal affective disorder and nonseasonal mood disorders are also suggested by neurobiological findings and possibly by a differential response to phototherapy (1, 20-23). The rates of family history of mood disorders in seasonal affective disorder assessed with the family history method appear lower than those found in studies using a clinician interview alone. This is most likely due to the more rigorous ascertainment and diagnostic criteria of the family history method. Rates of a family history of mood disorder were similar for seasonal affective disonden and nonseasonal mood disorders, but patients with seasonal affective disorder were more likely to have alcoholism in their family histories. The demonstration of a genetic contribution to seasonality in mood disorders will require rigorous methods for identifying seasonal Am J Psychiatry 1 50:3, March

6 SEASONAL AND NONSEASONAL MOOD DISORDERS patterns of mood disorders in family history data collection. As noted by Kendall (24), however, the validation of any clinical syndrome will require supportive evidence from several areas of study, including clinical descriptions, basic pathophysiology, course and prognosis, treatment response, and family studies. For seasonal affective disorder, as for many other psychiatric diagnoses, there are still only sparse data to substantiate its clinical validity. REFERENCES 1. Wehr TA, Rosenthal NE: Seasonality and affective illness. Am J Psychiatry 1989; 146: Rosenthal NE, Sack DA, Gillin JC, Lewy AJ, Goodwin FK, Dayenport Y, Mueller PS, Newsome DA, Wehr TA: Seasonal affective disorder: a description of the syndrome and preliminary findings with light therapy. Arch Gen Psychiatry 1984; 41: Eastwood MR. Peter AM: Epidemiology and seasonal affective disorder. Psychol Med 1988; 18: Mrosovsky N: Seasonal affective disorder, hibernation and annual cycles in animals: chipmunks in the sky. J Biol Rhythms 1988; 3: S. Spitzer RL, Williams JBW: The validity of seasonal affective disorder, in Seasonal Affective Disorders and Phototherapy. Edited by Rosenthal NE, Blehar MC. New York, Guilford Press, Rosenthal NE, Wehr TA: Seasonal affective disorders. Psychiatr Annals 1987; 17: Lam RW, Buchanan A, Remick RA: Seasonal affective disorder-a Canadian sample. Annals of Clin Psychiatry 1989; 1: Thompson C, Isaacs G: Seasonal affective disorder-a British sample. J Affective Disord 1988; 14: Wirz-Justice A, Buchili C, Gran P, Kielholz P, Fisch HU, Waggan B: Light treatment of seasonal affective disorder in Switzerland. Acta Psychiatn Scand 1 986; 16: Himmelhoch JM, Thase ME: The vagaries of the concept of atypical depression, in Modern Perspectives in the Psychiatry of the Affective Disorders. Edited by HowellsJG. New York, Brunnen/Mazel, 1989 I I. Garvey MJ, Wesner R, Godes M: Comparison of seasonal and nonseasonal affective disorders. Am J Psychiatry 1988; 145: White DM, Lewy AJ, Sack RI., Blood ML, Wesche DL: Is winter depression a bipolar disorder? Compr Psychiatry 1990; 31: Mendlewicz J, Fleiss JL, Cataldo M, Rainen JD: Accuracy of the family history method in affective illness: comparison with direct interviews in family studies. Arch Gen Psychiatry 1975; 32: Remick RA, Sadovnick AD, Gimbarzevsky B, Huggins MJ, Lam RW, Zis AP, Baird PA: Obtaining a family psychiatric historyis it worth the effort? in Abstracts of the 40th Annual Meeting of the Canadian Psychiatric Association. Ottawa, CPA, Andreasen NC, Endicott J, Spitzer RL, Winokur G: The family history method using diagnostic criteria: reliability and validity. Arch Gen Psychiatry 1977; 34: SPSSIPC Chicago, SPSS, McGrath R, Dickinson F, Yahia M, Keuhl E, Main M: A preliminary report on seasonal alcohol abuse and dependence, in Abstracts ofthe First Annual Meeting of the Society for Light Treatment and Biological Rhythms. Wilsonville, Ore, SLTBR, 1989 I 8. Satel SL, Gawin FH: Seasonal cocaine abuse. Am J Psychiatry 1989; 146: Jam AK, Kelwala 5, Campbell J, Powell B, Yerasi S, Khoury S: Seasonal change of melatonin rhythm in alcoholics, in New Research Program and Abstracts, 143rd Annual Meeting of the American Psychiatric Association. Washington, DC, APA, I Blehar MC, Lewy AJ: Seasonal mood disorders: consensus and controversy. Psychopharmacol Bull 1990; 26: Skwener RG, Jacobsen FM, Duncan CC, Kelly KA, Sack DA, Tamarkin L, Gaist PA, Kasper S, Rosenthal NE: Neurobiology of seasonal affective disorder and phototherapy. J Biol Rhythms 1988; 2: Kripke DF, Mullaney DJ, Savides TJ, Gillin JC: Phototherapy for nonseasonal major depressive disorders, in Seasonal Affective Disorders and Phototherapy. Edited by Rosenthal NE, Blehar MC. New York, Guilfond Press, Lam RW, Knipke DF, Gillin JC: Phototherapy for depressive disorders: a review. Can J Psychiatry 1989; 34: Kendall RE: Clinical validity. Psychol Med 1989; 19: AmJ Psychiatry 150:3, March 1993