Treatment of Bipolar Depression: A Survey of Canadian Psychiatrists

Verinder Sharma, MB, BS¹, Dwight S Mazmanian, PhD², Emmanuel Persad, MB, BS³, Karen M Kueneman, BA (Hons)⁴

Objective: This study was conducted to examine how Canadian psychiatrists manage bipolar depression.

Method: A questionnaire specific to the treatment of bipolar depression was mailed to 1639 active members of the Canadian Psychiatric Association.

Results: Seven hundred and sixty-six completed questionnaires were returned (46.7%). Most psychiatrists indicated that a combination of psychotherapy and somatic therapy was their preferred approach. For bipolar disorder, depressed, lithium carbonate and selective serotonin reuptake inhibitors (SSRIs) were the preferred treatment strategies. For substitution, tricyclic antidepressants (TCAs) were the favoured choice. Lithium carbonate was the preferred choice for augmentation and addition.

Conclusion: These findings indicate that a combination of psychotherapy and somatic therapy is the preferred treatment approach for bipolar depression. Lithium carbonate and SSRIs are the favoured somatic therapies.

(Can J Psychiatry 1997;42:298–302)

Key Words: bipolar depression, treatment, survey, Canadian psychiatrists

Bipolar disorder is a severe, recurrent illness with a lifetime prevalence of 1.2% in the adult population (1). Consequences of untreated illness are often devastating because of high rates of morbidity from loss of major life activity and normal health (2). In sharp contrast to symptoms of euphoria and grandiosity during mania, patients find themselves battling unrelenting dysphoria, paralyzing psychomotor retardation, and feelings of utter hopelessness during the depressed phase of the illness; 15% to 19% of those afflicted commit suicide (3,4).

Depressed bipolar patients have not generally benefitted from the gains made recently in the treatment of unipolar depression. All the currently available unimodal agents have been reported to induce mania. These drugs are known to accelerate cycle frequency in some patients and lead to the development of mixed states or continuous cycling in others (5–9). Mood stabilizers, by contrast, are clearly more effective in the acute treatment of mania than depression. In spite of these therapeutic challenges, treatment of bipolar depression remains largely unstudied. Results of the few controlled trials of somatic treatments have yielded limited or inconclusive information (10,11). Similarly, the data on the psychotherapeutic management of depression are rather sparse (11).

The purpose of this study was to examine the prescribing habits of Canadian psychiatrists in the treatment of bipolar disorder, depressed (bipolar I), and bipolar disorder not otherwise specified (NOS) (bipolar II) and to compare information obtained from this survey to that reported in the literature.

Method

A questionnaire developed by the authors was mailed with a stamped, return envelope to 1639 active members of the Canadian Psychiatric Association in March 1994. Four months later, a follow-up letter and a duplicate questionnaire were mailed to all members who had not returned the origi- nal questionnaire. In the first part of the questionnaire, respondents were asked to provide their age, sex, year of graduation from medical school, year of completion of postgraduate training, and type of practice. They were also asked to indicate whether they had an academic affiliation and whether they considered mood disorders to be their primary interest. The first 2 questions of the second part of the questionnaire asked the respondents to estimate the number of depressed patients (bipolar and unipolar) they treat per year and the percentage of those depressed patients who are bipolar. For the bipolar patients, respondents were asked to estimate the relative proportions of patients with bipolar type I versus type II disorder. Respondents were asked to rank the treatment options listed for the remaining 9 questions. In each case, separate responses were requested for bipolar disorder, depressed, and bipolar disorder NOS. Respondents were first asked to rank their general treatment approaches (psychotherapy alone, somatic therapy alone, or psychotherapy and somatic therapy combined) and the type of psychotherapy used (psychodynamic, cognitive–behavioural, interpersonal, eclectic, or “other”). They then ranked the types of somatic therapy they would employ for the acute treatment of bipolar depression. The listed options consisted of lithium carbonate, TCAs, monoamine oxidase inhibitors (MAOIs), SSRIs, anticonvulsants, electroconvulsive therapy, phototherapy, or “other.” Next, respondents were instructed to indicate what they would do if the initial trial of their preferred drug failed. Options provided for this question were: substitute with another drug from the same class, substitute with a drug from another class, augment with an agent, and add another drug. Finally, respondents rank-ordered options for substitution, augmentation, and addition of a drug. For substitution and addition, lithium carbonate, TCAs, MAOIs, anticonvulsants, and “other” were provided as options. Lithium carbonate, liothyronine sodium, tryptophan, anticonvulsants, and “other” were provided as options for augmentation. For all relevant questions, respondents were asked to specify the agent if they selected either “anticonvulsants” or the option “other.”

Data Reduction and Analysis

The mean rank for each treatment option and the percentage of psychiatrists selecting that option as their first choice were calculated. Mean rank was calculated using only the responses of psychiatrists who had chosen the drug (that is, if an option was not chosen by a psychiatrist, the response was treated as missing for computing the mean). In calculating the percentage of psychiatrists choosing a particular drug (option), the denominator comprised all respondents who had provided a response to any option of the question. Respondents who provided a rank of “1” for more than one option in a question were excluded from analyses involving that question. This resulted in an average loss of 17.5 respondents per question (range 1 to 72) or 2.3% of the sample. Statistical differences among the number of respondents selecting various options were assessed using the chi-square (goodness-of-fit) test with Yates’s correction (12). Since the comparisons between bipolar disorder, depressed, and bipolar disorder NOS violated the requirement of independence of events (or observations), McNemar’s procedure for correlated percentages was employed (13). These analyses were conducted on each treatment option in each question. For example, the number of respondents choosing versus not choosing lithium carbonate as the preferred treatment in bipolar disorder, depressed, were compared with the number choosing versus not choosing that drug for bipolar disorder NOS. Differences in treatment preferences between psychiatrists who indicated that mood disorders was their primary area of interest and those who did not were assessed using the z test for independent proportions. Because we anticipated making a large number of comparisons among the various percentages, we decided, a priori, to set our significance level at a conservative 0.001 in order to reduce the probability of type I errors. Finally, no statistical comparisons were conducted among cells when the frequency of observations in a cell was less than 10.

Results

Seven hundred and sixty-six of the 1639 questionnaires were returned (a return rate of 46.7%). The characteristics of the respondents are summarized in Table 1. As can be seen, the mean age of the respondents was 47.8 years (SD = 9.4, range 30 to 73 years). Five hundred and seventy-seven (75.3%) of the respondents were male and 183 (23.9%) were female. Six respondents did not indicate their sex on the form. The average length of time since completion of medical school was 23.6 years, and the average length of time since completion of postgraduate training was 16.4 years. Five hundred and fourteen respondents (67.1%) reported an academic affiliation. Two hundred and seven respondents (27.0%) indicated that their primary interest was mood disorders. The most frequent type of practice was private practice (356 or 46.5%), but a large proportion of respondents (38.2%) indicated 2 or more types of practice. Approximately half (45.4%) of the respondents treated more than 100 depressed patients (unipolar and bipolar) per year. Of the total number of depressed patients, the respondents estimated that an average of 18.6% were bipolar. Of the total number of bipolar patients, 49.5% had a diagnosis of bipolar disorder, depressed, and 36.2% had a diagnosis of bipolar disorder NOS.

Treatment preferences between bipolar disorder, depressed, and bipolar disorder NOS were compared using McNemar’s procedure, described previously. No statistically significant differences were observed between preferred treatments for the 2 diagnostic groups. Given this equivalence of treatment preferences, the remainder of the Results section will focus on the findings for bipolar disorder, depressed.

Table 2 presents preferred types of therapy. As can be seen in the table, 614 respondents (84.2%) preferred a combination of psychotherapy and somatic therapy. This number was significantly greater than that for somatic therapy alone (110 or 15.1% of respondents;  c²  =  349.5, P < 0.001). Only 4 respondents indicated a preference for psychotherapy alone.

Eclectic psychotherapy was the most frequent response for preferred type of psychotherapy (first choice for 368 or 54.2% of respondents); cognitive–behavioural therapy was the next most popular type (144 or 21.2%). The number of psychiatrists selecting eclectic therapy was significantly greater than the number preferring cognitive therapy, and the frequencies for both of these were significantly greater than those for interpersonal therapy (71 or 10.5%), psychodynamic therapy (57 or 8.4%), and “other” therapy (for example, supportive, educational [39 or 5.7%]) (c² ³ 24.1, P < 0.001). The frequencies for the latter 3 forms of therapy were statistically equivalent.

Somatic therapies for the acute treatment of bipolar depression are presented in Table 3. Lithium carbonate and SSRIs were the most frequent choices (ranked first by 42.0% and 38.7% of respondents, respectively) followed by TCAs (15.9%). The number of respondents selecting lithium carbonate did not differ significantly from the number selecting SSRIs, but both of these values were greater than that for TCAs (c²  =  75.5 and 61.4, P < 0.001).

Strategies employed if the initial trial of the preferred drug failed are listed in Table 4. The most frequent strategies were substitution with a drug from another class (42.0% of respondents), followed by augmentation (30.8%), substitution with a drug from the same class (20.0%), and addition of another drug (7.2%). The frequencies for all of these options differed significantly (c² values ³ 11.3, P < 0.001). Preferences for substitution are presented in Table 5. Tricyclic antidepressants, lithium carbonate, and anticonvulsants were the most frequent selections (chosen by 42.1%, 26.2%, and 15.0% of respondents, respectively; c² values ³ 18.8, P < 0.001). The frequencies for “other” treatments and MAOIs did not differ from each other, but both were significantly less than the frequencies for the other 3 options (c² values ³ 13.4, P < 0.001). Table 6 presents the preferences for augmentation. Lithium carbonate was by far the most frequent choice (c² values ³ 279.0, P < 0.001). The frequencies for the remaining agents did not differ when adjacent pairs were compared (for example, liothyronine versus anticonvulsants, anticonvulsants versus tryptophan). The number of respondents selecting liothyronine, however, was significantly greater than the number selecting tryptophan and “other,” and the number of respondents selecting anticonvulsants was greater than that for “other” (c² values ³ 12.8, P < 0.001). When preferences for addition of another drug were examined, lithium carbonate was once again the most frequently selected agent (ranked first by 324 respondents [56.2%], c² values ³ 122.7, P < 0.001). TCAs (96 or 16.6%) and anticonvulsants (94 or 16.3%) were selected with equivalent frequency, and the frequencies for both of these options were greater than those for “other” options (that is, SSRIs, neuroleptics [48 or 8.3%]), and MAOIs (14 or 2.4%) (c² values ³ 14.3, P < 0.001). The difference in frequency between “other” and MAOI was also significant (c² = 17.6, P < 0.001).

Additional comparisons were conducted after dividing the sample into 2 groups: those who indicated that their primary interest was in mood disorders (n = 207) and those who did not (n = 549). No significant differences were observed between these 2 groups, with the exception of a difference on the question concerning the acute somatic treatment of bipolar depression. The proportion of “specialized” psychiatrists choosing TCAs as their preferred somatic therapy was significantly less than the proportion for psychiatrists who did not indicate a primary interest in mood disorders (5.6% versus 20.0% for bipolar I, 5.6% versus 15.3% for bipolar II; z = 4.44 and 3.33, P < 0.001).

Discussion

While over half of the psychiatrists failed to return the questionnaires, a response rate of 46.7% compares favourably with those for similar studies. A survey of Canadian psychiatrists in 1988 aimed at determining strategies employed to treat resistant depression yielded a response rate of 24.5% (14). Completed surveys were returned by 28% of American psychiatrists who had a special interest in mood disorders when asked about their experience with carbamazepine (15). We believe that the second “reminder” mailing was one of the main reasons for our relatively high return rate. The design of our questionnaire may have contributed to the return rate as well. The format of our questionnaire permitted clinicians to choose from a variety of options and to rank their choices in order. Surveys involving vignettes have been criticized because written responses may not be an accurate reflection of actual clinical practice (16).

Treatment strategies reportedly used for treatment of bipolar disorder, depressed, and bipolar disorder NOS were not significantly different. Not surprisingly, an extremely large number of psychiatrists reported a combination of somatic therapy and psychotherapy as their preferred therapeutic intervention. The choice of eclectic psychotherapy by over half of respondents might indicate that psychiatrists use a variety of psychotherapeutic strategies rather than adhering to a specific technique. Of somatic therapies for depression, TCAs were the distant third choice after lithium and SSRIs. Lithium was also the preferred choice for augmentation and addition, but 58% of psychiatrists did not choose this drug as the first line of treatment. According to the American Psychiatric Association’s practice guidelines for the treatment of patients with bipolar disorder, lithium should be tried initially and antidepressants added only if depression is severe or there is a failure to respond to lithium with or without psychiatric management (11). A review of the controlled studies comparing lithium to placebo showed an antidepressant response in 79% of patients (10). Goodwin and Jamison have also recommended that the treatment of bipolar depression should begin with a trial of lithium (4).

Even though there are only limited data on the efficacy of SSRIs, these drugs were popular with clinicians, presumably because of their favourable side effect profile and the possibility of a lower risk of induction of mania (17,18). The use of electroconvulsive therapy, which remains the most effective treatment for bipolar depression, was surprisingly low. Similarly, MAOIs and anticonvulsants were the first choice of only a small number of physicians. Anticonvulsants, especially if used alone, do not appear to have a strong antidepressant effect. MAOIs, however, are considered safe and effective in the treatment of bipolar depression. Tranylcypromine was found to be significantly superior to imipramine in a study of the treatment of anergic bipolar depression (19). Another study showed moclobemide to be equally as effective as imipramine (20).

TCAs were the first choice of respondents for substitution and the second choice for addition. These drugs were not, however, as favoured by psychiatrists who had a special interest in mood disorders, presumably because of their heightened awareness of the risks (rapid cycling and/or induction of mania) associated with their use (9).

In case of failure of response with their favoured drug, 42% of respondents indicated a preference for substitution with another drug from another class, and 31% reported using augmentation strategies. Anticonvulsants were the third most common choice for each of substitution (following TCAs and lithium), augmentation (following lithium and liothyronine), and finally addition (following lithium and TCAs).

The results of our study may have been biased by factors such as type and setting of clinical practice, particular areas of interest, or whether respondents had academic affiliations. For example, more than two-thirds of physicians reported having an academic affiliation, while only 27% indicated mood disorders as their primary area of interest. The number and types of choices provided to questions are another factor likely to have introduced a bias. There were questions pertaining to the use of psychotherapy and other modalities, such as electroconvulsive therapy and phototherapy, but most of the items related to pharmacological management of depression.

To the best of our knowledge, this is the only study with such a large sample size that describes how bipolar depression is managed in clinical practice. Control studies on treatment of bipolar depression are urgently needed because this aspect of bipolar disorder is poorly understood and difficult to treat.

Acknowledgement

This project was supported by a Research Fund grant from the Department of Psychiatry, University of Western Ontario, London, Ontario, Canada.

References

1. Weissman MM, Leaf PJ, Tischler GL, Blazer DG, Karno M, Bruce ML, and others. Affective disorders in five United States communities. Psychol Med 1988;18:141–53.

2. Medical Practice Project. A state-of-the-science report for the office of the Assistant Secretary for the United States Department of Health, Education and Welfare. Baltimore: Policy Research; 1979.

3. Guze SB, Robins E. Suicide and primary affective disorders. Br J Psychiatry 1970;117:437–8.

4. Goodwin FK, Jamison KR. Manic-depressive illness. New York: Oxford University Press; 1990.

5. Koukopoulos A, Faedda G, Proietti R, D’Amico S, de Pisa E, Simonetto C. Un syndrome depressif mixte. L’Encephale 1922;18:19–21.

6. Post RM, Kramlinger KG, Altshuler LL, Ketter T, Denicoff K. Treatment of rapid cycling bipolar illness. Psychopharmacol Bull 1990;26:37–47.

7. Wehr TA, Goodwin FK. Do antidepressants cause mania? Psychopharmacol Bull 1987;23:61–5.

8. Akiskal HS. The milder spectrum of bipolar disorders: diagnostic, characterologic and pharmacologic aspects. Psychiatric Annals 1987;17:32–7.

9. Wehr TA, Goodwin FK. Rapid cycling in manic-depressives induced by tricyclic antidepressants. Arch Gen Psychiatry 1979;36:555–9.

10. Zornberg GL, Pope HG Jr. Treatment of depression in bipolar disorder: new directions for research. J Clin Psychopharmacol 1993;13:397–408.

11. American Psychiatric Association. Practice guideline for the treatment of patients with bipolar disorder. Am J Psychiatry 1994;151(12 Suppl):1S–36S.

12. Ferguson G, Takane Y. Statistical analysis in psychology and education. 6th ed. New York: McGraw-Hill; 1989.

13. McNemar Q. Note on the sampling error of the differences between correlated proportions or percentages. Psychometrika 1947;12:153–7.

14. Chaimowitz GA, Links PS, Padgett RW, Carr AC. Treatment-resistant depression: a survey of practice habits of Canadian psychiatrists. Can J Psychiatry 1991;36:353–5.

15. Denicoff KD, Meglathery SB, Post RM, Tandeciarz SI. Efficacy of carbamazepine compared with other agents: a clinical practice survey. J Clin Psychiatry 1994;55:70–6.

16. Jones TV, Gerrity MS, Earp J. Written case simulations: do they predict physicians’ behavior? J Clin Epidemiol 1990;43:805–15.

17. Peet M. Induction of mania with selective serotonin re-uptake inhibitors and tricyclic antidepressants. Br J Psychiatry 1994;164:549–50.

18. Cohn JB, Collins G, Ashbrook E, Wernicke JF. A comparison of fluoxetine, imipramine and placebo in patients with bipolar depressive disorder. Int Clin Psychopharmacol 1989;4:313–22.

19. Himmelhoch JM, Thase ME, Mallinger AG, Houck P. Tranylcypromine versus imipramine in anergic bipolar depression. Am J Psychiatry 1991;148:910–6.

20. Baumhacki U, Biziere K, Fischbach R, Geretstegger CH, Hebenstreit G, Radmayr E, and others. Efficacy and tolerability of moclobemide compared with imipramine in depressive disorder (DSM-III): an Austrian double-blind, multicentre study. Br J Psychiatry 1989;155(6 Suppl):78S–83S.

Résumé

Objectif : Cette étude visait à examiner comment les psychiatres canadiens traitent la dépression bipolaire.

Méthode : Un questionnaire sur le traitement de la dépression bipolaire a été envoyé par la poste à 1 639 membres actifs de l’Association des psychiatres du Canada.

Résultats : On a reçu sept cent soixante-six questionnaires dûment remplis  (46,7 %).  La plupart des psychiatres ont répondu que leur approche privilégiée consiste à associer la psychothérapie au traitement somatique.  En cas de trouble bipolaire déprimé, le carbonate de lithium et les inhibiteurs spécifiques du recaptage de la sérotonine (ISRS) constituent les stratégies de traitement de choix.  À des fins de substitution, on préfère les antidépresseurs tricycliques.  Le carbonate de lithium constitue le traitement de choix à des fins d’augmentation ou d’ajout.

Conclusion : Ces constatations révèlent qu’une association de la psychothérapie et du traitement somatique constitue l’approche privilégiée en cas de dépression bipolaire.  Le carbonate de lithium et les ISRS sont les traitements somatiques de choix.

Manuscript received May 1996, revised November 1996.

This paper was presented at the American Psychiatric Association 149th Annual Meeting, New York, May 1996.

¹Director, Mood Disorders Unit, London Psychiatric Hospital; Assistant Professor, Department of Psychiatry, University of Western Ontario, London, Ontario.

²Assistant Professor, Department of Psychology, Lakehead University, Thunder Bay, Ontario.

³Professor and Chairman, Department of Psychiatry, University of Western Ontario, London, Ontario.

⁴Project Coordinator (Research), Mood Disorders Unit, London Psychiatric Hospital, London, Ontario.

Address for correspondence: Dr V Sharma, Director, Mood Disorders Unit, London Psychiatric Hospital, 850 Highbury Avenue, PO Box 2532, Station A, London, ON N6A 4H1

Can J Psychiatry, Vol 42, April 1997