Prevalence of psychotropic treatment
Only 7% of the patients were treatment-free; antidepressants were the most prescribed treatment, especially in BD-II; lithium was administered to 62% of the bipolar I patients and only 18% of the bipolar II patients.
In comparing the findings of our study with the results from scientific literature, we noticed that the percentage of patients not given any psychotropics in our sample was similar to results taken from previous studies conducted within academic centres or teaching hospitals where percentages ranged from 2% to 8% [36–39] were noted. Verdoux et al.  found lower percentages (0.9%) in hospital settings. Higher percentages (ranging between 16% and 22%) were found by Blanco et al. , examining data on a nationally representative group of visits to psychiatrists in an office-based practice within the US.
In our study, the global percentage of patients given lithium was 31%, which is in the low range of percentages found in other studies, ranging from 30% to 53% [36–38, 40–46]. This low rate is comparable with that observed in the United States where the percentage of bipolar patients prescribed with lithium has decreased from 50.9% in 1992–95 to 30.1% in 1996–99, based on the National Ambulatory Medical care survey . By contrast, European countries show a higher percentage of lithium use compared with the United States [36–38, 40, 42–46]. Decrease in lithium use has paralleled the increased use of valproate (from 10.9% to 26.6%) , and of second generation antipsychotics, (from 1.2% to 17.0% in Blanco et al. ; see also Hayes et al.  Pillarella et al. ). The low percentage of lithium use found in our study may well reflect the general North-American trend. Our study actually indicates both second generation antipsychotic and valproate use rates as being relatively high (41.6% and 15.0%, respectively), and this was consistent with the increasing trends in the use of such medications over the past decade.
In conducting our study, the type of BD appeared to be an important factor in predicting the use of lithium, with BD-I recording a much higher proportion of lithium users. This result is at odds with the STEP-BD data  which indicates a similar percentage of lithium users in BD-I and II (39.4% and 35.7%, respectively). On the other hand, Arvilommi et al. (2007) found that bipolar I patients were prescribed with mood stabilizers more often than bipolar II patients, which is consistent with our results, where the percentage of bipolar I patients taking at least one mood stabilizer or atypical antipsychotic in the therapeutic dosage was almost double than in bipolar II patients. This might not only reflect the CANMAT guidelines, but also the perceived lower degree of severity in BD-II and low consistency diagnosis. In our sample, about 20% of patients taking lithium had levels below the therapeutic range, which is consistent with the results of previous studies [30, 36].
The percentage of bipolar patients given antidepressants in our study (62.8%) was similar to that found in similar studies conducted in North America. In the United States studies, corresponding percentages range from 40% to 74% [41, 44–46, 49, 50]. Studies by Blanco et al.  indicate a constant pattern in the use of antidepressants between 1992–1995 and 1996–1999 (45.8% and 45.7%, respectively). As expected, bipolar patients with depressive episodes seemed to be more likely to take antidepressants than other bipolar patients in our study. A similar trend was observed in the STEP-BD study .
Interestingly, there is also a transatlantic difference in antidepressant use, more frequent in the United States than in Europe  In UK studies, the percentage of antidepressant use varies from 14.3% to 38% [36–38, 43], whereas in France, Verdoux  et al. found 34.2% of bipolar patients prescribed with antidepressants.
In our sample, no bipolar patients were prescribed aripiprazole or ziprasidone, although CANMAT guidelines specifically recommended the two atypical antipsychotics in the treatment of BD-I since 2005 . This can be due to the fact that Health Canada approved their use in the treatment of bipolar disorder only in 2009.
In our sampling, monotherapy with mood stabilizers was not a frequent approach, consistent with literature on the subject where only a minority of bipolar patients benefited from monotherapy treatment with lithium or other mood stabilizer (Ahmed and Anderson : 23%; Frangou et al. : 23.8%; Lloyd et al. : 29%; Ghaemi et al. : 10.8%; Verdoux et al. : 14%).
In recent years, polytherapy has been used more frequently in clinical practice [44, 51]. Polytherapy is more commonly administered for depressed bipolar patients where monotherapy is much less efficacious , while guidelines for acute mania recommend monotherapy with mood stabilizers or antipsychotics. The high frequency of polytherapy noted in our study may well reflect the higher prevalence of depression; it also suggests that the severe state of patients referred to our tertiary care services are more likely to have comorbidities or disorders resistant to treatment.
Concordance with treatment guidelines
Treatment was congruent with guidelines in 68% of our sampling cases, when both type and dosage of the medications were considered. This percentage was quite high when compared with other studies. The larger spectrum of recommendations present in the CANMAT guidelines, compared with other guidelines, can partially explain this result .
Divergence from guidelines was mostly noted for patients suffering from hypomania, with 68% of hypomanic patients receiving non-concordant treatment to guidelines. In more than half of the cases, treatments not congruent with guidelines were due to inappropriate treatments with antidepressants; the remaining cases of non-concordance were due to the absence of psychotropic treatment or to inadequate dosage. Also noteworthy is that more than one-third of depressed bipolar II patients receiving treatment concordant with guidelines were prescribed antidepressants within monotherapy, which constitutes a third line recommendation under the CANMAT guidelines, and almost one third of all bipolar II patients were treated with antidepressant without any mood stabilizer.
The use of antidepressants for bipolar depression is still controversial. The major criticism regarding use of antidepressants in instances of bipolar depression is based on the risk of destabilizing mood, switching into mania or hypomania [54–56] or increasing the cycle frequency [57, 58]. Such negative effects appear to be limited to certain classes of antidepressants [59–61]. Factors which appear to increase the risk of destabilization are: presence of mixed states [62, 63], subthreshold manic symptoms at baseline  and BD-I (versus BD-II) [65, 66], which suggests that different subpopulations may present different vulnerabilities to affective switching. In the STEP-BD study, approximately 15% of bipolar patients receiving long-term treatment with antidepressants developed a chronic irritable dysphoric state (ACID syndrome). The study indicated that the risk of developing ACID symptoms was linked to a previous history of antidepressant-induced affective switch, and to female gender . Finally, the use of antidepressants may significantly increase the risk of suicidal ideation .
The efficacy of antidepressants in the acute treatment of bipolar depression is uncertain. The STEP-BD study showed similar efficacy comparing lithium or divalproex, alone or associated with an antidepressant . In his meta-analysis, Gijsman et al.  suggested antidepressants as being effective in the short-term treatment of bipolar depression, but in a more recent meta-analysis, antidepressants were found not to be superior in any way to placebos in the acute treatment of bipolar depression . Then again, Altshuler et al.  suggested the usefulness of maintaining antidepressant treatments within the subpopulation of patients who require both a mood stabilizer and an antidepressant to achieve complete remission from depressive episodes. Also controversial is the efficacy of long-term treatment with antidepressants in protecting patients from the recurrences of bipolar depression [55, 72]. A meta-analysis of seven trials involving long-term treatments concluded that long-term adjunctive antidepressant treatments were not superior to mood stabilizers alone in protecting patients from recurrences, and that antidepressant monotherapies were associated with higher risks of mania or hypomania relapses .
Antidepressants administered within monotherapy are contraindicated for BD-I, because of the high risk of inducing mania or rapid cycling . However, the risk/benefits may be different in BD-I and BD-II, and some studies suggest the efficacy and safety of antidepressant monotherapies in the treatment of BD-II [74–77]. CANMAT guidelines recommend antidepressants associated with mood stabilizers or atypical antipsychotics in the treatment of acute bipolar depression. Antidepressant monotherapy is only recommended as a third line treatment for BD-II, particularly for those with infrequent hypomanias .
Finally, though there is agreement on the different guidelines in the discontinuance of treatment with antidepressants during manic episodes, nevertheless they are at times prescribed in cases of acute mania in clinical practice [33, 78]. Within our sampling, antidepressants were prescribed to certain bipolar II patients suffering from current hypomanic symptoms with “mixed” depressive symptoms or with rapid cycling. In both of these conditions, antidepressant treatment is specifically contraindicated .
Non-concordant treatment with guidelines observed in our sample can be explained in different ways. First, the diagnosis of BD, especially type II, can be difficult at the onset of the disorder. In most cases, BD-II patients consult their doctor during the depressive phase, not during the hypomanic phase. Identifying the hypomanic episodes in BD-II can be difficult if appropriate questions are not asked, or if family members are not properly interviewed. Thus, before it is finally recognized, BD can be diagnosed as a recurrent depression over several years [79–82]. Some of the patients in our sample were misdiagnosed as having recurrent major depressive disorder. The second possible reason is that treatment with antidepressants, administered during episodes of depression or for purposes of preventing further episodes, is not suspended during a hypomanic phase, either because patients do not require medical intervention or because symptoms are not recognized. There is no consensus in the current guidelines on criteria governing time frames necessary to interrupt antidepressants , which can create confusion as to the appropriate discontinuation of antidepressants following depressive episodes. It is noteworthy that the CANMAT guidelines do not make explicit recommendations for BD-II hypomania, leaving some potential confusion for the reader whether recommendations for acute mania have to be adopted. Patients not adhering to treatment with mood stabilizers [83–85] is a third possible factor which explains the frequency of treatment with antidepressants, especially in monotherapy. Finally, there remains the possibility of general practitioners not following guidelines for lack of awareness, lack of agreement, or lack of familiarity.
Although the study focused on patients seen between the end of 2006 and beginning of 2009, we decided to use the 2009 guidelines rather than 2007, as some of the recommendations adopted in 2009 were already suggested in the previous version. For example, the possibility to use antidepressant monotherapy in bipolar II patients in depressive phase was already described in 2007, and more explicitly formulated in 2009. Other new recommendations adopted in the 2009 CANMAT version, which could potentially affect the percentage of concordance with guidelines in our sample, concerned with the use of quetiapine monotherapy as first line recommendation of BD-II depression and in BD-I maintenance; the use of divalproex monotherapy as second line recommendation in BD-I and BD-II depression; and the use of quetiapine in combination with lithium or divalproex in BD-I maintenance. In our sample, only two treatments would have been classified in a different way if we had adopted the 2007 guidelines: two BD-II depressive patients taking divalproex monotherapy would have been classified as “doesn’t follow guidelines”, instead of “second line treatment”.
Our study, however, dealt with limitations. Our sample was relatively small, most notably as far as stratification was concerned. Also, we had to contend with a strong representation of depressed patients, which corresponds to a higher prevalence of depressive symptoms in the bipolar population.