Switching concerns: Bipolar disorder and the antidepressant dilemma

In a serene, picturesque town, there lived a devoted gardener named Lily. Known for her vibrant and diverse garden, Lily was cherished by her community for her meticulous care of her plants, which brought a splash of color and joy to everyone around. One summer, inspired by an article in a gardening magazine, Lily decided to experiment with a new type of fertilizer reputed to enhance the vibrancy and speed of her flowers' bloom. Initially, the results were spectacular—her garden transformed into an extraordinary display of vivid hues, earning widespread admiration from her neighbors. However, as days passed, an unexpected and troubling pattern emerged. Some plants grew uncontrollably, sprawling beyond their intended spaces, while others, previously healthy, began to wilt and die. Lily, distressed and bewildered, questioned whether the new fertilizer was to blame or if other factors, like the unusual summer weather, were at play.

This scenario mirrors the complexities faced in psychiatry when treating depressive episodes with antidepressants. Just as Lily's garden experienced unforeseen consequences, patients treated with antidepressants may experience shifts from euthymia to manic symptoms, potentially leading to a diagnostic transition from unipolar depression to bipolar disorder. In cases of bipolar disorder, the use of antidepressants potentially carries the risk of switching patients from depression through euthymia to hypomania or mania. The causal relationship between antidepressant treatment and these mood changes remains a topic of ongoing debate.¹ Clinicians often grapple with whether these changes are direct effects of the medication or if they reflect a natural progression and fluctuations of the mood disorder itself. Are antidepressants directly accountable, does the treatment more frequently result in euthymia and perhaps a greater risk of mania or are the antidepressant treatment not linked to switch of polarity? Tools like the Naranjo scale can help assess causality by evaluating the emergence of symptoms in relation to timing of treatment initiation, dose escalation, or recurrence after repeated administration, alongside side-effect symptom reduction following discontinuation or dose reduction.² Randomized controlled trials (RCTs) on antidepressants for unipolar depression show significant improvements in symptoms, particularly in severely affected patients.^{3, 4} For bipolar disorder, RCTs have demonstrated the efficacy of treatments such as olanzapine plus fluoxetine, quetiapine, olanzapine, lurasidone, lumateperone, cariprazine, and lamotrigine on depressive symptoms during a depressive episode.⁵ Given the DSM-5 classification of bipolar disorder into types I and II based on the occurrence of mania, treatment responses and primary end-point of treatment may vary accordingly, with a greater focus on depressive symptoms in bipolar II and with a divided focus on minimizing depressive symptoms and avoiding manic episodes in bipolar I.

Studies have shown that antidepressants can effectively manage depressive symptoms in bipolar II disorder without a significant increase in the rate of switching to hypomania.^{6, 7} Long-term antidepressant treatment in bipolar patients has also shown a reduced risk of recurrent depressive episodes without a significant increase in manic or hypomanic episodes, with data again being most robust for bipolar II.⁸ In contrast, RCTs do not support a clinically meaningful immediate effect of antidepressant on depressive symptoms in patients with bipolar disorder in a meta-analysis including both bipolar I and II.⁹ Thus, while treatment with antidepressants remains an option, guidelines often recommend alternative treatments first, due to more substantial and prolonged data supporting their efficacy.¹⁰

Approximately 40% of patients with bipolar disorder present with depression as their initial affective episode resulting in a unipolar depressive disorder diagnosis.¹¹ Furthermore, a significant number of patients, particularly those with bipolar II disorder not presenting, not being questioned concerning or negatively affected by the hypomanic episodes, are often misdiagnosed with major depressive disorder, leading to a considerable diagnostic delay.¹² Consequently, given that antidepressants are the first-choice pharmacological treatment for patients with unipolar depressive disorder, many patients ultimately diagnosed with bipolar disorder have been exposed to antidepressants before the diagnosis of bipolar disorder, providing insight into the efficacy and tolerability of these treatments and the associated risk of switching for that individual. This historical exposure complicates the interpretation of RCTs on antidepressants efficacy in patients with bipolar disorder, as these studies most often include patients for whom there is clinical uncertainty about the suitability of antidepressant treatment, thereby excluding the group of patients previously exposed resulting in reduced generalizability of the results from RCTs.

In clinical practice, antidepressants are frequently used in patients diagnosed with bipolar disorder, and real-world data supports their tolerability and the absence of manic symptoms necessitating hospitalization, especially when combined with a mood stabilizer.^{13, 14} The descriptive nature of psychiatric classification is designed to minimize biases related to personal views on the significance of various etiological, psychological, psychodynamic, or familial factors, thereby enhancing diagnostic reliability among clinicians. Thus, studies like that by Tondo et al.¹⁵ may have limited implications for diagnostic practices but underscore the importance of recognizing the relatively low conversion rate from unipolar depressive disorder to bipolar disorder, and the variables associated with conversion. Tondo et al. utilized a sample of 3212 patients initially diagnosed with unipolar depressive disorder, followed over a mean of 12.7 years. They found that 6.69% of participants experienced a diagnostic change to bipolar disorder, predominantly to bipolar II.¹⁵ The long period on antidepressant treatment could suggest that the conversion is more likely related to the natural course of the condition rather than being a direct treatment-emergent side effect, although the temporal relationship between antidepressant exposure and the onset of (hypo)manic symptoms remains unclear. The study also highlights that those who experienced a diagnostic change had higher rates of rapid cycling of depressive episodes, unemployment, and mood-stabilizing treatment, with females more often having had post-partum depressive episodes and were younger age at onset of the affective disorder among other factors. These findings align with Angst et al.,¹⁶ who noted that conversion to bipolar I was associated with male sex and early onset, while conversion to bipolar II was linked to female sex, later onset, and a family history of mania. While Tondo et al. study¹⁵ does not directly alter diagnostic practices or predict individual patient outcomes, it underscores the nuanced risk of conversion over time. The study's insights can guide treatment choices, particularly in patients with early-onset depression, a family history of mood disorders, or those presenting with a postpartum depressive episode or previous psychotic depressive episodes. In such cases, augmenting treatment with lithium or an antipsychotic drug may be prudent, rather than choosing switching antidepressant medication as the next treatment choice, to reduce the long-term risk of mania.

In conclusion, the complex interplay between antidepressant treatment and mood disorders necessitates careful consideration and individualized treatment strategies. In particular, data supporting the use of antidepressants in bipolar II disorder is compelling. Both RCTs and real-world evidence indicate that the risk of switching to hypomania or mania is minimal, especially when appropriate adjunctive therapies are used. Studies have demonstrated that antidepressants can provide significant relief from depressive symptoms in bipolar II patients, contributing to an overall improved quality of life and reduced recurrence of depressive episodes. Furthermore, real-world data corroborates these findings, showing that antidepressants are well-tolerated and do not significantly increase the risk of manic episodes requiring hospitalization, especially when combined with a mood-stabilizer.

REN has received funding for research or been an investigator for Lundbeck Pharmaceuticals, Otsuka Pharmaceuticals, Compass Pharmaceuticals, Sage, Boehringer-Ingelheim and Janssen-Cilag. He has received speaking fees from Bristol-Myers Squibb, Astra Zeneca, Janssen Cilag, Lundbeck Pharmaceuticals, Otsuka Pharmaceuticals, Teva and Eli Lilly, and he has been part of an advisory board for Astra Zeneca, Eli Lilly, Lundbeck Pharmaceuticals, Otsuka Pharmaceuticals, Takeda, Janssen-Cilag and Medivir.