In 1974 David L. Dunner and Ronald R. Fieve reported the results of a study conducted in the Lithium Clinic of the New York State Psychiatric Institute with the aim of identifying clinical factors that predicted recurrence of affective episodes in lithium-treated patients with bipolar disorder.¹ The study included 55 patients who received lithium for a period of 6–66 months. Notably, the patients having frequent episodes of mania and depression immediately prior to their acceptance into the clinic were termed “rapid cyclers” if they had four or more affective episodes per year. A total of 27 (49%) of the patients experienced at least one depressive or hypomanic episode during the treatment period. Recurrences occurred disproportionately frequently in those who were rapid cyclers: of the 11 lithium-treated rapid cyclers, 9 (82%) had recurrences, whereas this was the case for only 18 (41%) of the 44 lithium-treated non-rapid cyclers. Other clinical factors that could potentially be associated with lithium prophylaxis failure were not identified in the overall study population. Importantly, serum lithium levels were comparable in those experiencing recurrence and those who did not.

The concept of rapid cycling as coined by Dunner and Fieve, although unresolved issues regarding its clinical correlates remain,² still stands today and is included in both the DSM-5 and the upcoming ICD-11 as a course specifier, defined by the occurrence of four or more distinct mood episodes within a 12-month period. Clinical characteristics associated with rapid cycling include a high proportion of women and a younger age of onset.³ Rapid cycling, while in some a more persistent state, is in most patients a transitory phenomenon.² For example, in the Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD) study it was found that of 359 patients with rapid cycling bipolar disorder, only 58 (5%) could be classified as rapid cyclers at the end of 12 months follow-up.⁴ As such, evidence indicates that rapid cycling does not constitute a discrete subtype of bipolar disorder but rather represents a transitory, severe presentation of the illness. Rapid cycling is thus associated with an increased total number of episodes,⁵ increased time spent in an affective episode,³ higher rates of comorbidity,⁶ higher rates of suicidal behavior,^7-9 lower levels of functioning,⁵ a greater symptom severity,⁴ and a worsened long-term illness course.³ Among patients with bipolar disorder, the point prevalence estimates of rapid cycling range from 5% to 33.3%, while lifetime prevalence estimates range from 25.8% to 43%.²

Based on the results of their landmark study, Dunner and Fieve suggested that lithium treatment is continued in non-rapid cycling patients who experience an affective episode during lithium treatment. Also, for several years to follow, partly based on the findings obtained by Dunner and Fieve, it was suggested that treatment with anticonvulsants might yield more favorable outcomes than treatment with lithium for prophylaxis in rapid cycling.¹⁰ Later evidence, however, based on largely open label studies indicated that lithium is not less effective than e.g. anticonvulsants for prophylaxis in patients with rapid cycling, while the prophylactic effect of lithium and anticonvulsants may generally be lower in patients with rapid cycling compared with non-rapid cycling patients.¹¹ A further consideration is that treatment with antidepressants might further destabilize the illness course,¹² which is of particular importance in rapid cycling, with practice guidelines advising against their use - or to use them cautiously.¹³

An important question thus remained: which treatments are optimal for patients with rapid cycling bipolar disorder? And, perhaps of particular importance, whether antidepressants should have a role in the treatment of depressive episodes and which relapse prevention treatment is more effective? There is unfortunately no clear consensus regarding the optimal treatment for rapid cycling, and the available evidence has not been quantitatively synthesized. To address this unmet need, Strawbridge et al., for the first time, conducted a systematic review with meta-analysis of treatments for rapid cycling bipolar disorder, published in the current issue of Acta Psychiatrica Scandinavica.¹⁴ They included a total of 30 randomized trials comprising 2266 patients with rapid cycling and analyzed the observed outcomes for each treatment arm separately as pooled estimates across studies, thus providing estimates of the change in illness severity within patients before and after treatment. While such a within-subject approach has the drawback that estimates do not directly reflect the effect of the treatment, because it is not possible to disentangle which proportion of the change was caused by the intervention or other factors,¹⁵ it allowed the authors to include more studies. Strawbridge et al. included 23 trials in their quantitative analysis, in which they compared the confidence intervals of the pooled change scores in the intervention arms with those of the control arms and considered instances where the confidence intervals did not overlap as an indication of an effect of the intervention. Their findings are telling of the still limited evidence for treatments in rapid cycling. Strawbridge et al. found that all antipsychotics appeared to be superior to the control interventions of placebo or treatment as usual in reducing manic symptoms, but there was a stark paucity of evidence for mood stabilizers: only one of the included studies investigated valproate, and none investigated lithium, which both have established efficacy in treating manic episodes.¹³ Neither antipsychotics nor mood stabilizers as a class appeared clearly more effective than placebo or treatment as usual for depressive episodes; the only individual antipsychotic with apparent efficacy was olanzapine (or the olanzapine/fluoxetine combination), while no individual mood stabilizer appeared to have clear efficacy in depression. Conversely, antidepressants as a class more clearly appeared effective in reducing depressive symptoms, with the confidence interval for the pooled estimate of within-subject changes during treatment barely overlapping with that of control arms. Despite this finding, the data presented by Strawbridge et al. does not seem to allow definitive conclusions regarding the potential role of antidepressants in the treatment of depression in the context of rapid cycling. This is particularly due to the absence of evidence regarding the risk of manic switch, which could not be assessed in the review due to lack of data, and as noted by the authors, a lack of long-term studies that would allow for estimating the potential risk of cycle acceleration posed by treatment with antidepressants. Also, there was no clear evidence of superiority of antidepressants over antipsychotics or mood stabilizers. Regarding relapse prevention, Strawbridge et al. did not calculate pooled estimates for within-subject effects of any drug or class, likely due to study heterogeneity and potentially because of scarcity of evidence. Individual relapse prevention trials in solely rapid cycling bipolar disorder patients have not demonstrated clear superiority of any intervention. The largest trial, which included a total of 182 patients with rapid cycling bipolar disorder, did not find a statistically significant difference in time to additional pharmacotherapy between lithium and lamotrigine.¹⁶ Two smaller trials including 60 and 31 patients with rapid cycling bipolar disorder, respectively, found no difference between either lithium and valproate¹⁷ or between the combination of lithium and valproate and lithium alone¹⁸ in rate or time to relapse. Evidence supporting clear recommendations for relapse prevention in rapid cycling is thus very limited, but few trials involving exclusively patients with rapid cycling bipolar disorder have been conducted. Therefore, most trials (n = 19, 63%) included by Strawbridge et al. included also non-rapid cycling bipolar disorder patients, with rapid cycling patients comprising 37% or less of the total study population in 12 of the trials. Comparison between patients with rapid cycling and non-rapid cycling, respectively, in trials with mixed populations did not provide clear answers regarding optimal treatment strategies in rapid cycling either, although many trials were likely not powered to detect such differences. Specifically, of the 14 studies reporting results for each group of patients, nine did not find statistically significant differences in response between the groups, while four trials found inferior response and one trial superior response, respectively, in patients with rapid cycling.

Notwithstanding the limitations related to within-subject meta-analysis and the inclusion of trials with a substantial proportion of non-rapid cycling bipolar disorder patients, which had the benefit of allowing for inclusion of more studies, the systematic review by Strawbridge et al. provided an important overview of the evidence base for treatment of rapid cycling bipolar disorder. As noted by the authors, no clear answers emerged from their findings, largely due to a scarcity of evidence, an important finding of itself. Their review thus points to an urgent need for further research to establish optimal treatment strategies for patients with rapid cycling bipolar disorder. Given that a primary concern should be the long-term goal of achieving and sustaining remission, trials in patients with acute affective episodes should ideally include sufficiently long follow-up periods to assess not only remission but, importantly, also the impact of treatments on the cycle pattern. Further, trials in patients with rapid cycling bipolar disorder should also consider non-pharmacological interventions and combination therapy, which in bipolar disorder in general is often necessary. Such trials, while difficult and costly to conduct, should be a priority given the high prevalence of rapid cycling and its negative impact on outcomes for patients. In the meantime, the question remains what the optimal treatment strategy in rapid cycling bipolar disorder should be. In the absence of clear evidence to suggest that the treatment of rapid cycling bipolar disorder should differ from that of non-rapid cycling bipolar disorder, the National Institute for Health and Care Excellence (NICE) has recommended that it should merely follow that recommended for bipolar disorder in general.¹⁹ While somewhat unsatisfying given the severity of rapid cycling bipolar disorder, this might currently be the most parsimonious approach to its treatment.