Rapid cycling: Are we pedalling in the right direction?

The term rapid cycling bipolar disorder was coined in 1974 by Dunner and Fieve to differentiate those patients in their study with a ‘classic course’ of manic-depressive illness, in which episodes are typically separated by 1.5–3 years of euthymia, from those patients experiencing episodes more frequently. They chose to define rapid cycling as the occurrence of four or more episodes in the past year because this number gave them a suitable sample size. Twenty years later, this theoretical and admittedly arbitrary definition was promulgated clinically upon its incorporation into the Diagnostic and Statistical Manual of Mental Disorders (4th Edition; DSM-IV), which also stipulated that episodes must meet minimum duration criteria (2 weeks for depression, 1 week for mania, 4 days for hypomania), and that individual episodes should be separated by an intermediate period of full remission or a switch to an episode of another polarity. This same definition was retained 19 years later in the Diagnostic and Statistical Manual of Mental Disorders (5th Edition; DSM-5), with additional clarification that intervening periods of remission must be of minimum 2 months duration, and it is this definition which remains in use today. This means that rapid cycling has been operationally defined for close to half a century but, remarkably, it is still unclear what this specifier means in clinical practice and indeed whether it is a useful construct for research.

Certainly, when first conceived it seemed to be a useful notion, and it was a reasonable scientific initiative to ‘recognise’ a rapid cycling variant of bipolar disorder for the purposes of further investigation, with the aim of potentially identifying an important subtype of the illness. However, in practice, the definition has been problematic to apply and its use in research has been sparse. The main reason for this is a lack of specificity, which is a product of two key problems.

First, the frequency criterion of four or more (⩾ 4) episodes per year does not capture any information with regard to the specific number of episodes (e.g. 4, 5, …, 9) that occur in the space of the year nor their particular pattern during this time, such as the duration of each episode and the spacing between them. Remark-ably, the criterion does not even specify the polarity of the episodes – mania, depression or mixed. This means that the specifier of rapid cycling can be applied equally to an individual who has experienced four manic episodes of 1-month duration separated by 2-month periods of remission, and an individual who switches continually between mania and depression every 2 weeks for several months. Similarly, an individual who experiences four alternating episodes of mania and depression at the beginning of a 12-month period followed by a sustained period of remission also qualifies as rapid cycling. These variable clinical patterns, along with many others, are clearly very different and suggest that diverse and possibly distinct aetiological processes are at play; and yet, the information that may help explain such variation is not captured.

Second, the stipulation for episodes to satisfy full duration criteria is restrictive because it means that those individuals who cycle over short periods of time, such as the weekly or daily alternation that occurs in ultra-rapid and ultradian cycling, cannot be diagnosed as rapid cycling. Research has shown that these groups are more similar to DSM-defined rapid cycling than they are to non-rapid cycling bipolar disorder, and yet they are not included in studies of mechanisms and treatment effects.

Ultimately, these issues with the diagnostic criteria for rapid cycling mean that knowledge concerning clinical variability that could potentially inform our understanding of why rapid cycling occurs and provide insights into how best to treat such presentations is ignored. This feeds into a vicious cycle, where a better appreciation of rapid cycling is needed to refine its definition, but the current definition is so meaningless that it hinders any deeper examination. Indeed, the definition fails to sufficiently inform researchers wishing to pursue Feighner et al.’s (1972) framework for validation of diagnoses on the basis of basic clinical description, exclusion of other disorders, follow up, and more sophisticated family or laboratory studies. This is perhaps why one review of the DSM-IV definition of rapid cycling concluded that, although the current definition was clearly arbitrary and lacked empirical support, there was insufficient evidence to promote a different model (Bauer et al., 2008). Surely this has to change?

The problems with rapid cycling’s definition stem from the fact that it is not based on an understanding of biology. Research examining the mechanisms of rapid cycling is virtually non-existent, although an emerging body of evidence implicates circadian rhythm dysfunction, and other, albeit as yet inconclusive research, suggests the involvement of specific genes (e.g. serotonin transporter and brain-derived neurotrophic factor) and dysregulation of the hypothalamic–pituitary axis (Bauer et al., 2008; Bellivier et al., 2015). Nevertheless, the specific processes that initiate and maintain rapid cycling are not known, and as a result its definition is necessarily arbitrary.

Conceptually, if we consider rapid cycling in the context of recurrent mood disorders, it seems intuitively to be a different kind of problem. A crude but useful analogy is to consider a water heating system, with a source of heat and a thermostat that regulates the temperature. Mood disorders in general can be viewed as the result of emotion processing problems, with episodes of illness occurring because of dysfunction in key neural networks that provide emotional context to our reasoning and experiences. In our analogy, this corresponds to problems with the heat generating system. However, in rapid cycling, it is likely that the principal problem lies with regulatory control of emotion, i.e., the thermostat. Thus, models that consider rapid cycling as part of the spectrum of mood disorders that extends from unipolar depression through to mixed states and bipolar disorder (i.e. as a more severe variant of bipolar disorder) are inapt. Rather, rapid cycling indicates the occurrence of a separate set of abnormalities involving specific processes concerned with the timing of mood episodes and their onset and offset. We can speculate that this is likely to involve regulatory networks and centres within the brain and, indeed, that dysfunction in these may stem from different aetiologies to those causing mood disorders more generally. However, to confirm this, rapid cycling must be defined with much greater precision.

To address this issue, we propose that rapid cycling be considered with greater specificity in research and clinical practice. Essentially, this means fully characterising participants and patients according to the exact number, polarity, duration and frequency of their episodes. Mapping the patterns observed will then allow for discernment of naturally occurring subsets of variations as well as the examination of the biological mechanisms involved and how these may differ between such subsets. For instance, there may be a fundamental difference between patterns of cycling that occur over periods of a day or several days as opposed to those that occur over weeks and months, or these patterns may simply be points on a continuum of severity. Furthermore, the pattern of oscillating between both poles may signify a different problem to that of frequent episodes of one polarity. Collecting specific information will allow us to answer such questions, and it is important that considerable investment and a concerted effort are made to inculcate this approach.

Recent resurgence of interest in mixed states perhaps offers another approach to furthering our understanding of rapid cycling as it is not clear how rapid cycling differs from mixed states, and whether the two phenomena represent separate pathophysiological processes. In fact, in DSM-5, ultra-rapid and ultradian cycling are diagnosed as mixed states or bipolar disorder not otherwise specified – not as rapid cycling and, historically, these shorter frequency cycling patterns have been referred to as ‘unstable’ mixed states. Similar processes, such as circadian dysfunction, appear to be involved in both (Malhi et al., 2019), with further evidence for shared mechanisms stemming from the fact that both mixed states and rapid cycling can be induced or exacerbated by treatment with antidepressants, while the mixed state of agitated depression has explicitly been associated with rapid cycling (Bauer et al., 2008). Thus, determining the mechanisms underlying mixed states could improve our understanding of rapid cycling.

As a final note, it is important that the current definition of rapid cycling that has failed to advance our understanding is addressed due to its clinical importance and prevalence. Indeed, prevalence estimates of rapid cycling are between 13% and 20% of patients with bipolar disorder (Bauer et al., 2008), and such patients are more likely to have poorer functioning than those with non-rapid cycling bipolar disorder, including more attempts at suicide and abuse of drugs and alcohol. Sadly, those with rapid cycling are less responsive to the treatments currently available for those with bipolar disorder (Bauer et al., 2008). It is thus essential that the definition of rapid cycling is improved to facilitate research into understanding its mechanisms and improving its treatment.