The largest meta-analysis to date of randomized controlled trials of mindfulness-based cognitive therapy (MBCT) for relapse prevention in recurrent depression was published in JAMA Psychiatry recently. Here two of the co-authors on the paper, Catherine Crane and Zindel Segal reflect on its findings and ask “What do we know? What does it mean? Where to next?”

A relatively large number of studies have now examined the effects of mindfulness-based cognitive therapy for people at different stages in the course of a depressive illness. However the new meta-analysis focuses on trials addressing the original intention of MBCT, to prevent relapse in recurrent depression and includes data from nine trials recruiting adults with a history of recurrent depression, comparing the effectiveness of MBCT with a range of comparators in preventing depression relapses over a sixty week period. Unlike previous meta-analyses, which have summarized effects at the study level, this study uses individual participant data. This enables us not only to ask the question, “What is the overall effectiveness of MBCT on depression relapses?,” but also to consider “What works for whom, how do participant characteristics (such as age and gender) influence these outcomes?” The meta-analysis focuses on nine trials of MBCT identified up to November 2014, with data available from 1258 participants. One Australian trial was identified during the search process as eligible for inclusion, but the authors could not release individual study data for ethical/legal reasons.

What do we know?

Is MBCT for recurrent depression effective? What about compared with other (active) treatments?
Consistent with the findings of previous meta-analyses, our work indicates that MBCT provides clear benefit over control conditions (comprising usual care in some studies and active controls in others), through reducing rates of relapse to depression over 60 weeks follow-up. Extending these findings, MBCT also appears to provide comparable benefits to active treatment controls, in general, and antidepressant controls in particular, although the reduction in risk of relapse to depression was larger for the comparison between MBCT and all control conditions than for MBCT and active treatment controls.

For whom is it most effective?
Across the studies there was no evidence to suggest that the relative benefit of MBCT compared to control conditions was significantly influenced by participant age, gender, sociodemographic status, educational level, participant age of onset of depression or number of prior depressive episodes. Unfortunately, information on ethnicity was collected inconsistently, but across studies the majority of participants were Caucasian and so it is not clear whether the benefits of MBCT would be similar in samples with greater ethnic and racial diversity. Participants who were experiencing more depression at entry to treatment showed a greater benefit of MBCT compared with other treatments than those who were less depressed, although those who were less depressed were not disadvantaged by receiving MBCT.

What about safety? Is MBCT safe?
Along with efficacy, it is important to consider the safety indications for MBCT. This is a subject Ruth Baer and Willem Kuyken discuss in a separate blog post. It is also being researched thoughtfully by researchers such as Willoughby Britton and raised by several commentators in the media. The meta-analysis gathered data on serious adverse events either from the original trial papers or if this data were not reported, directly from the authors. All the trials had well-trained MBCT teachers and included careful baseline assessment of patients. As with psychotherapy trials more generally information about safety and adverse events has only been collected more recently and the ways in which this information is collected varies somewhat from trial to trial. Nonetheless, where it was collected these data suggest that rates of serious adverse events were comparable across both MBCT and comparator groups and in no instance was a serious adverse event attributable to MBCT. This suggests that when MBCT is delivered according to the manual by well-trained MBCT therapists it is safe.

In Summary
The results of this meta-analysis are promising and suggest that MBCT can provide a viable relapse prevention intervention for people with a history of recurrent depression. The findings that there was no evidence to suggest a range of factors such as age and gender significantly influenced the effectiveness of MBCT, and that there was no evidence for an association between MBCT and the occurrence of serious adverse events increases our confidence that MBCT is acceptable for a broad range of people with recurrent depression.

The results of this meta-analysis are promising and suggest that MBCT can provide a viable relapse prevention intervention for people with a history of recurrent depression.

So What Does It All Mean? What Are the Remaining Questions

Although the largest meta-analysis of MBCT for depressive relapse prevention to date, the dataset still only reflects a relatively small number of trials. It is important to consider how this research fits into the broader research literature, what we don’t yet know and which questions future research might fruitfully address. Given the prevalence of depression and the fact that the current standard of care for relapse prevention is maintenance medication, there has been a lot of interest in the relative efficacy of MBCT’s preventive effects. This meta-analysis provides evidence that MBCT (combined with antidepressants or delivered alongside antidepressant tapering/discontinuation) is comparable to maintenance antidepressants alone in preventing subsequent relapse. It is, however, important to take into account design differences in the four trials that directly compared MBCT to antidepressant arms.

In a 2010 study (Segal, et al.), participants who had been treated to remission with antidepressants were randomized to either stay on these, were switched from their antidepressants to a placebo (with participants blinded to which of these two arms they were in), or were randomized to discontinue their medication prior to receiving MBCT.  In two other trials (Kuyken et al., 2008; 2015), individuals were randomized to continued antidepressant medication or to receive MBCT plus antidepressant tapering, which occurred during treatment. In the larger trial (2015), tapering or discontinuation of antidepressants occurred in 87% of participants, with 71% stopping antidepressant medication altogether.  Finally a 2015 study (Huijbers et al.) allowed all participants to stay on their antidepressants with half additionally randomized to receive MBCT. In this trial 70% of those randomized to receive MBCT and stay on antidepressants adhered to both interventions. Of those randomized to stay on maintenance antidepressants and NOT receive MBCT, only 60% adhered, with almost a quarter actually accessing and attended 4 or more sessions of MBCT during the trial period.

The results of the meta-analysis tell us that there is a small but significant benefit of MBCT when delivered alongside or as an alternative to antidepressants, in terms of reduced rates of relapse. However each trial employed a different design and recruited participants with different expectations and preferences, with none reaching statistical significance individually in favor of MBCT. This means that we need to be careful about drawing premature conclusions concerning the relative effectiveness of MBCT and medication. We know from clinical experience and qualitative work that people have strong views and preferences around antidepressants. Indeed this powerfully shaped recruitment to the studies, and in the Huijbers study the design was changed from the original intention to run a trial with three arms (MBCT alone, maintenance antidepressants alone, and their combination) to two 2-arm trials to accommodate patient preferences (Huijbers et al., 2015; Huijbers et al., 2016). The findings also cannot address the question many people find themselves asking of whether, as someone who is on maintenance antidepressants and chooses to enroll in MBCT in addition, they would then be better off, or worse off, if they chose to come off their antidepressants, all else being equal.

The results of the meta-analysis tell us that there is a small but significant benefit of MBCT when delivered alongside or as an alternative to antidepressants, in terms of reduced rates of relapse.

This latter issue is addressed by the second Huijbers two arm pragmatic trial, recently published in the British Journal of Psychiatry. The study compared rates of relapse in people who were randomized to receive MBCT and then either stay on or come off their antidepressants. In this trial, those who were randomized to receive MBCT and come off antidepressants fared significantly worse, in terms of relapse to depression, than those who were randomized to receive MBCT and stay on antidepressants i.e. the combination of the two treatment approaches was superior to MBCT alone. This finding is not necessarily contradictory to the findings of the meta-analysis—it addresses a different scenario. It is also complicated by low levels of protocol adherence in both trial arms (52% in the MBCT + discontinuation arm and 56% in the MBCT + maintenance arm) and very variable quality of MBCT training delivered (only half of the MBCT therapists were rated as competent).

The other issue that is relevant to the interpretation of all tapering trials is the possibility that people tapering or discontinuing antidepressants during and immediately after treatment with MBCT may be experiencing withdrawal symptoms after discontinuation of selective serotonin reuptake inhibitors (SSRIs) resulting in elevated symptom levels. This context highlights the complexity of any conclusions we might want to draw from the evidence to date. We certainly should not be claiming that “MBCT is better than antidepressants.” Our view is that this would be premature. Indeed even a large five arm trial comparing all possible MBCT/antidepressant combinations would be unlikely to give a definitive answer to the question of relative superiority because it appears from existing research, including the most recent meta-analysis, that the relative efficacy of MBCT (and hence its likely benefit over alternative treatments) may depend in part on patient preferences and the risk of relapse at the point of entry to the trial (in terms of residual symptoms and/or the presence of other vulnerability factors such as childhood trauma). Such a trial would be so large and difficult to carry out it is unlikely to ever take place. Instead, what is needed is careful reading of the trials in the context of the broader research, both quantitative and qualitative, consideration of patient views, preferences and needs and clinical consensus.

The JAMA Psychiatry meta-analysis did not have the power to consider whether residual symptoms moderated the comparison between MBCT and maintenance antidepressants specifically. Of course we might expect that this would be the case—people showing partial remission (e.g. significant residual symptoms) might be expected to benefit more from an additional, psychological, approach to relapse prevention than those in full remission.  It also did not consider how MBCT compares to other psychological interventions since only a single study, the Staying Well After Depression Trial, compared MBCT to an active psychological control treatment. However there are at least two more recently published studies that speak to this question. The first, Shallcross et al., 2015, compared MBCT to the closely matched Health Enhancement Programme and found no difference in rates of relapse to depression. The second, a trial by Meadows et al., (2014) compared MBCT combined with ‘depression relapse active monitoring’ (DRAM) to DRAM alone. In this study, as in the Staying Well trial, antidepressant use was uncontrolled but balanced across trial arms and the analysis showed no significant benefit of MBCT+DRAM over DRAM alone in terms of time to depressive relapse.

We are at a relatively early stage in our understanding of the mechanisms through which MBCT has its therapeutic effects. That is to say, “What are people actually learning in MBCT and how does this help them stay well in the long-term?” We have done a recent review of this area (van der Velden et al., 2015), which identified a number of potential mechanisms of change including alterations in mindfulness, rumination, worry, compassion and meta-awareness. However there is still a great deal of work to be done to answer these questions and until we do it is difficult to formulate clear hypotheses predicting which individuals might benefit more from MBCT than other psychological interventions. Nonetheless, on the basis of these trials in the round there is no convincing evidence that MBCT is superior to plausible alternative psychological interventions for individuals with a history of recurrent depression as a whole. They are all about as effective (or ineffective) as each other.

There is no convincing evidence that MBCT is superior to plausible alternative psychological interventions for individuals with a history of recurrent depression as a whole. They are all about as effective (or ineffective) as each other.

In many ways this is not surprising. No single treatment approach will be right for all and most psychological treatments will share common features. Indeed whilst we can look at the findings of clinical trials to provide evidence, at a population level, of which treatment approaches might most usefully be provided within clinical services, each person remains an individual and a range of predictive variables will need to be taken into account to optimize treatment planning. Treatment approaches differ not only in their average effects across a population, but also in the relative spread of benefits and harms they produce, what they require in terms of personal commitment, both in the short term and the long term, the specificity of their impact and their consequences for a person’s sense of agency and control over their condition.  We still need to know much more about the individual characteristics that determine the personal benefit to be gained from receiving one treatment rather than (or in combination with) another (e.g. DeRubeis and Colleagues). Additionally we need to know more about the broad range of potential unexpected and unwanted experiences that may arise in response to any psychological intervention, including MBCT, but which would not be classified according to the adverse/serious adverse event categories utilized in most clinical trials.  As for psychological therapies more broadly (e.g. Duggan et al., 2014), the issue of such experiences in response to MBCT requires careful thought and focused research. Our blog on safety of MBCT addresses this question in more depth.

Where to next?

What is notable is that across most, if not all trials of acute phase treatments for depression, a sizable proportion of patients, despite receiving high quality care from well-trained therapists or other medical professionals, suffer relapse/recurrence at unacceptably high rates. With increased understanding of the demographic and clinical history predictors of treatment response and MBCT’s mechanisms of action we will be better positioned to personalize, sequence and optimize care pathways that will, in turn, optimize prevention.