Mindfulness, antidepressants match on worsening depression risk

Clinical takeaway: MBCT has some guideline support for relapse prevention in recurrent depression, but what it lacked was controlled safety data. For the remitted patient considering it in place of medication, randomized comparison found no added risk of symptom worsening. 

Patients in remission from recurrent depression may have read that mindfulness can backfire. That concern has a real basis. Two recent studies documented patient-reported harm during mindfulness-based cognitive therapy (MBCT), but neither included a control. Without one, there was no way to separate what the treatment did from what recurrent depression does on its own. 

Recurrent major depressive disorder is a relapsing illness, and relapse prevention is where most of the long-term management effort goes. MBCT earned a place in that work on efficacy grounds. UK guidelines recommend it outright, and the VA/DoD depression guideline suggests it as one of three psychotherapy options during the continuation phase. Safety, however, remained untested against a comparator, which left clinicians answering patient questions about meditation-related harm from uncontrolled case reports. A preregistered secondary analysis of individual participant data from nine randomized relapse-prevention trials, covering 1,258 adults in remission or partial remission from recurrent depression, supplies the missing comparison. 

About one in three patients assigned to a comparison group, most often antidepressants or another structured therapy, reported worsened depressive symptoms by the end of treatment, compared with roughly one in five who received MBCT. In raw terms that is 33.3% versus 22.5%, a gap of roughly 11 percentage points, or about one fewer patient worsening for every nine treated. The preregistered primary model did not find that difference statistically significant. 

The medication comparison held up on its own. Patients randomized to MBCT worsened at rates statistically indistinguishable from those continuing antidepressants, and the same held across all active controls, which also included cognitive psychoeducation. Only four trials informed the antidepressant analysis, and the range of plausible results ran from a substantial advantage for MBCT to a modest disadvantage. The defensible reading is absence of added risk rather than demonstrated equivalence. 

Analytic method shaped how strong the finding looked. The preregistered approach left the overall difference short of significance, while a second modeling approach applied to the same data crossed it. Results varied considerably across the nine trials, and the two approaches are known to diverge when that variation is high. Three separate definitions of worsening produced a consistent pattern, which strengthens the null more than it supports a benefit. 

The pooled analysis drew individual participant data from nine randomized MBCT relapse-prevention trials, the same set behind the 2016 meta-analysis that established the treatment's efficacy. Participants were 1,258 adults in remission or partial remission from recurrent depression, mean age 47, three-quarters women. Worsening was defined as a within-patient rise on the Beck Depression Inventory by the end of treatment, with the analysis plan registered in advance. 

Two limits shape how far the finding travels. Every trial enrolled patients already in remission, so nothing here speaks to mindfulness training during an acute episode. And the outcome was movement on a depression scale, which does not capture the anxiety, dissociation, and distress that the meditation-harm literature describes. 

The authors call for randomized trials that assess a broader range of adverse experiences. Symptom worsening on a depression scale was the only harm endpoint the pooled trials offered, and the experiences that raised the concern, including dissociation and acute anxiety during practice, have never been measured against a control arm. Until they are, MBCT does not appear to push remitted depression in the wrong direction, and anything beyond that is uncharacterized rather than ruled out. 

Source: Goldberg SB, et al. (2026 Aug 20) JAMA Netw Open. Symptom Worsening in Mindfulness-Based Cognitive Therapy