Treatment effect variation in brain stimulation across psychiatric disorders

Noninvasive brain stimulation methods such as transcranial magnetic stimulation (TMS) and transcranial direct current stimulation (tDCS) are promising add-on treatments for a number of psychiatric conditions. Yet, some of the initial excitement is wearing off. Randomized controlled trials (RCT) have found inconsistent results. This inconsistency is suspected to be the consequence of variation in treatment effects and solvable by identifying responders in RCTs and individualizing treatment. However, is there enough evidence from RCTs that patients do indeed respond differently to treatment? This question can be addressed by comparing the variability in the active stimulation group with the variability in the sham group across studies. We searched MEDLINE/PubMed and included all double-blinded, sham-controlled RCTs and crossover trials that used TMS or tDCS in adults with a unipolar or bipolar depression, bipolar disorder, schizophrenia spectrum disorder, or obsessive compulsive disorder. In accordance with the PRISMA guidelines to ensure data quality and validity, we extracted a measure of variability of the primary outcome. A total of 114 studies with 5005 patients were considered in the analysis. We calculated variance-weighted variability ratios for each comparison of active versus sham stimulation and entered them into a random-effects model. We hypothesized that treatment effect variation in TMS or tDCS would be reflected by increased variability after active compared with sham stimulation, or in other words, a variability ratio greater than one. Across diagnoses, we found a slight increase in variability after active stimulation compared with sham (variability ratio = 1.05; 95% CI, 1.01-1.11, P = 0.012). This effect was likely driven by studies in patients with schizophrenia who received rTMS compared with sham (variability ratio = 1.11; 95% CI, 1.03-1.2, P = 0.007). In conclusion, this study found evidence for treatment effect variation in brain stimulation, particularly for studies in schizophrenia. The extent of this variation, however, was modest, suggesting that the need for personalized or stratified medicine is still an open question.