Meta-Analysis: Do Electroencephalographic Biomarkers Predict Treatment Response in Depression?
The meta-analysis consisted of a review of 53 articles covering 57 distinct biomarkers.
At present, there is insufficient evidence to support the predictive use of qualitative electroencephalographic (QEEG) biomarkers for treatment of depression, according to a meta-analysis published in the American Journal of Psychiatry.
The analysis indicates that, in contrast to its commercial promotion, QEEG, as has been studied to date, is not well supported as a predictive biomarker for cases of depression. "Inappropriate use of invalid 'predictive' tests could easily increase health care costs without benefiting patients," the researchers noted. As it stands, the use of research-grade or commercial QEEG in routine clinical practices "would not be a wise use of health care dollars."
The meta-analysis consisted of a review of 53 articles covering 57 distinct biomarkers. Considering them all together, analysis of the biomarkers produced predictive power greater than chance, implying that, in general, QEEG may be a viable way to predict treatment response for depressive illness. The meta-analytic estimate of sensitivity was 0.72 (95% CI, .67-0.76), specificity was 0.68 (95% CI, 0.63-0.73), and log(diagnostic odds ratio) was 1.89 (95% CI, 1.56-2.21). However, funnel-plot analysis suggested that the apparent predictive power QEEG demonstrated was largely driven by small studies with stronger positive results, a lack of publication of many weak or negative studies, and insufficient cross-validation of results.
The researchers noted that as a result of designing the meta-analysis for maximum sensitivity, the results are overly optimistic. In contrast, the omnibus treatment of QEEG could obscure its ability to successfully predict a single specific treatment or response in a well-defined subpopulation. In addition, studies were not considered negative if they found significant change in the "wrong" direction. Finally, because of the heterogenous nature of depression, the primary studies' lumping together of many different neurobiological entities makes the rationale for QEEG-based prediction less clear.
The researchers stated that there is still plenty of work to be done understanding the predictive powers of QEEG, and although research is currently insufficient to support its clinical use in this way, they advocate for its continued study.
Widge AS, Bilge MT, Montanta R, et al. Electroencephalographic biomarkers for treatment in response prediction in major depressive illness: a meta-analysis [published online October 3, 2018]. Am J Psychiatry. doi: 10.1176/appi.ajp.2018.17121358