Comparative Effectiveness of Pharmacogenomics for Treatment of Depression.

Proprietary combinatorial pharmacogenomic test returning an interpretation of gene-drug interactions to the clinician to guide antidepressant selection.

This trial compared, in patients with major depressive disorder, combinatorial pharmacogenomic-guided plus guideline-informed treatment with guideline-informed treatment alone. The 201 eligible participants were randomised between the two arms, clinicians in the pharmacogenomic arm receiving combinatorial test results within two business days to inform medication change decisions. Well-being (WHO-5), depression severity (PHQ-9) and physical and social functioning (PROMIS) were measured every two weeks for two months, then every two months for the remaining ten months. Both groups improved their average well-being over 12 months (model-based change in WHO-5 per log(week): 4.1 [95% CI 3.3-5.0] with testing versus 4.8 [4.0-5.5] without), with no superiority for the pharmacogenomic arm (model-based difference -0.6 [-1.8, 0.5], p = 0.270) or on any secondary outcome. The effect of randomised treatment was not moderated by depression severity, number of previously failed medications, or presence of a comorbid condition; the authors raise a possible ceiling effect of guideline-informed treatment.

Synthesis written by Geno'X. For the full original abstract, please refer to the source publication.

The comparator is what makes this trial valuable: the test is not pitted against ordinary care but against genuinely guideline-informed treatment, and it adds nothing to it over twelve months. The absence of moderation by number of previously failed medications removes the last argument for compassionate use of the test in treatment-resistant patients, an argument nonetheless often heard in clinic. Caution remains on scope: with 201 patients the trial is underpowered for a small effect, and the result condemns a proprietary combinatorial panel, not the individually documented gene-drug pairs such as CYP2C19 and escitalopram.

Analysis by Dr Thibaut Benquey