Pharmacogenomic-guided antidepressant choice raised short-term remission with one test, but not across panels

Design

Systematic review and platform-stratified meta-analysis of randomised trials

Population

8 trials in adults with major depressive disorder

Primary outcome

Response and remission at 8 to 12 weeks

Effect

One platform: remission RR 1.53 (95% CI 1.19 to 1.97); other panels: remission RR 1.18 (0.77 to 1.80), I² 82%

Combinatorial pharmacogenomic tests use several genes, mostly CYP2D6, CYP2C19 and others, to sort antidepressants by predicted metabolism and tolerability. This meta-analysis deliberately analysed platforms separately because they differ in genes and decision rules. It included eight randomised trials in adults with major depressive disorder.

For the platform with the most trial data (GeneSight, three trials), guided care increased response (RR 1.30, 95% CI 1.09 to 1.56) and remission (RR 1.53, 1.19 to 1.97) at 8 to 12 weeks, with no heterogeneity. For other multigene panels (four to five trials), results were heterogeneous and not significant: response RR 1.18 (0.91 to 1.54), remission RR 1.18 (0.77 to 1.80).

The authors warn against reading this as one product being better than others; the difference may reflect how much good-quality trial data each has. For prescribers, the honest position is that pharmacogenomic guidance may modestly improve short-term outcomes, mainly in people who have not responded to or tolerated earlier antidepressants, and that the evidence depends on which test is used. CYP2D6 and CYP2C19 results already have established dosing guidance for several antidepressants, and that is the most solid part of any report.

• Consider pharmacogenomic testing for patients with depression who have failed or not tolerated one or more antidepressants.
• Act first on CYP2D6 and CYP2C19 results, where dosing guidance for several antidepressants is established.
• Ask which test and decision rules a report uses; evidence is not interchangeable between panels.
• Do not delay starting treatment while waiting for a result in moderate or severe depression.

Why it matters

It explains why trials of pharmacogenomic testing disagree: pooling different tests blurs any effect.

Don't overread it

Few, short trials with some risk of bias; this does not show one commercial platform is superior.

The statistics, in plain English

A remission risk ratio of 1.53 means about half as many again reached remission with guided care, from a small number of trials. For other panels the intervals cross 1 and I² is around 80%, so results varied too much between trials to draw a conclusion.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.