CYP2D6 poor metabolisers were more likely to abandon venlafaxine

Design

Retrospective cohort study in two independent naturalistic cohorts

Population

5,443 genotyped patients in Norway and 3,624 UK Biobank participants on venlafaxine

Primary outcome

Switch to another antidepressant within 1 year

Effect

CYP2D6 poor vs normal metabolisers OR 2.05 (95% CI 1.53–2.71) Norway; OR 1.58 (1.14–2.19) UK Biobank

This retrospective study used two naturalistic cohorts: 5,443 genotyped patients from a psychopharmacology centre in Oslo and 3,624 UK Biobank participants with venlafaxine prescriptions. People on strong CYP2D6 or CYP2C19 inhibitors or enzyme inducers were excluded. Treatment failure was defined as switching to another antidepressant within a year.

In the Norwegian cohort, CYP2D6 poor metabolisers had about twice the odds of switching compared with normal metabolisers (OR 2.05, 95% CI 1.53 to 2.71); intermediate (OR 1.25) and ultra-rapid metabolisers (OR 1.78) were also raised. The UK Biobank cohort repeated the poor-metaboliser finding (OR 1.58, 1.14 to 2.19). People who were poor metabolisers for both enzymes had the highest odds (OR 6.11), though from small numbers and a wide interval (1.81 to 18.77).

Switching is a proxy: it captures side effects, non-response and patient preference together. The finding supports checking genotype where it is already known, and therapeutic drug monitoring where venlafaxine is failing, but routine pre-emptive genotyping has not been tested here as a strategy, and testing access in India is limited.

• Check the record for a known CYP2D6 result before starting venlafaxine; poor metabolisers switched more often.
• Consider a venlafaxine plus O-desmethylvenlafaxine level when response is poor or side effects are early and heavy.
• Look for co-prescribed CYP2D6 inhibitors such as fluoxetine, paroxetine or bupropion, which mimic poor metaboliser status.
• Treat routine pre-emptive genotyping as unproven; this study did not test it as a strategy.
• Record the reason for any switch, so that tolerability and non-response can be told apart.

Why it matters

A drug judged ineffective may in some patients be one that was never reaching the intended exposure.

Don't overread it

This was observational, and switching mixes side effects with non-response; it does not show genotype-guided dosing improves outcomes.

The statistics, in plain English

An odds ratio of 2.05 means the odds of switching were about double in poor metabolisers. That two separate cohorts found the same direction makes chance less likely. The dual poor-metaboliser estimate of 6.11 has an interval from 1.81 to 18.77 — the true effect could be modest or very large, which is what happens when a group is small.

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.