- Design
- Randomised placebo-controlled Bayesian adaptive trial
- Population
- 399 Brazilian adults with fatigue 3–24 months after COVID-19
- Primary outcome
- Change in Fatigue Severity Scale score
- Effect
- Fluvoxamine vs placebo −0.43 at day 60 and −0.58 at day 90; posterior probability of superiority 99%
This Bayesian adaptive trial at outpatient sites in Brazil randomised 399 adults with fatigue starting 3 to 24 months after confirmed COVID-19 to fluvoxamine 100 mg twice daily, extended-release metformin or placebo for 60 days, with follow-up to day 90.
Fluvoxamine lowered the Fatigue Severity Scale score by about 0.4 points more than placebo at day 60 and 0.6 at day 90, from a baseline near 5.7 on a 7-point scale, with a 99% posterior probability of superiority. Quality of life also improved. Adverse events were less frequent with fluvoxamine (24%) than placebo (36%). Metformin had to be reduced in dose for side effects, and its arm was stopped for futility, so its result is inconclusive.
Long COVID fatigue leaves physicians with little to offer. This is a modest effect in a single trial, but from a cheap, familiar drug.
Fluvoxamine is a potent CYP1A2 inhibitor and interacts with theophylline, clozapine, tizanidine and others, so it is not a casual prescription. Benefit beyond 90 days is unknown.
- Consider discussing fluvoxamine as an option for persistent long COVID fatigue, noting the evidence comes from one trial.
- Check the drug list for CYP1A2 substrates such as theophylline, clozapine and tizanidine before prescribing fluvoxamine.
- Do not offer metformin for long COVID fatigue on this evidence.
- Keep pacing advice and screening for depression, sleep disorder and anaemia as the core of assessment.
Why it matters
It gives a first randomised signal for a symptom where most patients currently leave with advice only.
Don't overread it
One 90-day trial cannot show lasting benefit or that the effect holds beyond this population.
The statistics, in plain English
A posterior probability of 99% means that, given the data and the model's assumptions, fluvoxamine was very likely better than placebo. It does not say the benefit is large: about half a point on a 7-point scale is a modest change. Bayesian results should be read alongside the effect size, not instead of it.
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.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for internal medicine, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free