Fluvoxamine reduced fatigue severity compared with placebo among patients with long COVID, while metformin showed no evidence of benefit, according to a randomized, placebo-controlled adaptive trial published in Annals of Internal Medicine.
The REVIVE-TOGETHER trial enrolled 399 patients across 22 outpatient sites in Brazil between October 2023 and February 2025. Eligible patients had prior SARS-CoV-2 infection confirmed by testing, clinician diagnosis, or self-report when testing was unavailable, along with new or worsening fatigue persisting 90 to fewer than 365 days following acute illness. All patients had moderate to severe fatigue, defined as a mean Fatigue Severity Scale (FSS) score of at least four.
Patients were randomly assigned in a 1:1:1 ratio to receive fluvoxamine, metformin extended-release, or placebo for 60 days. The primary outcome was change in FSS score from baseline to day 60.
Key efficacy findings
- Fluvoxamine reduced fatigue scores compared with placebo at day 60 and day 90
- Metformin showed no evidence of benefit compared with placebo
- Recovery (FSS score < 3) was more frequent with fluvoxamine than placebo at all-time points evaluated
At day 60, patients who received fluvoxamine had greater reductions in fatigue scores compared with those who received placebo, with sustained benefit at day 90. Metformin did not show improvement at either time point. Early stopping of the metformin group reduced precision and a clinically important benefit cannot be excluded.
The Data Safety and Monitoring Committee recommended stopping the metformin group early for futility in September 2024 and stopping the fluvoxamine group for efficacy in February 2025. Two interim analyses were conducted.
Baseline characteristics were similar across groups. The median age was 44 years, and most patients were female. Median baseline fatigue scores were approximately six across treatment groups.
On secondary outcomes, fluvoxamine improved health-related quality of life at days 30, 60, and 90. Metformin showed a small improvement at day 30 that did not persist.
Recovery occurred more often among patients who received fluvoxamine compared with those who received placebo at days 30, 60, and 90. Metformin did not show consistent benefit across these time points.
Prespecified subgroup analyses showed consistent benefit with fluvoxamine across age, sex, vaccination status, obesity, and diabetes subgroups, with no meaningful effect modification. Metformin results were generally consistent with no effect.
In exploratory analyses using the anxiety/depression dimension of the EQ-5D-5L, baseline symptoms did not appear to modify the treatment effect of fluvoxamine on fatigue. Patients with major depressive disorder or those taking antidepressant medications were excluded, although patients with depressive symptoms without a formal diagnosis were included.
Adverse events occurred in 20% of patients who received fluvoxamine, compared with 29% of those who received metformin and 30% of those who received placebo. Serious adverse events were uncommon across all groups, and no treatment-related deaths were reported.
The trial used a Bayesian adaptive design with analyses adjusted for baseline fatigue score, site, age, and sex. Missing data were addressed using multiple imputation. All analyses were conducted in the intention-to-treat population.
The researchers noted several limitations. All patients were enrolled in Brazil, which may limit generalizability. Fatigue was measured using a self-reported scale and may be influenced by patient perception or mood. The study did not assess biological markers or structured psychiatric outcomes, leaving underlying mechanisms unclear. In addition, the 90-day follow-up period limits conclusions about long-term effects.
The researchers concluded that fluvoxamine may reduce fatigue and improve quality of life among patients with long COVID and called for further trials to confirm these findings and identify which patients are most likely to benefit.
Disclosures The trial was funded by the Latona Foundation, which had no role in the design, conduct, analysis, or submission of the study. Disclosure forms are available with the article online.
Source: Annals of Internal Medicine
