Long COVID supplement evidence — what the 2025 trials actually show
Long COVID is the hardest article we publish. The pathophysiology is still being mapped, the symptom clusters are heterogeneous, and the supplement industry has been aggressive about marketing into the gap left by the absence of established treatments. This article is a careful read of what the controlled trial literature through 2025 actually supports — and an explicit list of what does not.
The supplements with at least small RCT signal
CoQ10 (or ubiquinol)
200 mg/day ubiquinol or 300 mg/day ubiquinone, with a fatty meal, for 12 weeks
A small Spanish RCT in long COVID published in 2023 reported improvements in fatigue scores at 12 weeks compared with placebo. Effect size was modest, the trial was single-centre, and replication is limited. Mechanism rationale (mitochondrial dysfunction is a leading hypothesis for long COVID fatigue) is plausible. Generally well tolerated. The most defensible "first try" supplement based on current data, but please calibrate expectations.
Omega-3 EPA + DHA (high-EPA leaning)
2 g/day combined EPA+DHA with food
Small RCT and observational data suggest modest improvements in self-reported cognitive symptoms at 2 g/day. Mechanism likely involves resolution of post-acute inflammation and direct neuronal effects. Mind the high-dose AFib paradox if dosing above 1 g/day long-term — see our EPA vs DHA vs ALA comparison. Reasonable to add, particularly if dietary fish intake is low.
Vitamin B12 (high-dose, in deficient or borderline patients)
1,000–5,000 mcg methylcobalamin or hydroxycobalamin daily, sublingual or IM
The autonomic and POTS-overlap subgroup of long COVID is the part of the literature with the most consistent supplement signal. Vitamin B12 deficiency is more common than expected in this population (some of which is explained by restricted diet during prolonged illness, some by absorptive issues). Repletion improves symptoms in deficient patients; benefit in non-deficient patients is unclear. Always test serum B12 and ideally methylmalonic acid before supplementing high-dose long-term.
Melatonin (low to moderate dose)
2–6 mg, 30 min before bed, for 8–12 weeks
A few small trials have reported improvements in sleep and global symptoms at 2 to 6 mg nightly. Mechanism may extend beyond circadian — melatonin has independent antioxidant and anti-inflammatory effects. Generally well tolerated; morning sedation is the main side effect, more common at higher doses.
Honourable mentions — mechanism without trials
The compounds below have plausible mechanistic rationale and small uncontrolled or open-label series, but lack controlled-trial data in long COVID specifically. Reasonable to consider in an individualised plan with a clinician who follows the literature; not yet ready for a confident recommendation.
- NAC (N-acetylcysteine) — antioxidant and glutathione precursor, with mechanistic rationale for the oxidative-stress hypothesis. A handful of small trials, mixed results.
- Quercetin (with bromelain or vitamin C for absorption) — anti-inflammatory and mast-cell-stabilising in vitro, popular in the post-viral community. Human evidence in long COVID is limited to uncontrolled series.
- Famotidine — H2 blocker, not a supplement but worth flagging because it's frequently combined with antihistamines in MCAS-overlap protocols. Some early COVID-acute data, no long COVID RCT.
- Low-dose naltrexone — prescription-only and outside this article's scope, but the most-discussed off-label intervention in long COVID clinics. Small trials, plausible mechanism, requires a prescriber familiar with it.
- L-arginine + vitamin C combination — one Italian RCT in 2022 reported endothelial-function improvement; replication is limited.
What to skip — the marketed-but-unevidenced category
- "Spike protein detox" protocols — typically combine high-dose nattokinase, bromelain, and assorted herbs. The premise (residual spike protein causing symptoms) is not established, the evidence for the proposed agents is preclinical at best, and the cost of the typical protocol is substantial. We do not recommend.
- "Long COVID complex" multi-ingredient products — same pattern as PCOS or anxiety complexes: sub-therapeutic doses of multiple plausible-sounding ingredients in one expensive capsule.
- Ivermectin (as a supplement) — multiple large RCTs have not supported efficacy in either acute COVID or long COVID. The agent is not a supplement; the supplement industry has marketed pharmaceutical-derived versions inappropriately.
- Hydroxychloroquine — same as above, plus more documented harms.
- "Heavy metal detox" panels and chelators — long COVID is not a heavy-metal toxicity syndrome.
- Methylene blue — promising mechanism, virtually no human long COVID data, narrow safety window, MAOI activity that creates real serotonergic-medication interactions.
The non-supplement layer that matters more
The interventions with the strongest evidence for long COVID symptom reduction are not in the supplement category. They include: pacing and energy management (reducing the post-exertional malaise cycle), graded exposure to upright posture for POTS-overlap symptoms, sleep optimisation, evidence-based cognitive rehabilitation for cognitive symptoms, and treatment of identified comorbidities. Vaccination status is also relevant — repeated infection appears to compound symptom burden in many cases. Supplements are at best an adjunct here.
What to track
Long COVID symptom tracking is hard because symptoms fluctuate dramatically. Useful approaches: a weekly fatigue scale (the Chalder Fatigue Scale is short and validated), a weekly post-exertional malaise log, and any specific symptom that's a personal priority. Don't change two things at once. Don't keep a supplement going past 12 weeks without an honest reassessment of effect.