Low-dose naltrexone for long COVID: zero trials found

Long COVID is described as a debilitating chronic condition, and with no established treatment, off-label options circulate quickly through patient communities. Low-dose naltrexone is one of the most talked-about, and it is already being prescribed despite an unclear evidence base. The core question this review asked is simple: does the published evidence actually show that low-dose naltrexone improves long COVID symptoms, and how confident can anyone be in that answer? Establishing what evidence exists — and what does not — is the necessary step before dosing, duration or candidate subgroups can even be debated.

The team ran a systematic review and meta-analysis, searching PubMed, Embase and the Cochrane Library for published studies, plus ClinicalTrials.gov and the WHO ICTRP for registered ongoing studies, through 5 May 2026. They included randomised controlled trials and pre-post studies of people with long COVID reporting fatigue, quality of life, cognitive symptoms or function, or other long COVID symptoms. Two independent reviewers screened and selected studies, risk of bias was assessed using the Newcastle-Ottawa Scale, and pooling used random effects models. Of 397 titles and abstracts screened, four observational pre-post studies from the USA and Ireland, totalling 155 participants, met inclusion criteria; doses ranged from 1 mg/day to 6 mg/day.

The headline finding is a gap: no randomised controlled trials were identified. Pooled pre-post analyses favoured low-dose naltrexone, with moderate effects for reducing fatigue (Hedges' g = -0.74; 95% CI -1.11 to -0.37; p<0.001), brain fog (g = -0.53; 95% CI -1.01 to -0.05; p=0.03) and improving sleep quality (g = -0.60; 95% CI -0.91 to -0.30; p=0.0001). Larger effects appeared for pain (g = -0.93; 95% CI -1.29 to -0.57; p<0.001) and daily functioning (g = -0.93; 95% CI -1.29 to -0.57; p<0.0001). Heterogeneity ranged from 0% to 62%, and no serious adverse events were reported in the two studies that assessed safety.

This puts a number on something that was previously anecdotal, but it also formally documents that the enthusiasm has run ahead of the trial evidence. Pre-post designs measure change over time in people who chose or were offered treatment, which cannot distinguish drug effect from placebo response, regression to the mean or spontaneous recovery — a particular concern in a fluctuating condition. The wide dose range across studies means no optimal dose or duration can be inferred. The authors are explicit that well-powered trials are needed to confirm efficacy, settle dosing and duration, and identify who is most likely to benefit.

If you have seen low-dose naltrexone discussed as a long COVID treatment, this review is the honest state of play: the symptom improvements reported so far are real measurements, but they come from uncontrolled studies in a small total number of people, so they cannot establish that the drug caused them. Low certainty is not the same as no promise — it means the question is genuinely open. Safety signals were reassuring in the limited studies that looked, though only two assessed it. This is a prescription medicine and a matter for a clinician, not a self-experiment.