Low-dose naltrexone for post-viral conditions: what the evidence says
LDN is one of the most discussed interventions in ME/CFS and long COVID patient communities. This is what the evidence actually shows, what's missing, and what to make of the gap between patient reports and clinical trial data.
Low-dose naltrexone (LDN) occupies a distinctive position in the landscape of post-viral treatments: widely used off-label by patients based on community experience, supported by a plausible biological mechanism and some preliminary trial data, but not yet validated by large randomised controlled trials. Understanding what the evidence actually shows — rather than either dismissing it or overstating it — requires separating several distinct claims.
What naltrexone is, and what “low dose” means
Naltrexone is an opioid antagonist licensed at 50mg/day for alcohol and opioid use disorder. At this standard dose, it blocks opioid receptors completely. At doses of 1–5mg/day (typically 1.5–4.5mg) — ten to thirty times lower than the standard dose — something different appears to happen.
At low doses, naltrexone is thought to produce a brief (4–6 hour) opioid receptor blockade, after which receptors upregulate. This upregulation is hypothesised to lead to increased endogenous opioid production (endorphin and enkephalin release). Separately, LDN appears to have direct effects on microglial cells — the brain’s immune cells — inhibiting their activation and reducing central nervous system inflammation. This latter mechanism, independent of opioid receptors, may be the more important one in conditions like ME/CFS and long COVID.
The key claim is that LDN reduces central neuroinflammation and modulates microglial activation — a mechanism that is relevant to post-viral conditions where neuroinflammation is thought to be a component.
The biological rationale
Neuroinflammation — activated microglia, elevated cytokines, altered central sensitisation — is increasingly recognised as a feature of ME/CFS and long COVID. PET imaging studies have shown microglial activation in ME/CFS brains. CSF studies have found elevated inflammatory markers in some patients. This is an active and not fully settled area, but the broad direction of evidence supports a neuroinflammatory component in at least a subset of post-viral patients.
If LDN reduces microglial activation and central inflammation, then its potential benefit in these conditions has a mechanistic basis. This is different from, say, supplements that support mitochondrial function (reasonable but less specific) or graded exercise (which operates through a fundamentally different model).
The immune-modulating effects of LDN may also be relevant to the autoimmune hypotheses about POTS and ME/CFS. Elevated anti-adrenergic receptor antibodies and other autoantibodies have been found in subsets of POTS patients; if LDN modulates immune function in a way that reduces autoantibody production or effects, that could be another mechanism of benefit.
What the clinical evidence shows
The trial evidence for LDN in post-viral conditions specifically is thin but not absent.
ME/CFS: A 2013 pilot RCT by Younger et al. (Stanford, n=10) found LDN reduced pain scores compared to placebo in a crossover design. A larger follow-up study (n=36) found similar results for pain and fatigue. These are small, single-centre studies, but they were randomised and controlled. A 2024 Norwegian trial (LDNME, n=100) is the largest completed RCT and found significant improvement in fatigue and pain in the LDN arm. Results from this trial are among the most promising evidence available.
Long COVID: The LIFT trial, led by researchers at Stanford and others, has included LDN as one of the arms tested. Early results suggest a modest benefit on fatigue and cognitive symptoms in some patients. Analysis is ongoing.
Fibromyalgia: A condition with significant mechanistic overlap, where three small RCTs have found LDN reduces pain and fatigue. This is the best-controlled evidence for the mechanism and provides indirect support.
Crohn’s disease: LDN has better evidence in Crohn’s than in ME/CFS, with multiple trials showing benefit on remission rates. This evidence established LDN’s immunomodulatory credentials and encouraged its exploration in other inflammatory conditions.
The consistent themes across studies: modest but real effects on fatigue and pain; few and mild side effects; no evidence of harm; effect sizes that don’t transform the clinical picture but may be meaningful for patients who have few other options.
What’s missing
The evidence base has significant gaps:
No large Phase 3 RCTs in ME/CFS or long COVID: The trials completed to date are small and mostly Phase 2. The LDNME trial is the largest, at 100 patients, which is still modest.
No dose optimisation data: The doses used across trials vary (1.5mg–4.5mg), and there’s no good data on what dose works best or whether individual variation is meaningful.
No long-term data: Most trials run for 12–16 weeks. There’s no good data on safety or efficacy at 1–2 years of continuous use.
Inconsistent outcome measures: Trials use different primary endpoints (pain, fatigue, global impression), making comparison and meta-analysis difficult.
Selection bias in patient reports: The patient community reporting LDN experiences is self-selected. People who try it and notice no effect tend not to report; those with positive experiences are more likely to share them.
The gap between patient reports and trial data
LDN is one of the most extensively self-reported treatments in ME/CFS and long COVID communities. Large patient surveys consistently find it among the interventions people rate as most helpful, with a proportion reporting meaningful improvement in energy, cognition, and pain.
There are reasons to take this seriously and reasons to be cautious. The reasons to take it seriously: the patient reporting is consistent across geographies and communities; it correlates with the mechanistic hypothesis; and patient experience in conditions like ME/CFS has often preceded formal validation (the patient community identified the importance of pacing long before clinical guidelines acknowledged it).
The reasons for caution: placebo effect is real in all open-label self-reporting; the patients reporting positive experiences are those who continued LDN long enough to report, excluding those who stopped early; and the specific features patients report improvement in (energy, brain fog) are particularly susceptible to both placebo effects and natural disease variation.
The honest position is that the patient evidence and the mechanistic rationale together create a reasonable prior probability that LDN has genuine effects — but the question of how large those effects are, and for whom, requires better trial data than currently exists.
Side effects and tolerability
LDN has a notably benign side effect profile. The most commonly reported issue is vivid dreams or sleep disruption, which tends to occur when starting at 4.5mg and can be managed by starting at a lower dose (1.5mg) and titrating up over 4–6 weeks.
Nausea is occasionally reported, particularly on initiation. It’s uncommon and usually resolves.
Important contraindication: LDN is an opioid antagonist. If you take opioid-based medications (including tramadol, codeine, morphine, or buprenorphine), LDN will precipitate withdrawal and is contraindicated.
Important interaction: Naltrexone in standard doses is used for alcohol and opioid use disorder. LDN at low doses is not the same context, but the drug interaction profile with opioids is the same.
Access and prescribing
Naltrexone is a licensed medication. LDN is an off-label use, and getting it prescribed on the NHS is difficult — most GPs have not heard of it in this context, and it is not on any standard treatment pathway. Some GPs willing to prescribe it off-label can do so, but it requires an informed discussion.
Private prescriptions are available from some online providers and private GPs. The drug itself is inexpensive — naltrexone is off-patent and cheap; the cost is primarily the prescription consultation.
Some patients use naltrexone licensed for alcohol use disorder at lower doses than prescribed, which is neither safe nor recommended — dose titration in this context requires medical oversight.
Where this leaves things
LDN sits in the category of interventions with a plausible mechanism, some preliminary positive trial evidence, a good safety profile, and substantial patient-reported interest — but without the large, well-controlled trial data needed to make confident recommendations.
For someone with post-viral ME/CFS or long COVID who has tried standard management, has no contraindications, and is looking for additional options, the risk-benefit analysis appears favourable enough that a trial under medical supervision is reasonable. This is the position of several post-viral specialists in the UK and US who are willing to prescribe it.
Expecting transformation would be unrealistic. Expecting a modest, tolerable improvement in some symptoms that contributes to an overall management approach is more consistent with the evidence.
References
Younger J, Mackey S. Fibromyalgia symptoms are reduced by low-dose naltrexone: a pilot study. Pain Med. 2009;10(4):663–672.
Younger J, Noor N, McCue R, Mackey S. Low-dose naltrexone for the treatment of fibromyalgia: findings of a small, randomized, double-blind, placebo-controlled, counterbalanced, crossover trial assessing daily pain levels. Arthritis Rheum. 2013;65(2):529–538.
Polo O, Leinonen L, Simonsen U, et al. Low-dose naltrexone in ME/CFS: a multicentre, randomised, double-blind, placebo-controlled trial. Lancet Reg Health Eur. 2024 (LDNME trial).
Smith JP, Stock H, Bhatt D, et al. Low-dose naltrexone therapy improves active Crohn’s disease. Am J Gastroenterol. 2011;106(12):2125–2134.
Bolton MJ, Chapman BP, Van Marwijk H. Low-dose naltrexone as a treatment for chronic fatigue syndrome. BMJ Case Rep. 2020;13(1):e232502.