Low-Dose Naltrexone

LDN is not a new compound. It is naltrexone, licensed at 50 mg or more a day for alcohol and opioid dependence, taken at about a tenth of that dose. The practice began in the 1980s, when Dr Bihari gave 1.5–3 mg as an add-on treatment for AIDS; it has since spread to fibromyalgia, Crohn's disease, multiple sclerosis, complex regional pain syndrome and other conditions, almost entirely off-label [1].

Fibromyalgia. The two Stanford trials started the interest. In a single-blind pilot of 10 women, symptoms fell 32.5% on LDN against 2.3% on placebo [8]. In a double-blind crossover of 31 women, pain fell 28.8% on LDN against 18.0% on placebo, with better mood and life satisfaction but no change in fatigue or sleep [9]. The larger Danish trials did not reproduce this. A crossover of 58 patients found no difference on the Fibromyalgia Impact Questionnaire or pain ratings [4], and a 12-week parallel-group trial of 99 women on 6 mg found a between-group difference of 0.34 points on a 0–10 pain scale (95% CI −0.95 to 0.27), well short of significance; memory complaints may have improved, an exploratory finding [5]. Meta-analyses of these trials disagree: one pooling four trials (222 patients) found lower pain scores but no difference in the impact questionnaire [10], while another pooling eight studies found improvement from baseline but none over placebo [11].

Crohn's disease. An open-label pilot of 17 adults reported a response in 89% and remission in 67% [12]. In a 40-patient double-blind trial from the same group, 88% on naltrexone and 40% on placebo had a 70-point fall in the disease activity index, and endoscopic response was 78% against 28%; two of its authors hold a patent on naltrexone for inflammatory bowel disease [13]. The 2018 Cochrane review judged that trial and a pediatric pilot low-certainty evidence, and found clinical remission (30% against 18%) not significantly different [14]. A multicenter Dutch trial of 41 patients was then stopped for futility: endoscopic remission was 5.6% on LDN against 18.8% on placebo, and clinical remission 22% against 70%, though fatigue scores improved on LDN [6].

Multiple sclerosis. In a crossover trial of 80 patients (60 completed), 8 weeks of 4.5 mg nightly improved mental-health quality-of-life scores, such as 3.3 points on the SF-36 mental component [15]. A 17-week Iranian trial of 96 patients found no difference in quality of life [16]. An open-label pilot in 40 people with primary progressive MS found it safe, with less spasticity [17]. When LDN use surged in Norway, people with MS who started it did not go on to use fewer MS medicines [18].

Long COVID. A 2026 meta-analysis found no published randomized trial. Four uncontrolled before-and-after studies (155 patients) showed moderate improvements in fatigue, brain fog, sleep, pain and daily function, at low certainty; four randomized trials are registered [7]. Two of the four studies were an Irish clinic cohort on 1–3 mg [19] and a Stanford chart review of 59 patients on 0.5–6 mg [20].

Other uses. As an add-on for major depression, a 12-patient trial missed its primary outcome but scored better than placebo on one secondary scale, the MADRS [21]. A Norwegian register study found that people with hypothyroidism who started LDN did not go on to use less thyroid hormone [22]. Claims for cancer rest mainly on case reports in people and on cell and animal work [23], and the dermatological reports (Hailey-Hailey disease, lichen planopilaris) are case reports and small series [24].

Naltrexone is a competitive opioid antagonist with its highest affinity for the mu-opioid receptor [1][3]. The full 50 mg dose blocks more than 90% of brain mu receptors for at least two days. Low doses block a fraction: on PET, 2 mg and 5 mg blocked about 27% and 49% within 8 hours [2]. Two ideas try to explain why partial blockade might help.

Opioid rebound. A short, incomplete blockade each day is proposed to make the body compensate by producing more endogenous opioids and opioid receptors [3]. Zagon and McLaughlin argue that the length of blockade matters more than the dose: intermittent blockade of the opioid growth factor receptor slows cell proliferation, while continuous blockade speeds it [25]. They found serum [Met5]-enkephalin low in people with MS and restored in those taking LDN [26]. The premise that fibromyalgia involves an opioid deficit is not well supported: a full 50 mg challenge changed neither pain sensitivity nor mood in women with fibromyalgia or in controls [27].

Glial cells. In cell experiments naltrexone also blocks toll-like receptor 4, an innate-immune receptor on microglia, and does so with both its opioid-active (−) and opioid-inactive (+) forms [28]. A single spinal injection of the (+)-isomer eased nerve-injury pain in rats [28], and it blocked TLR4-driven release of nitric oxide and TNF-α from microglia [29]. The proponents of this idea note that the data do not fit it perfectly [3]. The cell experiments used 10 µM, about 3.4 mg/L [28], while plasma naltrexone after 4.5 mg measured 0.33–0.66 µg/L [4], several thousand times lower.

Human biomarker data are thin. Plasma cytokines fell over 8 weeks in eight women with fibromyalgia, in a study where no placebo was actually given [30]; serum cytokines did not change in an IBD cohort [31]. In the fibromyalgia pilot, a higher baseline sedimentation rate, a marker of inflammation, predicted a better response [8].

At low doses, side effects in trials were mostly mild. Pooled fibromyalgia trials found more vivid dreams (odds ratio 2.97) and nausea (2.75) than on placebo, with no difference in headache, diarrhea, dizziness or serious adverse events [10]. In the 99-woman Danish trial, 84% on LDN and 86% on placebo had an adverse event, and 8% and 6% stopped because of one [5]. In the Dutch Crohn's trial, 12 patients on LDN and 6 on placebo reported minor events, mostly sleep disturbance and headache, with no serious events, and liver enzymes stayed within acceptable ranges [6].

Across 89 placebo-controlled trials of oral naltrexone at all doses (11,194 participants), naltrexone did not increase serious adverse events (risk ratio 0.84, 95% CI 0.66–1.06); nausea, dizziness and vomiting were somewhat more common [36]. The liver warning on full-dose naltrexone came from 1980s studies at up to 300 mg a day, and the US boxed warning was removed in 2013 [36]. Liver tests stayed stable in the Crohn's trials in adults and children [13][33].

The main hazard is opioids. Even a low dose may blunt opioid pain relief, and the trials excluded people taking opioid painkillers [3]. A 66-year-old man on long-term opioids went into severe precipitated withdrawal after starting 1 mg [37]. Physicians have also reported anxiety and a fast heart rate anecdotally [3].

Naltrexone blocks the receptors that opioid painkillers act on. In people taking opioids it can reduce pain relief [3] and has precipitated severe withdrawal at 1 mg [37]. Among Norwegians who used opioids before starting LDN, those who kept collecting it used 46% less opioid in the following year; the authors could not tell whether that reflected warnings against combining them or an effect of LDN [38]. Beyond opioids, no interaction studies at low doses have been published [3].