Low-dose naltrexone: cheap, intriguing, and under-studied

Naltrexone is an old, well-understood drug: at its standard 50 mg dose it blocks opioid receptors and is used for opioid and alcohol use disorder. Low-dose naltrexone (LDN) is something else entirely — the same molecule at roughly a tenth to a twentieth of that dose, 1.5 to 4.5 mg, used off-label for chronic pain, fibromyalgia, and a scattering of inflammatory and autoimmune conditions. It has a devoted following, a genuinely interesting mechanistic story, and — this is the honest part — a thin evidence base that hasn’t caught up to the enthusiasm.

The reason LDN is worth an entry rather than a dismissal is its risk/benefit shape: it’s cheap, it’s well tolerated, and the downside of a trial is low. But “low downside” is not the same as “proven,” and the whole value of an honest guide is keeping those apart.

The mechanism — plausible, not proven

The proposed mechanism is genuinely clever. At low doses, naltrexone briefly and incompletely blocks opioid receptors; the body responds to that transient blockade by upregulating its own endorphins and receptors — an endorphin rebound that outlasts the drug. Separately, and probably more importantly for pain, naltrexone appears to modulate glial cells in the central nervous system, dampening the neuroinflammatory signaling (via pathways including TLR4) that can amplify chronic pain.

That’s a coherent, appealing story — but it is mostly mechanistic and preclinical reasoning. The Younger 2014 review lays it out well, and it’s the best available account of why LDN might work. It is not evidence that it does work in a given patient. Grade the mechanism D: an interesting hypothesis, not a clinical fact.

What the evidence actually shows

Here’s the honest inventory.

• Fibromyalgia is the best-studied indication, and it rests largely on the work of Jarred Younger and colleagues at Stanford. Their 2013 crossover RCT — the single strongest datapoint — randomized 31 women and found LDN 4.5 mg reduced pain more than placebo (a 28.8% vs 18.0% reduction, p=0.016), with good tolerability. That’s a real, statistically significant result. It’s also a tiny, single-site study. Encouraging; nowhere near definitive.
• Crohn’s disease has a couple of small pilot RCTs. The 2018 Cochrane review pooled them and found just two trials, 46 patients total, and concluded there is insufficient evidence to say whether naltrexone helps. “Insufficient evidence” is exactly the phrase to sit with — not negative, but not a green light.
• Multiple sclerosis and other autoimmune conditions have generated small pilots and a lot of patient reports, but again nothing approaching a large, well-powered trial.

And the skeptical counterweight matters: a 2024 systematic review of LDN for chronic pain did not find clear efficacy across the literature. When you step back from the individual encouraging pilots and look at the body of evidence as a whole, the honest summary is that no large confirmatory RCT establishes that LDN works. That’s not a reason to sneer at it — it’s a reason to be precise about what we do and don’t know.

Who it’s for, and how to think about a trial

Given all that, where does LDN reasonably fit? It’s most defensible for someone with stubborn fibromyalgia or chronic pain that hasn’t responded to first-line options, who wants to try something with a very low risk profile. The tolerability is good — the most common issue is vivid dreams or sleep disruption early on, often managed by starting low or dosing in the morning — and the cost is modest.

Two safety points are non-negotiable:

• It blocks opioids. If you take opioid pain medication, naltrexone — even at low doses — can precipitate withdrawal or blunt the medication. This has to be sorted out with a clinician before starting.
• It’s off-label and compounded. Standard naltrexone comes as 50 mg tablets, so the 1.5–4.5 mg dose has to be made by a compounding pharmacy to prescription. That’s the legitimate path; it’s also why a prescriber is involved.

The honest bottom line

Low-dose naltrexone is a fair example of a favorable risk/benefit profile that outruns its evidence. The mechanism is interesting, the safety and cost are genuinely attractive, and the small fibromyalgia data are encouraging. But the trials are tiny, the systematic evidence doesn’t yet establish efficacy, and there is no large confirmatory RCT. For the right person — stubborn chronic pain, low expectations, a willing clinician — a monitored trial is a reasonable, low-stakes experiment. Just go in clear-eyed: this is a promising, under-studied option, not an established treatment, and we don’t yet know how well it works.