Low-Dose Naltrexone for Central Sensitization: Emerging Option or Premature Prescription
A woman whose fibromyalgia hasn't responded to any guideline-first-line drug wants to try low-dose naltrexone after reading about it online, and the real evidence behind it is genuinely promising and genuinely small.
K.S., a 47-year-old woman, has had widespread musculoskeletal pain, unrefreshing sleep, and profound fatigue for six years, formally diagnosed as fibromyalgia three years ago after a rheumatologist ruled out inflammatory arthritis and lupus. She works as a high school librarian, a job she has kept only by structuring her week around rest days she didn't used to need, and describes the exhaustion as harder to live with, most days, than the pain itself. She has tried, in sequence, pregabalin (stopped for weight gain and dizziness), duloxetine (modest benefit, then plateaued), and a supervised aerobic exercise program (genuinely helpful, but not enough on its own) — the three interventions with the strongest guideline backing for fibromyalgia, all tried in good faith, none producing the level of relief she was hoping for. A friend in an online fibromyalgia support group told her about low-dose naltrexone, and she brings a printed article about it to this visit, asking directly whether it's something she should try.
Low-dose naltrexone's proposed mechanism in fibromyalgia is genuinely distinct from naltrexone's standard opioid-blocking use — at the very low doses studied (typically 4.5mg, roughly a tenth of the standard dose), the drug is thought to act primarily through transient microglial modulation and proposed TLR4 antagonism rather than sustained opioid receptor blockade, a mechanism that maps plausibly onto fibromyalgia's central sensitization model. The clinical evidence behind that mechanism has since been tested twice, in opposite directions. Younger and colleagues' 2013 randomized, placebo-controlled crossover trial found a real, statistically significant reduction in pain with LDN — in 31 patients. Due Bruun and colleagues' 2024 FINAL trial was built to confirm that result at scale, randomized 99 women to 6mg naltrexone or placebo, and did not: LDN was not superior to placebo on its primary pain endpoint (mean difference −0.34, p=0.27), though more women on LDN did reach a 30% pain reduction (45% versus 28%). There is no FDA approval or fibromyalgia indication of any kind. The article K.S. is holding almost certainly describes the 2013 result and not the 2024 one, which means the gap this visit has to close is not between hope and evidence but between the evidence she has read and the evidence that came after it.
Discussing an off-label request
She's genuinely exhausted the strongest guideline-backed options — pregabalin, duloxetine, structured exercise — all tried appropriately, none sufficient. LDN's proposed mechanism, low-dose microglial modulation rather than classical opioid blockade, fits the central sensitization model fibromyalgia is increasingly understood through, and Younger's 2013 trial found a real, significant pain reduction — with the FINAL trial's own responder analysis showing more women on LDN reaching a 30% pain reduction than on placebo. I think it's a reasonable next step to offer.
I'm not against trying it — I want to be precise about what "the evidence" actually is before we frame it for her. Younger's trial was 31 patients in a crossover design — and the larger trial built to confirm it, Due Bruun's 2024 FINAL study, missed its primary endpoint outright. Quoting the responder analysis without saying the primary endpoint was negative is exactly the move I want us not to make with her. That's a positive early signal followed by a failed replication, not an unconfirmed promise, and it is nowhere near the evidentiary weight duloxetine or pregabalin carried when she tried them.
Mechanistic plausibility is a reason to study a drug further, not, on its own, evidence that it works — plenty of biologically plausible mechanisms haven't panned out once tested at scale.
I think you're both right, and I don't think the disagreement is actually about whether to prescribe it — it's about how much evidence should be required before offering a trial of therapy, and that answer changes with the drug's risk profile. LDN at these doses has an unusually favorable safety record; the downside of a properly consented trial is genuinely low. The real question isn't "is 31 patients enough," it's whether she walks in understanding that the one adequately sized trial of this drug in her condition was negative, and that she would be trying it anyway because it is cheap and safe — not because the evidence favors it.
Withholding it until a larger trial exists would make sense for a higher-risk drug; for one this safe, the more honest path is offering it with accurate expectations attached, not gatekeeping it behind an evidence bar we don't apply this strictly to every low-risk option.
Agreed: an 8-week trial of low-dose naltrexone at 4.5mg nightly, continued alongside her existing duloxetine, with an explicit discussion documenting that the therapy is offered on the basis of one small positive trial rather than established guideline evidence, and a defined follow-up to assess whether it's actually helping before continuing indefinitely.
Not agreed: whether the pain medicine specialist's framing at the start of the visit — before the pharmacologist's correction — would have accurately conveyed the evidence's real weight to K.S. if the conversation had stopped there. All three voices treated the correction as necessary, but the primary care physician's own note, unresolved: how often an off-label option gets offered to a patient with the mechanistic story told first and the trial's actual size mentioned only once someone else in the room raises it.