Chronic CRPS After Two Failed Infusion Courses: Does Naltrexone's Fibromyalgia Evidence Transfer?
Low-dose naltrexone has a real but contested randomized record in fibromyalgia. In complex regional pain syndrome specifically, what exists is a single case series — and the question is whether a shared sensitization mechanism is enough to justify borrowing evidence built for a different disease.
Foster T., a 46-year-old man who has coached his son's youth baseball team every spring for the past six years, has been through the standard escalation for CRPS Type I without lasting benefit since an ankle inversion injury sixteen months ago. An intravenous bisphosphonate course produced no measurable change in his pain scores at three-month follow-up. A subsequent ketamine infusion series gave him roughly three weeks of meaningful relief before his allodynia and burning pain returned to baseline. He still can't tolerate a sock on that foot without significant distress, and handed the coaching duties off to another parent this spring for the first time in six years because he can no longer count on being able to stand through a full practice. He brings up naltrexone today after reading a patient forum post describing it as "the same drug that works for fibromyalgia."
That description isn't wrong about fibromyalgia, but it glosses over a real gap in what's actually been shown for his own diagnosis. Naltrexone's proposed mechanism at microdose — toll-like receptor 4 antagonism on microglia, quieting the neuroinflammatory signaling thought to drive central sensitization — is not disease-specific in principle, and CRPS shares real sensitization biology with fibromyalgia. But the fibromyalgia evidence being borrowed is less settled than a forum post makes it sound: Younger and Mackey's 2009 and 2013 crossover trials were positive, while the larger FINAL trial (Due Bruun and colleagues, 2024) found 6mg no better than placebo on its primary pain endpoint. Whatever would be transferring here is a contested signal, not a settled one — and for CRPS it has no equivalent at all. What exists for his diagnosis specifically is Chopra and Cooper's 2013 case series, a handful of patients followed openly without a control arm, reporting benefit but incapable of ruling out placebo response, natural fluctuation, or the coincidence of starting a new treatment after two others had already failed. Spinal cord stimulation sits at the other evidentiary extreme: Kemler and colleagues' randomized trial, published in 2000, found real pain reduction from spinal cord stimulation in CRPS specifically — though the same cohort's five-year follow-up in 2008 found the advantage over physical therapy alone had faded to non-significance, so what it demonstrates is a genuine early effect rather than a durable one. Discounted that far, it is still a population match LDN cannot offer at all, purchased at the cost of an implanted device and a genuinely invasive procedure Foster has said outright he wants to avoid if any reasonable alternative is left.
Pain clinic, after two failed infusion courses
I want to be direct about where the evidence actually sits. Kemler and colleagues' randomized trial of spinal cord stimulation in CRPS, published in 2000, found real pain reduction in exactly this diagnosis. I'll concede the five-year follow-up showed that advantage narrowing to non-significance, so I'm not going to sell it as permanent. Naltrexone's case for CRPS specifically is Chopra and Cooper's 2013 case series — open-label, no control arm, unable to separate a real drug effect from placebo response or from the simple fact that starting something new after two failures often produces a temporary bump regardless of what it is. After sixteen months and two failed courses, I think the honest recommendation is the option that's actually been proven in his diagnosis.
I don't disagree about the evidence tiers — Kemler's trial is real, randomized, disease-specific evidence, and Chopra and Cooper's case series plainly is not in the same category.
Then we're arguing about a narrower gap than you opened with. And the mechanism naltrexone is proposed to act through, toll-like receptor 4 antagonism quieting microglial neuroinflammatory signaling, isn't specific to fibromyalgia's diagnostic label — it's a sensitization pathway CRPS shares. I'll grant the fibromyalgia evidence is itself split now that FINAL has reported; what I won't grant is that a mechanism stops being plausible because one trial of one dose missed. And Foster has told us plainly, more than once, that he wants to avoid an implanted device if there's any reasonable alternative left. A reasonable alternative doesn't have to mean an equally proven one; it has to mean a genuine option, and low-dose naltrexone at these doses has an extremely benign side-effect profile even where its efficacy is unproven.
I think both of you are right about the piece you're each emphasizing, which is exactly why I'd bound this rather than pick one side permanently. A defined eight-week naltrexone trial, with a pre-agreed threshold — say, no meaningful change on his pain diary by week eight — set today, not renegotiated later, gives his stated preference one genuine attempt without indefinitely deferring stimulation if it doesn't work.
That's not evidence-tier confusion. It's sequencing a low-risk option first when a patient has a clearly stated preference, with a real, pre-committed off-ramp to the higher-evidence option rather than an open-ended trial that quietly becomes permanent.
Low-dose naltrexone started at 3mg nightly with a pre-agreed eight-week checkpoint on his pain diary; spinal cord stimulation referral placed in parallel today so the evaluation process runs concurrently rather than sequentially, activated only if the LDN trial doesn't clear the agreed threshold.
Not fully agreed: how much weight the neuroinflammatory-mechanism argument should carry going forward if this specific trial fails.
The neurologist reads a clean failure as confirmation the mechanism doesn't meaningfully apply outside fibromyalgia in this patient, and would not revisit LDN at a higher microdose.
The pain medicine specialist would still consider a dose adjustment before fully abandoning the mechanism, given how thin the existing evidence is either way — a position not adopted today, but not foreclosed either.