A Fentanyl Patch That Stopped Working Isn't Always a Tolerance Problem
A severely cachectic hospice patient's pain has crept back despite an unchanged fentanyl patch dose, and the reflexive answer — opioid tolerance — misses a simpler, absorption-driven explanation sitting in what forty-six pounds of wasting has done to the skin the patch actually delivers through.
P.N. has kept a bird feeder log in the same spiral notebook for nineteen years, entries getting shorter over the last two months as end-stage heart failure has taken more of her strength than her attention — she still notes the finches, just no longer the exact count. Cardiac cachexia has stripped nearly all her subcutaneous fat over the past four months; her weight has dropped from 142 to 96 pounds, and the transdermal fentanyl patch that controlled her chronic back and joint pain reliably for the first six weeks of hospice enrollment has, over the last ten days, stopped holding through its full three-day interval — pain breaking through by day two, sometimes earlier.
The reflexive read on a fentanyl patch losing effect is opioid tolerance, the same neuroadaptive process behind escalating doses in any chronic opioid course. But fentanyl's transdermal delivery depends on a physical mechanism unrelated to receptor biology, and one that is routinely misdescribed at the bedside: after application the drug accumulates as a reservoir in the upper skin layers — the stratum corneum and dermis, not the fat beneath them — and diffuses from there into the circulation across the labeled 72-hour period. So the question cachexia raises isn't whether she still has fat to hold the drug; it's whether the skin and the perfusion underneath it still behave the way a fixed 25mcg/hr label assumes. Heiskanen and colleagues measured that directly, finding significantly lower serum fentanyl concentrations in cachectic cancer patients than in non-cachectic ones on identical patch doses, and a 2024 scoping review of nine such studies found the same direction of effect without settling its size. P.N. has lost forty-six pounds in four months, and her breakthrough arrives by day two of three — a failure clocked to patch age rather than to exertion, which is what an absorption problem looks like and is not what tolerance looks like.
Her notebook itself carries a quieter version of the same evidence: the shortened entries she attributes to fatigue cluster in the final day of each patch cycle, the same window where her pain breaks through, rather than being spread evenly across the week the way a generally declining energy level would show up. She mentioned this pattern to the visiting hospice nurse almost as an aside, framing it as an observation about her handwriting rather than her medication — but it lines up closely enough with the breakthrough-pain timing that the team is now treating it as real clinical data rather than an incidental detail from a bird log.
At the weekly hospice check-in
I don't think this is tolerance. Her breakthrough pain arrives on a predictable schedule tied to patch age — day two of a three-day patch, consistently — rather than the more diffuse, activity-driven pattern true escalating tolerance usually produces. Transdermal fentanyl forms its reservoir in the upper skin layers, not the fat below, and releases from there into whatever perfusion is underneath. Heiskanen's cachexia series found significantly lower serum fentanyl on identical doses in exactly this kind of patient. She's lost forty-six pounds in four months. I think the patch simply isn't delivering the dose the label assumes anymore.
I want to push back gently — six weeks of continuous opioid exposure is real time for genuine tolerance to develop, and I worry we're reaching for the more elegant mechanistic explanation because it's more satisfying than 'she needs more drug.' I don't want an absorption theory to become the reason we under-dose her.
I'm not saying the mechanism you're describing is wrong — I've seen it before in genuinely cachectic patients — I'm saying I want us to hold both possibilities honestly rather than settle on the more interesting one.
I think you're both right, and the good news is it doesn't actually matter which of you is right for what we do next. Whether it's a failing depot or genuine tolerance, the fix is the same first move: come off a delivery route whose reliability depends on skin she doesn't have anymore, onto one that doesn't. If we switch to subcutaneous hydromorphone and she needs escalating doses over the following weeks, that's real evidence for tolerance. If a modest, appropriately-calculated equianalgesic dose holds steady, that's evidence it really was the patch.
Agreed: transdermal fentanyl discontinued, converted to scheduled subcutaneous hydromorphone at an equianalgesic starting dose with breakthrough dosing available, reassessed at 72 hours.
Not resolved, and deliberately left open rather than smoothed over: whether the original mechanism was distribution failure or true tolerance. The team agreed to read the following week's dose-escalation pattern as the actual answer — a stable requirement supporting the Clinical Pharmacologist's reading, a climbing one supporting the Hospice Medical Director's — rather than deciding it from today's evidence alone.