The One Drug With Real Evidence, and a Liver Already Carrying Two Burdens
Duloxetine is the only adjuvant with a genuinely positive trial specifically for this kind of pain, and her liver is already managing both metastatic disease and the chemotherapy causing the pain in the first place. The disagreement isn't which drug has the best evidence — it's whether that evidence still applies to a liver already under two kinds of strain.
Sylvia N., 55, has knitted scarves and hats for a craft-fair table she runs most weekends for the better part of a decade, work that has quietly become a real second income since she cut back her hours after her diagnosis. Over the past two cycles of FOLFOX for metastatic colorectal cancer, the burning and tingling in her fingertips has gone from an annoyance she worked around to something that makes her drop stitches without feeling it happen, and twice now she's had to unravel and restart a whole row because her fingers told her nothing was wrong until she looked down.
Her exam is consistent with grade 2 oxaliplatin-associated peripheral neuropathy, symmetric and stocking-glove in distribution, worse in her hands than her feet. CALGB 170601 (Smith et al., 2013, JAMA) remains the only placebo-controlled trial to show real benefit from an adjuvant agent for this specific condition, and ASCO's own guideline names duloxetine preferentially for exactly that reason — gabapentin and pregabalin, despite wide use in neuropathic pain generally, have never shown the same CIPN-specific signal. What complicates reaching for the best-evidenced option is her liver: two small, stable metastases on her most recent scan, and transaminases running roughly one and a half to two times the upper limit of normal, a combination of disease and the chemotherapy already treating it. Duloxetine's own label carries a real hepatotoxicity caution, and her chart is not the clean slate that trial's own average patient may have started from. Her craft-fair table is more than a hobby to her at this point — the income has covered her copays twice this year already — which is part of why she has already asked, unprompted, how much of the feeling in her fingers she should expect to get back.
In clinic, choosing an adjuvant for a liver already carrying two burdens
Start duloxetine. CALGB 170601 is the only placebo-controlled trial that's actually shown benefit for an adjuvant agent in CIPN specifically, and ASCO's guideline reflects that by naming it preferentially. Gabapentin is what most people reach for out of habit, but it's never demonstrated the same CIPN-specific efficacy — we'd be choosing the more familiar option over the better-evidenced one.
The evidence point is fair, but her liver isn't a neutral starting point. She has two stable metastases and transaminases already running one and a half to two times normal from disease and chemotherapy combined. Duloxetine carries a real, labeled hepatotoxicity caution. Layering a hepatically metabolized drug with its own liver signal onto a liver already managing two things doesn't strike me as a risk worth taking when gabapentin, renally cleared, avoids it entirely.
The trial data is real, but it wasn't generated in patients with liver metastases and elevated transaminases — that's exactly the population the label's caution exists for.
I'd separate "elevated" from "impaired" here. The label's strongest hepatotoxicity caution is built around significant hepatic impairment, not a mild, one-and-a-half-to-two-times elevation with normal synthetic function. That's a real distinction, not a technicality — and her neuropathy has a genuine functional cost already, income and daily function both.
Start duloxetine at a reduced dose, with baseline and interval liver-function monitoring, and switch to gabapentin if her LFTs move meaningfully in the wrong direction. If her pain remains truly refractory despite an adequate trial, ketamine is a reasonable next step later — its own evidence base is weaker still, which is exactly why it isn't the first adjuvant to reach for here.
Agreed: start duloxetine at a reduced dose with baseline and interval liver-function testing, with an explicit plan to switch to gabapentin if her AST/ALT trend meaningfully upward rather than remaining stable.
Not agreed: the clinical pharmacologist would have preferred starting gabapentin now and reserving duloxetine only if it proved genuinely necessary, arguing that accepting a labeled hepatotoxicity risk for a functional-quality-of-life indication, rather than a life-threatening one, sets the risk tolerance too high even at a mild elevation. The oncologist and pharmacist held that the efficacy gap between the two drugs was too large to default away from the better-evidenced option without a clearer sign her liver couldn't tolerate it. Neither position was overruled; the duloxetine trial proceeded, with the switch threshold for LFTs specified explicitly rather than left to individual judgment at her next visit.