A Rebuttal Letter Can't Settle What a Trial's Pain Data Actually Showed

A reply published in the October 2026 issue of the Journal of the American Academy of Dermatology shows how little a single correspondence exchange can settle. The piece, by Martina L. Porter, M. Celeste Ferreira-Cornwell, and Melinda J. Gooderham, responds to a letter from Yang and Li titled "Beyond abscess and inflammatory counts: Reflecting on placebo and pain in the ruxolitinib trial." Both documents sit behind ScienceDirect's paywall, and neither is the trial itself. The response is a rebuttal to a critique of a 2024 phase 2 study, which means the specific data point under dispute has to be reconstructed from the trial record rather than from either side's framing.

That trial tested ruxolitinib 1.5% cream, Incyte's Opzelura, against vehicle in 69 adults with mild to moderate hidradenitis suppurativa over 16 weeks. On the primary endpoint, abscess and inflammatory nodule count, the drug won: a least-squares mean reduction of 3.61 versus 2.42 for vehicle, with 79.2% of ruxolitinib patients hitting the AN50 response threshold against 56.3% on vehicle. Flares were rarer on the drug, 2.9% versus 25.7%. Those numbers, first presented at AAD 2024 and later detailed by HCPLive and Practical Dermatology, are what the original paper led with.

Buried in the secondary outcomes is the detail that appears to have prompted Yang and Li's letter: on the Skin Pain Numeric Rating Scale, the vehicle arm improved more than the active drug, a reduction of 2.61 points against 1.85. A trial can clear its primary endpoint on objective lesion counts while losing on a patient-reported symptom, and that gap is the kind of result a single paper's topline framing can bury. Without the text of the letter itself, it is not possible to say here whether Yang and Li treated that pain result as a flaw in this specific trial or as an illustration of something broader in the field. The rebuttal exists specifically because that question was raised, but the rebuttal alone does not answer what the letter actually claimed.

That ambiguity matters because a vehicle arm outperforming active treatment on pain is not unprecedented in hidradenitis suppurativa research. Trials of adalimumab, including the PIONEER 1 and PIONEER 2 studies, documented unusually high placebo response rates years earlier, and a PubMed Central review of placebo response in HS trials attributes the pattern partly to high natural variability in lesion counts and partly to the binary outcome measures the field has standardized on. Whether this particular pain discrepancy belongs to that known phenomenon, or signals something specific to ruxolitinib's mechanism or this trial's design, is exactly the kind of question a correspondence exchange is built to raise and not resolve. Settling it needs either the original trial's full patient-reported-outcome data or a prospective comparison designed to test the placebo effect directly, not two more letters in a journal.

The exchange also lands at a specific moment for the drug. Incyte has since registered two phase 3 trials, TRuE-HS1 (NCT06959225) and TRuE-HS2 (NCT06958211), to push ruxolitinib cream toward an actual HS indication; it is currently approved only for atopic dermatitis and vitiligo. Three of the original trial's co-authors, Ferreira-Cornwell among them, are Incyte employees, which is standard for sponsor-funded trials but relevant to how any defense of the pain data gets read. If the phase 3 protocols carry forward the same pain measurement approach without accounting for what the vehicle arm did in phase 2, the question Yang and Li raised will resurface in a much larger, and much more expensive, dataset.