Design: randomised, double-blind, placebo-controlled phase 3 trial with placebo crossover at week 16
Population: 670 adults with active psoriatic arthritis, biologic-naive, C-reactive protein 3 mg/L or more and at least one hand or foot erosion
Primary outcome: ACR20 response at week 16
Effect: 54.2% versus 34.1% (P<0.001); serious adverse events 1.8% versus 2.4% at week 16, no deaths and no class imbalances through week 52

Six hundred and seventy adults with active psoriatic arthritis, naive to biologic disease-modifying drugs, with high-sensitivity C-reactive protein of 3 mg/L or more and at least one radiographic hand or foot erosion, were randomised 1:1 to oral deucravacitinib 6 mg once daily or placebo to week 16, after which the placebo group crossed over and everyone continued to week 52.

ACR20 at week 16 was 54.2% with deucravacitinib against 34.1% with placebo (P<0.001). Responses rose further by week 52, and patients who crossed over from placebo reached improvements similar to those on continuous treatment. Structural damage inhibition was seen at weeks 16 and 52, though by post-hoc rank analysis of covariance without imputation. Serious adverse events at week 16 were 1.8% with the drug and 2.4% with placebo, discontinuations for adverse events 2.4% and 1.8%, and both stayed low to week 52. There was no imbalance in cardiovascular events, malignancies or opportunistic infections, and no deaths.

The safety data are the reason this trial matters more than its efficacy numbers. Tyrosine kinase 2 inhibition is mechanistically distinct from JAK1/2/3 inhibition, and the question everyone has about an oral drug in this space is whether it carries the JAK class warnings — the cardiovascular, thrombotic and malignancy signals that constrain tofacitinib and its relatives, particularly in older patients with cardiovascular risk. At 52 weeks, in 670 patients, nothing appeared. That is not the same as proof, but it is the right kind of evidence in the right direction.

An effective oral option matters disproportionately in Indian practice, where biologic cost and the logistics of parenteral therapy exclude many patients from adequate treatment. The relevant unknowns are price and CDSCO status, neither of which this trial addresses.

Consider this where an oral option is needed in biologic-naive psoriatic arthritis, once availability and cost are known
Note the entry criteria: raised C-reactive protein and radiographic erosion — this was not mild disease
Do not carry the JAK class cardiovascular and malignancy warnings across automatically, but do not assume their absence either
Record baseline radiographs; structural inhibition was assessed post hoc and needs local verification over time
Reassure patients who start late: the placebo crossover group caught up by week 52

Why it matters

It addresses the question that actually limits oral targeted therapy in this disease, which is safety rather than efficacy.

The statistics, in plain English

A 20 percentage point absolute difference in ACR20 means about five patients treated for one extra responder — solid for psoriatic arthritis but below what the best interleukin-17 and interleukin-23 biologics achieve. ACR20 is also a low bar: it means one-fifth improvement, not remission. The structural damage finding is weaker than it sounds, being a post-hoc rank analysis with no imputation for missing data, so it should be read as supportive rather than as a demonstrated radiographic benefit. And a 52-week safety record in 670 patients cannot exclude rare events; the JAK signals emerged from far larger and longer datasets.