JAK Inhibitor Safety in Rheumatology: Navigating the ORAL Surveillance Aftermath

ORAL Surveillance: What the Data Actually Showed

Few clinical trials have reshaped prescribing patterns as dramatically as ORAL Surveillance reshaped the use of JAK inhibitors in rheumatology. For the rheumatologist who had come to rely on tofacitinib as a convenient, effective oral alternative to biologic DMARDs, the safety signal was jarring — and the FDA's decision to extend black box warnings to the entire JAK inhibitor class raised questions that are still being debated. The reality is more nuanced than either the alarm or the dismissal suggests. Understanding what ORAL Surveillance actually showed, where the risk lies, and which patients can still benefit from JAK inhibitors is essential for anyone managing RA in current practice.

The ORAL Surveillance trial compared tofacitinib to TNF inhibitors in RA patients aged 50 or older with at least one cardiovascular risk factor[2]. Tofacitinib was associated with higher rates of MACE (HR 1.33, 95% CI 0.91-1.94) and malignancy (HR 1.48, 95% CI 1.04-2.09) compared to TNF inhibitors[1]. These results led to FDA black box warnings for all JAK inhibitors, impacting prescribing in IBD management as well.

Context Matters: The Study Population

The most important detail about ORAL Surveillance — and one that is frequently lost in clinical conversations — is the study population. This trial specifically enrolled patients aged 50 or older with at least one cardiovascular risk factor. This was not a general RA population. It was a population enriched for the very comorbidities that the safety signal identified. Whether the same cardiovascular and malignancy risks apply to a 35-year-old woman with RA and no cardiovascular risk factors is simply not known from these data. The FDA's decision to apply the black box warning to all JAK inhibitors in all patient populations was a regulatory precaution, not a data-driven conclusion for every individual patient.

It is also worth noting the confidence intervals. The MACE HR of 1.33 had a confidence interval of 0.91-1.94 — crossing 1.0, meaning the MACE signal did not achieve statistical significance in isolation. The malignancy HR of 1.48 (95% CI 1.04-2.09) was statistically significant. The trial was designed as a non-inferiority study with a non-inferiority margin of 1.8 for the hazard ratio, and tofacitinib failed to meet non-inferiority for both endpoints. This is different from a trial proving harm — it is a trial that failed to prove safety equivalent to the comparator in a high-risk population.

Risk Stratification: Who Can Safely Receive JAK Inhibitors

The key insight from ORAL Surveillance is that the excess risk was concentrated in patients with pre-existing cardiovascular risk factors. For patients under 50 without cardiovascular risk factors or malignancy history, the absolute risk increase is likely small. Current ACR guidelines recommend JAK inhibitors after failure of at least one biologic DMARD for patients with cardiovascular risk factors, but allow earlier use in lower-risk patients.

A Practical Risk Assessment Framework

When considering a JAK inhibitor for an RA patient, the risk assessment should address four domains: cardiovascular risk (age, hypertension, diabetes, dyslipidemia, smoking, established CVD), malignancy history (any prior malignancy, particularly lymphoma or lung cancer), VTE risk (prior DVT or PE, obesity, hormonal therapy, recent surgery or immobilization), and infection risk (age, diabetes, chronic lung disease, prior serious infections). For the patient who scores low across all four domains — the young, otherwise healthy patient who has failed methotrexate and one biologic — the absolute risk of JAK inhibitor therapy is low, and the convenience of oral dosing and the robust efficacy data make it a reasonable choice. For the 65-year-old with hypertension, diabetes, and a 20-pack-year smoking history, the risk-benefit calculus shifts substantially toward biologic alternatives.

Comparing JAK Inhibitors: Tofacitinib, Baricitinib, Upadacitinib

The three FDA-approved JAK inhibitors for RA (see also inflammatory arthritis differential) differ in selectivity: tofacitinib (JAK1/JAK3), baricitinib (JAK1/JAK2), and upadacitinib (JAK1-selective)[4]. SELECT-COMPARE demonstrated upadacitinib superiority to adalimumab for ACR50 response[3], while head-to-head data between JAK inhibitors remain limited.

Does JAK Selectivity Matter for Safety?

This is one of the most debated questions in rheumatology. The hypothesis is that more JAK1-selective agents like upadacitinib might have a better safety profile than less selective agents like tofacitinib, because off-target inhibition of JAK2 (involved in hematopoiesis) and JAK3 (involved in lymphocyte signaling) could contribute to some of the safety signals seen in ORAL Surveillance. The problem is that ORAL Surveillance studied tofacitinib specifically, and extrapolating those results to more selective agents is hypothesis-driven rather than evidence-driven. There is no equivalent large-scale, long-term safety study for upadacitinib or baricitinib versus TNF inhibitors. Until such data exist, the FDA has applied the same black box warnings to all three agents, and clinicians should apply the same risk assessment framework regardless of which JAK inhibitor is being considered.