Janus Kinase inhibitors (JAKi) tofacitinib and upadacitinib are emerging inflammatory bowel disease and rheumatology therapies. However, postmarketing studies such as ORAL Surveillance have led to a category-wide black-box warning for increased venous thromboembolism (VTE) risk with JAKi use, perhaps as high as 1% per year. There is a subset of patients that may benefit from a JAKi but have a history of VTE, and the risk of recurrent VTE for these patients is unclear. Our goal was to evaluate rates of new VTE events in JAKi-treated patients with prior VTE, and whether pharmacologic anticoagulation (AC) concurrent with JAKi impacts recurrent VTE incidence.

We conducted a retrospective review of electronic medical records from a large healthcare network to identify adults with a history of VTE prescribed a JAKi between 1/1/2000 and 6/30/2023. Patient charts were manually reviewed to collect demographic data, disease type, VTE date(s), and verify dates of JAKi use along with any concurrent AC. VTEs were identified by ICD10 code and then verified with physician documentation and imaging.

We identified 79 patients with a history of VTE who received a JAKi for inflammatory bowel (n=10) or rheumatologic disease (n=63), or other conditions (n=6). Patients were divided into two groups: JAKi-AC patients (n=47) received AC during JAKi treatment (median time on JAKi 1.16 years) and included 14 patients with hypercoagulable diagnosis, while JAKi-no AC patients (n=32) received no AC during JAKi treatment (median time on JAKi 0.83 years) and included 3 patients with a hypercoagulable diagnosis. There were more patients with a hypercoagulable diagnosis in the JAKi-AC group (n=14) than the JAKi-no AC group (n=3) (p=0.049; Table 1). 15 JAKi-AC patients discontinued AC while remaining on a JAKi. In total there were 65.22 patient-years of follow-up for those on AC and JAKi, and 55.42 patient-years on JAKi alone (Table 2). Overall, in patients with a history of prior VTE there were 5 new VTE events while taking a JAKi but no concurrent AC, accounting for 9.0 VTE events per 100 patient-years (9% risk of VTE per year; Fig. 1). The median time to VTE event was 144 days (range 10-239). There were no new VTE events for any patient while on AC with JAKi. The incidence rate of VTE events was significantly lower in the JAKi-AC group compared to the JAKi-no AC group with a calculated incidence rate ratio of 0 (95% CI 0-0.7; mid-p exact test, p=0.020).

While the increased VTE risk with JAKi over comparative treatments is uncertain, cautionary labeling may limit their use in patients who could benefit. Our study shows a thrombosis risk of 9% per year in patients with prior VTE taking a JAKi, indicating cautious prescribing is needed. However, we also found that concurrent AC appears protective in at-risk patients on a JAKi. If supported, this may enable JAKi use in these patients.

Tables 1 and 2.

<b>Figure 1</b>. Visual representation of time until venous thromboembolism (VTE) and start of anticoagulation for the patients that experienced a VTE. Time is calculated as number of days since initiation of JAK inhibitor. Dot symbols with line segment represent time of medication start and continued use, cross symbol represents a VTE event. The x-axis is truncated to 250 days for visualization purposes.

Figure 1. Visual representation of time until venous thromboembolism (VTE) and start of anticoagulation for the patients that experienced a VTE. Time is calculated as number of days since initiation of JAK inhibitor. Dot symbols with line segment represent time of medication start and continued use, cross symbol represents a VTE event. The x-axis is truncated to 250 days for visualization purposes.

Jeffrey Lowell

Appearances