The cardiovascular risk of JAK inhibitors in treating rheumatic diseases

The cardiovascular risk of JAK inhibitors (JAKi) is a contentious issue. The ORAL surveillance study—a randomized, non-inferiority, safety endpoint trial—demonstrated that the JAKi tofacitinib (TOF) was not non-inferior to tumor necrosis factor inhibitor (TNFi) regarding major adverse cardiovascular events (MACE) in patients with rheumatoid arthritis (RA).¹ Since patients in ORAL surveillance were older with cardiovascular risk factors, it is difficult to generalize this result to other patients. To clarify this issue, we performed a systematic literature search of the Cochrane library and MEDLINE (Ovid) databases with key terms related to JAKi, MACE, and rheumatic diseases. We included English studies investigating adults with RA, psoriatic arthritis (PsA), or axial spondyloarthritis (axSpA) who were receiving TOF, baricitinib (BARI), or upadacitinib (UPA), which involved MACE or changes in cardiometabolic factors as outcome parameters (Table 1). While most studies explored TOF and RA, some explored the impact of JAKi on cardiometabolic factors, and how this may subsequently alter cardiovascular risk.^2-5

One review explored hypothetical mechanisms through which JAKi may alter MACE.² In both a mouse model and human studies, blockade of interleukin 6 downstream pathways (which plays a pivotal role in RA pathogenesis) may positively or negatively modify cardiovascular outcomes, depending on the surrounding pathophysiological milieu. However, regarding PsA and axSpA, JAKi may provide cardioprotective effects via blockade of JAK2, which is involved in the pathogenesis of these diseases. JAKi may also modulate LDL catabolism, which seems consistent with observations in early JAKi studies.³

In 2013, TOF use in RA was associated with a mean rise in LDL, with LDL serum levels stabilizing from the third month.³ Nevertheless, strong associations between serum lipid level and increased risk of cardiovascular events cannot be demonstrated through early JAKi trials alone. Firstly, such studies were of shorter duration and would only detect MACE that occurred within 1 year. Additionally, cardiovascular safety is rarely a primary endpoint in these trials; high-risk patients are often excluded from such studies.

These challenges are partly what prompted the ORAL surveillance study. Here, fasting LDL and HDL levels rose and stabilized from the second month for patients receiving TOF 5 mg and TOF 10 mg relative to TNFi.¹ However, increased numerical incidence of MACE in TOF users relative to TNFi users may be partially attributable to background cardiovascular risk rather than changes in lipid levels alone.⁴ Therefore, the link between JAKi, serum lipid levels, and cardiovascular risk requires further enquiry.

The impact of JAKi on other cardiometabolic parameters has also been explored. One review found that TOF was not associated with body weight alteration in females with RA.⁵ Additionally, in a cohort study by Novikova and colleagues, patients receiving TOF experienced a reduction in visceral adiposity despite a concurrent increase in BMI, suggesting that lean mass was the main contributor of weight gain in these patients.⁶ However, further studies are needed to characterize such body composition changes, especially those with larger cohorts and the use of bioelectrical impedance.

JAKi may not worsen glycemic control. One cohort study found a crude incidence rate for RA patients initiating TOF and developing T2DM of 8.4 per 1000 patient-years, but small sample sizes limit this result.⁷ Findings from another cohort study suggested that the risk of diabetes management intensification may in fact be lower for TOF initiators relative to other bDMARDs.⁸ This year, a prospective study demonstrated improvement in insulin resistance over 6 months in RA patients receiving TOF, 35% of which were receiving glucocorticoids.⁹ However, the number of patients still receiving glucocorticoids at the conclusion of the study is not reported, so it is unclear whether JAKi alone, rather than possible glucocorticoid dose reduction, contributed to this result.⁹

The impact of JAKi class on cardiovascular risk remains unclear due to a lack of head-to-head studies; however, several population studies provide some insight. The Swedish ARTIS program demonstrated no increased MACE risk for RA patients receiving TOF or BARI and associated BARI use with a lower risk of acute coronary syndrome relative to those receiving a bDMARD (etanercept).¹⁰ However, 95% confidence intervals were wide due to the small number of events. Similarly, Hoisnard and colleagues reported low incidence rates for MACE for RA patients receiving TOF (2.8 per 1000 patient-years) or BARI (5.2 per 1000 patient-years).¹¹

A systematic review and meta-analysis by Xie et al.¹² demonstrated that TOF, BARI, and UPA were not associated with any significant increased risk of MACE relative to placebo across 26 randomized controlled trials. However, many of these trials were of short duration. In a recently published systematic review and meta-analysis, there was no statistically significant difference in MACE incidence across several JAKi including TOF, BARI, and UPA.¹³

The impact of JAKi dose on MACE remains unclear due to a paucity of adequately powered long-term studies. Due to short follow-up times and minimal numbers of recorded MACE, no significant dose-dependent concerns for MACE emerged in the clinical trial programs for JAKi across RA, PsA, and axSpA.^14-16

Although the ORAL surveillance study included both a TOF 5 mg twice daily and a TOF 10 mg twice daily arm, many patients who were initially categorized in the TOF 10 mg twice daily group were switched to the lower dose group due to concerns of venous thromboembolism (VTE) risk, limiting the ability to associate higher TOF dose with increased incidence of MACE.¹

In one pooled analysis, the incidence rate of MACE in RA patients receiving TOF monotherapy was 0.53 patients per 100 patient-years (95% confidence interval 0.23–1.04) compared to TOF combination therapy with csDMARDs, with an incidence rate of 0.62 (0.35–1.02).¹⁷ In a cohort study of RA patients, the hazard ratio for cardiovascular disease for TOF monotherapy relative to csDMARD was 0.23 (albeit with a wide 95% confidence interval of 0.03–1.62).¹⁸ Combination therapy with glucocorticoids may be associated with increased cardiovascular risk, which may be due to ongoing disease activity.¹⁹ However, glucocorticoids themselves may be responsible for increasing this risk.

Ongoing disease activity may contribute to increased cardiovascular risk in patients receiving JAKi. Navarro-Millan and colleagues found higher rates of cardiovascular events in a federally insured cohort of RA patients (who had potentially more disabling disease activity) relative to a commercially insured cohort (which may represent employed patients with lower ongoing disease activity).²⁰ However, the reasons for disability were not recorded for many patients, and it is therefore unclear whether disease activity, or confounding factors such as socioeconomic status, altered cardiovascular risk.²⁰

Type of rheumatic disease seems to have minimal impact on the incidence of MACE with JAKi use. For those treated with TOF, a post-hoc analysis performed by Burmester and colleagues demonstrated no significant difference in MACE incidence across all disease cohorts, including those diagnosed with RA and PsA.²¹ However, the sample sizes of the non-RA cohorts in this study were small, so these results should be interpreted with caution.²¹ Similar findings are reported in post-hoc analyses of studies involving patients receiving BARI and UPA.^{22, 23}

Since JAKi reduce systemic inflammation, clinicians would expect that this confers a reduction in cardiovascular risk. In this sense, the results of the ORAL surveillance study seem unexpected. As others have recognized, one possible explanation for these results is that while TNFi and TOF may both reduce cardiovascular risk, TNFi may have reduced this risk more than TOF relative to the baseline.²⁴ However, this may only form part of the picture.

Recent studies suggest that JAKi may in fact increase cardiovascular risk but only in the presence of certain baseline risk factors. One post-hoc analysis of the ORAL surveillance study found that patients who were at least 65 years old with a positive smoking history had a higher associated MACE risk when receiving TOF while those without these factors did not, even in the presence of other cardiovascular risk factors.²⁵ Additionally, another post-hoc analysis of the same study associated prior history of atherosclerotic disease with increased risk of MACE.⁴ However, since the ORAL surveillance study was not powered to investigate MACE risk according to baseline characteristics, these results only serve as hypothesis-generating.

Nevertheless, similar results have been observed in other studies. In the STAR-RA cohort study, RA patients receiving TOF with cardiovascular risk factors or a cardiovascular disease history experienced a numerically higher number of cardiovascular events relative to individuals without these characteristics.²⁶ Several other studies for both RA and PsA patients also allude to the detrimental impact of advanced age and the presence of cardiovascular risk factors.^{23, 27-29}

These studies re-affirm existing warnings regarding JAKi use and re-emphasize the importance of managing traditional cardiovascular risk factors in patients with rheumatic disease. A threshold effect may exist where mainly older patients with cardiovascular risk factors or smoking history experience a higher cardiovascular risk with JAKi use.³⁰ Since randomized controlled studies investigating the cardiovascular risk of JAKi are time consuming and expensive, future cohort studies and post-marketing data analyses are perhaps better suited for assessing JAKi cardiovascular risk. These should stratify patients into subgroups according to cardiovascular risk, cardiometabolic parameters, and other characteristics (such as smoking history) to determine when the benefit of JAKi use is outweighed by the risk of MACE.

Alexander Kwan: Conceptualization, Writing – Original Draft; Elvina Ingrid, Matthew Jiang, Keith KT Lim: Writing – Reviewing and Editing; Keith KT Lim – Supervision.

A/Prof Keith KT Lim chaired a related Janssen meeting in 2023. The authors declare no other conflicts of interest.

This editorial did not receive any special grants from funding agencies in public, commercial, or not-for-profit sectors.