Cardiovascular comorbidity in rheumatoid arthritis : studies on venous thromboembolism

Individuals with rheumatoid arthritis (RA) are at around 50-100% increased risk of venous thromboembolism (VTE). In 2019, signals emerged suggesting an increased risk of VTE in RA patients treated with JAK inhibitors (JAKi), a new and highly efficacious class of disease- modifying antirheumatic drug (DMARD). In this thesis, we aimed to further disentangle the associations between RA, its treatment, and VTE risk in a series of observational studies. We used data from the Swedish rheumatology quality register and other national registers, thereby including a majority of all Swedish patients with RA, as well as a general population comparator cohort.

Study I was a study validating International classification of diseases (ICD) 10 codes for VTE registered in the National patient register specifically in patients with RA, through manual review of medical records. Based on review of 269 registered VTE events at hospitals in Region Stockholm from 2009 through 2018, 235 were confirmed as incident VTE. This resulted in a positive predictive value (PPV) of 87% for incident VTE, which ranged from 80% to 98% when we applied various restrictions to our definition of VTE. The majority of the confirmed events were diagnosed independently of RA (only 17 cases involved initial work- up by a rheumatologist, and in 3 cases there was an initial suspicion af an RA-related symptom). This study demonstrates a high validity for registered ICD-10 codes for VTE specifically in patients with RA, and suggests a low degree of surveillance or diagnostic bias.

Study II was a cohort study on the association between RA disease activity, measured by the Disease activity score 28 (DAS28), and the risk of VTE. Based on 322,601 rheumatologist visits in 46,316 patients with RA from 2006 through 2018, we identified 2,241 incident VTE events during the year following the rheumatologist visit. This resulted in a cumulative 1- year incidence of 0.52% for DAS28 remission, and 1.08% for high disease activity, with a corresponding adjusted risk ratio of 2.03 (95% CI 1.73-2.38) for high DAS28 vs. remission. Compared to the general population, the risk ratio of VTE for the overall RA population was 1.88 (95% CI 1.65-2.05). The study demonstrates a significant association between RA disease activity and VTE risk, thereby underscoring the importance of VTE risk stratification in patients with RA, and that treating RA to remission is important to minimize VTE risk.

Study III was an active comparator, new user design cohort study on the risk of VTE in patients with RA treated with JAKi, TNFi and other biologic DMARDs (bDMARDs) from 2010 through 2021. Based on 32,737 treatment episodes and 559 incident VTE events, the adjusted hazard ratio (HR) for the risk of VTE for JAKi vs. TNFi was 1.73 (95% CI 1.24-2.42). Regarding VTE subtypes, the corresponding HR was 3.21 (95% CI 2.11-4.88) for pulmonary embolism and 0.83 (95% CI 0.47-1.45) for deep vein thrombosis. The incidence of VTE was doubled with use of JAKi compared to the entire RA population, and tripled compared to the general population. The study demonstrates an increased risk of VTE for patients with RA treated with JAKi in clinical practice, an increase confined to pulmonary embolism rather than deep vein thrombosis.

Study IV was a cohort study on the potential impact of channeling bias for biologic and targeted synthetic DMARD (b/tsDMARD) drug classes when new on the market, compared to when established, in patients with RA from 2006 through 2022. We identified 33,550 b/tsDMARD initiations and 5,862 composite safety events (first of: major adverse cardiovascular event, VTE, cancer or serious infection). b/tsDMARD treatments initiated >5 years since market entry of the drug class in question (vs. < 2 years) were associated with lower rates of the composite safety outcome (unadjusted HR 0.74 (95% CI 0.67-0.81)). This association was attenuated after adjusting for patient characteristics (HR 0.93 (95% CI 0.84- 1.03)). By contrast, rates of the composite safety outcome decreased gradually for b/tsDMARD initiations during the study period, although these rates were stable in bionaïve RA and the general population. The study demonstrates a relatively modest level of channeling, although it makes b/tsDMARD drug classes appear more harmful when new on the market. Since the rates of these composite safety outcomes gradually decreased over calendar time for RA patients treated with b/tsDMARD, this suggests that the bar that defines acceptable rates of safety events for new and future b/tsDMARD should be lower.

List of scientific papers

I. Validation and characterization of venous thromboembolism diagnoses in the Swedish National Patient Register among patients with rheumatoid arthritis. Viktor Molander, Hannah Bower, Johan Askling. Scand J Rheumatol 2023;52:111-117. https://doi.org/10.1080/03009742.2021.2001907

II. Risk of venous thromboembolism in rheumatoid arthritis, and its association with disease activity: a nationwide cohort study from Sweden. Viktor Molander, Hannah Bower, Thomas Frisell, Johan Askling. Annals of the Rheumatic Diseases 2021;80:169-175. https://doi.org/10.1136/annrheumdis-2020-218419

III. Risk of venous thromboembolism with JAK inhibitors and other immune- modulatory drugs: a Swedish comparative safety study among patients with rheumatoid arthritis. Viktor Molander, Hannah Bower, Thomas Frisell, Bénédicte Delcoigne, Daniela Di Giuseppe, Johan Askling. Annals of the Rheumatic Diseases 2023;82:189-197. https://doi.org/10.1136/ard-2022-223050

IV. Do newly approved drugs have a worse observed safety profile than once established? A study on time-trends in observed risks of key safety outcomes with immune-modulatory drugs against rheumatoid arthritis. Viktor Molander, Hannah Bower, Thomas Frisell, Johan Askling. [Manuscript]