Incidence of a first venous thrombotic event in people with HIV in the Netherlands: a retrospective cohort study

Background: The risk of venous thrombotic events is elevated in people with HIV, but overall risk estimates and estimates specific to immune status and antiretroviral medication remain i mprecise. In this study, we aimed to estimate these parameters in a large cohort of people with HIV in the Netherlands. Methods: In this retrospective cohort study, we used the Dutch ATHENA cohort to estimate crude, age and sex standardised, and risk period-specific incidences of a first venous thrombotic event in people with HIV aged 18 years or older attending 12 HIV treatment centres in the Netherlands. Crude and standardised incidences were compared with European population-level studies of venous thrombotic events. We used time-updated Cox regression to estimate the risk of a first venous thrombotic event in association with HIV-specific factors (CD4 cell count, viral load, recent opportunistic infections, antiretroviral medication use) adjusted for traditional risk factors for venous thrombotic events. Findings: With data collected from Jan 1, 2003, to April 1, 2015, our study cohort included 14 389 people with HIV and 99 762 person-years of follow-up, with a median follow-up of 7·2 years (IQR 3·3–11·1). During this period, 232 first venous thrombotic events occurred, yielding a crude incidence of 2·33 events per 1000 person-years (95% CI 2·04–2·64) and an incidence standardised for age and sex of 2·50 events per 1000 (2·18–2·82). CD4 counts less than 200 cells per μL were independently associated with higher risk of a venous thrombotic event: adjusted hazard ratio (aHR) 3·40 (95% CI 2·28–5·08) relative to counts of 500 cells per μL. A high viral load (aHR 3·15, 95% CI 2·00–5·02; >100 000 copies per mL vs <50 copies per mL) and current or recent opportunistic adverse events (2·80, 1·77–4·44) were also independently associated with higher risk of a venous thrombotic event. There were no associations between any specific antiretroviral drugs and risk of a venous thrombotic event. Rates associated with pregnancy (9·4, 95% CI 4·6–17·3), malignancy (16·7, 10·6–25·1), and hospitalisation (24·4, 19·1–30·6) were lower than primary thromboprophylaxis thresholds suggested by the respective guidelines. Interpretation: Our findings support neither prescribing primary outpatient thromboprophylaxis nor avoiding any type of antiretroviral medication in people with HIV at high risk of a venous thrombotic event. Funding: Dutch Ministry of Health, Welfare and Sport.