Background: Multiple studies have described a higher incidence of venous thromboembolism (VTE) in people living with an HIV infection (PWH). However, data on the risk of recurrent VTE in this population are lacking, although this question is more important for clinical practice. This study aims to estimate the risk of recurrent VTE in PWH compared to controls and to identify risk factors for recurrence within this population.

Methods and findings: PWH with a first VTE were derived from the AIDS Therapy Evaluation in the Netherlands (ATHENA) cohort (2003-2015), a nationwide ongoing cohort following up PWH in care in the Netherlands. Uninfected controls were derived from the Multiple Environmental and Genetic Assessment of risk factors for venous thrombosis (MEGA) follow-up study (1999-2003), a cohort of patients with a first VTE who initially participated in a case-control study in the Netherlands who were followed up for recurrent VTE. Selection was limited to persons with an index VTE suffering from deep vein thrombosis in the lower limbs and/or pulmonary embolism (PE). Participants were followed from withdrawal of anticoagulation to VTE recurrence, loss to follow-up, death, or end of study. We estimated incidence rates, cumulative incidence (accounting for competing risk of death) and hazard ratios (HRs) using Cox proportional hazards regression, adjusting for age, sex, and whether the index event was provoked or unprovoked. When analyzing risk factors among PWH, the main focus of analysis was the role of immune markers (cluster of differentiation 4 [CD4]+ T-cell count). There were 153 PWH (82% men, median 48 years) and 4,005 uninfected controls (45% men, median 49 years) with a first VTE (71% unprovoked in PWH, 34% unprovoked in controls) available for analysis. With 40 VTE recurrences during 774 person-years of follow-up (PYFU) in PWH and 635 VTE recurrences during 20,215 PYFU in controls, the incidence rates were 5.2 and 3.1 per 100 PYFU (HR: 1.70, 95% CI 1.23-2.36, p = 0.003). VTE consistently recurred more frequently per 100 PYFU in PWH in all predefined subgroups of men (5.6 versus 4.8), women (3.6 versus 1.9), and unprovoked (6.0 versus 5.2) or provoked (3.1 versus 2.1) first VTE. After adjustment, the VTE recurrence risk was higher in PWH compared to controls in the first year after anticoagulant discontinuation (HR: 1.67, 95% CI 1.04-2.70, p = 0.03) with higher cumulative incidences in PWH at 1 year (12.5% versus 5.6%) and 5 years (23.4% versus 15.3%) of follow-up. VTE recurred less frequently in PWH who were more immunodeficient at the first VTE, marked by a better CD4+ T-cell recovery on antiretroviral therapy and during anticoagulant therapy for the first VTE (adjusted HR: 0.81 per 100 cells/mm3 increase, 95% CI 0.67-0.97, p = 0.02). Sensitivity analyses addressing potential sources of bias confirmed our principal analyses. The main study limitations are that VTEs were adjudicated differently in the cohorts and that diagnostic practices changed during the 20-year study period.

Conclusions: Overall, the risk of recurrent VTE was elevated in PWH compared to controls. Among PWH, recurrence risk appeared to decrease with greater CD4+ T-cell recovery after a first VTE. This is relevant when deciding to (dis)continue anticoagulant therapy in PWH with otherwise unprovoked first VTE.