HIV patients are at increased risk for first venous thrombotic events (VTE) compared to the general population. What causes this increased risk and whether HIV patients are also at increased risk for recurrent VTE is unclear. An assessment of VTE risk factors and a reliable recurrence estimate are essential to determine the optimal duration of anticoagulant therapy (ACT). We assessed the risk factors for a first VTE and evaluated VTE recurrence rates in HIV patients.
Observational study using data of the AIDS Therapy Evaluation in the Netherlands (ATHENA) cohort. To identify VTE, we systematically reviewed charts of patients who initiated ACT in 2003-2014 in 12 participating centers, covering 70% of the Dutch HIV population. The sensitivity of this case finding strategy was confirmed by full chart review of all patients in 2 centers. We analyzed risk factors for first VTE (regardless of location) by time-updated Cox regression models. VTE recurrence after ACT withdrawal was analyzed in patients with a first VTE in proximal leg veins or pulmonary arteries only. VTE associated with estrogen use, pregnancy, surgery, cancer or trauma were considered provoked.
229 first VTE occurred in 14,386 HIV patients (80% males) during 97,556 person-years (py) of follow up (2.3 VTE/1000 py). Lower CD4 T-cell counts were independently associated with a higher first VTE risk in HIV patients and this association remained significant after adjustment for the high hospitalization rates in severely immunocompromised HIV patients (Table). HIV patients with >500 CD4 T-cells/mm3 had 1.3 VTE/1000 py while HIV patients with <200 CD4 T-cells/mm3 had 7.1 VTE/1000 py. 153 of 202 HIV patients with first VTE localized in proximal leg veins or pulmonary arteries withdrew ACT, including 108 with unprovoked VTE. VTE recurred in 31 HIV patients (57 VTE/1000 py; 95%CI: 39-79). Kaplan-Meier recurrence rates at 1, 2 and 5 years of follow up were 16%, 19%, and 28% following unprovoked first VTE and 5%, 9%, and 15% following provoked first VTE.
The increased risk for a first VTE in HIV patients was strongly driven by lower CD4 T-cell counts. VTE incidence in those with high CD4 T-cell counts approached the incidence in non-HIV patients. The VTE recurrence rate was high in patients with unprovoked first VTE and clustered in the first year after ACT withdrawal. Our results suggest that ACT withdrawal in HIV patients with unprovoked VTE and low CD4 T-cell counts should be cautiously considered.