Timing of recurrent venous thromboembolism early after the index event: A meta-analysis of randomized controlled trials

Introduction Studies suggest a decreasing risk of recurrent venous thromboembolism (rVTE) in relation to time since the index event. We sought to conduct a meta-analysis examining the time course of rVTE over the first 3-months of anticoagulation. Materials and Methods A literature search of MEDLINE, EMBASE and CENTRAL (through 4/2013) was conducted to identify randomized trials of acute pharmacologic treatment and prevention of rVTE, enrolling ≥ 200 subjects/treatment arm, requiring anticoagulation for ≥ 3-months and reporting time-to-objectively-confirmed rVTE. Trials assessing extended-duration treatment, randomizing only cancer patients or not in English were excluded. Treatment arms were divided into monthly and weekly time periods for comparison (months 1-3 and weeks 1-12 after the index event). Treatment arm rVTE rates (per person-year) were pooled using a random-effects approach. Results Fifteen trials (31 treatment arms; n = 27,237) were included. Higher rVTE rates were observed during the first month after the index event (0.19, 95%CI = 0.16-0.23) compared to the second (0.05, 95%CI 0.04-0.06; p < 0.001 vs. first month) and third months (0.02, 95%CI = 0.02-0.03; p < 0.001 vs. first month). While the highest rate of rVTE was in week 1 (0.29, 95%CI = 0.21-0.37; p < 0.01 vs. week 2), rates remained high through the fourth week (between 0.15 and 0.10 events/person-year) before decreasing and stabilizing at week 5 (≤ 0.05 events/person-year; p < 0.01 vs. week 4). Conclusions Our findings demonstrate a significant interaction between rVTE rates and time after the index event. High rVTE rates during the 3-4 weeks following the index event emphasize the importance of frequent surveillance during this time and the early optimization of pharmacologic therapy.