Systematic Review and Meta-Analysis of Thromboprophylaxis with Heparins Following Intracerebral Hemorrhage

Abstract

Background The efficacy and safety of pharmacological thromboprophylaxis in patients with intracerebral hemorrhage (ICH) remains unclear. Methods A literature search was performed to collect studies comparing the effect of thromboprophylaxis in patients with ICH. The primary endpoints were deep vein thrombosis (DVT), pulmonary embolism (PE), and hematoma expansion or rebleeding. A meta-analytic approach was employed to estimate the relative risk (RR) by fitting fixed-effects (FE) and random-effects (RE) models. Results A total of 28 studies representing 3,697 hospitalized patients with ICH were included. Thromboprophylaxis was initiated within 4 days following hospital presentation and continued for 10 to 14 days in most of studies. Compared with control, thromboprophylaxis was associated with a reduced risk of DVT (47/1,399 [3.4%] vs. 202/1,377 [14.7%]; FE: RR, 0.24; 95% CI, 0.18–0.32; RE: RR, 0.27; 95% CI, 0.19–0.39) as well as PE (9/953 [0.9%] vs. 37/864 [4.3%]; FE: RR, 0.33; 95% CI, 0.19–0.57; RE: RR, 0.37; 95% CI, 0.21–0.66). Thromboprophylaxis was not associated with increased risk of hematoma expansion or rebleeding (32/1,319 [2.4%] vs. 37/1,301 [2.8%]; FE: RR, 0.75; 95% CI, 0.48–1.18; RE: RR, 0.80; 95% CI, 0.49–1.30) or mortality (117/925 [12.6%] vs. 139/904 [15.4%]; FE: RR, 0.82; 95% CI, 0.65–1.03; RE: RR, 0.83; 95% CI, 0.66–1.04). Conclusion Thromboprophylaxis was effective in preventing DVT and PE without increasing the risk of hematoma expansion or bleeding among ICH patients. Future studies should explore the long-term effects of thromboprophylaxis in this population, particularly on the functional outcomes.

Anticoagulation for Thromboprophylaxis in Patients with Intracerebral Hemorrhage: Less Room for Skepticism

“…Furthermore, pharmacological thromboprophylaxis (RR: 0.33, 95% CI: 0.19–0.57 for the fixed-effects model; RR: 0.37, 95% CI: 0.21–0.66 for the random-effects model) was associated with a lower risk of PE without any heterogeneity ( I 2 = 0%), with the PI substantially overlapping the estimates (CI: 0.20–0.69), strengthening the results of the meta-analysis. 16 Moreover, pharmacological thromboprophylaxis was not associated with increased risk for hematoma expansion or rebleeding (RR: 0.75, 95% CI: 0.48–1.18 for the fixed-effects model; RR: 0.80, 95% CI: 0.49–1.30 for the random-effects model), with no heterogeneity ( I 2 = 0%) and overlapping PI. Also, a trend of reduced mortality was identified in patients treated with pharmacological thromboprophylaxis (RR: 0.82, 95% CI: 0.65–1.03 for fixed-effects model; RR: 0.83, 95% CI: 0.66–1.04 for the random-effects model; I 2 = 0%; PI: 0.60–1.15).…”

“…Also, the additional analyses showed that among randomized controlled trials, patients receiving LMWH/UFH were associated with a lower risk of hematoma expansion or rebleeding (RR: 0.53, 95% CI: 0.28–0.99; I 2 = 0%; p = 0.13 for test for subgroup differences). 16 Finally, it is noted that there was a significant overall risk of bias for most of the studies enrolled, with eight studies being at high risk of bias.…”

“…The outcomes assessed included DVT, PE, hematoma expansion or rebleeding, major disability, and mortality. 16…”

“…14,15 In this issue of Thrombosis and Haemostasis, Chi and colleagues present a systematic review of the association between pharmacological thromboprophylaxis with LMWH or UFH and the risk of VTE in patients with ICH. 16 After a methodologically robust systematic search and study selec-tion, the authors included 28 studies and a total of 3,697 hospitalized patients with ICH. The mean patient age ranged between 50 and 72 years.…”

“…The outcomes assessed included DVT, PE, hematoma expansion or rebleeding, major disability, and mortality. 16 The investigators show that the use of pharmacological thromboprophylaxis was associated with a significant reduction in the risk of DVT, both in fixed-effects model (risk reduction [RR]: 0.24, 95% confidence interval [CI]: 0.28-0.32) and in the random-effects model (RR: 0.27, 95% CI: 0.19-0.39). These estimates were corroborated by low grade of heterogeneity (I 2 ¼ 25%) and by a strict prediction interval (PI; 0.11-0.66), which is a statistical tool used to examine the possible variation of pooled estimates according to future studies performed in different clinical scenarios and with different clinical characteristics.…”