Ten RCTs (n=1,335) were included, one with two publications.
Randomisation methods were described in only 6 studies. None of the 5 studies comparing heparin with LMWH provided information on randomisation. There was heterogeneity between the studies with respect to heparin dosage, definitions of VTE and the reporting of bleeding outcomes. One trial reported an adequate power size calculation and another two were powered to detect differences in bleeding but not VTE.
Heparin versus control.
The pooling of 2 RCTs (n=489) showed a statistically significant decrease in the risk of DVT in women receiving heparin over a 30-day follow-up (RR 0.58, 95% CI: 0.35, 0.95). No statistically significant difference was found between the groups in the risk of PE or VTE. Data for bleeding outcomes could not be pooled, but no statistically significant difference was found between the groups in the individual studies.
Heparin versus LMWH.
The pooling of 5 RCTs (n=320) found no statistically significant difference between the groups in the risk of PE or VTE. Similarly, no statistically significant difference was found for bleeding outcomes (1 RCT, n=102).
SCD versus control (1 RCT, n=107).
There was a statistically significant decrease in the risk of DVT in women receiving SCDs (RR 0.28, 95% CI: 0.11, 0.66). No statistically significant difference was found between the groups in the risk of PE.
SCD versus heparin or LMWH.
No statistically significant difference was found between SCD and heparin in the risk of DVT or PE (1 RCT, n=208). Significantly more blood transfusions were necessary in the heparin group. No statistically significant difference was found between SCD and LMWH in the risk of DVT, PE or bleeding (1 RCT, n=211).