Cause of death was classified as autopsy proven, clinically certain or unknown.
Deaths were grouped into cardiopulmonary excluding pulmonary embolism (PE), PE, bleeding-related, gastrointestinal, central nervous system, and others (miscellaneous). Meta-analysis based on fixed effects or random effects models was used for pooling incidence data.
In all, 70 studies were included (99 441 patients; 373 deaths). The mortality was lowest in the MM (0.2%) and WC (0.2%) groups.
The most frequent cause of death was cardiopulmonary (47.9%), followed by PE (25.4%) and bleeding (8.9%). The proportion of deaths due to PE was not significantly affected by the thromboprophylaxis regimen (PA, 35.5%; PAC, 28%; MM, 23.2%; and NRC, 16.3%). Fatal bleeding was higher in groups relying on the use of anticoagulation (W, 33.8%; PA, 9.4%; PAC, 10.8%) but the differences were not statistically significant.
Our study demonstrated that the routine use of PA does not reduce the overall mortality or the proportion of deaths due to PE.
DOI - The Journal of bone and joint surgery. British volume (DOI)