[Skip to Navigation]
October 11, 1993

Prophylactic Agents for Venous Thrombosis in Elective Hip Surgery: Meta-analysis of Studies Using Venographic Assessment

Author Affiliations

From the Division of Area Medicine (Dr Mohr), Section of Clinical Epidemiology (Dr Silverstein), and Section of Biostatistics (Dr Murtaugh and Mr Harrison), Mayo Clinic and Mayo Foundation, Rochester, Minn.

Arch Intern Med. 1993;153(19):2221-2228. doi:10.1001/archinte.1993.00410190057007

Background:  We determined the relative efficacy of various agents or combinations of agents in the prophylaxis of deep venous thrombosis after elective hip arthroplasty.

Methods:  Peer-reviewed, English-language, human studies articles from 1975 through 1991 were obtained through a MEDLINE database search. Additional references were obtained from bibliographies. Articles that compared the effect of two or more prophylactic agents or placebo in preventing deep venous thrombosis as assessed by venography were selected for further review. Only studies of elective hip surgery in which all patients had venographic screening for thrombosis were included. Twenty-three of 101 studies met these criteria. Data were abstracted by one of us. Methodologic criteria and outcome data from each study were recorded and analyzed.

Results:  There was significant heterogeneity in the deep venous thrombosis rate among studies. Although the rates were lowest for low-molecular-weight heparin with or without the use of stockings, adjusted-dose heparin, and warfarin, many agents had similar low rates. There was less heterogeneity when the relative risk was used as a summary statistic for studies in which two agents were compared. With pairwise comparisons, low-molecular-weight heparin performed better than every agent with which it was compared. Other agents performed well but were not consistently better.

Conclusions:  Multiple agents or combinations are effective prophylaxis for deep venous thrombosis, but none decreases the rate to zero. There was overlap in the 95% confidence intervals for the probability of deep venous thrombosis for various agents and especially for the probabilities for proximal thrombi. Many agents have not been compared directly with each other, but low-molecular-weight heparin consistently performed well.(Arch Intern Med. 1993;153:2221-2228)

Add or change institution