JBJS | A Meta-Analysis of Thromboembolic Prophylaxis Following Elective Total Hip Arthroplasty*

The Journal of Bone & Joint Surgery, Volume 82, Issue 7

Articles | July 01, 2000

A Meta-Analysis of Thromboembolic Prophylaxis Following Elective Total Hip Arthroplasty*

Kevin B. Freedman, M.D., M.S.C.E.†; Keith R. Brookenthal, M.D.†; Robert H. Fitzgerald, Jr, M.D.†; Sankey Williams, M.D.†; Jess H. Lonner, M.D.†

View Disclosures and Other Information

Investigation performed at the University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
†Department of Orthopaedic Surgery (K. B. F., K. R. B., R. H. F., Jr., and J. H. L.), Division of General Internal Medicine (S. W.), Center for Clinical Epidemiology and Biostatistics (K. B. F. and S. W.), and Department of Biostatistics and Epidemiology (K. B. F. and S. W.), University of Pennsylvania School of Medicine, 2 Silverstein Pavilion, 3400 Spruce Street, Philadelphia, Pennsylvania 19104. E-mail address for K. B. Freedman: kfreedma@mail.med.upenn.edu.

The Journal of Bone & Joint Surgery. 2000; 82:929-929

5 Recommendations (Recommend) | 3 Comments | Saved by 3 Users Save Case

Article

Tables

References

text A A A

Abstract

Background: Although several agents have been shown to reduce the risk of thromboembolic disease, there is no clear preference for thromboembolic prophylaxis in elective total hip arthroplasty. The purpose of this study was to define the efficacy and safety of the agents that are currently used for prophylaxis against deep venous thrombosis - namely, low-molecular-weight heparin, warfarin, aspirin, low-dose heparin, and pneumatic compression.

Methods: A Medline search identified all randomized, controlled trials, published from January 1966 to May 1998, that compared the use of one of the prophylactic agents with the use of any other agent or a placebo in patients undergoing elective total hip arthroplasty. For a study to be included in our analysis, bilateral venography had to have been performed to confirm the presence or absence of deep venous thrombosis. Fifty-two studies, in which 10,929 patients had been enrolled, met the inclusion criteria and were included in the analysis. The rates of distal, proximal, and total (distal and proximal) deep venous thrombosis; symptomatic and fatal pulmonary embolism; minor and major wound-bleeding complications; major non-wound bleeding complications; and total mortality were determined for each agent in each study. The absolute risk of each outcome was determined by dividing the number of events by the number of patients at risk. A general linear model with random effects was used to calculate the 95 percent confidence interval of risk. A crosstabs of study by outcome was performed to test homogeneity (ability to combine studies). The risk of each outcome was compared among agents and between each agent and the placebo.

Results: With prophylaxis, the risk of total (proximal and distal) deep venous thrombosis ranged from 17.7 percent (low-molecular-weight heparin) to 31.1 percent (low-dose heparin); the risk with prophylaxis with any agent was significantly lower than the risk with the placebo (48.5 percent) (p < 0.0001). The risk of proximal deep venous thrombosis was lowest with warfarin (6.3 percent) and low-molecular-weight heparin (7.7 percent), and again the risk with any prophylactic agent was significantly lower than the risk with the placebo (25.8 percent) (p < 0.0001). Compared with the risk with the placebo (1.51 percent), only warfarin (0.16 percent), pneumatic compression (0.26 percent), and low-molecular-weight heparin (0.36 percent) were associated with a significantly lower risk of symptomatic pulmonary embolism. There were no significant differences among agents with regard to the risk of fatal pulmonary embolism or of mortality with any cause. The risk of minor wound-bleeding was significantly higher with low-molecular-weight heparin (8.9 percent) and low-dose heparin (7.6 percent) than it was with the placebo (2.2 percent) (p < 0.05). Compared with the risk with the placebo (0.28 percent), only low-dose heparin was associated with a significantly higher risk of major wound-bleeding (2.56 percent) and total major bleeding (3.46 percent) (p < 0.0001).

Conclusions: The best prophylactic agent in terms of both efficacy and safety was warfarin, followed by pneumatic compression, and the least effective and safe was low-dose heparin. Warfarin provided the lowest risk of both proximal deep venous thrombosis and symptomatic pulmonary embolism. However, there were no identifiable significant differences in the rates of fatal pulmonary embolism or death among the agents. Significant risks of minor and major bleeding complications were observed with greater frequency with certain prophylactic agents, particularly low-molecular-weight heparin (minor bleeding) and low-dose heparin (both major and minor bleeding).

Figures in this Article

Introduction

Introduction | Materials and Methods | Results | Discussion | References

Venous thromboembolism is the most commonly seen complication following total hip arthroplasty. In the absence of prophylaxis, the risk of deep venous thrombosis has been reported to range from 39 to 74 percent, and the risk of fatal pulmonary embolism has ranged from 0.19 to 3.4 percent^8,18. Several prophylactic agents, including warfarin, low-molecular-weight heparin, low-dose heparin, aspirin, and pneumatic compression stockings, have been shown to reduce the risk of deep venous thrombosis⁶². However, there is considerable controversy over the best available agent for thromboembolic prophylaxis in elective total hip surgery.

Several factors are responsible for the lack of a clear choice of thromboembolic prophylaxis. The most common reason cited for not using prophylaxis is uncertainty regarding the safest and most effective agent⁵¹. A surgeon's choice of prophylactic agent is determined by weighing the risk of thromboembolic disease against the risk of bleeding complications due to the prophylaxis. However, the most serious adverse consequences of total hip arthroplasty, such as fatal pulmonary embolism and death, may be too rare to study as an outcome in clinical trials. Although the ultimate goal of any thromboembolic prophylactic regimen is the prevention of death, several less consequential outcomes may be salient for patients undergoing elective total hip arthroplasty. Therefore, the decision regarding prophylaxis against thromboembolic disease depends primarily on which events one is attempting to prevent: all deep venous thromboses, proximal deep venous thromboses, all pulmonary emboli, fatal pulmonary emboli, death, or all of the above. When considering the issue of safety, one must determine which adverse consequences are important: minor wound-bleeding, major wound-bleeding, or major non-wound bleeding (gastrointestinal or intracerebral hemorrhage). The currently available evidence on the efficacy and safety of thromboembolic prophylaxis may be inadequate because the relative efficacy and safety of each agent has not been defined in a single study, to our knowledge.

Meta-analysis is a technique that statistically combines or integrates the results of several independent clinical trials to increase statistical power. Meta-analysis is an attractive alternative for answering clinically important questions when a large, expensive, and logistically difficult trial would otherwise be necessary. Previous meta-analyses of thromboembolic prophylaxis in patients undergoing total joint arthroplasty have had several shortcomings, including exclusion of clinically important agents, use of standards other than venography for detection of deep venous thrombosis, and failure to include all of the outcomes that are important factors in the decision about prophylaxis^{1,36,45,47,48,50,57}. We addressed these problems in the current meta-analysis by including only randomized, controlled trials that involved venographic assessment of deep venous thrombosis as well as by calculating the risks for all major outcomes.

The purpose of our study was to determine and compare the efficacy and safety of the currently accepted methods of thromboembolic prophylaxis in patients undergoing total hip arthroplasty. Specifically, the study was designed to assimilate the best estimate of risk, for each major prophylactic agent, of distal and proximal deep venous thrombosis, symptomatic and fatal pulmonary embolism, minor and major wound-bleeding complications, minor and major non-wound bleeding complications, and death.

View Larger

Table I : Description of Studies Included in the Meta-Analysis

Treatment Arm	No. of Studies	Citations	No. of Patients	Mean Age (yrs.)	Mean Percent Male	Mean No. of Days Postop. Venography Performed
Low-molecular-weight heparin	21	5, 6, 9, 10, 14, 15, 22, 23, 25, 26, 31, 34, 38-40, 42, 52-54, 60, 61	5512	66	43	10
Warfarin	12	2, 17, 20-22, 26, 27, 34, 37, 44, 49, 54	1493	63	52	9
Aspirin	8	27-30, 35, 43, 44, 59	687	62	45	9
Low-dose heparin	11	5, 7, 15, 16, 24, 25, 27, 40, 41, 52, 56	1859	66	42	10
Pneumatic compression	5	2, 21, 33, 37, 49	431	64	53	9
Placebo	13	3, 6, 19, 28, 31-33, 38, 39, 55, 58, 60, 61	947	67	42	10

Add to My JBJS

Table I : Description of Studies Included in the Meta-Analysis

Treatment Arm	No. of Studies	Citations	No. of Patients	Mean Age (yrs.)	Mean Percent Male	Mean No. of Days Postop. Venography Performed
Low-molecular-weight heparin	21	5, 6, 9, 10, 14, 15, 22, 23, 25, 26, 31, 34, 38-40, 42, 52-54, 60, 61	5512	66	43	10
Warfarin	12	2, 17, 20-22, 26, 27, 34, 37, 44, 49, 54	1493	63	52	9
Aspirin	8	27-30, 35, 43, 44, 59	687	62	45	9
Low-dose heparin	11	5, 7, 15, 16, 24, 25, 27, 40, 41, 52, 56	1859	66	42	10
Pneumatic compression	5	2, 21, 33, 37, 49	431	64	53	9
Placebo	13	3, 6, 19, 28, 31-33, 38, 39, 55, 58, 60, 61	947	67	42	10

View Larger

Table II: Risk of Deep Venous Thrombosis*

*CI = confidence interval.†The agent differed significantly from the placebo.‡Aspirin did not differ significantly from the placebo for distal deep venous thrombosis.§The p value was <0.0001 for the overall test of the difference between prophylaxis with an agent and the placebo.

Agent	No. of Patients	Distal Deep Venous Thrombosis		Proximal Deep Venous Thrombosis		Total Deep Venous Thrombosis
No. of Events	Risk (95 Percent CI) (percent)	No. of Events	Risk (95 Percent CI) (percent)	No. of Events	Risk (95 Percent CI) (percent)
Low-molecular- weight heparin	5512	566	9.6 (7.7, 12.0)	342	7.7 (5.7, 10.3)	918	17.7 (14.5, 21.4)
Warfarin†	1493	244	17.1 (13.3, 21.6)	90	6.3 (4.7, 8.4)	334	23.2 (19.3, 27.7)
Aspirin†‡	687	134	19.7 (14.2, 26.8)	82	11.4 (7.3, 17.5)	214	30.6 (21.2, 41.8)
Low-dose heparin†	1859	201	10.8 (8.4, 13.9)	222	19.0 (12.9, 27.2)	429	31.1 (22.7, 40.9)
Pneumatic compression†	431	36	7.7 (4.5, 12.9)	58	13.3 (9.9, 17.7)	94	20.7 (14.6, 28.7)
Placebo§	947	211	22.4 (18.8, 26.6)	245	25.8 (21.4, 30.7)	456	48.5 (43.4, 53.7)

Add to My JBJS

Table II: Risk of Deep Venous Thrombosis*

Agent	No. of Patients	Distal Deep Venous Thrombosis		Proximal Deep Venous Thrombosis		Total Deep Venous Thrombosis
No. of Events	Risk (95 Percent CI) (percent)	No. of Events	Risk (95 Percent CI) (percent)	No. of Events	Risk (95 Percent CI) (percent)
Low-molecular- weight heparin	5512	566	9.6 (7.7, 12.0)	342	7.7 (5.7, 10.3)	918	17.7 (14.5, 21.4)
Warfarin†	1493	244	17.1 (13.3, 21.6)	90	6.3 (4.7, 8.4)	334	23.2 (19.3, 27.7)
Aspirin†‡	687	134	19.7 (14.2, 26.8)	82	11.4 (7.3, 17.5)	214	30.6 (21.2, 41.8)
Low-dose heparin†	1859	201	10.8 (8.4, 13.9)	222	19.0 (12.9, 27.2)	429	31.1 (22.7, 40.9)
Pneumatic compression†	431	36	7.7 (4.5, 12.9)	58	13.3 (9.9, 17.7)	94	20.7 (14.6, 28.7)
Placebo§	947	211	22.4 (18.8, 26.6)	245	25.8 (21.4, 30.7)	456	48.5 (43.4, 53.7)

View Larger

Table III: Multiple Comparisons of Agents for Risk of Deep Venous Thrombosis

*The agent associated with the lower risk appears first in each entry.†Significant at the p < 0.0083 level.

Comparison*	P Values
Total Deep Venous Thrombosis	Distal Deep Venous Thrombosis	Proximal Deep Venous Thrombosis
Low-molecular-weight heparin vs. warfarin	0.1916	0.0047†	0.3513†
Low-molecular-weight heparin vs. aspirin	0.1121	0.0005	0.8862
Low-molecular-weight heparin vs. low-dose heparin	0.0004†	0.5185	0.0023†
Low-molecular-weight heparin vs. pneumatic compression	0.8031	0.0673	0.0059†
Warfarin vs. aspirin	0.4819	0.1880	0.7955
Warfarin vs. low-dose heparin	0.6356	0.0219	0.0047†
Warfarin vs. pneumatic compression	0.1962	0.0007†	0.0004†
Aspirin vs. low-dose heparin	0.7495	0.0018	0.1281
Pneumatic compression vs. aspirin	0.0970	0.0001	0.0831
Low-dose heparin vs. pneumatic compression	0.1491	0.0447	0.8936

Add to My JBJS

Table III: Multiple Comparisons of Agents for Risk of Deep Venous Thrombosis

*The agent associated with the lower risk appears first in each entry.†Significant at the p < 0.0083 level.

Comparison*	P Values
Total Deep Venous Thrombosis	Distal Deep Venous Thrombosis	Proximal Deep Venous Thrombosis
Low-molecular-weight heparin vs. warfarin	0.1916	0.0047†	0.3513†
Low-molecular-weight heparin vs. aspirin	0.1121	0.0005	0.8862
Low-molecular-weight heparin vs. low-dose heparin	0.0004†	0.5185	0.0023†
Low-molecular-weight heparin vs. pneumatic compression	0.8031	0.0673	0.0059†
Warfarin vs. aspirin	0.4819	0.1880	0.7955
Warfarin vs. low-dose heparin	0.6356	0.0219	0.0047†
Warfarin vs. pneumatic compression	0.1962	0.0007†	0.0004†
Aspirin vs. low-dose heparin	0.7495	0.0018	0.1281
Pneumatic compression vs. aspirin	0.0970	0.0001	0.0831
Low-dose heparin vs. pneumatic compression	0.1491	0.0447	0.8936

View Larger

Table IV: Risk of Pulmonary Embolism*

*CI = confidence interval.†The agent differed significantly from the placebo at p < 0.05 for the risk of symptomatic pulmonary embolism.

Agent	No. of Patients	Symptomatic Pulmonary Embolism		Fatal Pulmonary Embolism
No. of Events	Risk (95 Percent CI) (percent)	No. of Events	Risk (95 Percent CI) (percent)
Low-molecular-weight heparin†	5238	19	0.36 (0.22, 0.57)	2	0.04 (0.01, 0.14)
Warfarin†	1232	2	0.16 (0.02, 0.59)	2	0.16 (0.02, 0.59)
Aspirin	625	8	1.28 (0.55, 2.51)	2	0.32 (0.04, 1.15)
Low-dose heparin	1760	24	1.36 (0.88, 2.02)	3	0.17 (0.04, 0.50)
Pneumatic compression†	388	1	0.26 (0.01, 1.43)	1	0.26 (0.01, 1.43)
Placebo	860	13	1.51 (0.81, 2.57)	0	0.00 (0.00, 0.43)

Add to My JBJS

Table IV: Risk of Pulmonary Embolism*

*CI = confidence interval.†The agent differed significantly from the placebo at p < 0.05 for the risk of symptomatic pulmonary embolism.

Agent	No. of Patients	Symptomatic Pulmonary Embolism		Fatal Pulmonary Embolism
No. of Events	Risk (95 Percent CI) (percent)	No. of Events	Risk (95 Percent CI) (percent)
Low-molecular-weight heparin†	5238	19	0.36 (0.22, 0.57)	2	0.04 (0.01, 0.14)
Warfarin†	1232	2	0.16 (0.02, 0.59)	2	0.16 (0.02, 0.59)
Aspirin	625	8	1.28 (0.55, 2.51)	2	0.32 (0.04, 1.15)
Low-dose heparin	1760	24	1.36 (0.88, 2.02)	3	0.17 (0.04, 0.50)
Pneumatic compression†	388	1	0.26 (0.01, 1.43)	1	0.26 (0.01, 1.43)
Placebo	860	13	1.51 (0.81, 2.57)	0	0.00 (0.00, 0.43)

View Larger

Table V: Risk of Minor Bleeding*

*CI = confidence interval.†The agent differed significantly from the placebo at p < 0.05 for both minor wound-bleeding and total minor bleeding.‡The agent differed significantly from the placebo at p < 0.05 for total minor bleeding only.

Agent	Minor Wound-Bleeding			Total Minor Bleeding
No. of Patients	No. of Events	Risk (95 Percent CI) (percent)	No. of Patients	No. of Events	Risk (95 Percent CI) (percent)
Low-molecular-weight heparin†	3468	277	8.9 (4.2, 17.9)	4064	382	10.5 (5.9, 17.8)
Warfarin‡	1381	42	4.7 (2.1, 10.2)	1381	64	5.7 (3.1, 10.4)
Aspirin	687	6	1.2 (0.4, 3.3)	687	6	1.2 (0.4, 3.3)
Low-dose heparin	1244	65	7.6 (4.2, 13.5)	1453	149	13.5 (8.3, 21.2)
Pneumatic compression	338	3	1.1 (0.2, 6.1)	388	13	4.1 (1.7, 9.8)
Placebo	663	14	2.2 (0.6, 7.6)	663	18	3.0 (1.1, 8.2)

Add to My JBJS

Table V: Risk of Minor Bleeding*

Agent	Minor Wound-Bleeding			Total Minor Bleeding
No. of Patients	No. of Events	Risk (95 Percent CI) (percent)	No. of Patients	No. of Events	Risk (95 Percent CI) (percent)
Low-molecular-weight heparin†	3468	277	8.9 (4.2, 17.9)	4064	382	10.5 (5.9, 17.8)
Warfarin‡	1381	42	4.7 (2.1, 10.2)	1381	64	5.7 (3.1, 10.4)
Aspirin	687	6	1.2 (0.4, 3.3)	687	6	1.2 (0.4, 3.3)
Low-dose heparin	1244	65	7.6 (4.2, 13.5)	1453	149	13.5 (8.3, 21.2)
Pneumatic compression	338	3	1.1 (0.2, 6.1)	388	13	4.1 (1.7, 9.8)
Placebo	663	14	2.2 (0.6, 7.6)	663	18	3.0 (1.1, 8.2)

View Larger

Table VI: Risk of Major Bleeding*

*CI = confidence interval.†The agent differed significantly from the placebo at p < 0.05 for both major wound-bleeding and total major bleeding.

Agent	Major Wound-Bleeding			Total Major Bleeding
No. of Patients	No. of Events	Risk (95 Percent CI) (percent)	No. of Patients	No. of Events	Risk (95 Percent CI) (percent)
Low-molecular-weight heparin	4049	61	1.51 (1.15, 1.93)	5412	120	2.22 (1.84, 2.64)
Warfarin	1185	9	0.76 (0.35, 1.44)	1381	23	1.67 (1.06, 2.49)
Aspirin	687	2	0.29 (0.04, 1.05)	687	5	0.73 (0.24, 1.69)
Low-dose heparin†	1991	51	2.56 (1.91, 3.35)	1992	69	3.46 (2.70, 4.36)
Pneumatic compression	388	0	0.00 (0.00, 0.95)	388	0	0.00 (0.00, 0.95)
Placebo	713	2	0.28 (0.03, 1.01)	713	4	0.56 (0.15, 1.43)

Add to My JBJS

Table VI: Risk of Major Bleeding*

*CI = confidence interval.†The agent differed significantly from the placebo at p < 0.05 for both major wound-bleeding and total major bleeding.

Agent	Major Wound-Bleeding			Total Major Bleeding
No. of Patients	No. of Events	Risk (95 Percent CI) (percent)	No. of Patients	No. of Events	Risk (95 Percent CI) (percent)
Low-molecular-weight heparin	4049	61	1.51 (1.15, 1.93)	5412	120	2.22 (1.84, 2.64)
Warfarin	1185	9	0.76 (0.35, 1.44)	1381	23	1.67 (1.06, 2.49)
Aspirin	687	2	0.29 (0.04, 1.05)	687	5	0.73 (0.24, 1.69)
Low-dose heparin†	1991	51	2.56 (1.91, 3.35)	1992	69	3.46 (2.70, 4.36)
Pneumatic compression	388	0	0.00 (0.00, 0.95)	388	0	0.00 (0.00, 0.95)
Placebo	713	2	0.28 (0.03, 1.01)	713	4	0.56 (0.15, 1.43)

View Larger

Table VII : Risk of Death*

*CI = confidence interval.

Agent	No. of Patients	No. of Events	Risk (95 Percent CI) (percent)
Low-molecular-weight heparin	5918	12	0.20 (0.10, 0.35)
Warfarin	1555	7	0.45 (0.18, 0.93)
Aspirin	687	2	0.29 (0.04, 1.05)
Low-dose heparin	1992	10	0.50 (0.24, 0.92)
Pneumatic compression	388	4	1.03 (0.28, 2.62)
Placebo	900	3	0.33 (0.07, 0.97)

Add to My JBJS

Table VII : Risk of Death*

*CI = confidence interval.

Agent	No. of Patients	No. of Events	Risk (95 Percent CI) (percent)
Low-molecular-weight heparin	5918	12	0.20 (0.10, 0.35)
Warfarin	1555	7	0.45 (0.18, 0.93)
Aspirin	687	2	0.29 (0.04, 1.05)
Low-dose heparin	1992	10	0.50 (0.24, 0.92)
Pneumatic compression	388	4	1.03 (0.28, 2.62)
Placebo	900	3	0.33 (0.07, 0.97)

Materials and Methods

Introduction | Materials and Methods | Results | Discussion | References

Identification and Selection of Studies

We performed a literature search of the Medline computerized database for articles published from January 1966 to May 1998. Our purpose was to identify all English-language literature included under the key words hip, hip arthroplasty, hip replacement, thrombosis, thromboembolism, pulmonary embolism, and prophylaxis. We also performed a manual search of our own reference files and the bibliographies of all included articles. Additionally, previous meta-analyses and review articles^{1,4,36,45-47,57} were cross-referenced to ensure inclusion of all relevant articles. The contents of 112 pertinent abstracts or full-text articles that were identified during our literature search were reviewed to determine whether they met the criteria for inclusion in our meta-analysis. No attempt was made to solicit unpublished data or to retrieve additional information from any of the authors of the studies.

Studies were selected for inclusion on the basis of strict methodological criteria: (1) the study had to be a prospective, randomized, controlled trial comparing one of the prophylactic agents of interest with any other method of prophylaxis or with a placebo, (2) it had to have involved patients treated with elective total hip arthroplasty, (3) the prophylaxis had to have been started within twenty-four hours after the surgery and to have been continued for at least seven days, and (4) deep venous thrombosis had to have been confirmed by bilateral venography. Only studies in which lower-extremity venography had been performed on all patients were included. If a trial included patients other than those undergoing elective total hip arthroplasty, it had to be possible to segregate the patients with elective total hip arthroplasty from those with a fracture or those undergoing total knee arthroplasty. If the patients of interest could not be clearly distinguished, the trial was excluded. We also excluded trials that included a combination of the agents evaluated in this study. We allowed studies in which treatment with graduated compression stockings had been combined with pharmacological prophylaxis; however, we excluded those in which pneumatic mechanical compression had been combined with pharmacological treatment. Anesthetic techniques were not considered. Studies were searched for duplicate data, and when any such data was identified it was deleted. Two investigators independently assessed all retrieved articles for eligibility and extracted the results. Fifty-two studies met the described criteria and were included in the meta-analysis.

The prophylactic agents that were evaluated in our study included (1) low-molecular-weight heparin, thirty to forty-five milligrams twice daily; (2) warfarin, either fixed low-dose or adjusted-dose; (3) aspirin, 150 to 900 milligrams twice daily; (4) low-dose heparin, 3500 to 7500 units two or three times daily; (5) intermittent lower-extremity pneumatic compression stockings; and (6) placebo. Low-dose heparin with dihydroergotamine, low-dose heparin with antithrombin III, low-dose heparin with sulfinpyrazone, adjusted-dose heparin, dextran, and aspirin-sulfinpyrazone were not considered because of an inadequate number of studies or infrequent clinical use. If an agent that was included in our study was compared with one of the above excluded agents, only the treatment arm of the included agent was used in the analysis.

Data Abstraction

All data was abstracted in duplicate with use of a standardized protocol and reporting form, and disagreements were resolved by consensus. No authors of the collected studies were contacted to request additional information. The study characteristics that were recorded included the prophylactic agents that had been compared, the dose of each agent, the time that the prophylaxis had been started, the duration of the prophylaxis, the number of patients in each arm of the trial, and patient demographics.

The primary outcomes of efficacy were distal deep venous thrombosis, proximal deep venous thrombosis, total (distal and proximal) deep venous thrombosis, symptomatic pulmonary embolism, and fatal pulmonary embolism. All diagnoses of deep venous thromboses had to have been confirmed venographically. The diagnosis of symptomatic pulmonary embolism was based on the determination of the authors of each study. Asymptomatic pulmonary embolism was not considered since ventilation-perfusion scans and pulmonary angiography were not routinely performed on asymptomatic patients.

The primary outcomes for safety were minor wound-bleeding (wound hematoma not requiring operative decompression), total minor bleeding (including minor wound-bleeding, bleeding at the injection site, and epistaxis), major wound-bleeding (wound hematoma requiring operative decompression), major non-wound bleeding (gastrointestinal, retroperitoneal, intracerebral, and epidural hematoma), total major bleeding (major wound-bleeding and major non-wound bleeding), and mortality due to any cause.

Appropriateness of Pooling

The trials were reviewed to determine if they could be combined. The trials were disassembled into treatment arms in order to combine each agent across studies since all individual agents had not been directly compared. The separation of trials into treatment arms was based on the premise that the treatment groups were clinically homogeneous in composition; such separation has been done previously in several studies^36,45-47. The similarity among treatment groups was determined on the basis of inclusion and exclusion criteria, the dose of the agent that was administered, when the prophylaxis was started, the duration of the prophylaxis, and patient demographics. Within each outcome event (for example, distal deep venous thrombosis), a crosstabs of study by outcome was performed to determine if there was a significant difference among studies for each individual agent (test of homogeneity).

Analysis

The absolute risk of each outcome event was determined for each agent. The point estimate of absolute risk was determined by calculating the number of events that had occurred in all studies and dividing that value by the number of patients at risk. If one of the outcomes had not been specifically determined in a study, the patients in that study were not considered to be at risk for that particular outcome and were therefore eliminated from the denominator. For example, if an author reported on minor bleeding without distinguishing minor wound-bleeding from minor non-wound bleeding, the data was used only to calculate the risk of total minor bleeding.

In order to accommodate for the pooling of multiple trials, the 95 percent confidence interval for the risk with each agent was determined with use of a general linear model, with study as a random effect¹¹. This model accounts for the increased variability in event rates when studies are combined. The practical effect of this approach is to increase the width of the estimated confidence interval as compared with that with a standard approach, which ignores the correlation between observations within a study¹³.

To compare the risk of each outcome among agents, an overall test for difference was performed. Agents were considered to be significantly different from the placebo when the value for alpha was less than 0.05. In addition, comparisons were made among agents for each outcome, with an adjustment for multiple comparisons. Due to the number of comparisons (all combinations among the six agents), only a p value of <0.05/6, or p < 0.0083, was considered to be significant.

All statistical analysis was performed with Intercooled Stata 5.0 software (Stata, College Station, Texas).

Results

Introduction | Materials and Methods | Results | Discussion | References

Appropriateness of Pooling

Fifty-two trials with 10,929 participants were included in our analysis. Three trials, which met the other inclusion criteria, were excluded because of duplicate data. The descriptive results are presented in Table I.

The trials had similar study populations. The inclusion criteria consisted of patients undergoing elective total hip arthroplasty, and the exclusion criteria consisted of prior thromboembolic disease or contraindications to anticoagulation.

When the crosstabs of study by outcome was carried out, there was some evidence of heterogeneity among studies with regard to several outcomes: distal deep venous thrombosis⁵⁶, total deep venous thrombosis⁵⁶, minor wound-bleeding⁴², and total minor bleeding^24,42. To determine the effect of heterogeneity on the parameter estimates and confidence intervals, all of the analyses were rerun and compared with the corresponding parameter estimates, with the studies expressing heterogeneity excluded. There was minimal change in the parameter estimates when these studies were removed, so they were retained in the analysis.

Risk of Deep Venous Thrombosis

The absolute risks of distal, proximal, and total deep venous thrombosis (as diagnosed by venography) with each agent are presented in Table II. Compared with the placebo, all agents significantly decreased the risk of deep venous thrombosis (p < 0.0001).

The results of the statistical analysis for the multiple comparisons among agents for the risk of total, distal, and proximal deep venous thrombosis are presented in Table III. Only low-molecular-weight heparin and low-dose heparin differed significantly with regard to the risk of total deep venous thrombosis (p = 0.0004). Otherwise, there were no significant differences among the agents with regard to this outcome. Pneumatic compression was associated with the lowest risk of distal deep venous thrombosis (7.7 percent); this risk was significantly lower than the risks with warfarin (17.1 percent) and aspirin (19.7 percent) (p = 0.0007 and p = 0.0001, respectively). The risk of distal deep venous thrombosis with low-molecular-weight heparin (9.6 percent) was also significantly lower than the risks with warfarin and aspirin (p = 0.0047 and p = 0.0005, respectively). Warfarin and low-molecular-weight heparin were associated with the lowest risks of proximal deep venous thrombosis (6.3 and 7.7 percent, respectively); these risks were significantly lower than the risks with pneumatic compression (13.3 percent) (p = 0.0004 and p = 0.0059, respectively) and low-dose heparin (19.0 percent) (p = 0.0047 and p = 0.0023, respectively), but they were not significantly lower than the risk with aspirin. There were no other significant differences among agents.

Risk of Pulmonary Embolism

The absolute risks of symptomatic pulmonary embolism and fatal pulmonary embolism with each agent are presented in Table IV. Each fatal pulmonary embolism was considered symptomatic by definition for the analysis.

The rates of symptomatic pulmonary embolism were lowest with warfarin (0.16 percent), pneumatic compression (0.26 percent), and low-molecular-weight heparin (0.36 percent); all of these risks were significantly lower than that with the placebo (1.51 percent) (p < 0.0001). In addition, the risks of symptomatic pulmonary embolism with warfarin and low-molecular-weight heparin were significantly lower than the risks with aspirin (1.28 percent) and low-dose heparin (1.36 percent) (p < 0.0083).

Compared with the placebo, no agent was associated with a significantly different risk of fatal pulmonary embolism.

Risk of Minor Bleeding

The absolute risks of minor bleeding with each agent are presented in Table V. The risks of minor wound-bleeding were highest with low-molecular-weight heparin (8.9 percent) and low-dose heparin (7.6 percent); these risks were significantly higher than the risks with the placebo (2.2 percent) (p < 0.05) and pneumatic compression stockings (1.1 percent) (p < 0.0083). In addition, the risks of total minor bleeding with low-molecular-weight heparin, low-dose heparin, and warfarin were higher than the risk with the placebo (p < 0.05). The risks of total minor bleeding with low-molecular-weight heparin (10.5 percent) and low-dose heparin (13.5 percent) were significantly higher than the risk with pneumatic compression stockings (4.1 percent) (p = 0.0013 and p = 0.0022, respectively). There were no other significant differences among agents with regard to the risk of minor wound-bleeding or total minor bleeding.

Risk of Major Bleeding

The absolute risks of major bleeding with each agent are presented in Table VI. Low-dose heparin was associated with the highest risks of major wound-bleeding (2.56 percent) and total major bleeding (3.46 percent); both risks were significantly higher than the risks with the placebo and all other agents (p < 0.0001). There were no significant differences between any of the other agents and the placebo, or among the agents, with regard to the risk of major wound-bleeding, major non-wound bleeding, or total major bleeding.

There was only one documented spinal hematoma²⁵, which occurred in a patient who had undergone spinal anesthesia combined with continuous epidural anesthesia for postoperative pain control. Administration of low-molecular-weight heparin had been initiated on the evening prior to surgery. The risk of spinal hematoma with low-molecular-weight heparin was one of 4049, or 0.025 percent.

There were no reported intracerebral hemorrhages.

Risk of Mortality Due to Any Cause

The absolute risks of mortality due to any cause with each agent are presented in Table VII. The risk of mortality was not significantly different between any agent and the placebo or among the agents. The deaths included those due to fatal pulmonary embolism. The overall risk of mortality ranged from 0.20 percent (with low-molecular-weight heparin) to 1.03 percent (with pneumatic compression).

Discussion

Introduction | Materials and Methods | Results | Discussion | References

The prevention of thromboembolic disease is one of the major controversies surrounding elective total hip arthroplasty. Because of the rarity of the most serious outcomes, previous randomized, controlled trials have lacked the statistical power to define the optimum agent for thromboembolic prophylaxis. A rigidly performed meta-analysis can provide an objective appraisal of the evidence in order for clinicians to make a well informed decision. The present study demonstrated that, compared with a placebo, low-molecular-weight heparin, warfarin, aspirin, low-dose heparin, and pneumatic compression can effectively reduce the risk of total (proximal and distal) deep venous thrombosis and proximal deep venous thrombosis as determined by routine venography.

The risks of total deep venous thrombosis were lowest with low-molecular-weight heparin (17.7 percent), pneumatic compression (20.7 percent), and warfarin (23.2 percent). These were the only agents that significantly decreased the risk of symptomatic pulmonary embolism. Of the three agents, warfarin was associated with the lowest risk of proximal deep venous thrombosis (6.3 percent) and symptomatic pulmonary embolism (0.16 percent). Warfarin therefore seems to be the most effective agent for the prevention of thromboembolism. It should be recognized, however, that there is substantial overlap of the confidence intervals when warfarin is compared with low-molecular-weight heparin, indicating that low-molecular-weight heparin may be as efficacious as warfarin in the prevention of proximal deep venous thrombosis and symptomatic pulmonary embolism. In addition, the risk of distal deep venous thrombosis with low-molecular-weight heparin was significantly lower than the risk of that outcome with warfarin.

Examination of the safety of prophylaxis demonstrated that low-dose heparin, low-molecular-weight heparin, and warfarin were associated with the highest risks of minor wound-bleeding, major wound-bleeding, and total major bleeding. However, the risks of major bleeding with warfarin and low-molecular-weight heparin were not significantly different from that with the placebo. This data suggests that the combined risk of thromboembolic disease and potential for bleeding complications associated with the use of low-dose heparin may make it an undesirable agent for thromboembolic prophylaxis.

Since the agents that reduce the risk of proximal deep venous thrombosis most significantly are also the agents that are associated with the highest risk of bleeding, one must weigh the combined value of these events in order to determine the optimum agent for thromboembolic prophylaxis. Low-molecular-weight heparin causes a significant reduction in thromboembolic disease, but the risk of minor bleeding complications may be unacceptably high. The risks of minor wound-bleeding, total minor bleeding, major wound-bleeding, and major non-wound bleeding with low-molecular-weight heparin are nearly double the risks with warfarin. Although this difference in bleeding risk did not reach significance, it cannot be interpreted as clinically unimportant.

The data indicates that warfarin and pneumatic compression provide the best balance of efficacy and safety. Warfarin affords the best protection against proximal deep venous thrombosis and symptomatic pulmonary embolism and increases the risk of minor and major wound-bleeding only moderately (if at all). Pneumatic compression provides a low rate of total deep venous thrombosis and symptomatic pulmonary embolism without any increased bleeding risk; however, the data indicates a significantly higher risk of proximal deep venous thrombosis compared with the risks with warfarin and low-molecular-weight heparin. Since there were fewer studies (a total of five) evaluating the efficacy of pneumatic compression than studies evaluating the other agents, the results with regard to this agent must be interpreted with greater caution.

The present study has several major strengths. It provides a comprehensive comparison of all major and minor outcomes, including efficacy and safety, associated with the use of agents for thromboembolic prophylaxis in patients undergoing elective total hip arthroplasty. The meta-analysis included only randomized, controlled trials of patients treated with elective total hip arthroplasty and included only those in which venography had been performed to confirm deep venous thrombosis. Previous meta-analyses have had several limitations, such as including data on non-venographically confirmed deep venous thrombosis³⁶ or not providing information concerning the safety of the prophylaxis⁴⁶. By using strict criteria for inclusion, our study helps to ensure the appropriateness of data-pooling. In addition, the inclusion of both thromboembolism and bleeding risks provides critical information with which clinicians can make an evidence-based decision regarding a prophylactic agent.

Limitations of the Study

There are several potential limitations of this study. Like all meta-analyses, it is dependent on the scientific validity of the studies that were incorporated in the analysis. Also, breaking down the individual studies into treatment arms may have violated the assumption of randomization. However, there is no other appropriate way to compare all prophylactic agents, since there is insufficient comparison of each of these individual agents in randomized, controlled trials. One could also argue that, because the treatment arms were disassembled, prospective studies in which only one agent was examined could have been included in the analysis. We excluded such studies, however, because they did not meet the strict inclusion criteria of a randomized, controlled trial and including them may have violated the assumption that the studies can be combined.

Another potential limitation concerns the inclusion of studies over a broad period of time (1974 to 1997). During this period, there have been many changes in the management of patients undergoing total joint arthroplasty, including shorter operative times, earlier mobilization, use of regional anesthesia, and shorter hospital stays. Each of these factors may contribute to the rate of thromboembolic disease, regardless of the prophylactic agent, and they were not controlled for in the analysis because there was too much variability among studies. These factors could bias the results against the agents that were evaluated primarily in older studies (aspirin) and could make the results look more favorable for the agents that were evaluated predominantly in more recent studies (low-molecular-weight heparin). In addition, an inclusion requirement was that the prophylaxis had to have been administered for at least seven days; this may make the results less generalizable to current prophylactic regimens, given the constantly decreasing duration of hospital stays.

An additional potential limitation of the technique of meta-analysis is the heterogeneity of study parameters that can be included when studies are combined. For example, although every patient in each study had to have had venography for the diagnosis of deep venous thrombosis, venography requires a radiologist's interpretation, which may differ among studies. This meta-analysis attempted to limit heterogeneity among study groups by requiring each study to meet strict inclusion criteria. However, the heterogeneity that may exist among study groups may affect the results and can account for differences in results between meta-analyses and subsequent randomized, controlled trials¹².

There remain several unanswered questions with regard to thromboembolic prophylaxis. Even though our analysis combined a large number of studies, it did not provide proof that any agent provides protection against fatal pulmonary embolism or death. There may be several reasons for this result. For one, the current risk of fatal pulmonary embolism has been reported to be as low as 0.19 percent¹⁸. A meta-analysis that uses strict guidelines for inclusion of studies may not have enough patient data to provide the power necessary to detect a significant decrease in this risk. It was impossible to find a decreased risk of fatal pulmonary embolism in our meta-analysis of randomized, controlled trials since none of the cases of symptomatic pulmonary embolism in the placebo group were fatal. Because of the strict selection criteria for inclusion in this study, there were only 860 patients in the placebo group. Given this small number of patients and the low risk of fatal pulmonary embolism, it is not surprising that there were no fatal pulmonary emboli in this group. Historically, however, the risk of fatal pulmonary embolism in the absence of prophylaxis has been as high as 3.4 percent⁸. Although our meta-analysis lacked the statistical power to detect a significant decrease in the rate of fatal pulmonary embolism with thromboembolic prophylaxis, it cannot be concluded that prophylaxis provides no benefit with regard to the prevention of fatal pulmonary embolism or death.

In our meta-analysis, the overall risk of fatal pulmonary embolism for all agents combined was 0.1 percent (ten cases of fatal pulmonary embolism in 10,103 patients). For a future randomized, controlled trial to demonstrate a 50 percent reduction in the risk of fatal pulmonary embolism (a risk of 0.05 percent), nearly 100,000 patients would be required. The feasibility of such a study is questionable, even at the multicenter level. Finally, it is possible that the proxy events of thromboembolic disease - namely, deep venous thrombosis - may not translate into fatal pulmonary embolism in a predictable manner.

A surgeon's choice of prophylactic agent is determined by weighing the risk of thromboembolic disease against the risk of bleeding complications due to the prophylaxis. The ability to trade off thromboembolic and bleeding complications is the decision of the physician and the patient, and it can be made only after considerable thought has been given to the seriousness of each consequence. Since the value placed upon, and the preferences with regard to, each of these events is physician and patient-dependent, there remains no absolutely correct choice of thromboembolic prophylaxis.

In conclusion, this meta-analysis provides important information with regard to thromboembolic prophylaxis in patients undergoing elective total hip arthroplasty. Warfarin offers the best combination of efficacy and safety. It provides the lowest rate of both proximal deep venous thrombosis and symptomatic pulmonary embolism. However, an increase in minor bleeding is associated with the use of this agent. Compared with warfarin, pneumatic compression provides lower risks of symptomatic pulmonary embolism and minor wound-bleeding, but this is at the expense of a higher risk of proximal deep venous thrombosis. Perhaps future studies that incorporate physician and patient preferences with regard to these outcomes, such as decision analyses, can help with the clinically difficult decision regarding the best means of prophylaxis.

References

Introduction | Materials and Methods | Results | Discussion | References

Anderson, D. R.; O'Brien, B. J.; Levine, M. N.; Roberts, R.; Wells, P. S.; and Hirsh, J.: Efficacy and cost of low-molecular-weight heparin compared with standard heparin for the prevention of deep vein thrombosis after total hip arthroplasty. Ann. Intern. Med.,119: 1105-1112, 1993.1191105 1993 [PubMed]

Bailey, J. P.; Kruger, M. P.; Solano, F. X.; Zajko, A. B.; and Rubash, H. E.: Prospective randomized trial of sequential compression devices vs low-dose warfarin for deep venous thrombosis prophylaxis in total hip arthroplasty. J. Arthroplasty,6 (Supplement): 29-S35, 1991.6 (Supplement)29 1991

Beisaw, N. E.; Comerota, A. J.; Groth, H. E.; Merli, G. J.; Weitz, H. H.; Zimmerman, R. C.; Discrio, F. J.; and Sasahara, A. A.: Dihydroergotamine/heparin in the prevention of deep-vein thrombosis after total hip replacement. A controlled, prospective, randomized multicenter trial. J. Bone and Joint Surg.,70-A: 2-10, Jan 1988.70-A2 1988

Clagett, G. P.; Anderson, F. A., Jr.; Heit, J.; Levine, M. N.; and Wheeler, H. B.: Prevention of venous thromboembolism. Chest,108 (Supplement 4): 312S-334S, 1995.108 (Supplement 4)312 1995

Colwell, C. W., Jr.; Spiro, T. E.; Trowbridge, A. A.; Morris, B. A.; Kwaan, H. C.; Blaha, J. D.Comerota, A. J.; Skoutakis, V. A. ; and Enoxaparin Clinical Trial Group: Use of enoxaparin, a low-molecular-weight heparin, and unfractionated heparin for the prevention of deep venous thrombosis after elective hip replacement. A clinical trial comparing efficacy and safety. J. Bone and Joint Surg.,76-A: 3-14, Jan 1994.76-A3 1994

Colwell, C. W., Jr., and Spiro, T. E.: Efficacy and safety of enoxaparin to prevent deep vein thrombosis after hip arthroplasty. Clin. Orthop.,319: 215-222, 1995.319215 1995 [PubMed]

Colwell, C. W., Jr.Spiro, T. E.Trowbridge, A. A.Stephens, J. W.Gardiner, G. A., Jr.; Ritter, M. A. ; and Enoxaparin Clinical Trial Group: Efficacy and safety of enoxaparin versus unfractionated heparin for prevention of deep venous thrombosis after elective knee arthroplasty. Clin. Orthop.,321: 19-27, 1995.32119 1995

Coventry, M. B.Nolan, D. R., and Beckenbaugh, R. D.: "Delayed" prophylactic anticoagulation: a study of results and complications in 2,012 total hip arthroplasties. J. Bone and Joint Surg.,,55-A: 1487-1492, Oct 1973.55-A1487 1973

Danish Enoxaparin Study Group: Low-molecular-weight heparin (enoxaparin) vs dextran 70: the prevention of post-operative deep vein thrombosis after total hip replacement. Arch. Intern. Med.,151: 1621-1624, 1991.1511621 1991 [PubMed]

Dechavanne, M.Ville, D.Berruyer, M.Trepo, F.Dalery, F.Clermont, N.Lerat, J. L.Moyen, B.Fischer, L. P.Kher, A., and Barbier, P.: Randomized trial of a low-molecular-weight heparin (Kabi 2165) versus adjusted-dose subcutaneous standard heparin in the prophylaxis of deep-vein thrombosis after elective hip surgery. Haemostasis,19: 5-12, 1989.195 1989 [PubMed]

DerSimonian, R., and Laird, N.: Meta-analysis in clinical trials. Controlled Clin. Trials,7: 177-188, 1986.7177 1986 [PubMed]

DerSimonian, R., and Levine, R. J.: Resolving discrepancies between a meta-analysis and a subsequent large controlled trial. J. Am. Med. Assn.,282: 664-670, 1999.282664 1999

Diggle, P. J.; Liang, K.-Y.; and Zeger, S. L.: Analysis of Longitudinal Data. New York, Oxford University Press, 1995.

Eriksson, B. I.Zachrisson, B. E.Teger-Nilsson, A. C., and Risberg, B.: Thrombosis prophylaxis with low molecular weight heparin in total hip replacement. British J. Surg.,75: 1053-1057, 1988 .751053 1988

Eriksson, B. I.Kalebo, P.Anthymyr, B. A.Wadenvik, H.Tengborn, L., and Risberg, B.: Prevention of deep-vein thrombosis and pulmonary embolism after total hip replacement. Comparison of low-molecular-weight heparin and unfractionated heparin. J. Bone and Joint Surg.,73-A: 484-493, April 1991.73-A484 1991

Eriksson, B. I.Ekman, S.Lindbratt, S.Baur, M.Bach, D.Torholm, C.Kalebo, P., and Close, P.: Prevention of thromboembolism with use of recombinant hirudin. Results of a double-blind, multicenter trial comparing the efficacy of desirudin (Revasc) with that of unfractionated heparin in patients having a total hip replacement. J. Bone and Joint Surg.,79-A: 326-333, March 1997.79-A326 1997

Feller, J. A.Parkin, J. D.Phillips, G. W.Hannon, P. J.Hennessy, O., and Huggins, R. M.: Prophylaxis against venous thrombosis after total hip arthroplasty. Australian and New Zealand J. Surg.,62: 606-610, 1992.62606 1992

Fender, D.Harper, W. M.Thompson, J. R., and Gregg, P. J.: Mortality and fatal pulmonary embolism after primary total hip replacement. Results from a regional hip register. J. Bone and Joint Surg.,79-B(6): 896-899, 1997.79-B(6)896 1997

Fordyce, M. J. F., and Ling, R. S. M.: A venous foot pump reduces thrombosis after total hip replacement. J. Bone and Joint Surg.,74-B(1): 45-49,, 1992.74-B(1)45 1992

Francis, C. W.Marder, V. J.Evarts, C. McC., and Yaukoolbodi, S.: Two-step warfarin therapy. Prevention of postoperative venous thrombosis without excessive bleeding. J. Am. Med. Assn.,249: 374-378, 1983.249374 1983

Francis, C. W.Pellegrini, V. D., Jr.Marder, V. J.Totterman, S.Harris, C. M.Gabriel, K. R.Azodo, M. V., and Leibert, K. M.: Comparison of warfarin and external pneumatic compression in prevention of venous thrombosis after total hip replacement. J. Am. Med. Assn.,267: 2911-2915, 1992.2672911 1992

Francis, C. W.Pellegrini, V. D., Jr.Totterman, S.Boyd, A. D., Jr.Marder, V. J.Liebert, K. M.Stulberg, B. N.Ayers, D. C.Rosenberg, A.Kessler, C., and Johanson, N. A.: Prevention of deep-vein thrombosis after total hip arthroplasty. Comparison of warfarin and dalteparin. J. Bone and Joint Surg.,79-A: 1365-1372, Sept 1997.79-A1365 1997

Freick, H., and Haas, S.: Prevention of deep vein thrombosis by low-molecular-weight heparin and dihydroergotamine in patients undergoing total hip replacement. Thromb. Res.,63: 133-143, 1991.63133 1991 [PubMed]

Gallus, A. S.Cade, J. F.Mills, K. W., and Murphy, W.: Apparent lack of synergism between heparin and dihydroergotamine in prevention of deep vein thrombosis after elective hip replacement: a randomized double-blind trial reported in conjunction with an overview of previous results. Thromb. and Haemost.,68: 238-244, 1992.68238 1992

German Hip Arthroplasty Trial (GHAT) Group:: Prevention of deep vein thrombosis with low molecular-weight heparin in patients undergoing total hip replacement. A randomized trial. Arch. Orthop. and Trauma Surg.,111: 110-120, 1992.111110 1992

Hamulyak, K.Lensing, A. W.van der Meer, J.Smid, W. M.van Ooy, A., and Hoek, J. A.: Subcutaneous low-molecular weight heparin or oral anticoagulants for the prevention of deep-vein thrombosis in elective hip and knee replacement? Fraxiparine Oral Anticoagulant Study Group. Thromb. and Haemost.,,74: 1428-1431, 1995.741428 1995

Harris, W. H.Salzman, E. W.Athanasoulis, C.Waltman, A. C.Baum, S., and DeSanctis, R. W.: Comparison of warfarin, low-molecular-weight dextran, aspirin, and subcutaneous heparin in prevention of venous thromboembolism following total hip replacement. J. Bone and Joint Surg.,56-A: 1552-1562, Dec 1974.56-A1552 1974

Harris, W. H.Salzman, E. W.Athanasoulis, C. A.Waltman, A. C., and DeSanctis, R. W.: Aspirin prophylaxis of venous thromboembolism after total hip replacement. New England J. Med.,297: 1246-1249, 1977.2971246 1977

Harris, W. H.Athanasoulis, C. A.Waltman, A. C., and Salzman, E. W.: High and low-dose aspirin prophylaxis against venous thromboembolic disease in total hip replacement. J. Bone and Joint Surg.,,64-A: 63-66, Jan 1982.64-A63 1982

Harris, W. H.Athanasoulis, C. A.Waltman, A. C., and Salzman, E. W.: Prophylaxis of deep-vein thrombosis after total hip replacement. Dextran and external pneumatic compression compared with 1.2 or 0.3 gram of aspirin daily. J. Bone and Joint Surg.,67-A: 57-62, Jan 1985.67-A57 1985

Hoek, J. A.Nurmohamed, M. T.Hamelynck, K. J.Marti, R. K.Knipscheer, H. C.ten Cate, H.Buller, H. R.Magnani, H. N., and ten Cate, J. W.: Prevention of deep vein thrombosis following total hip replacement by low molecular weight heparinoid. Thromb. and Haemost.,67: 28-32, 1992.6728 1992

Hui, A. C. W.Heras-Palou, C.Dunn, I.Triffitt, P. D.Crozier, A.Imeson, J., and Gregg, P. J.: Graded compression stockings for prevention of deep-vein thrombosis after hip and knee replacement. J. Bone and Joint Surg.,78-B(4): 550-554, 1996.78-B(4)550 1996

Hull, R. D.Raskob, G. E.Gent, M.McLoughlin, D.Julian, D.Smith, F. C.Dale, N. I.Reed-Davis, R.Lofthouse, R. N., and Anderson, C.: Effectiveness of intermittent pneumatic leg compression for preventing deep vein thrombosis after total hip replacement. J. Am. Med. Assn.,263: 2313-2317, 1990.2632313 1990

Hull, R.Raskob, G.Pineo, G.Rosenbloom, D.Evans, W.Mallory, T.Anquist, K.Smith, F.Hughes, C.Green, D.Elliott, G.Panju, A., and Brant, R.: A comparison of subcutaneous low-molecular-weight heparin with warfarin sodium for prophylaxis against deep-vein thrombosis after hip or knee implantation. New England J. Med.,329: 1370-1376, 1993.3291370 1993

Huo, M. H.Salvati, E. A.Sharrock, N. E.Brien, W. W.Sculco, T. P.Pellicci, P. M.Mineo, R., and Go, G.: Intraoperative heparin thromboembolic prophylaxis in primary total hip arthroplasty. A prospective, randomized, controlled, clinical trial. Clin. Orthop.,274: 35-46, 1992.27435 1992 [PubMed]

Imperiale, T. F., and Speroff, T.: A meta-analysis of methods to prevent venous thromboembolism following total hip replacement. J. Am. Med. Assn.,271: 1780-1785, 1994.2711780 1994

Kaempffe, F. A.Lifeso, R. M., and Meinking, C.: Intermittent pneumatic compression versus Coumadin. Prevention of deep vein thrombosis in lower-extremity total joint arthroplasty. Clin. Orthop.,269: 89-97, 1991.26989 1991 [PubMed]

Kalodiki, E. P.; Hoppensteadt, D. A.; Nicolaides, A. N.; Fareed, J.; Gill, K.; Regan, F.; al-Kutoubi, A.; Cunningham, D. A.; Birch, R.; Harris, N.; Hunt, D.; Johnson, J.; and Marx, C.: Deep venous thrombosis prophylaxis with low molecular weight heparin and elastic compression in patients having total hip replacement. A randomised controlled trial. Internat. Angiol.,15: 162-168, 1996.15162 1996

Lassen, M. R.; Borris, L. C.; Christiansen, H. M.; Boll, K. L.; Eiskjaer, S. P.; Nielsen, B. W.; Schott, P.; Olsen, A. D.; Rodenberg, J. C.; and Lucht, U.: Prevention of thromboembolism in 190 hip arthroplasties. Comparison of LMW heparin and placebo. Acta Orthop. Scandinavica,62: 33-38, 1991.6233 1991

Levine, M. N.; Hirsh, J.; Gent, M.; Turpie, A. G.; Leclerc, J.; Powers, P. J.; Jay, R. M.; and Neemeh, J.: Prevention of deep vein thrombosis after elective hip surgery. A randomized trial comparing low molecular weight heparin with standard unfractionated heparin. Ann. Intern. Med.,,114: 545-551, 1991.114545 1991

Leyvraz, P. F.; Richard, J.; Bachmann, F.; Van Melle, G.; Treyvaud, J. M.; Livio, J. J.; and Candardjis, G.: Adjusted versus fixed-dose subcutaneous heparin in the prevention of deep-vein thrombosis after total hip replacement. New England J. Med.,309: 954-958, 1983.309954 1983

Leyvraz, P.; Bachmann, F.; Bohnet, J.; Breyer, H. G.; Estoppey, D.; Haas, S.; Hochreiter, J.; Jakubek, H.; Mair, J.; Sorensen, R.; and Stiekema, J.: Thromboembolic prophylaxis in total hip replacement: a comparison between the low molecular weight heparinoid Lomoparan and heparin dihydroergotamine. British J. Surg.,79: 911-914, 1992.79911 1992

Lieberman, J. R.; Huo, M. M.; Hanway, J.; Salvati, E. A.; Sculco, T. P.; and Sharrock, N. E.: The prevalence of deep venous thrombosis after total hip arthroplasty with hypotensive epidural anesthesia. J. Bone and Joint Surg.,76-A: 341-348, March 1994.76-A341 1994

Lotke, P. A.; Palevsky, H.; Keenan, A. M.; Meranze, S.; Steinberg, M. E.; Ecker, M. L.; and Kelley, M. A.: Aspirin and warfarin for thromboembolic disease after total joint arthroplasty. Clin. Orthop.,324: 251-258, 1996.324251 1996 [PubMed]

Menzin, J.; Colditz, G. A.; Regan, M. M.; Richner, R. E.;; and Oster, G.: Cost-effectiveness of enoxaparin vs low-dose warfarin in the prevention of deep-vein thrombosis after total hip replacement surgery. Arch. Intern. Med.,155: 757-764, 1995.155757 1995 [PubMed]

Mohr, D. N.; Silverstein, M. D.; Murtaugh, P. A.; and Harrison, J. M.: Prophylactic agents for venous thrombosis in elective hip surgery. Meta-analysis of studies using venographic assessment. Arch. Intern. Med.,153: 2221-2228, 1993.1532221 1993 [PubMed]

O'Brien, B. J.; Anderson, D. R.; and Goeree, R.: Cost-effectiveness of enoxaparin versus warfarin prophylaxis against deep-vein thrombosis after total hip replacement. CMAJ: Canadian Med. Assn. J.,150: 1083-1090, 1994.1501083 1994

Oster, G.; Tuden, R. L.; and Colditz, G. A.: A cost-effectiveness analysis of prophylaxis against deep-vein thrombosis in major orthopedic surgery. J. Am. Med. Assn.,257: 203-208, 1987.257203 1987

Paiement, G.; Wessinger, S. J.; Waltman, A. C.; and Harris, W. H.: Low-dose warfarin versus external pneumatic compression for prophylaxis against venous thromboembolism following total hip replacement. J. Arthroplasty,2: 23-26, 1987.223 1987 [PubMed]

Paiement, G. D.; Wessinger, S. J.; and Harris, W. H.: Cost-effectiveness of prophylaxis in total hip replacement. Am. J. Surg.,161: 519-524, 1991.161519 1991 [PubMed]

Parker-Williams, J., and Vickers, R.:: Major orthopaedic surgery on the leg and thromboembolism. BMJ: British Med. J.,,303: 531-532, 1991.303531 1991

Planes, A.; Vochelle, N.; Mazas, F.; Mansat, C.; Zucman, J.; Landais, A.; Pascariello, J. C.; Weill, D.; and Butel, J.: Prevention of postoperative venous thrombosis: a randomized trial comparing unfractionated heparin with low molecular weight heparin in patients undergoing total hip replacement. Thromb. and Haemost.,60: 407-410, 1988.60407 1988

Planes, A.: Comparison of antithrombotic efficacy and haemorrhagic side-effects of Clivarin versus enoxaparin in patients undergoing total hip replacement surgery. Blood Coagulat. and Fibrinol.,4 (Supplement 1): 33-S35, 1993.4 (Supplement 1)33 1993

RD Heparin Arthroplasty Group: RD heparin compared with warfarin for prevention of venous thromboembolic disease following total hip or knee arthroplasty. J. Bone and Joint Surg.,76-A: 1174-1185, Aug 1994.76-A1174 1994

Rogers, P. H.; Walsh, P. N.; Marder, V. J.; Bosak, G. C.; Lachman, J. W.; Ritchie, W. G. M.; Oppenheimer, L.; and Sherry, S.: Controlled trial of low-dose heparin and sulfinpyrazone to prevent venous thromboembolism after operation on the hip. J. Bone and Joint Surg.,60-A: 758-762, Sept 1978.60-A758 1978

Santori, F. S.; Vitullo, A.; Stopponi, M.; Santori, N.; and Ghera, S.: Prophylaxis against deep-vein thrombosis in total hip replacement. Comparison of heparin and foot impulse pump. J. Bone and Joint Surg.,76-B(4): 579-583, 1994.76-B(4)579 1994

Sarasin, F. P., and Bounameaux, H.: Antithrombotic strategy after total hip replacement. A cost-effectiveness analysis comparing prolonged oral anticoagulants with screening for deep vein thrombosis. Arch. Intern. Med.,156: 1661-1668, 1996.1561661 1996 [PubMed]

Sautter, R. D.; Koch, E. L.; Myers, W. O.; Ray, J. R., III; Mazza, J. J.; Larson, D. E.; Chen, H. M.; Milbauer, J. P.; Treuhaft, P. S.; Plotka, E. D.; Wendel, F. J.; Nyez, G.; and Pierce, W. E.: Aspirin-sulfinpyrazone in prophylaxis of deep venous thrombosis in total hip replacement. J. Am. Med. Assn.,250: 2649-2654, 1983.2502649 1983

Sharrock, N. E.; Brien, W. W.; Salvati, E. A.; Mineo, R.; Garvin, K.; and Sculco, T. P.: The effect of intravenous fixed-dose heparin during total hip arthroplasty on the incidence of deep-vein thrombosis. A randomized, double-blind trial in patients operated on with epidural anesthesia and controlled hypotension. J. Bone and Joint Surg.,72-A: 1456-1461, Dec 1990.72-A1456 1990

Turpie, A. G.; Levine, M. N.; Hirsh, J.; Carter, C. J.; Jay, R. M.; Powers, P. J.; Andrew, M.; Hull, R. D.; and Gent, M.: A randomized controlled trial of a low-molecular-weight heparin (enoxaparin) to prevent deep-vein thrombosis in patients undergoing elective hip surgery. New England J. Med.,315: 925-929, 1986.315925 1986

Warwick, D.; Bannister, G. C.; Glew, D.; Mitchelmore, A.; Thornton, M.; Peters, T. J.; and Brookes, S.: Perioperative low-molecular-weight heparin. Is it effective and safe?. J. Bone and Joint Surg.,77-B(5): 715-719, 1995.77-B(5)715 1995

Zimlich, R. H.; Fulbright, M.; and Friedman, R. J.: Current status of anticoagulation therapy after total hip and total knee arthroplasty. J. Am. Acad. Orthop. Surgeons,4: 54-62, 1996.454 1996

Topics

View Larger

Table I : Description of Studies Included in the Meta-Analysis

Treatment Arm	No. of Studies	Citations	No. of Patients	Mean Age (yrs.)	Mean Percent Male	Mean No. of Days Postop. Venography Performed
Low-molecular-weight heparin	21	5, 6, 9, 10, 14, 15, 22, 23, 25, 26, 31, 34, 38-40, 42, 52-54, 60, 61	5512	66	43	10
Warfarin	12	2, 17, 20-22, 26, 27, 34, 37, 44, 49, 54	1493	63	52	9
Aspirin	8	27-30, 35, 43, 44, 59	687	62	45	9
Low-dose heparin	11	5, 7, 15, 16, 24, 25, 27, 40, 41, 52, 56	1859	66	42	10
Pneumatic compression	5	2, 21, 33, 37, 49	431	64	53	9
Placebo	13	3, 6, 19, 28, 31-33, 38, 39, 55, 58, 60, 61	947	67	42	10

Add to My JBJS

Table I : Description of Studies Included in the Meta-Analysis

Treatment Arm	No. of Studies	Citations	No. of Patients	Mean Age (yrs.)	Mean Percent Male	Mean No. of Days Postop. Venography Performed
Low-molecular-weight heparin	21	5, 6, 9, 10, 14, 15, 22, 23, 25, 26, 31, 34, 38-40, 42, 52-54, 60, 61	5512	66	43	10
Warfarin	12	2, 17, 20-22, 26, 27, 34, 37, 44, 49, 54	1493	63	52	9
Aspirin	8	27-30, 35, 43, 44, 59	687	62	45	9
Low-dose heparin	11	5, 7, 15, 16, 24, 25, 27, 40, 41, 52, 56	1859	66	42	10
Pneumatic compression	5	2, 21, 33, 37, 49	431	64	53	9
Placebo	13	3, 6, 19, 28, 31-33, 38, 39, 55, 58, 60, 61	947	67	42	10

View Larger

Table II: Risk of Deep Venous Thrombosis*

Agent	No. of Patients	Distal Deep Venous Thrombosis		Proximal Deep Venous Thrombosis		Total Deep Venous Thrombosis
No. of Events	Risk (95 Percent CI) (percent)	No. of Events	Risk (95 Percent CI) (percent)	No. of Events	Risk (95 Percent CI) (percent)
Low-molecular- weight heparin	5512	566	9.6 (7.7, 12.0)	342	7.7 (5.7, 10.3)	918	17.7 (14.5, 21.4)
Warfarin†	1493	244	17.1 (13.3, 21.6)	90	6.3 (4.7, 8.4)	334	23.2 (19.3, 27.7)
Aspirin†‡	687	134	19.7 (14.2, 26.8)	82	11.4 (7.3, 17.5)	214	30.6 (21.2, 41.8)
Low-dose heparin†	1859	201	10.8 (8.4, 13.9)	222	19.0 (12.9, 27.2)	429	31.1 (22.7, 40.9)
Pneumatic compression†	431	36	7.7 (4.5, 12.9)	58	13.3 (9.9, 17.7)	94	20.7 (14.6, 28.7)
Placebo§	947	211	22.4 (18.8, 26.6)	245	25.8 (21.4, 30.7)	456	48.5 (43.4, 53.7)

Add to My JBJS

Table II: Risk of Deep Venous Thrombosis*

Agent	No. of Patients	Distal Deep Venous Thrombosis		Proximal Deep Venous Thrombosis		Total Deep Venous Thrombosis
No. of Events	Risk (95 Percent CI) (percent)	No. of Events	Risk (95 Percent CI) (percent)	No. of Events	Risk (95 Percent CI) (percent)
Low-molecular- weight heparin	5512	566	9.6 (7.7, 12.0)	342	7.7 (5.7, 10.3)	918	17.7 (14.5, 21.4)
Warfarin†	1493	244	17.1 (13.3, 21.6)	90	6.3 (4.7, 8.4)	334	23.2 (19.3, 27.7)
Aspirin†‡	687	134	19.7 (14.2, 26.8)	82	11.4 (7.3, 17.5)	214	30.6 (21.2, 41.8)
Low-dose heparin†	1859	201	10.8 (8.4, 13.9)	222	19.0 (12.9, 27.2)	429	31.1 (22.7, 40.9)
Pneumatic compression†	431	36	7.7 (4.5, 12.9)	58	13.3 (9.9, 17.7)	94	20.7 (14.6, 28.7)
Placebo§	947	211	22.4 (18.8, 26.6)	245	25.8 (21.4, 30.7)	456	48.5 (43.4, 53.7)

View Larger

Table III: Multiple Comparisons of Agents for Risk of Deep Venous Thrombosis

*The agent associated with the lower risk appears first in each entry.†Significant at the p < 0.0083 level.

Comparison*	P Values
Total Deep Venous Thrombosis	Distal Deep Venous Thrombosis	Proximal Deep Venous Thrombosis
Low-molecular-weight heparin vs. warfarin	0.1916	0.0047†	0.3513†
Low-molecular-weight heparin vs. aspirin	0.1121	0.0005	0.8862
Low-molecular-weight heparin vs. low-dose heparin	0.0004†	0.5185	0.0023†
Low-molecular-weight heparin vs. pneumatic compression	0.8031	0.0673	0.0059†
Warfarin vs. aspirin	0.4819	0.1880	0.7955
Warfarin vs. low-dose heparin	0.6356	0.0219	0.0047†
Warfarin vs. pneumatic compression	0.1962	0.0007†	0.0004†
Aspirin vs. low-dose heparin	0.7495	0.0018	0.1281
Pneumatic compression vs. aspirin	0.0970	0.0001	0.0831
Low-dose heparin vs. pneumatic compression	0.1491	0.0447	0.8936

Add to My JBJS

Table III: Multiple Comparisons of Agents for Risk of Deep Venous Thrombosis

*The agent associated with the lower risk appears first in each entry.†Significant at the p < 0.0083 level.

Comparison*	P Values
Total Deep Venous Thrombosis	Distal Deep Venous Thrombosis	Proximal Deep Venous Thrombosis
Low-molecular-weight heparin vs. warfarin	0.1916	0.0047†	0.3513†
Low-molecular-weight heparin vs. aspirin	0.1121	0.0005	0.8862
Low-molecular-weight heparin vs. low-dose heparin	0.0004†	0.5185	0.0023†
Low-molecular-weight heparin vs. pneumatic compression	0.8031	0.0673	0.0059†
Warfarin vs. aspirin	0.4819	0.1880	0.7955
Warfarin vs. low-dose heparin	0.6356	0.0219	0.0047†
Warfarin vs. pneumatic compression	0.1962	0.0007†	0.0004†
Aspirin vs. low-dose heparin	0.7495	0.0018	0.1281
Pneumatic compression vs. aspirin	0.0970	0.0001	0.0831
Low-dose heparin vs. pneumatic compression	0.1491	0.0447	0.8936

View Larger

Table IV: Risk of Pulmonary Embolism*

*CI = confidence interval.†The agent differed significantly from the placebo at p < 0.05 for the risk of symptomatic pulmonary embolism.

Agent	No. of Patients	Symptomatic Pulmonary Embolism		Fatal Pulmonary Embolism
No. of Events	Risk (95 Percent CI) (percent)	No. of Events	Risk (95 Percent CI) (percent)
Low-molecular-weight heparin†	5238	19	0.36 (0.22, 0.57)	2	0.04 (0.01, 0.14)
Warfarin†	1232	2	0.16 (0.02, 0.59)	2	0.16 (0.02, 0.59)
Aspirin	625	8	1.28 (0.55, 2.51)	2	0.32 (0.04, 1.15)
Low-dose heparin	1760	24	1.36 (0.88, 2.02)	3	0.17 (0.04, 0.50)
Pneumatic compression†	388	1	0.26 (0.01, 1.43)	1	0.26 (0.01, 1.43)
Placebo	860	13	1.51 (0.81, 2.57)	0	0.00 (0.00, 0.43)

Add to My JBJS

Table IV: Risk of Pulmonary Embolism*

*CI = confidence interval.†The agent differed significantly from the placebo at p < 0.05 for the risk of symptomatic pulmonary embolism.

Agent	No. of Patients	Symptomatic Pulmonary Embolism		Fatal Pulmonary Embolism
No. of Events	Risk (95 Percent CI) (percent)	No. of Events	Risk (95 Percent CI) (percent)
Low-molecular-weight heparin†	5238	19	0.36 (0.22, 0.57)	2	0.04 (0.01, 0.14)
Warfarin†	1232	2	0.16 (0.02, 0.59)	2	0.16 (0.02, 0.59)
Aspirin	625	8	1.28 (0.55, 2.51)	2	0.32 (0.04, 1.15)
Low-dose heparin	1760	24	1.36 (0.88, 2.02)	3	0.17 (0.04, 0.50)
Pneumatic compression†	388	1	0.26 (0.01, 1.43)	1	0.26 (0.01, 1.43)
Placebo	860	13	1.51 (0.81, 2.57)	0	0.00 (0.00, 0.43)

View Larger

Table V: Risk of Minor Bleeding*

Agent	Minor Wound-Bleeding			Total Minor Bleeding
No. of Patients	No. of Events	Risk (95 Percent CI) (percent)	No. of Patients	No. of Events	Risk (95 Percent CI) (percent)
Low-molecular-weight heparin†	3468	277	8.9 (4.2, 17.9)	4064	382	10.5 (5.9, 17.8)
Warfarin‡	1381	42	4.7 (2.1, 10.2)	1381	64	5.7 (3.1, 10.4)
Aspirin	687	6	1.2 (0.4, 3.3)	687	6	1.2 (0.4, 3.3)
Low-dose heparin	1244	65	7.6 (4.2, 13.5)	1453	149	13.5 (8.3, 21.2)
Pneumatic compression	338	3	1.1 (0.2, 6.1)	388	13	4.1 (1.7, 9.8)
Placebo	663	14	2.2 (0.6, 7.6)	663	18	3.0 (1.1, 8.2)

Add to My JBJS

Table V: Risk of Minor Bleeding*

Agent	Minor Wound-Bleeding			Total Minor Bleeding
No. of Patients	No. of Events	Risk (95 Percent CI) (percent)	No. of Patients	No. of Events	Risk (95 Percent CI) (percent)
Low-molecular-weight heparin†	3468	277	8.9 (4.2, 17.9)	4064	382	10.5 (5.9, 17.8)
Warfarin‡	1381	42	4.7 (2.1, 10.2)	1381	64	5.7 (3.1, 10.4)
Aspirin	687	6	1.2 (0.4, 3.3)	687	6	1.2 (0.4, 3.3)
Low-dose heparin	1244	65	7.6 (4.2, 13.5)	1453	149	13.5 (8.3, 21.2)
Pneumatic compression	338	3	1.1 (0.2, 6.1)	388	13	4.1 (1.7, 9.8)
Placebo	663	14	2.2 (0.6, 7.6)	663	18	3.0 (1.1, 8.2)

View Larger

Table VI: Risk of Major Bleeding*

*CI = confidence interval.†The agent differed significantly from the placebo at p < 0.05 for both major wound-bleeding and total major bleeding.

Agent	Major Wound-Bleeding			Total Major Bleeding
No. of Patients	No. of Events	Risk (95 Percent CI) (percent)	No. of Patients	No. of Events	Risk (95 Percent CI) (percent)
Low-molecular-weight heparin	4049	61	1.51 (1.15, 1.93)	5412	120	2.22 (1.84, 2.64)
Warfarin	1185	9	0.76 (0.35, 1.44)	1381	23	1.67 (1.06, 2.49)
Aspirin	687	2	0.29 (0.04, 1.05)	687	5	0.73 (0.24, 1.69)
Low-dose heparin†	1991	51	2.56 (1.91, 3.35)	1992	69	3.46 (2.70, 4.36)
Pneumatic compression	388	0	0.00 (0.00, 0.95)	388	0	0.00 (0.00, 0.95)
Placebo	713	2	0.28 (0.03, 1.01)	713	4	0.56 (0.15, 1.43)

Add to My JBJS

Table VI: Risk of Major Bleeding*

*CI = confidence interval.†The agent differed significantly from the placebo at p < 0.05 for both major wound-bleeding and total major bleeding.

Agent	Major Wound-Bleeding			Total Major Bleeding
No. of Patients	No. of Events	Risk (95 Percent CI) (percent)	No. of Patients	No. of Events	Risk (95 Percent CI) (percent)
Low-molecular-weight heparin	4049	61	1.51 (1.15, 1.93)	5412	120	2.22 (1.84, 2.64)
Warfarin	1185	9	0.76 (0.35, 1.44)	1381	23	1.67 (1.06, 2.49)
Aspirin	687	2	0.29 (0.04, 1.05)	687	5	0.73 (0.24, 1.69)
Low-dose heparin†	1991	51	2.56 (1.91, 3.35)	1992	69	3.46 (2.70, 4.36)
Pneumatic compression	388	0	0.00 (0.00, 0.95)	388	0	0.00 (0.00, 0.95)
Placebo	713	2	0.28 (0.03, 1.01)	713	4	0.56 (0.15, 1.43)

View Larger

Table VII : Risk of Death*

*CI = confidence interval.

Agent	No. of Patients	No. of Events	Risk (95 Percent CI) (percent)
Low-molecular-weight heparin	5918	12	0.20 (0.10, 0.35)
Warfarin	1555	7	0.45 (0.18, 0.93)
Aspirin	687	2	0.29 (0.04, 1.05)
Low-dose heparin	1992	10	0.50 (0.24, 0.92)
Pneumatic compression	388	4	1.03 (0.28, 2.62)
Placebo	900	3	0.33 (0.07, 0.97)

Add to My JBJS

Table VII : Risk of Death*

*CI = confidence interval.

Agent	No. of Patients	No. of Events	Risk (95 Percent CI) (percent)
Low-molecular-weight heparin	5918	12	0.20 (0.10, 0.35)
Warfarin	1555	7	0.45 (0.18, 0.93)
Aspirin	687	2	0.29 (0.04, 1.05)
Low-dose heparin	1992	10	0.50 (0.24, 0.92)
Pneumatic compression	388	4	1.03 (0.28, 2.62)
Placebo	900	3	0.33 (0.07, 0.97)