Current issues of ACP Journal Club are published in Annals of Internal Medicine


Warfarin prophylaxis after knee or hip surgery was more cost-effective than tinzaparin in the United States but not in Canada

ACP J Club. 1997 Jul-Aug;127:24. doi:10.7326/ACPJC-1997-127-1-024

Source Citation

Hull RD, Raskob GE, Pineo GF, et al. Subcutaneous low-molecular-weight heparin vs warfarin for prophylaxis of deep vein thrombosis after hip or knee implantation. An economic perspective. Arch Intern Med. 1997 Feb 10;157:298-303.



To determine the cost and effectiveness of warfarin and low-molecular-weight heparin (LMWH) prophylaxis after total hip or knee arthroplasty.


Randomized, double-blind, controlled trial for 14 days or to discharge.


Clinical centers in North America.


1436 patients older than 18 years who had either elective hip (n = 795) or knee (n = 641) arthroplasty.


Patients were allocated to warfarin using an adjusted-dose protocol (n =721) or subcutaneous fixed-dose LMWH (tinzaparin) (75 antifactor Xa U/kg of body weight, n =715).

Main cost and outcome measures

Deep venous thrombosis, proximal venous thrombosis, and major bleeding complications. Direct costs of administering the active drug plus cost of diagnosis, treatment of complications, and "hotel costs" of hospitalization, were determined in 1992 Canadian and U.S. dollars. The cost of placebo and indirect and opportunity (time lost from work) costs were not included.

Main results

Cost differences varied by country and type of surgery (Table). Sensitivity analyses showed that the most important costs were for drugs, monitoring, and bleeding.


Warfarin prophylaxis after knee or hip surgery was more cost-effective than tinzaparin in the United States but less cost-effective in Canada. The cost of monitoring was more important in Canada; the cost of drugs was more important in the United States.

Source of funding: In part, Heart and Stroke Foundation of Alberta; Novo Nordisk; DuPont Pharma.

For article reprint: Dr. R.D. Hull, Department of Medicine, University of Calgary, Health Sciences Centre, Room 1741, 3330 Hospital Drive North West, Calgary, Alberta, T2N 4N1, Canada. FAX 403-283-0400.

Table. Total costs and effects/100 patients for hip or knee surgery*

Approach DVT, % PVT, % Bleeding, % 1992 Cdn $ 1992 US $
Total surgery
Warfarin 37 8 1 11 598 20 876 (4718)†
LMWH 31 6 3 9197 (2401)† 25 594
Hip surgery
Warfarin 23 4 2 5933 10 578 (2105)†
LMWH 21 5 3 4398 (1535)† 12 863
Knee surgery
Warfarin 55 12 1 5688 10 363 (2455)†
LMWH 45 8 3 4803 (855)† 12 818

*DVT = deep venous thrombosis; PVT = proximal venous thrombosis; LMWH = low-molecular-weight heparin.
†Cost savings per 100 patients.


Venous thrombosis and pulmonary embolism are major clinical problems in patients who have had hip or knee arthroplasty. U.S. and European consensus statements have supported the prophylactic use of warfarin and LMWH, respectively. In this study, Hull and colleagues addressed the relative costs and effects of these agents, using data from a randomized controlled trial done in North America.

One key finding was that LMWH therapy is less expensive than warfarin therapy when Canadian unit costs were used but is more expensive when U.S. unit costs were used. The main difference is the unit costs. For example, the unit cost for a complete blood cell count test was approximately 3 times higher in the United States than in Canada, and the unit cost for syringes, needles, and swabs was approximately 20 times higher. Because no cost had been declared in the United States for tinzaparin, the authors assumed parity with Canada (Cdn $6/d) and added a pharmacy dispensing cost, which raised the U.S. figure to $15/d or $13 500 per 100 patients—$6/d would have been $5400 per 100 patients, which in fact is the figure reported for the Canadian analysis (dispensing costs were already included). This difference alone is more than sufficient to explain the discrepant results.

Therefore, the authors might have considered whether the unit costs derived from just 1 hospital in each country were representative. Some measures of variance in the cost estimates, and of the significance of the reported differences, would also have assisted the reader in assessing the evidence. Finally, it would have been helpful if the volume and cost data had been separated by country. The treatment decision is a finely tuned tradeoff, and the result may well be health care dependent.

Alastair M. Gray, DPhil
University of OxfordOxford, England, UK

Author's Response

Our analysis provides the "direct prospective economic comparison of LMWH and oral anticoagulation..." requested in a recent commentary (1). We evaluated the effect of variance in cost by sensitivity analysis using a range of 40% to 300% to reflect the regional variation in costs in the United States and Canada.

Gary E. Raskob, MSc


1. Owens DK. ACP J Club. 1997 Mar-Apr; 126:51. Comment on: Bergqvist D, Lindgren B, Mätzsch T. Br J Surg. 1996;83:1548-52.