Total joint arthroplasty is an operation characterized by a high risk for thromboembolic complications with potentially life-threatening consequences. Charnley et al,14 in a series of 7959 total hip replacements from 1962 to 1973, identified the overall prevalence of pulmonary embolism to be 7.89% with a fatal outcome in 1.04%. He stated, “The possibility of fatal pulmonary embolism after total hip replacement is a hip surgeon's constant worry … no matter how rare this might be.”3 Similarly, in 1984 John Insall suggested “prevention of thrombophlebitis and pulmonary embolism should be a major goal of every orthopaedic surgeon who performs total knee arthroplasty.” He later suggested, “The positive venogram, although not itself associated with local symptoms, is of clinical significance.”27
During the last 30 years, substantial advances have been made in the technical aspects of performing total hip arthroplasty (THA) and total knee arthroplasty (TKA), in the convalescence from the operation, and in our understanding of the pathophysiology and prevention of venous thromboembolic disease (VTED) associated with total joint replacement.24-26 Though controversial, the prevalence of fatal pulmonary embolism (PE) with contemporary techniques of operation and perioperative management would appear to be on the decline. In the absence of anticoagulant prophylaxis, the rate of fatal PE has been reported to be 0.5% after 1162 total hip replacements in one study28 and even less after TKA with numbers approaching 0.1%. In North America, a combination of warfarin prophylaxis, regional anesthesia, and the collective use of predonated autologous blood, expeditious operation, and early mobilization would appear to have lowered the prevalence of fatal PE after THA even more to less than 0.1%.1,16,20 Likewise, the prevalence of deep venous thrombosis (DVT) after THA has been reduced to 15% to 20% with warfarin prophylaxis and general anesthesia, and less than 10% with newer fractionated heparins or when warfarin is combined with continuous epidural anesthetic techniques. However, DVT after THA has been much more refractory to conventional prophylactic agents with greater residual rates of 30% to 50% with warfarin and in the 30% range with fractionated heparins. Because of the increasingly small prevalence of fatal pulmonary emboli after TKA, many surgeons have adopted a nihilistic attitude toward VTED prophylaxis after total knee replacement and employ only aspirin or compressive stockings for the ever-shrinking period of hospitalization.
In consideration of this apparent reduction in the prevalence of fatal PE and our heightened awareness of the intense activation of the clotting cascade occurring during instrumentation of the femoral canal and the 24 hours immediately thereafter,26 it is not surprising many investigators routinely screen for DVT to guide a strategy of selective thromboprophylaxis to prevent fatal PE after total joint arthroplasty.19
We believed that a definitive knowledge of the presence or absence of deep venous thrombosis after TKA, as determined by a reliable screening modality such as contrast venography, would support a strategy of selective extended prophylaxis limited to only patients with documented DVT or those at extraordinarily high risk. In this way we hoped to reduce the number of patients exposed to outpatient anticoagulation and its attendant bleeding risk. We hypothesized patients stratified by negative venography at the time of hospital discharge would be at sufficiently low risk for VTED-related readmission so as to not require continued outpatient warfarin therapy.
MATERIALS AND METHODS
Over 20 years, 1321 consecutive patients undergoing primary or revision TKA in the practices of the senior authors (VDP and CME) at three different university teaching hospitals (University of Rochester, Pennsylvania State University, and University of Maryland) were approached for inclusion in a venous thromboembolism study database. A parallel cohort of 1972 consecutive patients undergoing total hip arthroplasty was concurrently collected and followed at the same three institutions; those data have been previously reported (Table 1).21 All participating patients were monitored for readmission events for a period of 6 months after operation. In addition, screening ascending contrast venography was performed with a nonionic agent (omnipaque) in consenting patients before discharge after TKA. Hospital re- admissions, with a primary or secondary diagnosis of symptomatic DVT, pulmonary embolism, or bleeding events possibly associated with anticoagulation, and death were tabulated. Deep vein thrombosis at the time of readmission was diagnosed by contrast venography whenever possible and secondarily by Doppler ultrasound examination. Pulmonary embolism was diagnosed by pulmonary angiography whenever possible before initiation of heparin anticoagulation and secondarily by high probability ventilation and perfusion scan. Readmissions were tracked for 6 months through the respective university hospital admitting databases. All patients were contacted by telephone if they had not personally presented for routine followup care during the first year after operation. If the patient did not return for followup and was unable to be contacted by phone, the primary care physician's office was contacted for relevant readmission information. Institutional Review Board Approval for the database was obtained and informed consent was secured from all participating patients.
At the University of Rochester Medical Center from 1984 through 1992, bilateral contrast venogram surveillance was performed only in those patients enrolled in Institutional Review Board-approved sponsored clinical trials studying efficacy of various thromboprophylaxis agents and modalities. These sponsored trials included evaluation of low molecular weight dextran,11 two-step warfarin, antithrombin III combined with unfractionated heparin,9,10 and routine warfarin with a target prothrombin time index of 1.5 (INR of 2.0). Warfarin, started the night before operation, was the default therapy in the absence of any sponsored trial. A total of 559 patients undergoing TKA were enrolled at the University of Rochester on three different sponsored protocols. Standard practice during this time period dictated patients who did not complete contrast venography before hospital discharge did not receive any additional anticoagulation therapy.
At the Milton S. Hershey Medical Center of Pennsylvania State University from 1993 through 2001, bilateral contrast venography surveillance was performed in all patients enrolled in Institutional Review Board-approved sponsored clinical trials studying efficacy of thromboprophylaxis agents and modalities. Dalteparin12 and pneumatic foot pumps were evaluated in two separate sponsored trials. All other patients undergoing primary or revision TKA not on industry sponsored protocols received warfarin starting the night before operation with a target INR of 2.0; these patients underwent contrast venography of only the operated limb before discharge as part of a clinical care protocol. A total of 707 consecutive patients undergoing TKA were enrolled. Because of an increasing appreciation of the extended risk of VTED related events, all patients during this period who did not complete venography before hospital discharge were empirically continued on outpatient warfarin therapy to maintain a target prothrombin time ratio of 1.5 or an INR of 2.0 for 6 weeks postoperatively.
At the University of Maryland from 2002 through 2003, screening contrast venography was abandoned as a routine clinical examination for total joint replacement. Warfarin was used almost exclusively for VTED prophylaxis, and all patients were empirically continued on warfarin after hospital discharge to complete a desired course of 6 weeks of VTED prophylaxis with a target INR of 2.0. A total of 55 consecutive patients undergoing TKA were enrolled.
Before 1988, patients had general anesthesia, intermittent bladder catheterization, and parenteral intramuscular narcotics after total joint arthroplasty with an average length of hospital stay ranging from 14 to 21 days. In 1989, epidural anesthesia with a continuous epidural catheter for 48 hours postoperatively and indwelling Foley bladder catheterization were adopted as the standard of care for patients undergoing total joint replacement of the hip and knee; hospital stay ranged from 7 to 14 days.6 After 1992, all patients undergoing total joint arthroplasty followed a specified protocol that included continuous epidural anesthesia/analgesia for 48 hours postoperatively, an indwelling bladder catheter, and two or three days of inpatient rehabilitation as part of the routine hospital stay; length of stay averaged 4 days after knee replacement. Screening contrast venography of the operative limb was part of the clinical pathway and was generally performed on the day before discharge.
Management of VTED was directed by accepted clinical practice guidelines. Patients with a venogram positive for thrombosis distal to the popliteal vein were treated with warfarin for 6 weeks with a target INR of 2.0. No alteration was made in activity level or physical therapy regimen and no followup venous imaging studies were undertaken. Patients with a venogram positive for thrombosis in the popliteal vein or more proximal deep venous system (superficial femoral vein, perforators, or iliac vein) were prescribed bed rest and treated with intravenous unfractionated heparin, started as a continuous drip with no loading dose, to maintain the partial thromboplastin time (PTT) at two times control for 5 days or until the INR was therapeutic on warfarin, whichever was longer. Warfarin was started one day after the heparin drip and continued for 3 months with a target INR of 2.0. Patients with a pulmonary embolism documented by pulmonary angiography were managed similarly to patients with proximal thrombi, with warfarin therapy continuing for a total of 6 months postoperatively. Patients who refused contrast venography, or consented but were unable to complete this study for any reason, were managed in the same manner as those patients with positive venograms for distal thrombi: they received warfarin for 6 weeks with a target INR of 2.0. All warfarin therapy was managed through the attending surgeon's office (VDP) in conjunction with the orthopaedic clinical practice nurse; blood draws for PT/INR were performed each Monday and Thursday and results were faxed in to the orthopaedic office. For the entire duration of the study, patients who successfully completed contrast venography with no identified deep system thrombi were discharged home without any additional pharmacologic prophylaxis for VTED. All patients were instructed to continue use of graduated antiembolism stockings during daytime hours for 6 weeks after discharge; stockings were removed each evening.
Data analysis was conducted with the SAS package using a two-sided Fisher's exact test or Chi square test for nonparametric data as appropriate to sample size and frequency of events. Statistical significance was defined as p < 0.05.
One thousand three hundred twenty one patients were enrolled in the study (Table 1); among those undergoing total knee arthroplasty between the three study institutions, 810 patients (61.3%) successfully completed contrast venography. The venogram completion rate was 35.6% in Rochester, where only patients on sponsored clinical trials were accessed, and 86.4% in Hershey, where all patients undergoing TKA were accessed for the study protocol. The most common reasons for failure to complete contrast venography were patient refusal, inability to gain venous access in the foot, and contrast allergy.
After TKA there were a total of 343 positive venograms (42.3%), with 86 of 199 (43.2%) positive in Rochester and 257 of 611 (42.1%) positive in Hershey. Of the 343 confirmed positive studies, 10 (2.9%) were initially incorrectly read as negative (false negative) and patients were managed with no further anticoagulation at the time of hospital discharge. Overall, negative venograms were completed in 477 patients and contrast venography was not performed in 511 patients.
The overall hospital readmission rate for venous thromboembolic disease after TKA was 0.6%; there were three readmissions in Rochester compared with five readmissions in Hershey (Table 2). Readmission after extended warfarin anticoagulation occurred in one patient with a positive venogram; this was a man with known anterior and posterior tibial vein thrombi treated with warfarin for 6 weeks after primary TKA who underwent emergent laparoscopic cholecystectomy 2 weeks after discontinuation of warfarin with no subsequent anticoagulant therapy. On the 16th day after cholecystectomy, bilateral symptomatic femoral DVT was diagnosed by duplex ultrasonography, and he was again placed on warfarin and treated with no further sequelae. There were two readmissions (0.4%) for VTED in the 511 patients who did not complete venography; the early cohort received no warfarin, while after 1992 warfarin was routinely prescribed when venography was not performed. Both readmissions (0.55%) occurred among the early 360 patients not receiving outpatient warfarin at hospital discharge, while no readmissions were observed in the later cohort of 151 patients who had no venogram but continued warfarin therapy for 6 weeks after operation.
Among the 477 TKA patients with venograms negative for DVT, all of whom were discharged without additional anticoagulant prophylaxis, five patients (1.05%) were re- admitted for clinically evident thromboembolic events (Table 2). Symptomatic PE was the admitting diagnosis in three patients (0.6%) who presented between 38 and 210 days after the index arthroplasty; one PE was fatal 38 days after the primary procedure when the patient collapsed at home. The two nonfatal pulmonary emboli occurred in patients 133 and 210 days after operation; both had received pneumatic foot pumps for VTED prophylaxis while hospitalized after knee replacement. Symptomatic proximal DVT was diagnosed in the remaining two patients (0.4%) who presented after the index arthroplasty with thigh pain and leg swelling from superficial femoral vein thrombosis.
If TKA patients are stratified by presence or absence of extended warfarin therapy, the specific cohort of patients managed with extended outpatient warfarin anticoagulation (333 positive venograms and 151 no venogram after 1992) had an observed readmission rate of 0.21% (1/484) (Table 2). In comparison, the readmission rate among all patients discharged from the hospital without continued warfarin therapy, specifically on the basis of a negative contrast venogram, was 1.05% (p = 0.12). Likewise, the readmission rate among patients discharged without extended warfarin therapy for any reason, because of no venogram between the years of 1984 and 1992 or a known negative venogram, was 0.84% (p = 0.27) (Table 2).
One patient experienced a major complication related to bleeding after TKA. The patient received antithrombin III- unfractionated heparin prophylaxis for VTED and suffered a wound hematoma resulting in skin necrosis and wound breakdown. She eventually required muscle flap coverage of the knee and staged removal-reimplantation for deep infection. There were no major bleeding complications requiring readmission or reoperation in 484 patients maintained on outpatient warfarin therapy.
The overall prevalence of venographic DVT was 42.3% (343/810) after TKA compared with 16.9% (175/1032; p < 0.001) after THA.21 Nonetheless, these prevalence figures represent the aggregate effect of several different prophylaxis regimens studied over the 20-year period of this report. Institution of a structured clinical pathway using regional epidural anesthesia (rather than general endotracheal anesthesia) with standard warfarin VTED prophylaxis resulted in no meaningful reduction in DVT prevalence after total knee arthroplasty compared with prior use of general anesthesia (43.2% versus 42%). Interestingly, concurrent use of the same clinical pathway for THA patients reduced the prevalence of venographic DVT after THA from 22.5% (78/347) to 14.2% (p = 0.0008).21 Readmission for VTED was observed in 1.62% of all patients after THA22 compared with 0.6% of all patients after TKA (p = 0.009). Moreover, when only those patients who did not receive extended warfarin therapy are compared, VTED related readmission after THA21 remained more frequent than after TKA (1.92% versus 0.84%; relative risk = 2.3; p = 0.039). Specifically, readmission after a negative venogram with no outpatient warfarin therapy was observed in 2.2% of patients after THA21 compared with 1.05% (p = 0.19) of patients after TKA (Table 2). Pulmonary embolism occurred in 0.7% of THA compared with 0.2% (p = 0.08) of TKA patients while symptomatic proximal DVT occurred in 0.9% of THA compared with 0.4% (p = 0.09) of TKA patients.
Notwithstanding an apparent reduction in fatal thromboembolic disease with more widespread use of routine anticoagulant prophylaxis and contemporary operative techniques, venous thromboembolism remains the most common reason for emergency readmission after total joint replacement.19 However, the increasingly small complication rate experienced by an individual practitioner attributable to VTED, the increasing effectiveness of anticoagulants, and the purported accuracy of duplex ultrasonography as a screening tool for DVT have all contributed to a renewed interest in a minimalist approach to venous thromboembolism risk management. Strategies of limited or no routine pharmacologic thromboprophylaxis after hospital discharge of patients undergoing total hip and knee arthroplasty have been rationalized on the basis of the rare occurrence of fatal PE, the risk of outpatient bleeding complications,15 shorter hospital stays with more rapid mobilization of the patient, and the ability to periodically and serially screen patients for DVT with noninvasive methods.17,18 Accordingly, we sought to investigate the clinical course of patients and their risk of symptomatic venous thromboembolic disease after TKA, whether they had known thrombi or a confident absence of thrombi as determined by ascending contrast venography. The observations in this study refute the hypothesis that the known absence of thrombi at the time of hospital discharge after total knee arthroplasty may exempt the patient from the need for additional anticoagulant prophylaxis. Our findings also dismiss the usefulness of routine screening for DVT as a guide for subsequent anticoagulant management. Our empirical rationale to continue warfarin for 6 weeks after hospital discharge was based on previous work that indicated readmission for VTED occurs within a 3-month window after THA. All readmissions for symptomatic DVT occurred within 8 weeks and readmissions related to PE all occurred within 3 months of operation.20 While the design of this study did not specifically address the variety of possible durations of prophylaxis after hospital discharge, there was a substantial reduction in read- mission rate for the combined THA and TKA population and nearly so for the isolated TKA group with our regimen of 6 weeks of warfarin therapy after discharge.
Limitations of the present study are related to the large numbers necessary to demonstrate statistical significance in low probability events, such as readmission for fatal PE. Readmission data were readily available for the aggregate THA and TKA analysis and our comparison with THA patients21 for symptomatic PE and all VTED related re- admissions over the two decades of this study. Similarly, while VTED readmission after TKA was five times greater in patients discharged without continued warfarin therapy compared with those receiving extended warfarin for any reason, even for a known fresh thrombus, the difference fell short of statistical significance (0.21% versus 1.05%; p = 0.12). Given we no longer perform venogram screening, standard treatment of only an additional 200 TKA patients with extended warfarin therapy who experienced no VTED readmission would achieve statistical significance for this comparison (p = 0.04). Interestingly, because there was no PE observed after THA or TKA in any patient maintained on extended warfarin prophylaxis, we would require an additional 2000 patients without a fatal PE on the same treatment regimen to demonstrate a reduction in fatal PE in the combined group of THA and TKA patients. In an effort to manage the statistical issues presented by the study of such infrequent events, our data span 20 years and three teaching institutions, during which the perioperative management of the total joint arthroplasty patient has substantially changed. Similarly, several prophylactic agents were used in various clinical trials during this period; nonetheless, the entry point for this analysis was the presence or absence of thrombi at the time of hospital discharge, at which point all patients were consistently managed on a single treatment schema. Exposure to variable anticoagulants or mechanical prophylaxis before the time of discharge may have some small impact on the subsequent course of thrombus formation or lysis and the clinical course of the patients over the ensuing weeks and months. This effect is difficult to anticipate or quantify but should be relatively trivial over the time period our patients were followed. Finally, the issue of venogram- induced thrombi must be addressed. While such contrast- induced clots are known to occur with older contrast agents, the risk of thrombus formation after use of nonionic contrast as employed in this study is exceedingly small and is largely restricted to superficial clots with little to no embolic potential.
In the largest previous venogram natural history study after TKA to date, Stulberg et al27 documented an 84% prevalence of deep venous thrombosis in patients without VTED prophylaxis. Isolated calf thrombi were the most common finding, proximal thrombi were always contiguous with calf disease and were present in roughly 10% of patients, and less than 5% of patients had DVT in the contralateral nonoperative limb. In their series of 638 arthroplasties, there were no fatal pulmonary emboli and a clinically evident pulmonary embolism rate of 1.7%. Consistent with Insall's previously stated concern about veno- graphic findings, the authors recommended “routine prophylactic screening… for most patients” after TKA.27 This latter position is refuted by our observations on veno- graphically screened patients with negative findings who were discharged without continued thromboprophylaxis.
Our data underscore the value of warfarin in preventing readmission for venous thromboembolic events after total knee arthroplasty, even when the same warfarin was ineffective in preventing formation of the initial clot as determined by a positive screening venogram.22 In only one TKA patient discharged on warfarin therapy for any reason, including 333 TKA patients who received extended warfarin therapy based on an initial venographic reading that was positive for deep system thrombi, did progression of clinically evident thrombosis become apparent as a re- admission event. Of even greater importance, those patients with a deep venous system known to be clear of thrombi as documented by a negative venogram at the time of discharge, and thereby exempted from continued anticoagulation on our protocol, experienced a readmission event nearly five times more commonly after TKA (p = 0.12) than fellow patients already harboring known thrombi but managed with continued warfarin therapy. In our experience, discontinuation of prophylactic anticoagulation at the time of discharge after TKA, even in the seemingly most reassuring situation with a confirmed negative venogram, would appear to be ill-advised in the absence of contraindications against anticoagulant therapy. Our present management of VTED risk after total knee arthroplasty includes 6 weeks of postoperative warfarin therapy for all patients with a target INR of 2.0, as in this series.
A strategy of periodic surveillance screening to direct duration and type of anticoagulation prophylaxis must be reevaluated in view of these data. Doppler ultrasound has been very effective in the detection of symptomatic DVT. Its application as a screening tool, however, for identification of nonocclusive asymptomatic thrombosis in the postoperative setting provides a greater challenge. Thigh thrombi have been successfully identified in several centers, but identification of thrombi in the calf is highly variable with sensitivities reported as low as 12% compared with venography.4 Similarly, visualization of pelvic veins is limited by the ileus, not uncommonly seen in the postoperative patient, and noncompressibility of the deep veins of the pelvis. In one large series comparing contrast venography to color Doppler ultrasound in 2000 patients having undergone orthopaedic operation, ultrasound was found to have a sensitivity of only 62% in detecting asymptomatic proximal DVT compared with venography.29 A multicenter study including 385 patients who had undergone unilateral hip or knee replacement also found poor sensitivity (38%) of color Doppler ultrasound in detecting proximal venous thrombosis.7 Advantages of color Doppler ultrasound include its noninvasive nature, relatively low cost, and portability. It is, however, highly operator dependent, with accuracy varying widely depending upon the expertise and experience of the technician. Ultrasound is also insensitive in the identification of isolated calf thrombi, which comprised 90% of our clots after TKA identified by contrast venography at the time of discharge; as such, it is a poor surveillance tool for identification of asymptomatic postoperative DVT in most institutions. Advances in magnetic resonance imaging of the venous system permit noninvasive identification of clots in the lower extremity, including the iliac veins, and these thrombi may explain the occurrence of PE in patients who have a negative venogram or ultrasound. Notwithstanding resolution of these technical issues, routine surveillance, even with a highly sensitive and perfectly accurate tool, would appear to represent an impractical and imprudent strategy for out- patient management of venous thromboembolic risk in view of a new thrombus formation rate after discharge approximating 20% as documented by serial venography2,23 and our compelling readmission data in patients receiving no further anticoagulation after hospital discharge on the basis of a negative contrast venogram.
The findings of this study compiled over two decades with venogram surveillance at the time of hospital discharge provide strong evidence in support of a strategy of continued outpatient warfarin prophylaxis for VTED after TKA. Moreover, these results were compiled with a target INR of 2.0, at the lower end of the acceptable range endorsed by the American College of Chest Physicians. Eikelboom et al8 has suggested there is a reduction in symptomatic venous thromboembolic events with extended fractionated heparin prophylaxis after THA, but those agents have been shown to be more effective at primary prevention of thrombus formation than warfarin while exposing patients to a greater bleeding risk.5,12,13 Our data demonstrate the efficacy of warfarin as an agent for secondary prophylaxis, even in patients in whom initial thrombi have formed in the face of warfarin, with a minimal risk of bleeding complications. The case for extended anticoagulant prophylaxis after TKA is compelling. Warfarin is a safe and effective agent in this clinical context, and management of venous thromboembolic risk by routine surveillance methods and selective prophylaxis would no longer appear to be justified.
1. Amstutz HC, Friscia DA, Dorey F, Carney BT. Warfarin prophylaxis to prevent mortality from pulmonary embolism after total hip replacement. J Bone Joint Surg Am
2. Bergqvist D, Benoni G, Bjorgell O. Fredin H, Hedlundh U, Nicolas S, Nilsson P, Nylander G. Low molecular weight heparin (enoxaparin) as prophylaxis against venous thromboembolism after total hip replacement. N Engl J Med
3. Charnley J. Low Friction Arthroplasty of the Hip: Theory and Practice
. Berlin, Germany: Springer-Verlag; 1979.
4. Ciccone WJ II, Fox PS, Neumeyer M, Rubens D, Parrish WM, Pellegrini VD Jr. Ultrasound surveillance for asymptomatic deep venous thrombosis after total joint replacement. J Bone Joint Surg Am
5. Colwell CW Jr, Collis D, Paulson R, McCutchen J, Bigler G, Lutz S, Hardwick M. Comparison of enoxaparin and warfarin for the prevention of venous thromboembolic disease after total hip arthroplasty: evaluation during hospitalization and three months after discharge. J Bone Joint Surg Am
6. Dalldorf P, Perkins F, Totterman S, Pellegrini VD Jr. Deep venous thrombosis following total hip arthroplasty: effects of prolonged postoperative epidural anesthesia. J Arthroplasty
7. Davidson BL, Elliott CG, Lensing AW. Low accuracy of color Doppler ultrasound in the detection of proximal leg vein thrombosis in asymptomatic high-risk patients. Ann Intern Med
. 1992;117: 735-738.
8. Eikelboom JW, Quinlan DJ, Douketis JD. Extended-duration prophylaxis against venous thromboembolism after total hip or knee replacement: a meta-analysis of the randomized trials. Lancet
. 2001; 358:9-15.
9. Francis CW, Pellegrini VD Jr, Harris C, Marder VJ. Antithrombin III prophylaxis of venous thromboembolic disease after total hip or total knee replacement. Am J Med
10. Francis CW, Pellegrini VD Jr, Harris CM, Stulberg B, Gabriel KR, Marder VJ. Prophylaxis of venous thrombosis following total hip and total knee replacement using antithrombin III and heparin. Semin Hematol
11. Francis CW, Pellegrini VD Jr, Marder VJ, Harris CM, Totterman S, Gabriel KR, Baughman DJ, Roemer S, Burke J, Goodman TL, Evarts MC. Prevention of venous thrombosis after total hip arthroplasty: antithrombin III and low-dose heparin compared with dextran-40. J Bone Joint Surg Am
12. Francis CW, Pellegrini VD Jr, Totterman S, Boyd AD, Marder VJ, Leibert KM, Stulberg BN, Ayers DC, Rosenberg A, Kessler C, Johanson N. Prevention of deep-vein thrombosis after total hip arthroplasty: comparison of warfarin and dalteparin. J Bone Joint Surg Am
13. Imperiale TF, Speroff T. A meta-analysis of methods to prevent venous thromboembolism following total hip replacement. JAMA
14. Johnson R, Green JR, Charnley J. Pulmonary embolism and its prophylaxis following the Charnley total hip replacement. Clin Orthop Relat Res
15. Landefeld CS, Goldman L. Major bleeding in outpatients treated with warfarin: incidence and prediction by factors known at the start of outpatient therapy. Am J Med
16. Lieberman JR, Wollaeger J, Dorey F, Thomas BJ, Kilgus DJ, Grecula MJ, Finerman GA, Amstutz HC. The efficacy of prophylaxis with low-dose warfarin for prevention of pulmonary embolism following total hip arthroplasty. J Bone Joint Surg Am
. 1997;79: 319-325.
17. Murray DW, Britton AR, Bulstrode CJK. Thromboprophylaxis and death after total hip replacement. J Bone Joint Surg Br
. 1996;78: 863-870.
18. Murray DW, Carr AJ, Bulstrode CJK. Pharmacological thromboprophylaxis and total hip replacement. J Bone Joint Surg Br
. 1995; 77:3-5.
19. Pellegrini VD Jr. Prevention of thromboembolic disease in patients after total joint arthroplasty by selective posthospitalization treatment. In: Seminars in Arthroplasty, Fitzgerald RH, ed. Philadelphia: W.B. Saunders Co; 1997:248-257.
20. Pellegrini VD Jr, Clement D, Lush-Ehmann C, Keller GS, Evarts CM. The natural history of thromboembolic disease following hospital discharge after total hip arthroplasty: the John Charnley Award. Clin Orthop Relat Res
21. Pellegrini VD Jr, Donaldson CT, Farber DC, Lehman EB, Evarts CM. The John Charnley Award: prevention of readmission for venous thromboembolic disease after total hip arthroplasty. Clin Orthop Relat Res
22. Pellegrini VD Jr, Langhans MJ, Totterman S, Marder VJ, Francis CW. Embolic complications of calf thrombosis following total hip arthroplasty. J Arthroplasty
23. Planes A, Vochelle N, Darmon JY, Fagola M, Bellaud M, Huet Y. Risk of deep venous thrombosis after hospital discharge in patients having undergone total hip replacement: double blind randomized comparison of enoxaparin versus placebo. Lancet
. 1996;348: 224-228.
24. Salvati EA, Pellegrini VD Jr, Sharrock NE, Lotke PA, Murray DW, Potter H, Westrich GH. Recent advances in venous thromboembolic prophylaxis during and after total hip replacement. J Bone Joint Surg
25. Sculco TP, Colwell CW, Pellegrini VD Jr, Westrich GH, Bottner F. Prophylaxis against venous thromboembolic disease in patients having a total hip or knee arthroplasty. J Bone Joint Surg Am
. 2002; 84:466-477.
26. Sharrock NE, Go G, Harpel PC, Ranawat CS, Sculco TP, Salvati EA. Thrombogenesis during total hip replacement. Clin Orthop Relat Res
27. Stulberg BN, Insall JN, Williams GW, Ghelman B. Deep-vein thrombosis following total knee replacement: an analysis of six hundred and thirty-eight arthroplasties. J Bone Joint Surg Am
. 1984; 66:194-201.
28. Warwick D, Williams MH, Bannister GC. Death and thromboembolic disease after total hip replacement: a series of 1162 cases with no routine chemical prophylaxis. J Bone Joint Surg Br
29. Wells PS, Lensing AW, Davidson BL, Prins MH, Hirsh J. Accuracy of ultrasound for the diagnosis of deep venous thrombosis in asymptomatic patients after orthopedic operation: a meta-analysis. Ann Intern Med