The study population consisted of 5832 patients undergoing elective primary or revision total hip or total knee replacement surgery at the authors’ institution from September 1997 to February 2003. The Institutional Review Board approved this study. Pre-, intra-, and postoperative variables from patient charts (Table 1) were analyzed to identify possible risk factors for the development of PE. Physical activity was rated using a modified version of the University of California Los Angeles (UCLA) Activity Level Scale [28], with 1 representing the lowest level of activity and 5 the highest.

Table 1

Risk factor characteristics for pulmonary embolism

^aTKA patients only.

A total of 457 patients were tested for suspected thromboembolic events during their acute inpatient stay or acute rehabilitation stay following surgery. Patients were considered to have a PE if they had a high-probability reading on a ventilation/perfusion scan (the test most frequently used for this purpose in our institution before 1999) and/or a positive result on a CT angiogram (1998-2003).  

Patients were divided into groups corresponding to the method of DVT prophylaxis administered them:

1. Warfarin (coumadin) 5-10 mg started the day before surgery and titrated to an INR of approximately 2.0 and continued for six weeks post-operatively;

2. Low-molecular-weight heparin (LMWH; Lovenox, enoxaparin sodium, Rhone Poulenc Rorer, Collegeville, PA) (subcutaneous injection of either 40mg once per day or  30 mg twice a day) started between 12 and 24 hr postoperatively and continued for 14 post-operative day;

3. Aspirin (325 mg each day) in conjunction with foot pneumatic compression devices (Kendall AV Impulse System, Mansfield, MA); compression devises were discontinued upon discharge and aspirin was continues for six weeks post-operatively.

Statistical Methods 

We considered all potential risk factors and confounders bivariately with PE status.  For continuous variables (age, American Society of Anesthesiologists [ASA] grade, body mass index, physical activity, operating room [OR] time, estimated blood loss [EBL], tourniquet time) the differences between PE cases and controls were compared using Student’s t-test, while for categorical variables (gender, diabetes and cardiac disease status, joint, anesthesia type, DVT prophylaxis), differences were compared using the Fisher’s exact test (with significance defined as p ≤ .05 for both tests). Multiple logistic regression analysis was then used to model the independent predictors of PE. Those variables found to be significant in the bivariate analysis, as well as any variables considered a priori as important confounders, were included in this logistic model. The logistic model did not, however, include the comorbidity variables (diabetes and cardiac disease), as these were assessed differently for the cases and controls.

Results: 

The study population was 67% female, had a mean age of 63 years, and comprised 2,665 THAs (1,679 unilateral and 986 bilateral procedures; this group included 399 revision procedures) and 3,167 TKAs (1,425 unilateral and 1,742 bilateral procedures; this group included 190 revision procedures). Thirty patients experienced a PE, 9 following THA and 21 following TKA; the mean age of these patients was 71 years (range: 59-88 years); 24 were females and 6 males. The mean postoperative day the PE was encountered was 4.2 (range 1-17). 

Student t-tests performed on continuous variables found that increasing patient age, increased American Society of Anesthesiologists (ASA) grade, and increased body mass index (BMI) were associated with an increased risk of PE (Table 2). No association for either increased or decreased risk for PE was found for patient preoperative physical activity level, procedure length (OR time), estimated blood loss (EBL), or length of tourniquet time (TKA patients only).

Table 2

Two-sample t-test bivariate analysis of characteristics with continuous variables.

^aTKA patients only; * p ≤ .05 for association with PE.

Fisher’s exact tests performed on categorical variables revealed an increased risk of PE for both female gender and TKA (Table 3). No association was found with preoperative cardiac disease, preoperative diabetes, type of anesthesia, or postoperative DVT prophylaxis.

Table 3

Fisher’s exact analysis of characteristics with categorical variables

* p ≤ .05 for association with PE.

The multiple logistic regression model showed that increasing age, increasing BMI, and female gender were associated with increased risk of PE (Table 4). Specifically, each year of advancing age was associated with a 7% increase in risk of PE (odds ratio = 1.07, p = .004). Women experienced 7.5 times the risk for developing a PE as men (odds ratio = 7.54, p = .05). An 11% increase in PE incidence was associated with each kg/m² increase in BMI (odds ratio = 1.11, p < .001). There were no statistically significant differences in PE risk among DVT prophylaxis types (p = .967); post hoc power analysis showed that 151 PEs would be required to demonstrate statistical significance between prophylaxis types.

Table 4

Multivariate logistic regression analysis

^aAspirin/compression boots is the referent value; * p ≤ .05 for association with PE.

The following logistic model can be transformed to estimate the probability of PE after specifying the covariate values for the individual patient:

Once transformed from the logit function, the following formula can then be used to represent the probability of PE for a given set of patient characteristics:

This equation allows the calculation of individual probabilities of PE given specific patient profiles.

Discussion :

This study of over 5000 total joint arthroplasties utilizing clinically detectable pulmonary embolism as an end point evaluated several preoperative, intraoperative, and postoperative potential risk factors for PE (Table 1). Increasing patient age, female gender, and increasing BMI were found to be independently associated with an increased risk for pulmonary embolism. While many studies have compared various DVT prophylaxes with the development of VTE, [10-13, 16] we found no association between any of three methods of VTE prophylaxis and greater or lesser risk of PE.

This retrospective observational study has both strengths and weaknesses. We were able to analyze data related to more than 5000 total hip and knee arthroplasties by retrieving the data gathered at the time of surgery and thereafter, which allowed for precise statistical analysis. On the other hand, medical comorbidities at the time of surgery were placed in categories, but not recorded in detail. Thus all preoperative cardiac conditions are classified as “systemic cardiac disease” and may include congestive heart failure, arrhythmias, and hypertension. Therefore, when assessing PE patients’ medical comorbidities, direct comparison with the control cohort was not possible. Our search of hospital records on ventilation perfusion scanning and CT angiograms gave us an accurate assessment of patients who experienced PE during the acute hospitalization or during their stay in the acute rehabilitation unit; it did not, however, capture any patient who had a PE after discharge to home or to a different facility. Since the majority of PE events occur within the first 14 days following arthroplasty [6, 29, 30], we are reasonably assured that we have not substantially underestimated the number of cases with clinically relevant PE. Furthermore, we examined our population by discharge to home versus to rehabilitative care to determine if there were substantive differences in risk factor profile between these two sub-populations (data not shown). While, those discharged to rehabilitative care were older (mean age 65 vs. 60 yrs), they were not significantly heavier, they were not in worse medical condition (according to the American Society of Anesthesiologists risk score), and they did not demonstrate any difference in length of hospital stay. Moreover, we do not feel that the difference in age between those discharged to home and those discharged to rehabilitative care represents potential bias in our results since all of our PE events were assessed during the acute patient stay rather than a combination of inpatient and rehabilitative patient accrual.

Our finding that increasing age is a risk factor for the development of PE after total joint arthroplasty is supported by White et al. [31], who noted that for each 10 years over the age of 50, the rate of PE increased by 15% (odds ratio = 1.15; 95% CI = 1.1-1.3). Lemos et al. showed that increasing age was significantly associated with the development of pulmonary embolism among patients undergoing a warfarin prophylactic regimen following arthroplasty[32]. Mantilla et al., in a retrospective study examining complications after total joint arthroplasty, also noted an increase in PE among older patients (p < .001) [33]. They found that embolic events were almost twice as likely in patients in their eighth decade as in patients in their sixth decade of life. This difference, however, was only apparent between the extreme quantiles; they did not find a linear increase in PE with increasing age. Increased venous stasis, which may be most relevant for VTE, is also a characteristic of increasing age [34].

It has not previously been shown that women have an increased risk for pulmonary embolism following arthroplasty. The study by Lemos et al. controlled for gender in their analysis of risk factors for PE after total joint arthroplasty [32]. In a separate study by Mantilla et al., the authors found no difference in incidence of PE in male and female patients [33]. In our study group, 24 of the 30 PE events were in female patients. BMI and age were both strong predictors of PE in our study, and women had a significantly higher BMI (p < .001) and were significantly older (p < .001) than men (data not shown). While BMI and age contribute to some of the excess risk experienced by women, gender differences nevertheless remained even after controlling for these two risk factors (Table 4).

The finding that higher BMI is a risk factor for the development of PE is supported by the work of Mantilla et al., who found a 50% increased risk of PE for each 5 kg/m² increase in BMI [30]. White et al. found that a BMI greater than 25 kg/m² was associated with an increased risk in patients rehospitalized for a VTE [35]. Obesity has been previously associated with complications following total joint arthroplasty [36, 37]. It has been postulated that obesity is associated with an increased inflammatory state as well as a restriction in venous outflow and blood propulsion [38]. These physiological alterations, particularly in the presence of subclinical unstable atherosclerotic lesions, may be the basis for the increase in embolic events, and thus we believe that this represents a strong patient risk factor for PE that should be considered by the patient and surgeon.

Several conclusions that can be drawn from our study regarding preoperative assessment must be further examined before they deserve to be made recommendations. A multicenter approach may allow investigators to acquire enough PE cases over a reasonable time period. This type of study should allow for additional stratification of various DVT prophylaxes and other risk factors that predispose total arthroplasty patients to embolic events. This is particularly important to determine the optimal intra- and postoperative thromboprophylactic regimen in the presence of specific risk factor profiles, something that our current lack of statistical power would not permit.

Based on the predictive models presented here using our data, one can tentatively ascribe an overall level of risk to patients prior to their total joint arthroplasty based on the patient BMI, age, ASA grade, and gender. We believe that these statistically significant risk factors can be utilized in the future to outline a more effective thromboprophylaxis following total hip and total knee arthroplasty.

Reference :

1. Haake, D.A. and S.A. Berkman, Venous thromboembolic disease after hip surgery. Risk factors, prophylaxis, and diagnosis. Clin Orthop Relat Res, 1989(242): p. 212-31.
2. Stamatakis, J.D., et al., Femoral vein thrombosis and total hip replacement. Br Med J, 1977. 2(6081): p. 223-5.
3. Turpie, A.G., et al., A randomized controlled trial of a low-molecular-weight heparin (enoxaparin) to prevent deep-vein thrombosis in patients undergoing elective hip surgery. N Engl J Med, 1986. 315(15): p. 925-9.
4. Geerts, W.H., et al., Prevention of venous thromboembolism. Chest, 2001. 119(1 Suppl): p. 132S-175S.
5. Lassen, M.R., et al., Prevention of thromboembolism in 190 hip arthroplasties. Comparison of LMW heparin and placebo. Acta Orthop Scand, 1991. 62(1): p. 33-8.
6. Johnson, R., J.R. Green, and J. Charnley, Pulmonary embolism and its prophylaxis following the Charnley total hip replacement. Clin Orthop Relat Res, 1977(127): p. 123-32.
7. Lotke, P.A., et al., Indications for the treatment of deep venous thrombosis following total knee replacement. J Bone Joint Surg Am, 1984. 66(2): p. 202-8.
8. Stulberg, B.N., et al., Deep-vein thrombosis following total knee replacement. An analysis of six hundred and thirty-eight arthroplasties. J Bone Joint Surg Am, 1984. 66(2): p. 194-201.
9. Stringer, M.D., et al., Deep vein thrombosis after elective knee surgery. An incidence study in 312 patients. J Bone Joint Surg Br, 1989. 71(3): p. 492-7.
10. Brookenthal, K.R., et al., A meta-analysis of thromboembolic prophylaxis in total knee arthroplasty. J Arthroplasty, 2001. 16(3): p. 293-300.
11. Freedman, K.B., et al., A meta-analysis of thromboembolic prophylaxis following elective total hip arthroplasty. J Bone Joint Surg Am, 2000. 82-A(7): p. 929-38.
12. Lieberman, J.R., et al., The efficacy of prophylaxis with low-dose warfarin for prevention of pulmonary embolism following total hip arthroplasty. J Bone Joint Surg Am, 1997. 79(3): p. 319-25.
13. Sarmiento, A. and A.D. Goswami, Thromboembolic prophylaxis with use of aspirin, exercise, and graded elastic stockings or intermittent compression devices in patients managed with total hip arthroplasty. J Bone Joint Surg Am, 1999. 81(3): p. 339-46.
14. Lieberman, J.R., et al., Low-dose warfarin prophylaxis to prevent symptomatic pulmonary embolism after total knee arthroplasty. J Arthroplasty, 1997. 12(2): p. 180-4.
15. Westrich, G.H., et al., Meta-analysis of thromboembolic prophylaxis after total knee arthroplasty. J Bone Joint Surg Br, 2000. 82(6): p. 795-800.
16. Stern, S.H., R.L. Wixson, and D. O'Connor, Evaluation of the safety and efficacy of enoxaparin and warfarin for prevention of deep vein thrombosis after total knee arthroplasty. J Arthroplasty, 2000. 15(2): p. 153-8.
17. Anderson, F.A., Jr. and F.A. Spencer, Risk factors for venous thromboembolism. Circulation, 2003. 107(23 Suppl 1): p. I9-16.
18. White, R.H. and M.C. Henderson, Risk factors for venous thromboembolism after total hip and knee replacement surgery. Curr Opin Pulm Med, 2002. 8(5): p. 365-71.
19. Goldhaber, S.Z., et al., A prospective study of risk factors for pulmonary embolism in women. Jama, 1997. 277(8): p. 642-5.
20. Kim, Y.H. and V.E. Kim, Factors leading to low incidence of deep vein thrombosis after cementless and cemented total knee arthroplasty. Clin Orthop Relat Res, 1991(273): p. 119-24.
21. Piccioli, A., et al., Cancer and venous thromboembolism. Am Heart J, 1996. 132(4): p. 850-5.
22. Daly, E., et al., Risk of venous thromboembolism in users of hormone replacement therapy. Lancet, 1996. 348(9033): p. 977-80.
23. Sharrock, N.E., et al., Effects of epidural anesthesia on the incidence of deep-vein thrombosis after total knee arthroplasty. J Bone Joint Surg Am, 1991. 73(4): p. 502-6.
24. Bruce, W., et al., Occurrence of pulmonary thromboembolism immediately after arthroplasty. Nucl Med Commun, 2001. 22(11): p. 1237-42.
25. Wolf, L.D., W.J. Hozack, and R.H. Rothman, Pulmonary embolism in total joint arthroplasty. Clin Orthop Relat Res, 1993(288): p. 219-33.
26. Kim, Y.H., S.H. Oh, and J.S. Kim, Incidence and natural history of deep-vein thrombosis after total hip arthroplasty. A prospective and randomised clinical study. J Bone Joint Surg Br, 2003. 85(5): p. 661-5.
27. Grady-Benson, J.C., et al., Postoperative surveillance for deep venous thrombosis with duplex ultrasonography after total knee arthroplasty. J Bone Joint Surg Am, 1994. 76(11): p. 1649-57.
28. Amstutz, H.C., et al., Treatment of primary osteoarthritis of the hip. A comparison of total joint and surface replacement arthroplasty. J Bone Joint Surg Am, 1984. 66(2): p. 228-41.
29. Westrich, G.H., et al., The incidence of venous thromboembolism after total hip arthroplasty: a specific hypotensive epidural anesthesia protocol. J Arthroplasty, 1999. 14(4): p. 456-63.
30. Mantilla, C.B., et al., Risk factors for clinically relevant pulmonary embolism and deep venous thrombosis in patients undergoing primary hip or knee arthroplasty. Anesthesiology, 2003. 99(3): p. 552-60; discussion 5A.
31. White, R.H., et al., Incidence and time course of thromboembolic outcomes following total hip or knee arthroplasty. Arch Intern Med, 1998. 158(14): p. 1525-31.
32. Lemos, M.J., et al., Pulmonary embolism in total hip and knee arthroplasty. Risk factors in patients on warfarin prophylaxis and analysis of the prothrombin time as an indicator of warfarin's prophylactic effect. Clin Orthop Relat Res, 1992(282): p. 158-63.
33. Mantilla, C.B., et al., Frequency of myocardial infarction, pulmonary embolism, deep venous thrombosis, and death following primary hip or knee arthroplasty. Anesthesiology, 2002. 96(5): p. 1140-6.
34. Lowe, G.D., Virchow's triad revisited: abnormal flow. Pathophysiol Haemost Thromb, 2003. 33(5-6): p. 455-7.
35. White, R.H., et al., Predictors of rehospitalization for symptomatic venous thromboembolism after total hip arthroplasty. N Engl J Med, 2000. 343(24): p. 1758-64.
36. Miric, A., et al., Perioperative morbidity following total knee arthroplasty among obese patients. J Knee Surg, 2002. 15(2): p. 77-83.
37. Jain, N.B., et al., Comorbidities increase complication rates in patients having arthroplasty. Clin Orthop Relat Res, 2005(435): p. 232-8.
38. Visser, M., et al., Elevated C-reactive protein levels in overweight and obese adults. Jama, 1999. 282(22): p. 2131-5.