When a bilateral total knee replacement is indicated, it is not clear whether both knees should be operated on during the same hospitalization (simultaneous bilateral total knee replacement) or whether it is better to wait until the patient has recovered from the first total knee replacement before the second procedure is carried out during a second hospital stay (staged bilateral total knee replacement). An important consideration in the choice of whether to perform a simultaneous or staged bilateral total knee replacement is the risk of serious complications such as pulmonary embolism. Some authors have reported a greater perioperative risk of pulmonary embolism after simultaneous bilateral knee replacements than after unilateral procedures1,2. Other large studies have demonstrated an increased risk of all thromboembolic and/or pulmonary complications following simultaneous replacement3-6. Several other investigations have shown no significant difference in the risk of perioperative thromboembolism between bilateral and unilateral total knee replacements7-11, but those studies had relatively small numbers of patients and therefore low power to detect differences between the two groups. In one single-center study, the complication rate associated with a simultaneous total knee replacement was found to be actually lower than that following a unilateral procedure12. Consequently, it is not clear whether a staged set of procedures is preferable to a simultaneous approach. Moreover, little is known regarding the characteristics of patients who receive each of the two types of intervention. Ritter et al. described the demographic characteristics of Medicare recipients who had undergone bilateral total knee replacement between 1985 and 199013, but to our knowledge no other national population-based description of bilateral total knee replacement has appeared in print.
In this study, we compared all Medicare recipients who had had a simultaneous bilateral procedure in 2000 with those who had had a staged bilateral procedure (with one of the operations performed in 2000). We compared the two groups with regard to their demographic attributes, comorbidities, geographic distribution, hospital and surgeon volume, and probability of pulmonary embolism in the first three months postoperatively.
Materials and Methods
This study focused on Medicare beneficiaries who had received a total knee replacement in 1999, 2000, or 2001. Medicare claims submitted by hospitals (Medicare Part A) and by surgeons (Medicare Part B) in 2000 were used to define the index primary total knee replacement. These procedures were identified by a hospital claim with International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM)14 procedure code 81.54 (total knee replacement) together with a surgeon's claim with Current Procedural Terminology (CPT) Fourth Edition15 code 27447 (arthroplasty, knee, condyle and plateau; medial and lateral components with or without patella resurfacing [total knee replacement]). To focus on elective total knee replacement, we excluded from the study cohort patients with claims indicating preexisting infection of the knee, metastatic cancer, or bone cancer. In order to obtain complete hospital claim histories for the years 1999, 2000, and 2001, we also excluded patients enrolled in health maintenance organizations, those not enrolled in both parts of Medicare, those who were not residents in the United States, and those who were less than sixty-six years of age at the time of the index surgery in 2000. If a patient had had a total knee replacement during more than one hospitalization in 2000, the first hospitalization was considered to be the index hospitalization.
When a patient had had an index total knee replacement in 2000, hospital claims were searched for any other total knee replacement in 1999, 2000, or 2001. When the index hospital claim had two ICD-9-CM procedure codes indicating total knee replacement (81.54), the event was considered to be a simultaneous bilateral procedure. We found it impossible to determine reliably from the claims data whether a given simultaneous bilateral total knee replacement had been done as a double procedure during a single anesthesia session or as sequential procedures during the same hospitalization with two separate anesthesia sessions. A patient was considered to have had a staged procedure if another total knee replacement had been done in 1999, 2000, or 2001 but not during the same hospitalization as the index procedure.
Characteristics of Medicare Patients with a Simultaneous
Total Knee Replacement and Those with a Staged Procedure Bivariate relationships between demographic variables and provider procedure volume on the one hand and staging status (simultaneous or staged less than a year apart) on the other were examined with contingency tables and the chi-square test. The demographic factors included age group, gender, race, census division of the United States, and eligibility for Medicaid supplement (a marker of low income). In addition, an adaptation of the Charlson index of comorbidity16,17 was calculated with use of inpatient claims data for the hospitalization for the first total knee replacement. The relationships between the staging status and the total number of total knee replacements performed in 2000 on Medicare beneficiaries by the hospital and by the surgeon were also examined. A multivariate logistic regression analysis was then performed to identify factors associated with having a simultaneous total knee replacement rather than two procedures staged less than a year apart, with adjustment for all of the above variables.
Pulmonary Embolism After Total Knee Replacement
We examined the probability of a symptomatic pulmonary embolism developing in the first three months after the first total knee replacement for all patients. A diagnosis of pulmonary embolism was ascertained by searching for hospital claims (including those for the index hospitalization) with an ICD-9-CM code of 415.1-415.19 (pulmonary embolism and infarction: iatrogenic or other). We used actuarial life-table techniques to estimate the proportion of patients in each group who had a pulmonary embolism in the first three months after surgery. Life tables tally the events of subjects during the separate short intervals of time that each subject is being followed. From these data, it is possible to estimate event rates over various time intervals while taking into account the fact that some subjects have died or have been lost to follow-up18,19.
In addition, we performed proportional hazards regression analysis to estimate the pulmonary embolism hazard for patients undergoing simultaneous bilateral procedures relative to all other patients, with adjustment for the demographic, comorbidity, and hospital and surgeon-volume characteristics listed above. All analyses were carried out with use of SAS statistical software (release 8.2; SAS Institute, Cary, North Carolina). The critical p value for significance was 0.05.
We identified 122,385 eligible patients as having had an index total knee replacement in 2000 (Table I). Of these patients, 8324 (6.8%) had two total knee replacements during the same hospitalization (simultaneous total knee replacement) and 13,039 (10.7%) had another staged total knee replacement within a year before or after the index hospitalization; 4957 patients (4.1%) had two total knee replacements between one and two years apart. The remaining 96,065 patients (78.5%), who had no other total knee replacement in 1999, 2000, or 2001, were considered to have had a unilateral total knee replacement.
Characteristics of Medicare Patients with a Simultaneous Total Knee Replacement and Those with a Staged Procedure
The bivariate analyses showed that the simultaneous total knee replacements were much more likely than the staged procedures to have been done in high-volume hospitals and by high-volume surgeons (Table II). The simultaneous procedures were also more likely than the staged procedures to have been carried out in men and in whites and were less likely to have been carried out in low-income patients (those with Medicaid coverage). Simultaneous procedures were much less common in the Pacific states than elsewhere. Very little difference in the Charlson comorbidity index scores was seen between the simultaneous and staged-procedure groups.
The strong associations with hospital and surgeon volume were confirmed by the logistic regression analysis, as was the finding that men had proportionately more simultaneous procedures than did women (Table III). When procedure volume and other covariates were taken into account, an east-west pattern became apparent. Hospitals in northeastern states were the most likely to perform simultaneous procedures, followed by those in other states east of the Mississippi River and by the Rocky Mountain states. Hospitals in the west-central and Pacific divisions were the least likely to perform simultaneous procedures.
Pulmonary Embolism After Total Knee Replacement
In the first three months after total knee replacement, the probability of having a symptomatic pulmonary embolism, as estimated from life tables (which take into account early deaths), was 1.44% for patients who had had a simultaneous bilateral procedure, 0.87% for patients who had had a unilateral procedure, and 0.54% for patients who had had the first of two staged procedures. The difference among the three groups was highly significant (log-rank p < 0.0001). In the group of patients who had undergone staged procedures, we found no clinically meaningful differences in the rate of pulmonary embolism after the first procedure according to how far apart the procedures had been staged.
The occurrence of a pulmonary embolism after a single procedure could well mean that a second procedure had been planned but had not taken place because of the complication, so that the patient appeared to have had a unilateral procedure in our dataset. When we grouped all 114,061 patients who had had a unilateral or staged procedure, we found, on the basis of life-table estimates, that the probability of a pulmonary embolism developing in the first three months after the first (or only) procedure was 0.81%. This was a little more than half of the 1.44% three-month probability of pulmonary embolism following the simultaneous procedures, a difference that was again highly significant (log-rank p < 0.0001). In the proportional hazards model, which took into account the demographic, geographic, comorbidity, and hospital and surgeon-volume variables listed in Table III that might affect the rate of pulmonary embolism within the first three months after surgery, the hazard ratio for the patients who had undergone simultaneous bilateral procedures compared with those who had had a single procedure was 1.81 (95% confidence interval, 1.49, 2.20). This finding suggests that the overall risk following two-staged procedures is comparable with that following simultaneous procedures.
In this analysis of Medicare patients who had undergone total knee replacement in 1999, 2000, or 2001, we found that high-volume hospitals and surgeons were much more likely to carry out simultaneous total knee replacements than were their low-volume counterparts. An east-west geographic pattern could also be seen, with hospitals in northeastern states most likely to perform simultaneous procedures, even when hospital and surgeon volume was taken into account. Simultaneous procedures were more likely to be carried out on men than on women, a finding that agrees with the earlier observations of Ritter et al.13.
Our estimates of the probability of a symptomatic pulmonary embolism developing within three months after staged surgery should be interpreted with caution. All patients who appear in our dataset as having had staged bilateral total knee replacement had to have survived the initial procedure to undergo the second. Any patient who died after the first planned procedure was listed in our dataset as having had a unilateral procedure and then died. Also, if a patient for whom a staged procedure had been planned had a pulmonary embolism, myocardial infarct, or other serious consequence after the first total knee replacement, the second procedure might well have been cancelled. Again, the adverse event would appear in our claims data to have happened to a patient treated with unilateral total knee replacement rather than to a patient for whom a staged procedure had been planned. We suspect that this selection bias largely explains why the pulmonary embolism rate for the patients who had undergone the first of two staged total knee replacements was lower than that for the patients who had undergone unilateral total knee replacement. In contrast, all outcomes after simultaneous procedures were counted as occurring after simultaneous procedures. Given that we do not know the initial intentions of the patients, it is more appropriate to compare those who had simultaneous bilateral total knee replacement with all of those who had a single total knee replacement during the first hospital stay, whether or not a second procedure was ever done. Our three-month pulmonary embolism probability estimates of 1.44% for simultaneous procedures compared with 0.81% for all single total knee replacements are similar to those in other large studies that demonstrated a higher probability of perioperative pulmonary embolism or thromboembolism after simultaneous procedures than after single total knee replacements1,3,4,6. Assuming that the probability of pulmonary embolism developing after the second staged procedure is no less than the probability of it developing after the first, this suggests that the sum of the two risks associated with a staged procedure may equal or exceed the risk associated with a simultaneous total knee replacement, although the absolute risk is low in either case.
It could be argued that differences between the simultaneous-procedure group and the other groups with regard to demographic variables and hospital and surgeon volume could account for part of the difference between the risk of pulmonary embolism after a simultaneous procedure and that after a single total knee replacement. However, the risk of pulmonary embolism in the three months after total knee replacement varies hardly at all by sex, race, or hospital or surgeon volume and only modestly by age20,21. Furthermore, our proportional hazards model confirmed that the hazard ratio of about 1.8 did not change substantially when those variables were taken into account. It is possible, however, that there was some residual confounding, particularly in our measure of comorbidity, which was derived from the diagnoses at the time of discharge from the hospitalization for the first total knee replacement.
The increased operative time involved in a simultaneous bilateral procedure, the surgical intervention in both lower extremities (where the emboli typically originate), and perhaps a more prolonged period of relative immobility constitute possible explanations for the increased risk of pulmonary embolism after the simultaneous procedures compared with the risk after the single total knee replacements.
The nature of Medicare data imposed several important limitations on our study. Reliable information is available only for patients who are sixty-five years of age or older, and about 30% of all total knee replacements are performed on people younger than sixty-five22. Also, it is impossible to determine whether, in a given case, the simultaneous total knee replacements were both performed in a single session in the operating room with a single anesthesia session or whether they were sequential procedures done during the same hospitalization with two sessions of anesthesia. There may well be different clinical outcomes in these different scenarios23,24. It is likely that pulmonary embolism rates calculated from Medicare and similar databases are underestimated. In particular, fatal pulmonary embolisms occurring outside of a hospital and asymptomatic pulmonary embolisms are not represented in hospital claims. Underreporting should, however, affect all groups in much the same way, leaving the differences between groups more or less unaffected. Furthermore, there is no information in the Medicare database about the use of prophylaxis for pulmonary embolism.
There are also important advantages of using the Medicare database. The program covers virtually all Americans over the age of sixty-five, so the referral bias of single-center studies is eliminated. Outcomes can be traced even if a patient moves from state to state. The large numbers of total knee replacements performed permit precise estimates of outcome risks for the entire United States.
We found that approximately 22% (26,320 of 122,385) of all Medicare patients who have a total knee replacement have a bilateral knee replacement, with either both procedures done during the same hospitalization or on separate occasions within two years. Among patients treated with bilateral total knee replacement, those who have simultaneous procedures are more likely than those who have staged procedures to be male, have a higher income, reside in the northeastern part of the United States, and have their procedures performed in a high-volume center by a high-volume surgeon. The risk of pulmonary embolism is about 80% higher in patients who have had a simultaneous bilateral procedure than it is in those who have had the first procedure of a staged bilateral replacement or have had a unilateral replacement, although the absolute risk is low in either case. This suggests that the sum of the two risks of a staged procedure may equal or exceed the risk of simultaneous total knee replacement. On the other hand, staging the total knee replacement results in a lower initial probability of pulmonary embolism and then allows the choice of proceeding with the second total knee replacement or not, depending on the result of the first. This approach may appeal to some patients.
These data have important implications for decision-making regarding the treatment of patients with two arthritic knees. Patients and physicians should use data such as these to discuss the relative advantages and drawbacks of the different strategies for undertaking a bilateral total knee replacement and make choices consonant with their preferences.
These findings may also have policy implications. A program that shifts patients from low to high-volume centers might correspondingly increase the relative frequency of simultaneous bilateral procedures as compared with staged bilateral procedures, which could result in a somewhat lower risk of pulmonary embolism. ▪
In support of their research for or preparation of this manuscript, one or more of the authors received grants or outside funding from the National Institute of Arthritis and Musculoskeletal and Skin Diseases and the National Institutes of Health. None of the authors received payments or other benefits or a commitment or agreement to provide such benefits from a commercial entity. No commercial entity paid or directed, or agreed to pay or direct, any benefits to any research fund, foundation, educational institution, or other charitable or nonprofit organization with which the authors are affiliated or associated.
Investigation performed at Dartmouth Medical School, Lebanon, New Hampshire and Brigham and Women's Hospital, Boston, Massachusetts
1. , Horlocker TT, Schroeder DR, Berry DJ, Brown DL. Frequency of myocardial infarction, pulmonary embolism, deep venous thrombosis, and death following primary hip or knee arthroplasty. Anesthesiology. 2002;96: 1140-6. Erratum in: Anesthesiology. 2002;97:531.
2. . Preoperative risk factors associated with symptomatic pulmonary embolism after total knee arthroplasty. Orthopedics. 1999;22: 1147-9.
3. , Harty LD, Davis KE, Meding JB, Berend M. Simultaneous bilateral, staged bilateral, and unilateral total knee arthroplasty. A survival analysis. J Bone Joint Surg Am. 2003;85: 1532-7.
4. , Burgess IA, Zicat B. The prevalence of venous thromboembolism after hip and knee replacement surgery. Med J Aust. 2005;182: 154-9.
5. , Cross M, Tribe KL, Lapsley HM, Courtenay BG, Cross MJ, Brooks PM, Cass C, Coolican M, Neil M, Pinczewski L, Quain S, Robertson F, Ruff S, Walter W, Zicat B; Arthritis C.O.S.T. Study Project Group. Two knees or not two knees? Patient costs and outcomes following bilateral and unilateral total knee joint replacement surgery for OA. Osteoarthritis Cartilage. 2004;12: 400-8.
6. , Mallory TH, Fada RA, Hartman JF, Capps SG, Kefauver CA, Dodds K, Adams JB. Simultaneous bilateral total knee arthroplasties: who decides? Clin Orthop Relat Res. 2001;392: 319-29.
7. , Kim JS. Incidence and natural history of deep-vein thrombosis after total knee arthroplasty. A prospective, randomised study. J Bone Joint Surg Br. 2002;84: 566-70.
8. , Sculco TP, Ranawat CS, Behr C, Tarrentino S. One-stage versus 2-stage bilateral total knee arthroplasty. Clin Orthop Relat Res. 1994;309: 94-101.
9. , Hozack WJ, Shah S, Rothman RH, Booth RE Jr, Eng K, Smith P. Simultaneous bilateral versus unilateral total knee arthroplasty. Outcomes analysis. Clin Orthop Relat Res. 1997;345: 106-12.
10. , Trousdale RT, Ilstrup DM. Complications after concomitant bilateral total knee arthroplasty in elderly patients. Mayo Clin Proc. 1997; 72: 799-805.
11. , Sporer SM, Shirreffs TG Jr. Comparison of simultaneous bilateral with unilateral total knee arthroplasty in terms of perioperative complications. J Bone Joint Surg Am. 2003;85: 1981-6.
12. , Jessup DE, Clelland C. Simultaneous bilateral total knee replacement versus unilateral replacement. Am J Orthop. 1996;25: 292-5.
13. , Mamlin LA, Melfi CA, Katz BP, Freund DA, Arthur DS. Outcome implications for the timing of bilateral total knee arthroplasties. Clin Orthop Relat Res. 1997;345: 99-105.
14. ICD-9-CM 2000: International classification of diseases 9th revision, clinical modification 5th edition: office edition, volumes 1, 2, and 3. Los Angeles: Practice Management Information Corporation (PMIC); 2000.
15. . CPT 2000: Current procedural terminology. Los Angeles: Practice Management Information Corporation; 2000.
16. , Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40: 373-83.
17. , Roos LL, Jollis JG. Adapting a clinical comorbidity index for use with ICD-9-CM administrative data: differing perspectives. J Clin Epidemiol. 1993;46: 1075-90.
18. . Statistical methods for survival data analysis. New York: Wiley-Interscience; 1992.
19. , Lemeshow S. Applied survival analysis: regression modeling of time to event data. New York: Wiley-Interscience; 1998.
20. , Barrett J, Katz JN, Baron JA, Wright J, Losina E. Epidemiology of total knee replacement in the United States Medicare population. J Bone Joint Surg Am. 2005;87: 1222-8.
21. , Barrett J, Mahomed NN, Baron JA, Wright RJ, Losina E. Association between hospital and surgeon procedure volume and the outcomes of total knee replacement. J Bone Joint Surg Am. 2004;86: 1909-16.
22. : H-CUPnet. hcup.ahrq.gov/HCUPnet.asp.2005.
23. , Callaghan JJ, Goetz DD, Taylor SG. Staggered bilateral total knee arthroplasty performed four to seven days apart during a single hospitalization. J Bone Joint Surg Am. 2005;87: 508-13.
24. , Hanssen AD. Bilateral total knee replacement using the same anesthetic is not justified by assessment of the risks. Clin Orthop Relat Res. 2004;428: 87-91.