For the purpose of building a parsimonious model restricted to strong predictive covariates only, the following steps were taken17: first, clinical judgment and significance at a P value of 15% level in the univariate analyses (as in Tables 1 and 2) were used to identify variables for the process of multivariable modeling. Second, additional variable selection and internal validation of the predictive performance of the model was achieved through a 2-step, nonparametric, bootstrapping process.18 In the bootstrap procedure, the original set of data of size n becomes a parent population from which samples of size n are randomly drawn with replacement. In the first step of internal validation, the bootstrapping technique was used for variable selection. One hundred bootstrap samples were created, and a stepwise procedure was applied to each sample using a forward selection method (with a selection entry level = 0.20). From this analysis, we calculated the percentage of samples for which each variable was included in the model from the 100 samples. Percent inclusion (80% cutoff decided a priori) was used to determine the prognostic importance of a variable, because it was expected that a prognostically important variable would be included in the model for a majority of the bootstrap samples. For variables that were not included, if the frequency of pair-wise combinations included in the model was >90%, then we included the one with the largest frequency in the final model. In the second internal validation step, optimism-corrected area under the receiver operating characteristic curve (also referred to as the C-statistic) was computed using the %BVAL macro from SAS.19 Models with a C-statistic >0.7 are deemed to have good discriminatory/predictive power. Third, the model finalized at the second stage was processed one more time using the SURVEYLOGISTIC procedure instead of the LOGISTIC procedure to obtain appropriate estimates of the variance for the weighted survey data. This step was necessary, because the SURVEYLOGISTIC does not allow for forward selection procedure. This methodology has previously been described by Hosmer and Lemeshow20 as appropriate.
We identified 42,003 entries fitting selection criteria, representing an estimated 206,573 elective BTKA procedures performed between 1998 and 2007 in the United States. Of those, 9.5% developed major complications or mortality during their hospitalization. Patients of older age, male gender, and white race were disproportionately more affected by major morbidity and mortality (Table 1). Cardiac complications made up the majority of adverse occurrences postoperatively (Fig. 1). Patients who had a major complication or mortality had a higher prevalence of comorbidities, including chronic lung diseases, congestive heart failure, and pulmonary hypertension (Table 2).
Age, sex, race, chronic lung disease, congestive heart failure, coagulopathy, neurologic disorders, peripheral vascular disease, renal disease, cardiac valvular disorders, electrolyte/fluid abnormalities, pulmonary hypertension, and obesity were selected for multivariate model building. Race, peripheral vascular disease, and obesity were excluded from the model because they did not meet the cutoff for inclusion. In the final model, increasing age emerged as an independent risk factor for major morbidity and mortality. Compared with patients in the age group between 45 and 64 years, those younger than 45 years were half as likely to have a major complication or mortality (odds ratio [OR]: 0.49 [confidence interval, CI: 0.30, 0.81]). Patients aged 65 to 74 and >75 years (OR: 1.81 [CI: 1.67, 2.30] and OR: 2.52 [CI: 2.30, 2.77], respectively) were more likely to have a major adverse outcome compared with those aged 45 to 65 years. Male gender was associated with increased odds for adverse outcome (OR: 1.50 [CI: 1.40, 1.61]). A number of comorbidities were identified as independent risk factors for major complications and mortality (Fig. 2). The presence of congestive heart failure and pulmonary hypertension were the most significant comorbidities associated with increased odds for adverse outcome. The optimism-corrected C-statistic from the final model was 0.7.
We were able to characterize the incidence of major complications and mortality in patients undergoing BTKA and identify risk factors. Independent risk factors included advanced age, male gender, and a number of comorbidities, with congestive heart failure and pulmonary hypertension carrying the highest odds for major adverse outcome. Our data can be used to help guide the selection of individuals who are considered candidates for this procedure.
Understanding the pathophysiology leading to increased morbidity and mortality in patients undergoing joint arthroplasty and especially bilateral procedures is important when interpreting our findings. Previous research suggests that complications after joint arthroplasty can be related to overall intraoperative embolic debris and cement load that gains access to the vascular system during surgery, on one hand, and the end-organ reserve, on the other.7,8,15,21 Embolic material entering the lungs causes lung injury and increases pulmonary vascular resistance,7,8 which in turn may lead to right ventricular and atrial strain promoting arrhythmias, hypotension, and venostasis.22 These events may be responsible for the increased rates of acute respiratory distress syndrome and thromboembolic events occurring in patients undergoing bilateral versus unilateral TKA.2 Similarly, increased exposure of the central nervous system and other organs, such as the kidney, to embolic material may explain the higher rates of delirium and renal complications in BTKA versus UTKA patients.2,3
In this study, we identified advanced age to be a risk factor for increased morbidity and mortality. This finding is consistent with previous research3,21 and could be explained by the known phenomenon of age-related physiologic decline in end-organ reserve.23 Thus, it is possible that although embolic load associated with surgery remains unchanged in the elderly, the capacity of organs to withstand the insult is decreased, resulting in worse outcome.
Male gender was found to be associated with increased odds of major morbidity and mortality. Although previously described in the arthroplasty population,21 the reasons for this finding have to remain speculative at this point, but potential reasons may include hormonal differences that may offer some degree of protection for female patients.24
A number of comorbidities independently increased the risk for major morbidity and mortality; congestive heart failure and pulmonary hypertension were associated with the highest odds. Preexisting decreased vascular reserve of the lungs is a likely factor associated with this finding. Indeed, the load-dependent response and capacity of the lungs to absorb the embolic insult was shown in healthy patients undergoing bilateral hip arthroplasty.8 Whereas no significant changes in pulmonary vascular resistance were seen after the first joint implantation, an increase in pulmonary vascular parameters was measured after the second implantation, suggesting that the ability of the pulmonary vascular bed to compensate may be overwhelmed by the larger embolic load of 2 joints. Interestingly, these derangements continued to be present on postoperative day 1, suggesting that the stresses are prolonged and not short-lived, as often assumed. Whereas these pulmonary hemodynamic changes may be of limited clinical consequence in otherwise healthy individuals,8 significantly increased rates of morbidity and mortality among patients with pulmonary hypertension undergoing even unilateral hip and knee replacement have been found.15 It is therefore not surprising that pulmonary hypertension was 1 of the 2 most significant risk factors for morbidity and mortality in this study.
Given these findings, it seems prudent to screen patients who are suspected of having increased pulmonary pressure or right heart dysfunction, including patients with sleep apnea25 and those with a history of pulmonary embolism,26 and consider them at high risk. Attempting the estimation of pulmonary pressures by echocardiography in patients at risk for pulmonary hypertension during preoperative testing may therefore be of benefit. Although no data are available to judge what level of pulmonary hypertension should be considered significant in this setting, taking a conservative approach in judgment when assessing suitability for BTKA may be advisable until more research is available. If preoperative pharmacologic treatment of abnormal parameters is of benefit has to remain speculative at this point and would warrant detailed investigations.
Equally important risk factors identified included a number of comorbidities suggesting decreased end-organ reserve, i.e., renal disease, neurologic disease, congestive heart failure, and chronic pulmonary disease. It is important to note that these comorbidities are not uniquely associated with adverse outcomes among BTKA patients,21 because many surgical procedures are defined by significant metabolic injury, fluid shifts, and other insults exposing various organ systems to a number of stresses. However, when considering the likely pathophysiology (i.e., intraoperative debris embolization) of morbidity and mortality in the BTKA population, physicians should be cautioned against worsening of organ function in this particular setting. Although many clinicians are well aware of the impact of cardiac and pulmonary disease on the outcome in surgical patients, our results serve to alert about the negative impact of diseases with low prevalence on perioperative morbidity and mortality. This, in turn, underlines the advantages of large database research, which allows for the study of low-incidence scenarios.
Our study is limited by a number of factors inherent to secondary data analysis of large administrative databases. As such, clinical information (i.e., type of anesthesia, amount of blood loss, length of surgery, etc.) available in the NIS is limited, and our analysis must be interpreted in this context. Because of the nature of the NIS, only inpatient data are available and thus complications and events after discharge are not captured. Furthermore, the need for readmission cannot be accounted for in this database. Thus, conclusions should be limited to the acute perioperative setting with the notion that mortality and complications are likely underestimated.
In this context, we are also unable to compare the outcomes of patients who have 2 TKAs performed during different hospitalizations. Whereas we have previously shown that staging procedures a few days apart during the same hospitalization offers no benefit in the risk for mortality and even may increase the risk for perioperative complications,3 Ritter et al.27 suggest that 30-day mortality rates of the BTKA performed 3 to 12 months apart is between 0.29% and 0.36%, compared with significantly increased rates for simultaneously performed BTKA (0.99%) and those scheduled 6 weeks apart (0.48%).
It must also be mentioned that the identification of comorbidities in this study was based on the validated method of Elixhauser et al., which is based on the bundling of ICD-9 codes to define various comorbidities in administrative databases.14 However, it is often not possible to determine from ICD-9 codes whether a comorbidity is preexisting or acquired during the hospitalization. It should be kept in mind, however, that this does not diminish the value of the comorbidities identified as risk factors as studied in this analysis, because they should alert clinicians of increased risk if they are encountered before surgery.
Furthermore, our modeling approaches are also somewhat limited by software availability and a gap in research in the arena of application of stepwise procedure in logistic regression analyses of survey data that take the survey design (stratification and clustering) and strata weights into account. We followed the latest recommendations published by Hosmer and Lemeshow20 on how to deal with this situation, but it is clear that further simulation-based research is needed for delineating the scenarios in which there might be deviation in the models fitted through “design-based” and “model-based” approaches. Software development incorporating procedures for fitting logistic regression in a stepwise manner for complex survey data is also necessary for easy implementation.
An additional limiting factor is the bias associated with the retrospective nature of our study. Nevertheless, because of the availability of data from a large, nationally representative sample, this type of analysis may provide a more accurate estimate of events surrounding BTKA than various prospective studies that are limited in sample size and thus lack the ability to capture low-incidence outcomes.
In conclusion, we were able to identify a number of risk factors for major morbidity and mortality in patients undergoing BTKA. These data can be used to aid in the selection of patients for this procedure, which otherwise may be associated with increased morbidity and mortality compared with a unilateral approach.
Although it is beyond the scope of this article to provide final and specific guidelines, we would urge institutions to engage in discussions to establish criteria to restrict BTKA procedures to patients with decreased reserve of the cardiopulmonary, vascular, renal, and central nervous system and contemplate exclusion of patients of advanced age and those with evidence of significant end-organ disease. Until detailed evaluation guidelines can be agreed on, it seems prudent to exclude the elderly and patients with ASA physical status ≥3. Patients at risk for occult derangements of pulmonary hemodynamics and right heart dysfunction (i.e., the obese and those with sleep apnea, chronic obstructive pulmonary disease, and previous pulmonary embolism) should undergo cardiopulmonary evaluation with echocardiography to rule out significant preexisting increases in pulmonary artery pressures, which may predispose patients to increased morbidity and mortality.
Given the controversy surrounding this issue and the fact that a number of studies have been published in recent years on this subject, it may be time for the establishment of national guidelines to aid physicians and patients with the decision of whether to proceed with BTKA.
Name: Stavros G. Memtsoudis, MD, PhD.
Contribution: This author helped design the study, conduct the study, analyze the data, write the manuscript, and secure funding.
Attestation: Stavros G. Memtsoudis has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.
Name: Yan Ma, PhD.
Contribution: This author helped design the study, conduct the study, and analyze the data.
Attestation: Yan Ma has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Ya-Lin Chiu, MS.
Contribution: This author helped analyze the data and write the manuscript.
Attestation: Ya-Lin Chiu has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Lazaros Poultsides, MD, PhD.
Contribution: This author helped design the study and write the manuscript.
Attestation: Lazaros Poultsides has reviewed the analysis of the data, and approved the final manuscript.
Name: Alejandro Gonzalez Della Valle, MD.
Contribution: This author helped design the study and write the manuscript.
Attestation: Alejandro Gonzalez Della Valle has reviewed the analysis of the data, and approved the final manuscript.
Name: Madhu Mazumdar, PhD.
Contribution: This author helped design the study, analyze the data, write the manuscript, and secure funding.
Attestation: Madhu Mazumdar has seen the original study data, approved the final manuscript, and is the author responsible for archiving the study files.
This manuscript was handled by: Sorin J. Brull, MD.
The flowchart illustrates the sample selection process and the incidence of cases with and without major complications and mortality.
1. Restrepo C, Parvizi J, Dietrich T, Einhorn TA. Safety of simultaneous bilateral total knee arthroplasty: a meta-analysis. J Bone Joint Surg Am 2007;89:1220–6
2. Memtsoudis SG, González Della Valle A, Besculides MC, Gaber L, Sculco TP. In-hospital complications and mortality of unilateral, bilateral, and revision TKA: based on an estimate of 4,159,661 discharges. Clin Orthop Relat Res 2008;466:2617–27
3. Memtsoudis SG, Ma Y, González Della Valle A, Mazumdar M, Gaber-Baylis LK, MacKenzie CR, Sculco TP. Perioperative outcomes after unilateral and bilateral total knee arthroplasty. Anesthesiology 2009;111:1206–16
4. Hooper GJ, Hooper NM, Rothwell AG, Hobbs T. Bilateral total joint arthroplasty: the early results from the New Zealand National Joint Registry. J Arthroplasty 2009;24:1174–7
5. Urban MK, Chisholm M, Wukovits B. Are postoperative complications more common with single-stage bilateral (SBTKR) than with unilateral knee arthroplasty: guidelines for patients scheduled for SBTKR. HSS J 2006;2:78–82
6. Memtsoudis SG, Besculides MC, Reid S, Gaber-Baylis LK, González Della Valle A. Trends in bilateral total knee arthroplasties: 153,259 discharges between 1990 and 2004. Clin Orthop Relat Res 2009;467:1568–76
7. Memtsoudis SG, Starcher B, González Della Valle A, Ma Y, Jules-Elysee K, Sculco TP. Urine desmosine as a marker of lung injury following total knee arthroplasty: a pilot study. HSS J 2009;5:154–8
8. Memtsoudis SG, Salvati SA, Go G, Ma Y, Sharrock NE. Perioperative pulmonary circulatory changes during bilateral total hip arthroplasty under regional anesthesia. Reg Anesth Pain Med 2010;35:417–21
9. HCUP Databases. Healthcare Cost and Utilization Project (HCUP) Agency for Healthcare Research and Quality, Rockville, MD, 2008. Available at: www.hcup-us.ahrq.gov/nisoverview.jsp
. Accessed October 15, 2010
11. Publications from the Healthcare Cost and Utilization Project Databases. Agency for Healthcare Research and Quality, Healthcare Cost and Utilization Project (HCUP), Rockville, MD, 2009. Available at: http://www.ahrq.gov/data/hcup/hcupref.htm
. Accessed October 15, 2010
12. Iezzoni LI, Daley J, Heeren T, Foley SM, Fisher ES, Duncan C, Hughes JS, Coffman GA. Identifying complications of care using administrative data. Med Care 1994;32:700–15
14. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care 1998;36:8–27
15. Memtsoudis SG, Ma Y, Chiu YL, Walz JM, Voswinckel R, Mazumdar M. Perioperative mortality in patients with pulmonary hypertension undergoing major joint replacement. Anesth Analg 2010;111:1110–6
16. Kozak LJ. Underreporting of race in the National Hospital Discharge Survey. Adv Data 1995;265:1–12
17. Harrell FE. Regression Modeling Strategies: With Applications to Linear Models, Logistic Regression, and Survival Analysis. 1st ed. New York: Springer-Verlag, 2001:215–65
18. Sauerbrei W, Schimacher M. A bootstrap resampling procedure for model building: application to the Cox regression model. Stat Med 1992;11:2093–109
19. Gonen M. Analyzing Receiver Operating Characteristic Curves with SAS. 1st ed. Cary, NC: SAS Institute, 2007
20. Hosmer DW, Lemeshow S. Applied Logistic Regression. 2nd ed. New York: Wiley, 2000
21. Memtsoudis SG, González Della Valle A, Besculides MC, Esposito M, Koulouvaris P, Salvati EA. Risk factors for perioperative mortality after lower extremity arthroplasty: a population-based study of 6,901,324 patient discharges. J Arthroplasty 2010;25:19–26
22. Urban MK, Sheppard R, Gordon MA, Urquhart BL. Right ventricular function during revision total hip arthroplasty. Anesth Analg 1996;82:1225–9
23. Loran DB, Hyde BR, Zwischenberger JB. Perioperative management of special populations: the geriatric patient. Surg Clin North Am 2005;85:1259–66
24. Choi BG, McLaughlin MA. Why men's hearts break: effects of sex steroids. Endocrinol Metab Clin North Am 2007;36:365–77
25. Bradley TD. Right and left ventricular functional impairment and sleep apnea. Clin Chest Med 1992;13:459–79
26. Kline JA, Steuerwald MT, Marchick MR, Hernandez-Nino J, Rose GA. Prospective evaluation of right ventricular function and functional status 6 months after acute submassive pulmonary embolism: frequency of persistent or subsequent elevation in estimated pulmonary artery pressure. Chest 2009;136:1202–10
© 2011 International Anesthesia Research Society
27. Ritter M, 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