The Journal of Arthroplasty 28 (2013) 732–735

Contents lists available at SciVerse ScienceDirect

The Journal of Arthroplasty

j ourna l homepage: www.ar throp lasty journa l .o rg

Immediate Complications Following Hip and Knee Arthroplasty: Does Race Matter?

Muyibat A. Adelani MD a, Kristin R. Archer DPT b, Yanna Song MS c, Ginger E. Holt MD b

a Department of Orthopaedic Surgery, Washington University School of Medicine, One Barnes-Jewish Hospital Plaza, St Louis, Missourib Vanderbilt Orthopaedic Institute; Medical Center East, South Tower, Nashville, Tennesseec Department of Biostatistics, Vanderbilt Medical Center, Nashville, Tennessee

The Conflict of Interest statement associated with thdx.doi.org/10.1016/j.arth.2012.09.015.

Reprint requests: Ginger E. Holt, MD, VanderbiltCenter East, South Tower, Suite 4200, Nashville, TN 372

0883-5403/2805-0005$36.00/0 – see front matter © 20http://dx.doi.org/10.1016/j.arth.2012.09.015

a b s t r a c t

a r t i c l e i n f o

Article history:Received 25 July 2012Accepted 22 September 2012

Keywords:racedisparitycomplicationsarthroplasty

Black race has been associated with a higher rate of complications following total joint arthroplasty,such as infection, deep vein thrombosis, pulmonary embolism, and death. We hypothesized that therewould be no significant association between black race and adverse outcome when medical conditionswere adjusted for. Data on 585,269 patients from the Nationwide Inpatient Samples were assessed bymultivariable logistic regression analysis. Black race was significantly associated with postoperativecomplication and death. Comorbidities do not account for racial differences in adverse events. Blackrace was an independent predictive factor for increased complications and death following hip andknee arthroplasty.

is article can be found at http://

Orthopaedic Institute, Medical32-8774.

13 Elsevier Inc. All rights reserved.

© 2013 Elsevier Inc. All rights reserved.

Black race has been associated with complications following totaljoint arthroplasty, including increased rates of intraoperative vascularinjury, infection, deep vein thrombosis, pulmonary embolism, anddeath [1–6]. Disproportionately poor outcomes experienced byminority patients undergoing total joint arthroplasty are of particularconcern, as racial and ethnic disparities in healthcare, includingorthopaedics, have become increasingly evident [7,8]. Prior literaturehas failed to adjust for co-morbidities and this consistent limitationwas the motivation for this article [1–8].

Potential reasons why Black race is associated with a higher rate ofcomplications following total joint arthroplasty have yet to beexplored. Race has complex relationships with other variables, suchas socioeconomic status and geographic segregation, making existingdata on disparities difficult to interpret [9–11]. The impact of racemayeven be confounded by the predominance of certain medicalcomorbidities amongst minorities, which has been speculated inother studies evaluating disparities in sepsis and breast cancersurvival [12–15]. In the arthroplasty literature, complications includ-ing wound complication, infection, and mortality, have been associ-ated with the presence of hypertension, diabetes mellitus, and obesity[5,16–20]. The disproportionate burden of these medical conditionsamong minorities, particularly African Americans, has long beendemonstrated in the literature; yet previous studies have notinvestigated the relationship between specific comorbidities, raceand complications after arthroplasty [21–25]. The purpose of the

current study is to evaluate the influence of selected comorbidities onthe relationship between black race and total joint arthroplastyoutcomes. Our hypothesis was that there would be no significantassociation between black race and adverse outcomewith adjustmentfor the presence selected medical conditions.

Methods

The Nationwide Inpatient Samples (NIS) database, an administra-tive database, was used for this study [26]. The NIS is part of theAgency for Healthcare Research and Quality's (AHRQ) Healthcare Costand Utilization Project (HCUP). It is the largest all-payer inpatient caredatabase that is publicly available in the United States, containing fiveto eight million hospitalization records per year. The database isdesigned to approximate a 20% stratified sample of all U.S. hospitalswith maximal representation based upon geographic region (North-east, Midwest, West, or South), location (urban or rural), teachingstatus (teaching or non-teaching), ownership (public, private non-profit, or private investor-owned), and size (small, medium, or large,based upon the number of beds). The data consist of dischargeabstracts from hospitals in participating states. Each dischargeabstract contains patient demographic information, including age,gender, race, insurance type, and median household income for thepatient's residential ZIP code. International Classification of Diseases,Ninth Edition (ICD-9) codes for up to fifteen inpatient procedures andfifteen diagnoses are reported for each hospitalization. The dischargedisposition of each patient, including in-hospital death, if applicable, isalso included. Reported hospital information includes a uniquehospital identification code as well as geographic region, urban orrural designation, teaching status, ownership, and size.

Table 1Summary of Patient Population.

White Black

N=544,209 N=41,060 P Valuea

Age (STD) 67.60 (11.22) 62.35 (12.23) b0.0001Female (%)b 330,833 (60.79) 29,022 (70.69) b0.0001ProcedureTotal Knee Arthroplasty (%) 351,053 (64.51) 27,646 (67.33) b0.0001Total Hip Arthroplasty (%) 193,156 (35.49) 13,414 (32.67) b0.0001ComorbiditiesHypertension (%) 287,566 (52.84) 26,258 (63.95) b0.0001Diabetes (%) 67,556 (12.41) 8476 (20.64) b0.0001Obesity (%) 47,728 (8.77) 5883 (14.33) b0.0001Medicaid (%) 18,841 (3.41) 6468 (15.78) b0.0001Urban (%)b 474,185 (87.15) 37,957 (92.54) b0.0001Teaching (%)b 216,824 (39.85) 22,887 (55.80) b0.0001Hospital cases/year (STD) 422 (437) 389 (408) b0.0001

a P values of b0.05 are considered significant.b Gender was missing in 0.01% of records. Urban/rural location and teaching status

were eachmissing in 0.03% of records. Insurance status was missing in 0.16% of records

Table 2Complication Rates Following Total Hip and Knee Arthroplasty.

ComplicationWhiteN (%)

BlackN (%) P Valuea

Any complication 15,060 (2.77) 1767 (4.30) b0.0001Infection 1089 (0.20) 100 (0.24) 0.06Wound complication 114 (0.02) 9 (0.02) 0.90Deep Vein Thrombosis 2938 (0.54) 278 (0.68) 0.0003Pulmonary Embolism 1856 (0.34) 152 (0.37) 0.33Myocardial Infarction 65 (0.01) 2 (0.005) 0.33Stroke 682 (0.13) 48 (0.12) 0.64Other 8831 (1.51) 1227 (2.99) b0.0001

Death 1029 (0.19) 92 (0.22) 0.12

a P values of b0.05 are considered significant.

733M.A. Adelani et al. / The Journal of Arthroplasty 28 (2013) 732–735

For this study, we utilized available data from 1998 through 2005.Non-Hispanic White and Black patients aged 18 to 90 years who had aprimary total hip or knee arthroplasty were identified by racialidentification and ICD-9 procedure codes for total hip arthroplasty(81.51) or total knee arthroplasty (81.54). Patients with multiplearthroplasty procedures in a single hospitalization were excluded.Also excluded were those patients with a diagnosis of bone infection,malignancy, pathologic fracture, or traumatic fracture. In order toensure adequate diversity in our patient population and to avoid anypotential bias, we excluded data from states where one percent or lessof eligible patients were identified as Black.

The primary variable of interest was race (White or Black). Otherpatient-related variables used in this analysis include age, gender,insurance type (Medicaid or insurance other than Medicaid), and thepresence of hypertension, diabetes mellitus, and obesity (based uponICD-9 diagnosis codes). Insurance type, specifically utilization ofMedicaid insurance, was used to approximate socioeconomic status.

Hospital-related variables used in this analysis were location(urban or rural), teaching status, and the average annual arthroplastyvolume. Annual volume was defined as the combined number ofprimary total hip and primary total knee arthroplasties performed in agiven institution per year, determined from the number of procedures(based upon ICD-9 procedure codes) for each institution with aunique hospital identification code in the database.

The outcomes of interest in this study were in-hospital postoper-ative complications and mortality. Complications were obtained fromsecondary diagnoses recorded in the NIS by ICD-9 diagnosis codes.These included postoperative infection, nonhealing or disruption ofsurgical wound, deep vein thrombosis, pulmonary embolism, myo-cardial infarction, and stroke.

Patient- and hospital-related data were collected and separated byracial group. Categorical variables were summarized with frequencyand compared between groups using a Chi-square test or Fisher'sexact test. Continuous variables were summarized with mean andstandard deviation; results were compared between races using a t-test or Wilcoxon rank sum test. The association of black race withpostoperative complication and death was assessed with a hierarchi-cal multiple variable logistic model with a Huber–White clustersandwich covariance estimator. Association with any nonfatalpostoperative complication was determined as an aggregate of allcomplications, while association with mortality was assessed sepa-rately. P values and 95% confidence intervals are reported with theodds ratios. A P value of less than 0.05 was considered statisticalsignificant. All statistical analyses were conducted with SAS, Version9.1.3 (SAS Institute, Inc., Cary, NC) and R Version 2.5.1.

This study was approved by our institution's InstitutionalReview Board.

Source of Funding

This study was funded by a grant from Nth DimensionsEducational Solutions, Inc. The funding source did not play any rolein our investigation.

Results

The patient population consisted of 585,269 patients—206,570patients (35%) who had undergone total hip arthroplasties and378,699 (65%)who had total knee arthroplasties. The average age was67 years. Women comprised 61% of the patient population. Ninety-three percent of patients were white and seven percent were Black.Hypertension was the most common medical comorbidity, affecting54% of the total patient population. The majority of patients (88%)were treated in an urban hospital; approximately 41% were treated inteaching hospitals. See Table 1 for a complete summary of the data.

.

Black patients were more likely to have both hypertension anddiabetes than whites (Pb0.0001). Obesity was nearly twice asprevalent among Black patients compared to whites (Pb0.0001).Black patients were more likely to have Medicaid insurance coveragethan whites, more likely to be treated in a teaching hospital, and morelikely to be treated in hospitals with significantly lower annual casevolumes (Pb0.0001 each).Overall, 1% of patients had postoperativecomplications and 0.2% died [Table 2]. The most common complica-tion was deep vein thrombosis (54% of all complications).

Multivariable logistic regression analysis demonstrated a signif-icant association between Black race and complication afterarthroplasty [OR 1.20, 95% CI 1.07–1.35]. Obesity [OR 1.23, 95% CI1.11–1.37] and treatment in a teaching hospital [OR 1.45, 95% CI1.23–1.71] were also associated with postoperative complication.Female patients were slightly less likely to have complications [OR0.94, 95% CI 0.88–1.00], as were those with hypertension [OR 0.91,95% CI 0.86–0.96].

Black race also had a significant association with death [OR 1.65,95% CI 1.33–2.05]. Medicaid insurance [OR 1.97, 95% CI 1.49–2.59],diabetes [OR 1.37, 95% CI 1.16–1.62] and treatment in a teachinghospital [OR 1.17, 95% CI 1.02–1.35], and treatment in an urbanhospital [OR 1.38, 95% CI 1.13–1.68] were also associated withpostoperative mortality. Female gender [OR 0.56, 95% CI 0.50–0.63]and hypertension [OR 0.45, 95% CI 0.39–0.50] were both negativelyassociated with death after joint arthroplasty. See Table 3 forcomplete results.

Discussion

This study showed that Black race was an independent predictivefactor for increased complications following hip and knee arthro-

Table 3Racial Disparities in Post-Operative Morbidity and Mortality.

Variable Reference Group

Mortality RiskOdds Ratio

(95% Confidence Interval) P Valuea

Aggregate Complication RiskOdds Ratio

(95% Confidence Interval) P Valuea

Black race White race 1.70 (1.37–2.10) b0.0001 1.54 (1.42–1.68) b0.0001Female gender Male gender 0.56 (0.50–0.63) b0.0001 0.82 (0.79–0.85) b0.0001Ageb b0.0001 b0.0001Hypertension Patients without Hypertension 0.45 (0.40–0.51) b0.0001 0.94 (0.91–0.97) 0.0004Diabetes Patients without Diabetes 1.41 (1.20–1.65) b0.0001 0.98 (0.94–1.03) 0.4421Obesity Nonobese patients 0.82 (0.61–1.10) 0.1886 1.12 (1.04–1.19) 0.0017Medicaid insurance Insurance other than Medicaid 1.86 (1.44–2.41) b0.0001 1.01 (0.93–1.10) 0.8226Urban location Rural location 1.34 (1.10–1.63) 0.0038 1.25 (1.11–1.42) 0.0004Teaching hospital Non-teaching hospital 1.14 (0.99–1.31) 0.0612 1.19 (1.06–1.34) 0.0040Hospital Volumeb b0.0001 b0.0001

a P values of b0.05 are considered significant.b Age and hospital volume were continuous variables, and thus, had no reference groups.

734 M.A. Adelani et al. / The Journal of Arthroplasty 28 (2013) 732–735

plasty. We hypothesized that black race was a proxy for our selectedcomorbidities and that the effect of race on total joint arthroplastyoutcomes would be insignificant after accounting for these comorbid-ities. On the contrary, we found that when hypertension, diabetes, andobesity were accounted for, Black race remained associated with bothpostoperative complications and mortality.

We demonstrated that black patients are at a significantlyincreased risk for postoperative complication and death followinghip or knee arthroplasty which is similar to prior studies [2–5,27].SooHoo et al demonstrated an association between black race andpulmonary embolism after total knee arthroplasty as well as anincreased risk for infection and thromboembolic disease after totalhip arthroplasty [2,5]. Mahomed et al showed higher rates of woundinfection and mortality among black patients following primarytotal knee arthroplasty [3]. A study by Ibrahim et al showed thatblack patients had higher rates of both infection-related and non-infection-related complications following total knee arthroplasty[4]. The association between black race and venous thromboembo-lism following total joint arthroplasty has been previously de-scribed [2–5,27].

Genetic factors may contribute to the finding of Black race as anindependent predictive factor for increased complications followinghip and knee arthroplasty. Consistent with previous literature, themost common complication in this study was deep vein thrombosisoccurring more often in black patients (P=0.0003). [2–5] Whencompared with Whites, Black-Americans may have a 40% higherincidence venous thromboembolism (VTE) incidence and whencompared with Whites, Blacks had a significantly higher proportionwith pulmonary embolism (PE), Blacks had a significantly higherfamily history of VTE, and documented thrombophilia attributed toreduced factor V Leiden and prothrombin G20210A prevalence.[28–30].

Limitations in this study, as in other similar studies, relate to theuse of an administrative database for our patient population. Codingin these data sets is known to be both incomplete and inaccurate, asseveral studies have demonstrated that administrative databases areoften inconsistent with physician records [31–33]. Such deficienciesin data can obviously have an impact on the results of any study. Thedatabase used in this study, in particular, adds the additionallimitation of lack of follow-up post hospital discharge. This leads toa small sampling of complications such as VTE, a complication thatholds its highest incidence post discharge. The NIS only includesevents up to hospital discharge; however, current literature statesthat complications typically occur up to six weeks postoperatively[34,35]. We attempted to address this by assessing immediatecomplications and not long term complications such as implantwear or failure. Another limitation is that data for educational leveland income were not available through this database and are,therefore, not reported.

Recognizing that Black race is an independent risk factor forimmediate complications and death following hip and knee arthro-plasty is the first step in elucidating causes and implementingsolutions. Potential practice improvements include preoperativelyassessing the risk of thromboembolism in this population at risk topersonalize treatment, or including Black race as a high riskthromboembolic category requiring more aggressive anticoagulation.This is an immediate and achievable goal. It is not the intent of thispaper to address racial disparities as a whole. Access to health care,the quality of health care, and health literacy are all beyond the scopeof this paper. It is, however a first step in unraveling the causes for theoverwhelming evidence that racial disparities contribute to immedi-ate complications associated with joint arthroplasties.

References

1. Nwachukwu BU, Kenny AD, Losina E, et al. Complications for racial and ethnicminority groups after total hip and knee replacement: a review of the literature.J Bone Joint Surg Am 2010;92:338.

2. SooHoo NF, Lieberman JR, Ko CY, et al. Factors predicting complications ratesfollowing total knee replacement. J Bone Joint Surg Am 2006;88:480.

3. Mahomed NN, Barrett J, Katz JN, et al. Epidemiology of total knee replacement inthe United States medicare population. J Bone Joint Surg Am 2005;87:1222.

4. Ibrahim SA, Stone RA, Han X, et al. Racial/ethnic differences in surgical outcomes inveterans following knee or hip arthroplasty. Arthritis Rheum 2005;52:3143.

5. SooHoo NF, Farng E, Lieberman JR, et al. Factors that predict short-termcomplication rates after total hip arthroplasty. Clin Orthop Relat Res 2010;468:2363.

6. Abularrage CJ, Weiswasser JM, Dezee KJ, et al. Predictors of lower extremity arterialinjury after total knee or total hip arthroplasty. J Vasc Surg 2008;43:803.

7. Nelson CL. Disparities in orthopaedic surgical intervention. J Am Acad Orthop Surg2007;15(Suppl 1):S13.

8. Pierce RO. Ethnic and racial disparities in diagnosis, treatment, and follow-up care.J Am Acad Orthop Surg 2007(Suppl 1):S8.

9. LaVeist TA. Disentangling race and socioeconomic status: a key to understandinghealth inequalities. J Urban Health 2005;82(2 Suppl):iii26.

10. Gibbons MC, Brock M, Alberg AJ, et al. The Sociobiologic Integrated Model (SBIM):enhancing the integration of sociobehavioral, environmental, and biomolecularknowledge in urban health and disparities research. J Urban Health 2007;84:198.

11. Skinner J, Weinstein JN, Sporer SM, et al. Racial, ethnic and geographic disparities inrates of knee arthroplasty among medicare patients. N Engl J Med 2003;349:1350.

12. Esper AM, Moss M, Lewis CA, et al. The role of infection and comorbidity: factorsthat influence disparities in sepsis. Crit Care Med 2006;34:2576.

13. Dombrovskiy VY, Martin AA, Sunderram J, et al. Occurrence and outcomes of sepsis:influence of race. Crit Care Med 2007;35:763.

14. Curtis E, Quale C, Haggstrom D, et al. Racial and ethnic differences in breast cancersurvival: how much is explained by screening, tumor severity, biology, treatment,comorbidities, and demographics? Cancer 2008;112:171.

15. Tammemagi CM, Nerenz D, Neslund-Dudas C, et al. Comorbidity and survivaldisparities among black and white patients with breast cancer. JAMA 2005;294:1765.

16. Jain NB, Guller U, Pietrobon R, et al. Comorbidities increase complication rates inpatients having arthroplasty. Clin Orthop Related Res 2005;435:232.

17. Namba RS, Paxton L, Fithian DC, et al. Obesity and perioperative morbidity in totalhip and total knee arthroplasty patients. J Arthroplasty 2005;20(7 Suppl):46.

18. Yang K, Yeo SJ, Lee BP, et al. Total knee arthroplasty in diabetic patients: a study of109 consecutive cases. J Arthroplasty 2001;16:102.

19. Lai K, Bohm ER, Burnell C, et al. Presence of medical comorbidities in patients withinfected primary hip or knee arthroplasties. J Arthroplasty 2007;22:651.

735M.A. Adelani et al. / The Journal of Arthroplasty 28 (2013) 732–735

20. Winiarsky R, Barth P, Lotke P. Total knee arthroplasty in morbidly obese patients. JBone Joint Surg Am 1998;80:1770.

21. Hajjar I, Kotchen TA. Trends in prevalence, awareness, treatment, and control ofhypertension in the United States, 1988–2000. JAMA 2003;290:199.

22. McBean AM, Gilbertson DT, Li S, et al. Differences in diabetes prevalence, incidence,and mortality among the elderly of four racial/ethnic groups: whites, blacks,Hispanics, and Asians. Diabetes Care 2004;27:2317.

23. Ogden CL, Carroll MD, Curtin LR, et al. Prevalence of overweight and obesity in theUnited States, 1999–2004. JAMA 2006;295:1549.

24. Hertz RP, Unger AN, Cornell JA, et al. Racial disparities in hypertension prevalence,awareness, and management. Arch Intern Med 2005;165:2098.

25. Seo DC, Torabi MR. Racial/ethnic differences in body mass index, morbidity andattitudes toward obesity among U.S. adults. J Natl Med Assoc 2006;98:1300.

26. HCUP Nationwide Inpatient Sample (NIS). Healthcare Cost and Utilization Project(HCUP). 1998–2005. Agency for Healthcare Research and Quality, Rockville, MD.www.hcup-us.ahrq.gov/nisoverview.jsp.

27. Mahomed NN, Barrett JA, Katz JN, et al. Rates and outcomes of primary and revisiontotal hip replacement in the United States Medicare population. J Bone Joint SurgAm 2003;85:27.

28. Heit JA, Beckman MG, Bockenstedt PL, et al. Comparison of characteristics fromWhite- and Black-Americans with venous thromboembolism: a cross-sectionalstudy. Am J Hematol 2010;85:467.

29. DeitelzweigSB Lin J, Johnson BH, Schulman KL. Venous thromboembolism in theUS: does race matter? J Thromb Thrombolysis 2011;31:133.

30. White RH, Keenan CR. Effects of race and ethnicity on the incidence of venousthromboembolism. Thromb Res 2009;123(Suppl 4):S11.

31. Mont MA, Mears SC, Jones LC, et al. Is coding of diagnoses, comorbidities, andcomplications in total knee arthroplasty accurate? J Arthroplasty 2002;17:767.

32. Mears SC, Bawa M, Pietryak P, et al. Coding of diagnoses, comorbidities,and complications of total hip arthroplasty. Clin Orthop Related Res 2002;402:164.

33. Holt G, Feurer ID, Schwartz HS. An analysis of concordance among hospitaldatabases and physician records. Ann Surg Oncol 1998;5:553.

34. Parvizi J, Mui A, Purtill JJ, et al. Total joint arthroplasty: when do fatal or near-fatalcomplications occur? J Bone Joint Surg 2007;89A(1):27.

35. Phillips CB, Barrett JA, Losina E, et al. Incidence rates of dislocation, pulmonaryembolism, and deep infection during the first six months after elective total hipreplacement. J Bone Joint Surg 2003;85A(1):20.