The Journal of Arthroplasty 30 (2015) 547–551

Contents lists available at ScienceDirect

The Journal of Arthroplasty journal homepage: www.arthroplastyjournal.org

Does Sickle Cell Disease Increase Risk of Adverse Outcomes Following Total Hip and Knee Arthroplasty? A Nationwide Database Study Dean C. Perfetti, BA a,b, Matthew R. Boylan, ScB a,b, Qais Naziri, MD a, Harpal S. Khanuja, MD c, William P. Urban, MD a a b c

Department of Orthopaedics, SUNY Downstate Medical Center, Brooklyn, New York Department of Epidemiology and Biostatistics, SUNY Downstate Medical Center, Brooklyn, New York Department of Orthopaedic Surgery, The Johns Hopkins University, Baltimore, Maryland

a r t i c l e

i n f o

Article history: Received 15 August 2014 Accepted 23 October 2014 Keywords: total hip arthroplasty total knee arthroplasty sickle cell disease complication length of stay cost

a b s t r a c t Sickle cell disease (SCD) is associated with impaired vascular function and progressive vaso-occlusive injury to bones. We used the Nationwide Inpatient Sample to identify all THA and TKA admissions between 1998 and 2010. After controlling for patient age, gender, insurance, race, and comorbidities, the risk of complication among admissions with SCD was 152% higher (P b 0.001) for THA and 137% higher (P = 0.001) for TKA. Patients with SCD had a length of stay that was 42% longer (P b 0.001) for THA and 20% longer for TKA (P b 0.001), and hospital charges that were 19% higher (P b 0.001) for THA and 16% higher (P = 0.001) for TKA. Orthopedic surgeons should counsel potential THA and TKA candidates with SCD of these risks prior to admission. © 2014 Elsevier Inc. All rights reserved.

Sickle cell disease (SCD), the most common congenital hemoglobinopathy in the United States, affects more than 90,000 Americans and 1 in every 500 African-Americans [1]. Sickle cell disease is an inheritable condition that is characterized by a single missense mutation on the beta globin gene and can present as homozygous hemoglobin S (HbSS) or compound heterozygous hemoglobin S (HbSC or HbSthalassemia) [2]. Irrespective of genotype, the expression of one HbS allele in compound heterozygotes makes individuals with SCD more prone to intracellular red blood cell polymerization in physiological states of low oxygen, which can ultimately result in vaso-occlusion, tissue ischemia and osteonecrosis [2,3]. Bone microcirculation is a common site for red blood cell sickling, and clinical presentation is characterized by severe localized pain with associated edema and erythema [4]. Infarction of the cancellous trabeculae of the femoral head is a frequent complication of SCD in adults, with as many as 50% of patients with SCD suffering from femoral head osteonecrosis by the age of 35 [5]. Sickle cell disease is also a risk factor for osteonecrosis of the knee, specifically at the femoral condyles and tibial plateau, with an incidence of 10% among patients with SCD [6]. Despite the use of joint-preserving surgical interventions including core decompression, osteotomy and bone grafting, many patients with SCD progress to advanced osteonecrosis requiring total joint arthroplasty [4,6].

The Conflict of Interest statement associated with this article can be found at http:// dx.doi.org/10.1016/j.arth.2014.10.035. Reprint requests: Dean C. Perfetti, BA, Department of Orthopaedics, SUNY Downstate, 16 Medical Center, 450 Clarkson Avenue, Box 30, Brooklyn, NY 11203. http://dx.doi.org/10.1016/j.arth.2014.10.035 0883-5403/© 2014 Elsevier Inc. All rights reserved.

Modern surgical techniques and hardware, drug therapy including hydroxyurea, and stem cell transplantation have brought dramatic improvements to the quality of life and life expectancy among patients with SCD, with many living comfortably into the seventh decade [6,7]. Despite these advances, however, patients with SCD remain high-risk surgical candidates. The stress from surgical procedures can stimulate cytokine release and lead to hypoxia, hypoperfusion and acidosis, an environment that promotes red blood cell sickling and subsequent occlusion of microcirculation [8,9]. Preoperative red blood cell transfusion has been investigated as prophylaxis against sickling crises, but postoperative complication rates among this population are estimated to be as high as 30% even after preoperative transfusion [10]. Current data on short-term outcomes following total hip arthroplasty (THA) and total knee arthroplasty (TKA) are limited to small cohort studies. Our study sought to examine differences in risk of perioperative complication, length of stay and hospital charges between patients with and without SCD who were admitted for THA or TKA between 1998 and 2010 in the United States. In secondary analyses, we studied the effects of concomitant osteonecrosis and transfusion during admission on these outcomes. Methods Study Population The Nationwide Inpatient Sample (NIS) is the largest all-payer inpatient care database in the United States. Each year, the NIS compiles

548

D.C. Perfetti et al. / The Journal of Arthroplasty 30 (2015) 547–551

records of roughly 8 million hospital stays across nearly 1000 hospitals, creating a 20% representative sample of annual U.S. hospital admissions [11]. Each patient discharge record contains demographic and clinical data, including ICD-9-CM (International Classification of Diseases, Ninth Revision, Clinical Modification) diagnosis and procedure codes. The NIS also provides discharge weights, based on the size and location of the hospital where each admission occurred, to allow researchers to calculate nationwide estimates of patient discharges. This database has been found to represent 96% of the U.S. population [12]. Cases Admissions who received a primary total hip arthroplasty (THA) or total knee arthroplasty (TKA) between the years of 1998 and 2010 were identified using ICD-9-CM procedure codes for THA (81.51, 00.74, 00.75, 00.76, 00.77) and TKA (81.54). Patients with diagnosis codes indicating pathological fracture, malunion of fracture, traumatic femoral neck fracture, and long-term mechanical loosening associated with revisions were excluded, as these admissions are predominantly non-elective [13]. Outcomes Perioperative complications were assessed using ICD-9-CM diagnosis codes as defined by a recent study of orthopedic-related complications [13]. We defined a “major complication” as having acute renal failure, death, myocardial infarction, pneumonia, pulmonary embolism, stroke, or tachycardia during admission. We defined a “minor complication” as having deep vein thrombosis, implant infection, implant dislocation, irrigation and debridement, sepsis, urinary tract infection, wound hemorrhage, wound disruption, or wound infection during admission [14]. For each admission, we also extracted data on length of stay and total hospital charges. Due to the right skew of these variables, we used a log transformation to increase the normality of their distributions. Covariates We used ICD-9-CM diagnosis codes to identify THA and TKA admissions that had sickle cell disease (282.41, 282.42, 282.60, 282.61, 282.62, 282.63, 282.64, 282.68, 282.69), and classified all other admissions as not having sickle cell disease [15]. Diagnosis codes were also used to identify patients with osteonecrosis (733.4, 733.40, 733.42, 733.43, 733.49) and procedure codes were used to identify admissions requiring transfusion of red blood cells (99.00, 99.01, 99.01, 99.03, 99.04). We extracted demographic data on each admission including age (in years), gender, race (white, black, Hispanic, other), insurance (Medicare, Medicaid, private, other), and year of admission (1998–2010). We assessed comorbidities using the Charlson and Deyo method for ICD-9-CM coding [16], which uses a weighted scoring system to predict the ten-year mortality for patients based on their comorbidities. One point is assigned to congestive heart failure, peripheral vascular disease, dementia, cerebrovascular disease, chronic pulmonary disease, rheumatologic disease, peptic ulcer disease, mild liver disease, past myocardial infarct, or uncomplicated diabetes. Two points are assigned to hemiplegia or paraplegia, renal disease, malignancy including leukemia and lymphoma, or diabetes with end organ damage. Three points are assigned to moderate or severe liver disease, while six points are assigned to metastatic solid tumor, or HIV/AIDS. Patients with no comorbidities are given a score of zero points [16].

calculate parameter estimates for mean length of stay and mean total charges of admission. We interpreted the results of our linear regressions as percent differences, using the formula 100*(e b − 1), where b is the regression coefficient of a log-transformed outcome variable [17]. Our regression models were adjusted for the confounding effects of age, gender, race, insurance type, and Deyo comorbidity score. Admissions with missing age (THA = 3,287; TKA = 4,194), gender (THA = 8,786; TKA = 13,467), race (THA = 678,604; TKA = 1,402,920), and insurance (THA = 4,568; TKA = 10,784) could not be included in multivariable regression models. However, to rule out the possibility that the exclusion of this subset of admissions with missing demographic data biased our results, we also performed sensitivity analyses on our data, where we removed each covariate from our regression model, thereby including all patients with missing demographic data. We also used regression models with SCD (no, yes) as the outcome variable, and age, sex, race, insurance, comorbidities, osteonecrosis and transfusion as independent predictor variables, to assess the effect size of each demographic variable. We used weighting variables in all our analyses to simulate national U.S. rates of THA and TKA admission. All statistical analyses were performed using SAS version 9.3 (SAS Institute Inc., Cary, North Carolina). All P-values were two-tailed, and P b 0.05 was interpreted as statistically significant. All figures were generated using Microsoft Excel 2010 (Microsoft Corporation, Redmond, Washington).

Results After exclusions, our study population consisted of 3,532 patients with SCD and 2,653,653 patients without SCD admitted for THA, as well as 724 patients with SCD and 5,660,896 patients without SCD admitted for TKA. Relative to the number of total admissions for THA and TKA, the proportion that had SCD remained steady during the study period. Between 1998 and 2010, there were 1.33 THA admissions with SCD per 1,000 total THA admissions (m = − 0.10, r = 0.64, P = 0.025), while there were 0.13 TKA admissions with SCD per 1,000 total TKA admissions (m = 0.01, r = 0.75, P = 0.003) (Fig. 1). For both THA and TKA, patients with SCD were more likely to be younger (THA: P b 0.001; TKA: P b 0.001), were more likely to be female (THA: P = 0.003; TKA: P = 0.002), were more likely to be black (THA: P b 0.001; TKA: P b 0.001), and were more likely to pay with Medicaid (THA: P b 0.001; TKA: P b 0.001). For THA but not TKA, fewer patients with SCD had a Charlson and Deyo score of 1 (P b 0.001) and ≥ 2 (P b 0.001). Osteonecrosis was observed in 87% of THA and 25% of TKA patients with SCD, compared to 11% of THA patients and 1% of TKA patients without SCD (THA: P b 0.001; TKA: P b 0.001). Blood transfusion was performed in 52% of THA and 47% of TKA admissions with SCD,

Statistical Analysis We used logistic regression (proc surveylogistic) to calculate the odds ratio (OR) of having a perioperative complication during the course of admission. We used linear regression (proc surveyreg) to

Fig. 1. Number of patients with sickle cell disease per 1000 admissions for THA and TKA (1998–2010).

D.C. Perfetti et al. / The Journal of Arthroplasty 30 (2015) 547–551

549

Table 1 Patient Characteristics of THA and TKA Admissions, According to Diagnosis of Sickle Cell Disease. THA

Admissions (N) Mean age (years) Gender (%) Male Female Race (%) White Black Hispanic Other Insurance (%) Medicare Medicaid Private Other Deyo Score (%) 0 1 ≥2 Osteonecrosis (%) No Yes Transfusion (%) No Yes

No

Yes

2,653,653 65

3,532 37

44 56

TKA OR (95% CI)

P

No

Yes

0.88 (0.87–0.88)

b0.001

5,660,896 67

724 54

OR (95% CI)

P

0.91 (0.90–0.92)

b0.001

38 62

1.00 [Reference] 1.26 (1.08–1.46)

0.003

36 64

24 76

1.00 [Reference] 1.81 (1.24–2.66)

0.002

88 6 3 3

4 90 3 3

1.00 [Reference] 328.96 (211.56–511.51) 27.61 (14.66–52.00) 21.64 (10.95–42.77)

b0.001 b0.001 b0.001

85 7 5 3

6 84 5 5

1.00 [Reference] 165.24 (76.56–356.61) 15.26 (5.12–45.50) 18.76 (5.95–59.20)

b0.001 b0.001 b0.001

54 3 40 3

29 30 35 6

1.00 [Reference] 18.55 (15.28–22.52) 1.65 (1.37–1.99) 3.94 (2.85–5.44)

b0.001 b0.001 b0.001

58 3 36 4

44 14 39 3

1.00 [Reference] 7.46 (4.55–12.24) 1.45 (1.01–2.07) 1.17 (0.47–2.92)

b0.001 0.043 0.111

68 23 9

76 19 5

1.00 [Reference] 0.72 (0.59–0.87) 0.53 (0.38–0.74)

b0.001 b0.001

63 27 10

57 29 14

1.00 [Reference] 1.17 (0.81–1.69) 1.62 (1.00–2.61)

89 11

13 87

1.00 [Reference] 55.06 (44.22–68.57)

b0.001

99 1

75 25

1.00 [Reference] 45.41(31.13–66.26)

b0.001

75 25

48 52

1.00 [Reference] 3.18 (2.74–3.69)

b0.001

82 18

53 47

1.00 [Reference] 3.92 (2.83–5.42)

b0.001

0.418 0.050

Abbreviations: CI = confidence interval; N = number of admissions; % = percent of admissions.

compared to 25% of THA and 18% of TKA admissions without SCD (THA: P b 0.001; TKA: P b 0.001) (Table 1). After adjusting our models for the confounding effects of age, sex, race, insurance, and comorbidities, we observed that there was a greater risk of complications in patients with SCD. For both THA and TKA, this was true for any complication (THA: OR = 2.52, P b 0.001; TKA: OR = 2.37, P = 0.001), major complication (THA: OR = 2.17, P = 0.003; TKA: OR = 1.85, P = 0.158), or minor complication (THA: OR = 2.53, P b 0.001; TKA: OR = 2.21, P = 0.012) (Table 2). For specific complications, THA patients with SCD were more likely to have pneumonia (OR = 6.79, P b 0.001), irrigation and debridement (OR = 4.00, P = 0.004), sepsis (OR = 17.12, P b 0.001), urinary tract infection (OR = 3.15, P b 0.001), and wound hemorrhage (OR = 1.87, P = 0.037), while TKA patients with SCD were more likely to have tachycardia (OR = 13.56, P b 0.001) and deep vein thrombosis (OR = 8.25, P b 0.001), and were more likely to die (OR = 11.22, P = 0.019) (Table 2). In our sensitivity analyses, the risks of any complication in patients with missing age (THA: OR = 1.62, P b 0.001; TKA: OR = 1.99, P = 0.001), gender (THA: OR = 2.54, P b 0.001; TKA: OR = 2.37, P = 0.001), insurance (THA: OR = 2.87, P b 0.001; TKA: OR = 2.43, P = 0.001), and race (THA: OR = 2.90, P b 0.001; TKA: OR = 2.55, P b 0.001) were similar to our full multivariable model, suggesting that the exclusion of the patients with missing demographic data did not bias our results.

There was a longer mean length of stay among patients with SCD. The mean length of stay was 6.92 days (95% CI, 6.28–7.55) for THA patients with SCD and 3.83 days (95% CI, 3.82–3.84) for THA patients without SCD. For TKA, mean length of stay was 5.33 days (95% CI, 4.01–6.64) for patients with SCD and 3.75 days (95% CI, 3.75–3.76) for patients without SCD. Patients with SCD had a multivariate-adjusted mean length of stay that was 42% (P b 0.001) longer for THA and 20% (P b 0.001) longer for TKA compared to patients without SCD (Table 3). We observed a negative linear trend in length of stay for both patients with and without SCD who were admitted for THA and TKA between 1998 and 2010 (Figs. 2 and 3). There were also higher total hospital charges among patients with SCD. The mean total charges were $51,240 (95% CI, $48,423–$54,057) for THA patients with SCD and $37,802 (95% CI, $37,740–$37,864) for THA patients without SCD. For TKA, mean total charges were $49,770 (95% CI, $42,491–$57,048) for patients with SCD and $36,673 (95% CI, $36,631–$36,716) for patients without SCD. Patients with SCD had multivariate-adjusted total charges that were 19% (P b 0.001) higher for THA and 16% (P = 0.001) higher for TKA as compared to patients without SCD (Table 3). We observed a positive linear trend in total charges for both patients with and without SCD who were admitted for THA and TKA between 1998 and 2010 (Figs. 4 and 5). When we restricted our models to patients who were admitted in 2010, the most recent year of data, we observed similar results to the entire sample. For THA, compared to patients without SCD, patients with SCD

Table 2 Risk of Perioperative Complications, According to Diagnosis of Sickle Cell Disease.

Table 3 Length of Stay and Total Hospital Charges, According to Diagnosis of Sickle Cell Disease.

No

Yesa

THA Any complication Major complication Minor complication

Reference Reference Reference

2.52 (1.90–3.34) 2.17 (1.31–3.58) 2.53 (1.85–3.45)

b0.001 0.003 b0.001

TKA Any complication Major complication Minor complication

Reference Reference Reference

2.37 (1.40–4.01) 1.86 (0.79–4.41) 2.21 (1.19–4.12)

0.001 0.158 0.012

a

No

Yesa

P

THA Length of stay estimateb Total charges estimateb

Reference Reference

42.12 (37.58–46.67) 19.13 (15.02–23.25)

b0.001 b0.001

TKA Length of stay estimateb Total charges estimateb

Reference Reference

20.01 (11.86–28.16) 16.39 (6.65–26.12)

b0.001 0.001

P

Models adjusted for age, gender, race, insurance, and Deyo comorbidity score.

a

Models adjusted for age, gender, race, insurance, and Deyo comorbidity score. Interpret as percentage difference, under the formula 100(eb − 1), where b is the estimated standardized regression coefficient of a log–transformed outcome variable. b

550

D.C. Perfetti et al. / The Journal of Arthroplasty 30 (2015) 547–551

Fig. 2. Mean length of stay of THA admissions, according to diagnosis of sickle cell disease (1998–2010).

Two prior studies have evaluated the relationship between SCD, length of stay and total cost for THA [8] and TKA [15] admissions using the NIS database. Our study expands on their work with: (1) the addition of perioperative complications as outcomes; (2) assessing length of stay and total hospital charges over a longer study period (1998–2010 versus 2002–2006); (3) assessing annual trends in length of stay and total hospital charges; and (4) examining outcomes according to subgroups of osteonecrosis and blood transfusion. These two prior studies reported a length of stay that was 56% longer for THA and 81% longer for TKA patients with SCD, as well as costs that were 25% higher for THA and 19% higher for TKA patients with SCD between 2002 and 2006. Our findings are generally consistent with these previous findings, although the difference in mean length of stay among TKA patients with and without SCD in our cohort was considerably smaller. We believe that this discrepancy can be attributed to the increased sample size of our study, as the year 2002 had an uncharacteristically high mean length of stay of 22 days for TKA admissions with SCD. Our observation that patients with SCD had an increased risk of perioperative complications and increased length of stay fits within the context of prior studies. Infections are a major cause of morbidity and mortality among SCD patients, secondary to persistent splenic infarctions with eventual auto-splenectomy [18]. Defective activation of complement's alternate pathway facilitated by a hypo-functioning or absent spleen increases the risk of infection, particularly by encapsulated bacteria [18]. Therefore, an overall weakened immune response in SCD patients may account for the increased risk of complications such as pneumonia, sepsis, wound infection, and irrigation and debridement among patients with SCD. Stroke is a significant cause of morbidity and mortality among patients with SCD [19], but we did not observe

Fig. 3. Mean length of stay of TKA admissions, according to diagnosis of sickle cell disease (1998–2010).

had a mean length of stay of 6.09 days (vs. 3.25 days) and mean total hospital charges of $69,928 (vs. $49,854). For TKA, compared to patients without SCD, those patients with SCD had a mean length of stay of 3.71 days (vs. 3.27 days) and mean total hospital charges of $54,990 (vs. $48,002). When restricting our THA models to patients who had a diagnosis of osteonecrosis in their medical record, THA admissions with SCD had a multivariate adjusted risk of complication that was 113% higher than that for patients without SCD (P b 0.001), a mean length of stay that was 38% longer (P b 0.001) and a mean total hospital charge that was 22% higher (P b 0.001). When we restricted our THA models to patients receiving transfusion during admission, THA admissions with SCD had a multivariate-adjusted risk of complication that was 86% higher (P b 0.001), a mean length of stay that was 38% longer (P b 0.001), and a mean total hospital charge that was 22% higher (P b 0.001). Among TKA patients, osteonecrosis and transfusion subgroup analyses were not feasible due to limited sample size.

Fig. 4. Mean total hospital charges for THA admissions, according to diagnosis of sickle cell disease (1998–2010).

Discussion In this large, nationwide, U.S. study of short-term outcomes following admission for THA and TKA, we found that sickle cell disease (SCD) was associated with an increased risk of perioperative complications, an increased length of stay, and increased total hospital charges during admission. These associations remained significant even after adjustment for potential confounding risk factors including age, gender, race, insurance, and comorbidities. In secondary analyses, we found that patients with SCD who had osteonecrosis and/or required red blood cell transfusion were at increased risk of adverse outcomes following THA. We also observed increased length of stay and total hospital charges among a subset of the most recent admissions, suggesting that this problem is ongoing.

Fig. 5. Mean total hospital charges for TKA admissions, according to diagnosis of sickle cell disease (1998–2010).

D.C. Perfetti et al. / The Journal of Arthroplasty 30 (2015) 547–551

any increased risk of stroke among patients with SCD as a short-term post-operative complication. We acknowledge that our study has several limitations. First, NIS data are restricted to diagnoses recorded during each patient's hospital stay, and provides no information on long-term outcomes such as readmission, future arthroplasty failure, and deep infection. Second, ICD-9-CM coding for diagnoses and procedures limits our ability to account for differences in implant type, severity of complications, and severity of joint degeneration. Third, variability among hospital coding practices may have led to undocumented cases of SCD. Fourth, patients with SCD who undergo an elective surgery such as THA or TKA may be healthier than the average patient with SCD. Therefore, our data should be interpreted within the context of SCD patients who are fit to undergo elective surgery. Fifth, we could not determine when transfusion occurred in relation to surgery, and patients may have also received transfusions prior to hospital admission. Therefore, our subset of patients receiving RBC transfusion most likely under-represents the number of patients receiving perioperative transfusions [8]. Sixth, although our data suggest that SCD itself can lead to longer and more costly admissions, we cannot rule out the possibility that these increases were due to additional precautions taken by the treating physician and hospital staff. Despite these limitations, the NIS remains a practical means to evaluate the epidemiology of topics in orthopedic surgery. This allpayer, multi-year database contains millions of U.S. patient admissions, which increases the generalizability of our findings. In summary, we found that between 1998 and 2010, patients with SCD who were admitted for THA and TKA had longer, more complicated, and more expensive hospital stays compared to other patients. Perioperatively, patients with SCD must be carefully monitored for early signs of adverse events including infection and hemodynamic compromise. However, we believe that the long-term benefits of joint replacement outweigh these moderate short-term risks in patients who are otherwise good surgical candidates. Prior to surgery, orthopedic surgeons should engage in a discussion with potential THA and TKA candidates with SCD to set practical expectations for the postoperative course.

551

References 1. Centers for Disease Control and Prevention. Sickle Cell Disease: Data & Statistics. Available at http://www.cdc.gov/ncbddd/sicklecell/data.html. [Accessed on 16 June 2014]. 2. Kassim AA, DeBaun MR. Sickle cell disease, vasculopathy, and therapeutics. Annu Rev Med 2013;64:451. 3. Kato GJ, Gladwin MT, Steinberg MH. Deconstructing sickle cell disease: reappraisal of the role of hemolysis in the development of clinical subphenotypes. Blood Rev 2007;21(1):37. 4. da Silva Junior GB, Daher Ede F, da Rocha FA. Osteoarticular involvement in sickle cell disease. Rev Bras Hematol Hemoter 2012;34(2):156. 5. Adam S, Jonassaint J, Kruger H, et al. Surgical and obstetric outcomes in adults with sickle cell disease. Am J Med 2008;121(10):916. 6. Mont MA, Marker DR, Zywiel MG, et al. Osteonecrosis of the knee and related conditions. J Am Acad Orthop Surg 2011;19(8):482. 7. Platt OS, Brambilla DJ, Rosse WF, et al. Mortality in sickle cell disease. Life expectancy and risk factors for early death. N Engl J Med 1994;330(23):1639. 8. Dinan MA, Chou CH, Hammill BG, et al. Outcomes of inpatients with and without sickle cell disease after high-volume surgical procedures. Am J Hematol 2009;84(11):703. 9. Hernigou P, Zilber S, Filippini P, et al. Total THA in adult osteonecrosis related to sickle cell disease. Clin Orthop Relat Res 2008;466(2):300. 10. Vichinsky EP, Haberkern CM, Neumayr L, et al. A comparison of conservative and aggressive transfusion regimens in the perioperative management of sickle cell disease. The Preoperative Transfusion in Sickle Cell Disease Study Group. N Engl J Med 1995; 333(4):206. 11. Steiner C, Elixhauser A, Schnaier J. The healthcare cost and utilization project: an overview. Effective Clinical Practice 2002;5(3):143. 12. Healthcare Cost and Utilization Project (HCUP). HCUP technical assistance. Agency for Healthcare Research and Quality (AHRQ). http://www.hcup-us.ahrq.gov/tech_ assist/nationalestimates/508_course/508course.htm#{044D5BEE-AA10-4D588E4C-9EF8BA1A3A8E}; 2012. [Accessed 2012 Oct 2]. 13. Lin CA, Kuo AC, Takemoto S. Comorbidities and perioperative complications in HIVpositive patients undergoing primary total hip and knee arthroplasty. J Bone Joint Surg Am 2013;95(11):1028. 14. Parvizi J, Mui A, Purtill JJ, et al. Total joint arthroplasty: when do fatal or near-fatal complications occur? J Bone Joint Surg Am 2007;89(1):27. 15. Kamble S, Telen MJ, Dinan MA, et al. Costs and length of stay for patients with and without sickle cell disease after hysterectomy, appendectomy, or knee replacement. Am J Hematol 2010;85(1):79. 16. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD9-CM administrative databases. J Clin Epidemiol 1992;45(6):613. 17. Glidden DV, Shiboski SC, McCulloch CE. Linear regression: checking model assumptions and fit. From: Regression Methods in Biostatistics: Linear, Logistic, Survival, and Repeated Measures Models. Springer Verlag; 2011. 18. Inati A, Koussa S, Taher A, et al. Sickle cell disease: new insights into pathophysiology and treatment. Pediatr Ann 2008;37(5):311. 19. Stanley AC, Christian JM. Sickle cell disease and perioperative considerations: review and retrospective report. J Oral Maxillofac Surg 2013;71(6):1027.