- Email: [email protected]
Race and ethnic differences in the epidemiology and risk factors for graft failure after heart transplantation
http://www.jhltonline.org
Race and ethnic differences in the epidemiology and risk factors for graft failure after heart transplantation Alanna A. Morris, MD,a Andreas P. Kalogeropoulos, MD, MPH, PhD,a Liping Zhao, MSPH,b Melissa Owen, RN,c S. Raja Laskar, MD,a J. David Vega, MD,a Andrew Smith, MD,a and Javed Butler, MD, MPHa From the Emory University aSchool of Medicine; bRollins School of Public Health; and the cSchool of Nursing, Atlanta, Georgia.
KEYWORDS: race/ethnicity; heart transplant; graft failure; cardiac allograft vasculopathy; population-attributable risk
BACKGROUND: Contemporary epidemiology of chronic graft failure (GF) after heart transplantation (HT) is not well described. Moreover, differences in the epidemiology of GF based on race/ethnicity remain poorly understood, despite clear evidence of inferior survival of ethnic minorities after HT. METHODS: The incidence of GF and the population-attributable risk (PAR) of independent risk factors for GF were assessed in 15,255 patients (76% men; mean age 52 ⫾ 12 years) who underwent primary HT from 2004 to 2012. RESULTS: During a median follow-up of 4.7 years (interquartile range, 2.3–7.1 years), GF developed in 2,926 patients (19.2%), corresponding to an incidence rate of 39.8/1,000 person-years (95% confidence interval, 38.4–41.3). Blacks were more likely to develop GF than Hispanics or whites, with incidence rates of 55.1, 42.2, and 36.5/1,000 person-years, respectively. After multivariable adjustment, black race was associated with a higher risk of GF (hazard ratio, 1.4; 95% confidence interval, 1.2–1.6; p o 0.001). Blacks and Hispanics were more likely to have risk factors for GF, including low education, public insurance, allosensitization, higher human leukocyte antigen mismatch, non-adherence, and history of rejection requiring hospitalization (all p o 0.001). Rejection requiring hospitalization carried the highest population-attributable risk in all groups, with the highest fraction in blacks (25.8%) compared with whites (18.6%) and Hispanics (15.6%). Socioeconomic and donor risk factors conferred relatively less risk of GF. CONCLUSIONS: Black HT recipients have the highest risk of GF, with immunologic factors conferring the greatest proportion of that risk. Racial differences in risk factors for GF after HT require further study. J Heart Lung Transplant 2015;34:825–831 Published by Elsevier Inc.
Heart transplantation (HT) remains the therapy of choice for select patients with end-stage heart failure. Median conditional survival now exceeds 10 years for recipients Reprint requests: Alanna A. Morris, MD, 1462 Clifton Rd Ste 520, Atlanta GA 30322. Telephone: 617-480-9262. Fax: 404-712-0149. E-mail address: [email protected] 1053-2498/$ - see front matter Published by Elsevier Inc. http://dx.doi.org/10.1016/j.healun.2014.12.012
who survive the first year after HT.1 However, approximately 20% of deaths after the first year are due to graft failure (GF), which often results from processes that include cellular or antibody-mediated rejection and cardiac allograft vasculopathy (CAV). Despite this overall trend toward improved outcomes after HT, disparities in survival based on race/ethnicity are
826
The Journal of Heart and Lung Transplantation, Vol 34, No 6, June 2015
well recognized. Several studies have demonstrated shorter graft and patient survival in HT recipients who are of racial/ ethnic minority background.2–4 The exact mechanisms for these disparities remain unclear; however, suggested explanations include socioeconomic, biological, and immunologic factors.5,6 Understanding the epidemiology of GF, including the influence of risk factors associated with GF, are imperative to further improve outcomes and bridge disparities after HT. The population-attributable risk (PAR) is often used to describe the proportion of disease risk in a population that can be attributed to the causal effects of a risk factor or set of factors.7 Thus, our objective for this study was to assess the epidemiology of incident GF by examining the race-related risk factor profile and PAR for GF in a contemporary cohort of HT recipients.
Methods Study population All individuals aged Z18 years who were listed for their first HT were identified in the Organ Procurement and Transplantation Network (OPTN) database, which includes deidentified data on all patients listed for a HT in the United States. Analysis was limited to patients listed between June 30, 2004 (when follow-up information on graft status was first reported to the OPTN) and September 2012. The Health Resources and Services Administration and the United States Department of Health and Human Services provide oversight to the activities of the OPTN contractor, the United Network of Organ Sharing. We compared baseline characteristics of patients listed for primary HT during the study period, excluding patients listed for a repeat HT or for multiorgan transplantation. Race was reported by the transplantation centers as white, black, Hispanic/Latino, Asian, American Indian/Alaska Native, Native Hawaiian/Pacific Islander, multiracial, or other. Ethnicity was reported as Hispanic or nonHispanic. For all Hispanic patients in the study, race and ethnicity variables were identical; therefore, all white patients in this analysis were non-Hispanic white, and all black patients were non-Hispanic black. Because of the small sample size of minority patients with race/ethnicity other than black or Hispanic (o4.5% of the sample), these patients were excluded.
Study definitions Incident graft failure Follow-up data on HT recipients is provided to the OPTN 6 months after transplant and then annually thereafter. On the follow-up form that is reported annually to the OPTN, transplant centers are asked to document allograft status as “functioning” or “failed.” GF is defined as having occurred when “an organ is removed, a recipient dies, or a recipient is placed on a chronic allograft support system.”8 We identified patients with GF that occurred at least 180 days after transplant to attempt to exclude patients with primary allograft dysfunction. Thus, GF includes patients living with allograft dysfunction, patients listed for repeat transplantation, and patients who died or did not have further information regarding their outcome of GF. Patients were monitored from the date of transplant until death, retransplantation, or date of last known
follow-up provided by United Network of Organ Sharing, with follow-up through September 2012.
Risk factors for incident graft failure Clinical, socioeconomic, and immunologic variables were defined at the time of listing for waiting list factors, whereas donor-related variables were defined at the time of transplant, and posttransplantation outcomes were defined at the time of follow-up. Variables compared between race/ethnic groups included clinical (recipient age, gender, creatinine), socioeconomic factors (education, primary insurance payer), immunologic (allosensitization, human leukocyte antigen [HLA] mismatch, non-adherence to medications, any episodes of rejection requiring hospitalization, CAV), and donor (donor age, ischemic time) factors. Transplant centers are asked to document “Was there evidence of noncompliance with immunosuppression medication during this follow-up period that compromised the patient’s recovery?” on the annual follow-up form. For the purposes of our analysis, any report of “yes” to this question during the total follow-up period for any patient was considered “non-adherence.”
Risk factor definitions for calculation of PAR For the PAR calculation, continuous predictors that were univariate predictors of GF were dichotomized using clinically relevant cutoff points. Age was dichotomized at the median age of the cohort, education at high school level or less, insurance as public (Medicare, Medicaid, Veterans Affairs) vs private, allosensitization as panel reactive antibody (PRA) level Z10%, HLA mismatch at Z 5 loci based on the cohort median, ischemic time as Z4 hours,9 and donor age as Z30 years.10 Medication non-adherence, history of rejection requiring hospitalization, and presence of CAV were collapsed into binary predictors (yes or no).
Statistical analysis Continuous variables were compared using the Mann-Whitney rank sum test, and categorical variables were compared using the chi-square test. Cumulative event rates were obtained using the Kaplan-Meier method and were compared using the log-rank statistic. Univariate and adjusted hazard ratios (HRs) and the corresponding 95% confidence intervals (CIs) were obtained using Cox proportional hazards regression models to identify significant predictors of GF. The proportional hazards assumption was tested and verified for all risk factors using Schoenfeld residual correlation analysis. Subsequently, race/ethnicity-stratified analysis was performed to obtain rate ratio (RR) estimates for the significant risk factors. We also calculated adjusted (multivariable) PARs using a Poisson regression model with incident GF as the outcome and the factors already described as predictors.11,12 Briefly, the predicted number of cases is calculated for the full model (Nfull), which equals the actual number of cases. Next, the effect of the risk factor of interest is “removed” by setting the value of the covariable to 0, and the predicted number of events is calculated (Nremoved). The adjusted PAR for the risk factor then becomes PARadjusted ¼ 1 (Nremoved/Nfull). PARs are not presented if the adjusted RR and corresponding 95% CI for the risk factor failed to reach a p-value of 0.2. Age and gender were included in regression models for multivariable PAR calculation, but the PARs of these variables were omitted from the tables because they cannot be
Morris et al.
Association of Race/Ethnicity with Cardiac Graft Failure
modified. Data were analyzed with SAS 9.3 software (SAS Institute Inc, Cary, NC).
827
levels of creatinine, PRA, and HLA mismatch. Hispanic HT recipients had the youngest median donor age, and black recipients had the shortest allograft ischemic time.
Results Risk factors for incident GF
Study population A total of 51,954 patients were listed for HT after June 30, 2004. Of these, 15,972 HT recipients had at least 180 days of follow-up with complete information on graft status and episodes of rejection. An additional 717 recipients were excluded for race/ethnicity other than white, black, or Hispanic, leaving 15,255 HT recipients for this analysis.
Clinical characteristics The pre-transplant characteristics of the 15,255 HT recipients who formed our analytic cohort are reported in Table 1. Compared with whites, blacks and Hispanics were younger, had a higher proportion of women, and were more likely to have a non-ischemic heart failure etiology. Blacks and Hispanics were less educated and were less likely to have private insurance. Black HT recipients had higher
During a median follow-up of 4.7 years (interquartile range, 2.3–7.1 years), 2,926 HT recipients (19.2%) developed GF (39.8/1,000 person-years; 95% CI, 38.4–41.3). The median time to GF was 4.18 years. In the entire cohort, age r44 years (HR, 1.2; 95% CI, 1.1–1.3; p o 0.001), high school education or less (HR, 1.3; 95% CI, 1.2–1.4; p o 0.001), public insurance (HR, 1.3; 95% CI, 1.2–1.4; p o 0.001), allosensitization (HR, 1.2; 95% CI, 1.1–1.3; p ¼ 0.001), HLA mismatch Z5 loci, (HR, 1.1; 95% CI, 1.0–1.2; p ¼ 0.01), history of rejection requiring hospitalization (HR, 2.6; 95% CI, 2.4–2.8; p o 0.001), history of non-adherence (HR, 2.3; 95% CI, 2.1–2.6; p o 0.001), ischemic time Z 4 hours (HR, 1.1; 95% CI, 1.0–1.2; p o 0.05), and donor age Z30 years (HR, 1.2; 95% CI, 1.2–1.3; p o 0.001) were associated with an increased risk of GF. There was a significant interaction of CAV and age (p ¼ 0.0006). CAV was associated with a lower risk of incident GF in patients
Table 1 Baseline Characteristics of White, Black, and Hispanic Heart Transplant Recipients Who Underwent Heart Transplantation From 2004 to 2012
Characteristics Age, years Female gender Heart failure etiology Dilated Ischemic Restrictive Hypertrophic Valvular Congenital Other Creatinine, mg/dl Education High school or less Some college College or graduate Insurance Private Medicare Medicaid Other PRA, % HLA mismatch, No. Ischemic time, hours Donor age, years Follow-up time, years
Whites
Blacks
(n ¼ 11,465)
(n ¼ 2,653)
56 (47–61) 2,454 (21) 4,555 5,586 224 268 261 315 254 1.2
(40) (49) (2) (2) (2) (3) (2) (1.0–1.4)
Hispanics (n ¼ 1,137) a,b
49 (39–57) 908 (34) 1,923 540 42 22 55 26 45 1.2
(73) (20) (1) (1) (2) (1) (2) (1.0–1.6)a,b
p-value c
51 (41–59) 289 (25) 625 406 16 21 35 22 12 1.1
(55) (36) (1) (2) (3) (2) (1) (0.9–1.4)c
4,208 (37) 2,431 (21) 2,667 (23)
1,230 (46) 607 (23) 347 (13)
597 (53) 174 (15) 121 (11)
7,217 2,849 915 483 0 5 3.2 29 5.0
1,246 726 533 148 0 5 3.1 29 4.0
459 279 278 121 0 5 3.2 26 4.1
o0.001 o0.001 o0.001
o0.001 o0.001
o0.001 (63) (25) (8) (4) (0–4) (4,5) (2.5–3.9) (21–41) (2.6–7.6)
HLA, human leukocyte antigen; PRA, panel reactive antibody. Data are presented as median (25th–75th percentile) or number (%). a p o 0.001 for white-black comparison. b p r 0.001 for black-Hispanic comparison. c p o 0.001 for white-Hispanic comparison.
(47) (27) (20) (6) (0–11)a,b (4–6)a,b (2.5–3.8)a (21–40)b (2.0–6.4)a
(40) (25) (25) (10) (0–4) (4–5) (2.4–3.9) (20–38)c (2.1–6.9)c
o0.001 o0.001 o0.001 o0.001 o0.001
(0.6–1.3) (0.7–1.4) (0.8–2.1) (0.7–1.5) (1.3–3.8)b (1.3–2.8)b (0.6–1.3) (0.7–1.7) (1.1–2.1)b 1.2 1.1 1.2 1.2 2.1 2.5 0.9 1.0 1.3 45 37 19 57 3 14 28 21 49 rHigh school education Public insurance PRA Z10% HLA mismatch Z5 loci Non-adherence Rejection requiring hospitalization Cardiac allograft vasculopathy Ischemic time Z4 hours Donor age Z30 years
CI, confidence interval; HLA, human leukocyte antigen; PAR, population-attributable risk; PRA, panel reactive antibody; RR, rate ratio. RRs are adjusted for age, gender, and the risk factors shown. a p r 0.001. b p r 0.05. c p o 0.2.
0.9 1.0 1.3 1.1 2.2 1.9 0.9 1.1 1.5 67 60 18 61 4 19 25 22 43 … … 5.3 … 7.6 25.8 … 3.6 5.7 (1.1–1.4) (1.0–1.3)b (1.0–1.3)b (1.1–1.3)b (1.7–2.5)a (2.2–2.8)a (0.8–1.0)b (0.9–1.1) (1.1–1.4)a
9.0 5.2 2.5 9.1 4.9 18.6 … … 10.6
56 53 26 68 6 24 26 18 48
1.1 1.0 1.3 0.9 1.8 2.2 0.8 1.2 1.1
(0.9–1.3) (0.8–1.2) (1.0–1.6)b (0.7–1.1) (1.4–2.4)a (1.8–2.7)a (0.7–1.0)b (0.9–1.5)c (0.9–1.4)c
Prevalence (%) PAR (%) RR (95% CI) Prevalence (%) PAR (%)
Blacks (n ¼ 2,653)
a
RR (95% CI) Prevalence (%)
Figure 1 Incident graft failure adjusted for age by race/ ethnicity. Graft failure was more likely to develop in black and Hispanic heart transplant recipients than in white recipients (stratified log-rank chi-square ¼ 105; p o 0.0001 for race).
Whites (n ¼ 11,465)
The prevalences and the adjusted (multivariable) PARs for the risk factors for GF are reported in Table 2. Low education level, public insurance, allosensitization, HLA mismatch at Z5 loci, reported non-adherence, and history of rejection requiring hospitalization were more prevalent in black and Hispanic HT recipients than in white recipients (p o 0.001 for all). In contrast, CAV (p ¼ 0.02) and older donor age (p ¼ 0.001) were more prevalent among white and black HT recipients, whereas longer ischemic time
Prevalence, Adjusted Rate Ratios, and Population-Attributable Risk of Risk Factors for Incident Graft Failure
Race-ethnic differences in the prevalence of risk factors for GF and PARs
Table 2
Kaplan-Meier estimates of incident GF in sub-groups based on race and ethnicity are displayed in Figure 1. Black recipients were more likely to develop GF than Hispanic and white recipients. The rates of GF were 55.1/1,000 person-years in black recipients (age-adjusted HR, 1.6; 95% CI, 1.4–1.7; p o 0.001) and 42.2/1,000 person-years in Hispanic recipients (age-adjusted HR, 1.2; 95% CI, 1.0–1.4; p ¼ 0.015) compared with 36.5/1,000 person-years in white participants. After adjustment for all risk factors in Table 1, black race was associated with an increased risk of GF (HR,1.4; 95% CI, 1.2– 1.6; p o 0.001). The median (interquartile range) ejection fraction was preserved among patients who developed GF: blacks, 57% (50%–60%); Hispanics, 60% (55%–64%); and whites, 59% (55%–61%; p o 0.001). However, among patients with GF, blacks and Hispanics were more likely to have an ejection fraction of r35% compared with whites (50 [8%], 9 [6.5%], 79 [3.8%]; p o 0.001).
Hispanics (n ¼ 1,137)
Differences in incident GF among race and ethnic sub-groups
RR (95% CI)
PAR (%)
with age 444 years (HR, 0.8; 95% CI, 0.8–0.9), but CAV was not associated with incident GF in patients aged r44 years (HR, 1.11; 95% CI, 0.96–1.3).
… … … … 6.5 15.6 … … 16.2
The Journal of Heart and Lung Transplantation, Vol 34, No 6, June 2015
Risk factor
828
Morris et al.
Association of Race/Ethnicity with Cardiac Graft Failure
(p ¼ 0.002) was more prevalent among white and Hispanic recipients. Rejection requiring hospitalization and non-adherence conferred the highest adjusted PARs in each race/ethnic group, with slightly higher PARs in black HT recipients compared with whites and Hispanics. Adjusted PARs for older donor age were higher in white and Hispanic HT recipients than in black recipients, whereas allosensitization was higher in black HT recipients than in white recipients. Although the prevalence of low education, public insurance, and HLA mismatch Z5 loci was higher in black and Hispanic HT recipients, the adjusted PAR for these factors was higher in white recipients.
Discussion In the present study, we found that the rate of incident GF is higher in black and Hispanic compared with white HT recipients. In addition, we found significant race-related differences in the risk factor profiles for incident GF. Although race/ethnic disparities in GF are multifactorial in their etiology, our findings add to the current literature on this topic because there is scant existing data that explain which risk factors for GF have the largest contribution to the observed disparities. Our data demonstrate that immunologic factors confer the greatest risk for incident GF for all race/ethnic groups, with a slightly higher PAR in black HT recipients. In our cohort, socioeconomic factors had a greater contribution to the development of GF in whites than in black and Hispanic HT recipients, whereas donor factors had a small effect on the risk of GF in all recipients. Chronic GF after HT may present as a wide spectrum of clinical or sub-clinical allograft dysfunction, including right ventricular, left ventricular, or biventricular dysfunction. Data from the registry of the International Society of Heart and Lung Transplant estimate that approximately 23% of recipients have developed GF 3 to 5 years after HT,13 similar to the rate of incident GF seen in our cohort. GF likely results from processes such as cellular and antibodymediated rejection and CAV; however, more research is needed to fully determine the mechanisms that contribute to chronic graft injury. Prior studies have examined individual risk factors for GF, including donor and recipient variables. Hong et al14 examined data of 11,703 HT recipients who received transplants between 2001 and 2007. Similar to our findings, these authors identified advanced donor and recipient age as well as prolonged donor ischemic time as risk factors for GF at 1 year. They also found poor recipient renal function was associated with the risk of GF at 1 year. In our cohort, pretransplant creatinine was not identified as an independent risk factor for incident GF. However, the incidence of chronic GF, as opposed to GF at 1 year, may be more strongly influenced by the development of chronic renal impairment related to nephrotoxicity and hypertension as a result of the use of calcineurin-inhibitors.15,16 CAV was associated with a lower risk of GF in older patients in our cohort. CAV that develops more than 2 years after transplant can be associated with a more indolent
829
course17 and may be associated with a better prognosis in the setting of preserved graft function than CAV with systolic dysfunction. Race-specific differences in the development of GF have been well described, with analyses consistently demonstrating worse long-term outcomes in blacks and Hispanics, even after adjustment for the higher clinical risk profile often seen in these groups pre-transplantation.3,6,18 We found that most of the socioeconomic and immunologic risk factors associated with GF were more common in blacks and Hispanics. Immunologic risk factors carried the highest risk of GF in all race/ethnic groups; however, immunologic factors played a larger role in the development of GF in blacks and Hispanics. Prior studies have documented higher levels of PRA and degrees of HLA mismatch in black and Hispanic HT recipients, leading to higher risk for acute rejection and allograft loss.2,6,19 Prior studies have confirmed the effect of socioeconomic status on the subsequent risk of graft loss, with lower household income, Medicare and Medicaid insurance, and less education all associated with worse outcomes.5,20,21 In our cohort, socioeconomic status, specifically lower education and public insurance, conferred a higher risk of GF in white HT recipients. Singh et al22 reported similar findings in an analysis of 490 pediatric HT recipients. In their study, low socioeconomic position was associated with a higher risk of graft loss among white children; however, socioeconomic status did not risk-stratify black children despite a higher prevalence of low socioeconomic position among blacks. These findings may be related to candidate selection, because whites of low socioeconomic position may also represent a higher risk group where social determinants of health amplify the biologic mechanisms that contribute to graft loss after HT. Few analyses have collected detailed information on medication non-adherence because this variable depends on multiple factors, including the intensity of the medication regimen, socioeconomic and family environment, the degree of education provided by health care providers, and other comorbid conditions. Although the OPTN database lacks detailed information on non-adherence, this was the only truly modifiable risk factor we examined, and it did carry a higher PAR in blacks and Hispanics than in whites. Prior data have shown higher reported rates of non-compliance among race/ethnic minorities, particularly blacks, compared with whites.18,23 However, medication non-adherence is often closely associated with socioeconomic status, with prior data showing that a lack of private insurance often decreases adherence to immunosuppressive therapy.24 Moreover, we lacked information on race/ethnic differences in genotypes that contribute to higher clearance and lower bioavailability of immunosuppressants. Black and Hispanic HT recipients are more likely to be “rapid metabolizers” of tacrolimus,25,26 placing these patients at higher risk of rejection after missing even a few doses or as being incorrectly classified as non-adherent if they are being underdosed. This analysis has some limitations. Risk factors that could have influenced the incidence of GF and/or the
830
The Journal of Heart and Lung Transplantation, Vol 34, No 6, June 2015
frequency of hospitalizations for rejection, such as positive retrospective crossmatch or the intensity of maintenance immune suppression, were not available in the OPTN database. Prior data have shown a slightly higher incidence of positive retrospective crossmatch in black HT recipients compared with whites.27 However, this variable was not associated with overall graft survival in adjusted models. In addition, a recent analysis of OPTN data has shown the proportion of white, black, and Hispanic HT recipients receiving mycophenolate was similar in recent eras, with a higher proportion of black recipients (vs whites) receiving tacrolimus.3 Thus, racial differences in graft survival are not likely to be due to disparities in the choice of immune suppression. We also lacked information on other factors that strongly influence socioeconomic status, including household income and employment status. Finally, the interpretation of PAR as the proportion of disease burden attributable to a factor (or a set of factors) is intended to estimate the percentage of disease that can be attributed to modifiable risk factors. However, PAR can greatly underestimate the proportion of disease burden that is etiologically related to the exposure.28 Moreover, many of the risk factors that were considered in our analysis are not necessarily “modifiable,” whereas other factors, including CAV and hospitalization for rejection, could be considered outcomes rather than risk factors. However, these variables have been associated with the risk of graft loss, so we favored including them in our model. Importantly, a significant proportion of disease risk is not accounted for by our models and remains to be understood. Still, our intent was to see to what extent risk factors affected the incidence of GF differently in diverse race/ethnic groups. We believe our analysis adds to the literature because it helps clarify which risk factors for GF might be targeted to improve transplant outcomes. Despite these limitations, our findings have many clinical implications. In 2006, the Diversity and Minority Affairs Committee of the American Society of Transplantation sponsored a symposium to examine disparities in transplantation for underserved and minority populations.29 Although the council noted that differences in socioeconomic status and barriers to access to care have an effect on disparities in outcomes, the council also stressed that the biologic mechanisms for outcome differences in diverse ethnic populations have been understudied and might provide novel and rational approaches for improvement in transplant outcomes. Thus, our findings that immunologic factors play a larger role in the development of GF in blacks and Hispanics argues that more research is needed that explores racial differences in the genetics and regulation of immune responses. For instance, prior data from kidney transplant recipients has suggested a more vigorous immune response in black patients partially related to the increased expression of costimulatory molecules that may contribute to enhanced T-cell activation.30 Furthermore, despite multiple observational studies confirming racial differences in pharmacogenetic profiles that influence tacrolimus metabolism, few clinical studies have incorporated this information in a prospective fashion nor have guidelines been issued that might guide transplant centers
how to systematically adopt this information in their clinical practice. To truly affect racial disparities in transplant outcomes, future research efforts must examine the biologic basis of outcome differences in order to well define alternative endpoints for clinical studies that may have greater applicability to diverse populations.29 In conclusion, we found race-related and ethnicity-related differences in the rate of incident GF and risk factor profiles for incident GF. Considering the inferior long-term survival in racial/ethnic minorities after HT and the worsening disparities in post-transplant outcomes,3 our study underscores the need for focused research and efforts attempting to eliminate these disparities and equalize post-transplant survival for all groups.
Disclosure statement None of the authors has a financial relationship with a commercial entity that has an interest in the subject of the presented manuscript or other conflicts of interest to disclose. This work was partly supported by funding from National Institutes of Health/National Heart, Lung and Blood Institute grant 1-U10-HL-110302 (J.B.), National Institutes of Health/National Institute of Nursing Research grant 5-P01-NR-011587-03 (J.B), and Health Resources and Services Administration contract 2342005-37011C. The content is the responsibility of the authors alone and does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the United States Government.
References 1. Stehlik J, Edwards LB, Kucheryavaya AY, et al. The Registry of the International Society for Heart and Lung Transplantation: 29th official adult heart transplant report—2012. J Heart Lung Transplant 2012;31:1052-64. 2. Allen JG, Weiss ES, Arnaoutakis GJ, et al. The impact of race on survival after heart transplantation: an analysis of more than 20,000 patients. Ann Thorac Surg 2010;89:1956-64. 3. Singh TP, Almond C, Givertz MM, Piercey G, Gauvreau K. Improved survival in heart transplant recipients in the United States: racial differences in era effect. Circ Heart Fail 2011;4:153-60. 4. Cohen O, De La Zerda D, Beygui RE, Hekmat D, Laks H. Ethnicity as a predictor of graft longevity and recipient mortality in heart transplantation. Transplant Proc 2007;39:3297-302. 5. Singh TP, Givertz MM, Semigran M, DeNofrio D, Costantino F, Gauvreau K. Socioeconomic position, ethnicity, and outcomes in heart transplant recipients. Am J Cardiol 2010;105:1024-9. 6. Morris AA, Cole RT, Veledar E, et al. Influence of race/ethnic differences in pre-transplantation panel reactive antibody on outcomes in heart transplant recipients. J Am Coll Cardiol 2013;62:2308-15. 7. Levin M. The occurrence of lung cancer in man. Acta Unio Int Contra Cancrum 1953;9:531-41. 8. Organ Procurement and Transplantation Network. Policies. Available at: http://optn.transplant.hrsa.gov/ContentDocuments/OPTN_Policies.pdf. Accessed August 15, 2014. 9. Young J, Naftel D, Bourge R, et al. Matching the heart donor and heart transplant recipient. Clues for successful expansion of the donor pool: a multivariable, multiinstitutional report. The Cardiac Transplant Research Database Group. J Heart Lung Transplant 1994;13:353-64. 10. Chen JM, Edwards NM. Management of high-risk donor. In: Edwards NM, Chen JM, Mazzeo PA, editors. Cardiac transplantation: the Columbia University Medical Center/New York-Presbyterian Hospital Manual. Totowa, NJ: Humana Press; 2004. p. 19-36.
Morris et al.
Association of Race/Ethnicity with Cardiac Graft Failure
11. Greenland S, Drescher K. Maximum likelihood estimation of the attributable fraction from logistic models. Biometrics 1993;49:865-72. 12. Kalogeropoulos A, Georgiopoulou V, Kritchevsky SB, et al. Epidemiology of incident heart failure in a contemporary elderly cohort: the health, aging, and body composition study. Arch Internal Med 2009;169:708-15. 13. Stehlik J, Edwards LB, Kucheryavaya AY, et al. The Registry of the International Society for Heart and Lung Transplantation: twentyeighth adult heart transplant report—2011. J Heart Lung Transplant 2011;30:1078-94. 14. Hong KN, Iribarne A, Worku B. Who Is the High-Risk Recipient? Predicting Mortality After Heart Transplant Using Pretransplant Donor and Recipient Risk Factors. The Annals of Thoracic Surgery 2011;92:520-7. 15. Naesens M, Kuypers DR, Sarwal M. Calcineurin inhibitor nephrotoxicity. Clin J Am Soc Nephrol 2009;4:481-508. 16. Khan UA, Williams SG, Fildes JE, Shaw SM. The pathophysiology of chronic graft failure in the cardiac transplant patient. Am J Transplant 2009;9:2211-6. 17. Mehra MR, Crespo-Leiro MG, Dipchand A, et al. International Society for Heart and Lung Transplantation working formulation of a standardized nomenclature for cardiac allograft vasculopathy—2010. J Heart Lung Transplant 2010;29:717-27. 18. Liu V, Bhattacharya J, Weill D, Hlatky MA. Persistent racial disparities in survival after heart transplantation. Circulation 2011;123:1642-9. 19. Kilic A, Weiss ES, Allen JG, et al. Simple score to assess the risk of rejection after orthotopic heart transplantation. Circulation 2012;125:3013-21. 20. Farmer SA, Grady KL, Wang E, McGee EC, Cotts WG, McCarthy PM. Demographic, psychosocial, and behavioral factors associated with survival after heart transplantation. Ann Thorac Surg 2013;95: 876-883.
831
21. Allen JG, Weiss ES, Arnaoutakis GJ, et al. Insurance and education predict long-term survival after orthotopic heart transplantation in the United States. J Heart Lung Transplant 2012;31:52-60. 22. Singh TP, Naftel DC, Addonizio L, et al. Association of race and socioeconomic position with outcomes in pediatric heart transplant recipients. Am J Transplant 2010;10:2116-23. 23. Oliva M, Singh TP, Gauvreau K, VanderPluym CJ, Bastardi HJ, Almond CS. Impact of medication non-adherence on survival after pediatric heart transplantation in the USA. J Heart Lung Transplant 2013;32:881-8. 24. Butkus DE, Meydrech EF, Raju SS. Racial differences in the survival of cadaveric renal allografts. N Engl J Med 1992;327:840-5. 25. Macphee IA, Fredericks S, Tai T, et al. Tacrolimus pharmacogenetics: polymorphisms associated with expression of cytochrome p4503A5 and P-glycoprotein correlate with dose requirement. Transplantation 2002;74:1486-9. 26. Staatz CE, Goodman LK, Tett SE. Effect of CYP3A and ABCB1 single nucleotide polymorphisms on the pharmacokinetics and pharmacodynamics of calcineurin inhibitors: part I. Clin Pharmacokinet 2010;49:141-75. 27. Kanter KR, Berg AM, Mahle WT, et al. Donor-recipient race mismatch and graft survival after pediatric heart transplantation. Ann Thorac Surg 2009;87:204-10. 28. Greenland S, Robins J. Conceptual problems in the definition and interpretation of attributable fractions. Am J Epidemiol 1988;128:1185-97. 29. Higgins RS, Fishman JA. Disparities in solid organ transplantation for ethnic minorities: facts and solutions. Am J Transplant 2006;6:2556-62. 30. Hutchings A, Purcell WM, Benfield MR. Increased costimulatory responses in African-American kidney allograft recipients. Transplantation 2001;71:692.
Race and ethnic differences in the epidemiology and risk factors for graft failure after heart transplantation Alanna A. Morris, MD,a Andreas P. Kalogeropoulos, MD, MPH, PhD,a Liping Zhao, MSPH,b Melissa Owen, RN,c S. Raja Laskar, MD,a J. David Vega, MD,a Andrew Smith, MD,a and Javed Butler, MD, MPHa From the Emory University aSchool of Medicine; bRollins School of Public Health; and the cSchool of Nursing, Atlanta, Georgia.
KEYWORDS: race/ethnicity; heart transplant; graft failure; cardiac allograft vasculopathy; population-attributable risk
BACKGROUND: Contemporary epidemiology of chronic graft failure (GF) after heart transplantation (HT) is not well described. Moreover, differences in the epidemiology of GF based on race/ethnicity remain poorly understood, despite clear evidence of inferior survival of ethnic minorities after HT. METHODS: The incidence of GF and the population-attributable risk (PAR) of independent risk factors for GF were assessed in 15,255 patients (76% men; mean age 52 ⫾ 12 years) who underwent primary HT from 2004 to 2012. RESULTS: During a median follow-up of 4.7 years (interquartile range, 2.3–7.1 years), GF developed in 2,926 patients (19.2%), corresponding to an incidence rate of 39.8/1,000 person-years (95% confidence interval, 38.4–41.3). Blacks were more likely to develop GF than Hispanics or whites, with incidence rates of 55.1, 42.2, and 36.5/1,000 person-years, respectively. After multivariable adjustment, black race was associated with a higher risk of GF (hazard ratio, 1.4; 95% confidence interval, 1.2–1.6; p o 0.001). Blacks and Hispanics were more likely to have risk factors for GF, including low education, public insurance, allosensitization, higher human leukocyte antigen mismatch, non-adherence, and history of rejection requiring hospitalization (all p o 0.001). Rejection requiring hospitalization carried the highest population-attributable risk in all groups, with the highest fraction in blacks (25.8%) compared with whites (18.6%) and Hispanics (15.6%). Socioeconomic and donor risk factors conferred relatively less risk of GF. CONCLUSIONS: Black HT recipients have the highest risk of GF, with immunologic factors conferring the greatest proportion of that risk. Racial differences in risk factors for GF after HT require further study. J Heart Lung Transplant 2015;34:825–831 Published by Elsevier Inc.
Heart transplantation (HT) remains the therapy of choice for select patients with end-stage heart failure. Median conditional survival now exceeds 10 years for recipients Reprint requests: Alanna A. Morris, MD, 1462 Clifton Rd Ste 520, Atlanta GA 30322. Telephone: 617-480-9262. Fax: 404-712-0149. E-mail address: [email protected] 1053-2498/$ - see front matter Published by Elsevier Inc. http://dx.doi.org/10.1016/j.healun.2014.12.012
who survive the first year after HT.1 However, approximately 20% of deaths after the first year are due to graft failure (GF), which often results from processes that include cellular or antibody-mediated rejection and cardiac allograft vasculopathy (CAV). Despite this overall trend toward improved outcomes after HT, disparities in survival based on race/ethnicity are
826
The Journal of Heart and Lung Transplantation, Vol 34, No 6, June 2015
well recognized. Several studies have demonstrated shorter graft and patient survival in HT recipients who are of racial/ ethnic minority background.2–4 The exact mechanisms for these disparities remain unclear; however, suggested explanations include socioeconomic, biological, and immunologic factors.5,6 Understanding the epidemiology of GF, including the influence of risk factors associated with GF, are imperative to further improve outcomes and bridge disparities after HT. The population-attributable risk (PAR) is often used to describe the proportion of disease risk in a population that can be attributed to the causal effects of a risk factor or set of factors.7 Thus, our objective for this study was to assess the epidemiology of incident GF by examining the race-related risk factor profile and PAR for GF in a contemporary cohort of HT recipients.
Methods Study population All individuals aged Z18 years who were listed for their first HT were identified in the Organ Procurement and Transplantation Network (OPTN) database, which includes deidentified data on all patients listed for a HT in the United States. Analysis was limited to patients listed between June 30, 2004 (when follow-up information on graft status was first reported to the OPTN) and September 2012. The Health Resources and Services Administration and the United States Department of Health and Human Services provide oversight to the activities of the OPTN contractor, the United Network of Organ Sharing. We compared baseline characteristics of patients listed for primary HT during the study period, excluding patients listed for a repeat HT or for multiorgan transplantation. Race was reported by the transplantation centers as white, black, Hispanic/Latino, Asian, American Indian/Alaska Native, Native Hawaiian/Pacific Islander, multiracial, or other. Ethnicity was reported as Hispanic or nonHispanic. For all Hispanic patients in the study, race and ethnicity variables were identical; therefore, all white patients in this analysis were non-Hispanic white, and all black patients were non-Hispanic black. Because of the small sample size of minority patients with race/ethnicity other than black or Hispanic (o4.5% of the sample), these patients were excluded.
Study definitions Incident graft failure Follow-up data on HT recipients is provided to the OPTN 6 months after transplant and then annually thereafter. On the follow-up form that is reported annually to the OPTN, transplant centers are asked to document allograft status as “functioning” or “failed.” GF is defined as having occurred when “an organ is removed, a recipient dies, or a recipient is placed on a chronic allograft support system.”8 We identified patients with GF that occurred at least 180 days after transplant to attempt to exclude patients with primary allograft dysfunction. Thus, GF includes patients living with allograft dysfunction, patients listed for repeat transplantation, and patients who died or did not have further information regarding their outcome of GF. Patients were monitored from the date of transplant until death, retransplantation, or date of last known
follow-up provided by United Network of Organ Sharing, with follow-up through September 2012.
Risk factors for incident graft failure Clinical, socioeconomic, and immunologic variables were defined at the time of listing for waiting list factors, whereas donor-related variables were defined at the time of transplant, and posttransplantation outcomes were defined at the time of follow-up. Variables compared between race/ethnic groups included clinical (recipient age, gender, creatinine), socioeconomic factors (education, primary insurance payer), immunologic (allosensitization, human leukocyte antigen [HLA] mismatch, non-adherence to medications, any episodes of rejection requiring hospitalization, CAV), and donor (donor age, ischemic time) factors. Transplant centers are asked to document “Was there evidence of noncompliance with immunosuppression medication during this follow-up period that compromised the patient’s recovery?” on the annual follow-up form. For the purposes of our analysis, any report of “yes” to this question during the total follow-up period for any patient was considered “non-adherence.”
Risk factor definitions for calculation of PAR For the PAR calculation, continuous predictors that were univariate predictors of GF were dichotomized using clinically relevant cutoff points. Age was dichotomized at the median age of the cohort, education at high school level or less, insurance as public (Medicare, Medicaid, Veterans Affairs) vs private, allosensitization as panel reactive antibody (PRA) level Z10%, HLA mismatch at Z 5 loci based on the cohort median, ischemic time as Z4 hours,9 and donor age as Z30 years.10 Medication non-adherence, history of rejection requiring hospitalization, and presence of CAV were collapsed into binary predictors (yes or no).
Statistical analysis Continuous variables were compared using the Mann-Whitney rank sum test, and categorical variables were compared using the chi-square test. Cumulative event rates were obtained using the Kaplan-Meier method and were compared using the log-rank statistic. Univariate and adjusted hazard ratios (HRs) and the corresponding 95% confidence intervals (CIs) were obtained using Cox proportional hazards regression models to identify significant predictors of GF. The proportional hazards assumption was tested and verified for all risk factors using Schoenfeld residual correlation analysis. Subsequently, race/ethnicity-stratified analysis was performed to obtain rate ratio (RR) estimates for the significant risk factors. We also calculated adjusted (multivariable) PARs using a Poisson regression model with incident GF as the outcome and the factors already described as predictors.11,12 Briefly, the predicted number of cases is calculated for the full model (Nfull), which equals the actual number of cases. Next, the effect of the risk factor of interest is “removed” by setting the value of the covariable to 0, and the predicted number of events is calculated (Nremoved). The adjusted PAR for the risk factor then becomes PARadjusted ¼ 1 (Nremoved/Nfull). PARs are not presented if the adjusted RR and corresponding 95% CI for the risk factor failed to reach a p-value of 0.2. Age and gender were included in regression models for multivariable PAR calculation, but the PARs of these variables were omitted from the tables because they cannot be
Morris et al.
Association of Race/Ethnicity with Cardiac Graft Failure
modified. Data were analyzed with SAS 9.3 software (SAS Institute Inc, Cary, NC).
827
levels of creatinine, PRA, and HLA mismatch. Hispanic HT recipients had the youngest median donor age, and black recipients had the shortest allograft ischemic time.
Results Risk factors for incident GF
Study population A total of 51,954 patients were listed for HT after June 30, 2004. Of these, 15,972 HT recipients had at least 180 days of follow-up with complete information on graft status and episodes of rejection. An additional 717 recipients were excluded for race/ethnicity other than white, black, or Hispanic, leaving 15,255 HT recipients for this analysis.
Clinical characteristics The pre-transplant characteristics of the 15,255 HT recipients who formed our analytic cohort are reported in Table 1. Compared with whites, blacks and Hispanics were younger, had a higher proportion of women, and were more likely to have a non-ischemic heart failure etiology. Blacks and Hispanics were less educated and were less likely to have private insurance. Black HT recipients had higher
During a median follow-up of 4.7 years (interquartile range, 2.3–7.1 years), 2,926 HT recipients (19.2%) developed GF (39.8/1,000 person-years; 95% CI, 38.4–41.3). The median time to GF was 4.18 years. In the entire cohort, age r44 years (HR, 1.2; 95% CI, 1.1–1.3; p o 0.001), high school education or less (HR, 1.3; 95% CI, 1.2–1.4; p o 0.001), public insurance (HR, 1.3; 95% CI, 1.2–1.4; p o 0.001), allosensitization (HR, 1.2; 95% CI, 1.1–1.3; p ¼ 0.001), HLA mismatch Z5 loci, (HR, 1.1; 95% CI, 1.0–1.2; p ¼ 0.01), history of rejection requiring hospitalization (HR, 2.6; 95% CI, 2.4–2.8; p o 0.001), history of non-adherence (HR, 2.3; 95% CI, 2.1–2.6; p o 0.001), ischemic time Z 4 hours (HR, 1.1; 95% CI, 1.0–1.2; p o 0.05), and donor age Z30 years (HR, 1.2; 95% CI, 1.2–1.3; p o 0.001) were associated with an increased risk of GF. There was a significant interaction of CAV and age (p ¼ 0.0006). CAV was associated with a lower risk of incident GF in patients
Table 1 Baseline Characteristics of White, Black, and Hispanic Heart Transplant Recipients Who Underwent Heart Transplantation From 2004 to 2012
Characteristics Age, years Female gender Heart failure etiology Dilated Ischemic Restrictive Hypertrophic Valvular Congenital Other Creatinine, mg/dl Education High school or less Some college College or graduate Insurance Private Medicare Medicaid Other PRA, % HLA mismatch, No. Ischemic time, hours Donor age, years Follow-up time, years
Whites
Blacks
(n ¼ 11,465)
(n ¼ 2,653)
56 (47–61) 2,454 (21) 4,555 5,586 224 268 261 315 254 1.2
(40) (49) (2) (2) (2) (3) (2) (1.0–1.4)
Hispanics (n ¼ 1,137) a,b
49 (39–57) 908 (34) 1,923 540 42 22 55 26 45 1.2
(73) (20) (1) (1) (2) (1) (2) (1.0–1.6)a,b
p-value c
51 (41–59) 289 (25) 625 406 16 21 35 22 12 1.1
(55) (36) (1) (2) (3) (2) (1) (0.9–1.4)c
4,208 (37) 2,431 (21) 2,667 (23)
1,230 (46) 607 (23) 347 (13)
597 (53) 174 (15) 121 (11)
7,217 2,849 915 483 0 5 3.2 29 5.0
1,246 726 533 148 0 5 3.1 29 4.0
459 279 278 121 0 5 3.2 26 4.1
o0.001 o0.001 o0.001
o0.001 o0.001
o0.001 (63) (25) (8) (4) (0–4) (4,5) (2.5–3.9) (21–41) (2.6–7.6)
HLA, human leukocyte antigen; PRA, panel reactive antibody. Data are presented as median (25th–75th percentile) or number (%). a p o 0.001 for white-black comparison. b p r 0.001 for black-Hispanic comparison. c p o 0.001 for white-Hispanic comparison.
(47) (27) (20) (6) (0–11)a,b (4–6)a,b (2.5–3.8)a (21–40)b (2.0–6.4)a
(40) (25) (25) (10) (0–4) (4–5) (2.4–3.9) (20–38)c (2.1–6.9)c
o0.001 o0.001 o0.001 o0.001 o0.001
(0.6–1.3) (0.7–1.4) (0.8–2.1) (0.7–1.5) (1.3–3.8)b (1.3–2.8)b (0.6–1.3) (0.7–1.7) (1.1–2.1)b 1.2 1.1 1.2 1.2 2.1 2.5 0.9 1.0 1.3 45 37 19 57 3 14 28 21 49 rHigh school education Public insurance PRA Z10% HLA mismatch Z5 loci Non-adherence Rejection requiring hospitalization Cardiac allograft vasculopathy Ischemic time Z4 hours Donor age Z30 years
CI, confidence interval; HLA, human leukocyte antigen; PAR, population-attributable risk; PRA, panel reactive antibody; RR, rate ratio. RRs are adjusted for age, gender, and the risk factors shown. a p r 0.001. b p r 0.05. c p o 0.2.
0.9 1.0 1.3 1.1 2.2 1.9 0.9 1.1 1.5 67 60 18 61 4 19 25 22 43 … … 5.3 … 7.6 25.8 … 3.6 5.7 (1.1–1.4) (1.0–1.3)b (1.0–1.3)b (1.1–1.3)b (1.7–2.5)a (2.2–2.8)a (0.8–1.0)b (0.9–1.1) (1.1–1.4)a
9.0 5.2 2.5 9.1 4.9 18.6 … … 10.6
56 53 26 68 6 24 26 18 48
1.1 1.0 1.3 0.9 1.8 2.2 0.8 1.2 1.1
(0.9–1.3) (0.8–1.2) (1.0–1.6)b (0.7–1.1) (1.4–2.4)a (1.8–2.7)a (0.7–1.0)b (0.9–1.5)c (0.9–1.4)c
Prevalence (%) PAR (%) RR (95% CI) Prevalence (%) PAR (%)
Blacks (n ¼ 2,653)
a
RR (95% CI) Prevalence (%)
Figure 1 Incident graft failure adjusted for age by race/ ethnicity. Graft failure was more likely to develop in black and Hispanic heart transplant recipients than in white recipients (stratified log-rank chi-square ¼ 105; p o 0.0001 for race).
Whites (n ¼ 11,465)
The prevalences and the adjusted (multivariable) PARs for the risk factors for GF are reported in Table 2. Low education level, public insurance, allosensitization, HLA mismatch at Z5 loci, reported non-adherence, and history of rejection requiring hospitalization were more prevalent in black and Hispanic HT recipients than in white recipients (p o 0.001 for all). In contrast, CAV (p ¼ 0.02) and older donor age (p ¼ 0.001) were more prevalent among white and black HT recipients, whereas longer ischemic time
Prevalence, Adjusted Rate Ratios, and Population-Attributable Risk of Risk Factors for Incident Graft Failure
Race-ethnic differences in the prevalence of risk factors for GF and PARs
Table 2
Kaplan-Meier estimates of incident GF in sub-groups based on race and ethnicity are displayed in Figure 1. Black recipients were more likely to develop GF than Hispanic and white recipients. The rates of GF were 55.1/1,000 person-years in black recipients (age-adjusted HR, 1.6; 95% CI, 1.4–1.7; p o 0.001) and 42.2/1,000 person-years in Hispanic recipients (age-adjusted HR, 1.2; 95% CI, 1.0–1.4; p ¼ 0.015) compared with 36.5/1,000 person-years in white participants. After adjustment for all risk factors in Table 1, black race was associated with an increased risk of GF (HR,1.4; 95% CI, 1.2– 1.6; p o 0.001). The median (interquartile range) ejection fraction was preserved among patients who developed GF: blacks, 57% (50%–60%); Hispanics, 60% (55%–64%); and whites, 59% (55%–61%; p o 0.001). However, among patients with GF, blacks and Hispanics were more likely to have an ejection fraction of r35% compared with whites (50 [8%], 9 [6.5%], 79 [3.8%]; p o 0.001).
Hispanics (n ¼ 1,137)
Differences in incident GF among race and ethnic sub-groups
RR (95% CI)
PAR (%)
with age 444 years (HR, 0.8; 95% CI, 0.8–0.9), but CAV was not associated with incident GF in patients aged r44 years (HR, 1.11; 95% CI, 0.96–1.3).
… … … … 6.5 15.6 … … 16.2
The Journal of Heart and Lung Transplantation, Vol 34, No 6, June 2015
Risk factor
828
Morris et al.
Association of Race/Ethnicity with Cardiac Graft Failure
(p ¼ 0.002) was more prevalent among white and Hispanic recipients. Rejection requiring hospitalization and non-adherence conferred the highest adjusted PARs in each race/ethnic group, with slightly higher PARs in black HT recipients compared with whites and Hispanics. Adjusted PARs for older donor age were higher in white and Hispanic HT recipients than in black recipients, whereas allosensitization was higher in black HT recipients than in white recipients. Although the prevalence of low education, public insurance, and HLA mismatch Z5 loci was higher in black and Hispanic HT recipients, the adjusted PAR for these factors was higher in white recipients.
Discussion In the present study, we found that the rate of incident GF is higher in black and Hispanic compared with white HT recipients. In addition, we found significant race-related differences in the risk factor profiles for incident GF. Although race/ethnic disparities in GF are multifactorial in their etiology, our findings add to the current literature on this topic because there is scant existing data that explain which risk factors for GF have the largest contribution to the observed disparities. Our data demonstrate that immunologic factors confer the greatest risk for incident GF for all race/ethnic groups, with a slightly higher PAR in black HT recipients. In our cohort, socioeconomic factors had a greater contribution to the development of GF in whites than in black and Hispanic HT recipients, whereas donor factors had a small effect on the risk of GF in all recipients. Chronic GF after HT may present as a wide spectrum of clinical or sub-clinical allograft dysfunction, including right ventricular, left ventricular, or biventricular dysfunction. Data from the registry of the International Society of Heart and Lung Transplant estimate that approximately 23% of recipients have developed GF 3 to 5 years after HT,13 similar to the rate of incident GF seen in our cohort. GF likely results from processes such as cellular and antibodymediated rejection and CAV; however, more research is needed to fully determine the mechanisms that contribute to chronic graft injury. Prior studies have examined individual risk factors for GF, including donor and recipient variables. Hong et al14 examined data of 11,703 HT recipients who received transplants between 2001 and 2007. Similar to our findings, these authors identified advanced donor and recipient age as well as prolonged donor ischemic time as risk factors for GF at 1 year. They also found poor recipient renal function was associated with the risk of GF at 1 year. In our cohort, pretransplant creatinine was not identified as an independent risk factor for incident GF. However, the incidence of chronic GF, as opposed to GF at 1 year, may be more strongly influenced by the development of chronic renal impairment related to nephrotoxicity and hypertension as a result of the use of calcineurin-inhibitors.15,16 CAV was associated with a lower risk of GF in older patients in our cohort. CAV that develops more than 2 years after transplant can be associated with a more indolent
829
course17 and may be associated with a better prognosis in the setting of preserved graft function than CAV with systolic dysfunction. Race-specific differences in the development of GF have been well described, with analyses consistently demonstrating worse long-term outcomes in blacks and Hispanics, even after adjustment for the higher clinical risk profile often seen in these groups pre-transplantation.3,6,18 We found that most of the socioeconomic and immunologic risk factors associated with GF were more common in blacks and Hispanics. Immunologic risk factors carried the highest risk of GF in all race/ethnic groups; however, immunologic factors played a larger role in the development of GF in blacks and Hispanics. Prior studies have documented higher levels of PRA and degrees of HLA mismatch in black and Hispanic HT recipients, leading to higher risk for acute rejection and allograft loss.2,6,19 Prior studies have confirmed the effect of socioeconomic status on the subsequent risk of graft loss, with lower household income, Medicare and Medicaid insurance, and less education all associated with worse outcomes.5,20,21 In our cohort, socioeconomic status, specifically lower education and public insurance, conferred a higher risk of GF in white HT recipients. Singh et al22 reported similar findings in an analysis of 490 pediatric HT recipients. In their study, low socioeconomic position was associated with a higher risk of graft loss among white children; however, socioeconomic status did not risk-stratify black children despite a higher prevalence of low socioeconomic position among blacks. These findings may be related to candidate selection, because whites of low socioeconomic position may also represent a higher risk group where social determinants of health amplify the biologic mechanisms that contribute to graft loss after HT. Few analyses have collected detailed information on medication non-adherence because this variable depends on multiple factors, including the intensity of the medication regimen, socioeconomic and family environment, the degree of education provided by health care providers, and other comorbid conditions. Although the OPTN database lacks detailed information on non-adherence, this was the only truly modifiable risk factor we examined, and it did carry a higher PAR in blacks and Hispanics than in whites. Prior data have shown higher reported rates of non-compliance among race/ethnic minorities, particularly blacks, compared with whites.18,23 However, medication non-adherence is often closely associated with socioeconomic status, with prior data showing that a lack of private insurance often decreases adherence to immunosuppressive therapy.24 Moreover, we lacked information on race/ethnic differences in genotypes that contribute to higher clearance and lower bioavailability of immunosuppressants. Black and Hispanic HT recipients are more likely to be “rapid metabolizers” of tacrolimus,25,26 placing these patients at higher risk of rejection after missing even a few doses or as being incorrectly classified as non-adherent if they are being underdosed. This analysis has some limitations. Risk factors that could have influenced the incidence of GF and/or the
830
The Journal of Heart and Lung Transplantation, Vol 34, No 6, June 2015
frequency of hospitalizations for rejection, such as positive retrospective crossmatch or the intensity of maintenance immune suppression, were not available in the OPTN database. Prior data have shown a slightly higher incidence of positive retrospective crossmatch in black HT recipients compared with whites.27 However, this variable was not associated with overall graft survival in adjusted models. In addition, a recent analysis of OPTN data has shown the proportion of white, black, and Hispanic HT recipients receiving mycophenolate was similar in recent eras, with a higher proportion of black recipients (vs whites) receiving tacrolimus.3 Thus, racial differences in graft survival are not likely to be due to disparities in the choice of immune suppression. We also lacked information on other factors that strongly influence socioeconomic status, including household income and employment status. Finally, the interpretation of PAR as the proportion of disease burden attributable to a factor (or a set of factors) is intended to estimate the percentage of disease that can be attributed to modifiable risk factors. However, PAR can greatly underestimate the proportion of disease burden that is etiologically related to the exposure.28 Moreover, many of the risk factors that were considered in our analysis are not necessarily “modifiable,” whereas other factors, including CAV and hospitalization for rejection, could be considered outcomes rather than risk factors. However, these variables have been associated with the risk of graft loss, so we favored including them in our model. Importantly, a significant proportion of disease risk is not accounted for by our models and remains to be understood. Still, our intent was to see to what extent risk factors affected the incidence of GF differently in diverse race/ethnic groups. We believe our analysis adds to the literature because it helps clarify which risk factors for GF might be targeted to improve transplant outcomes. Despite these limitations, our findings have many clinical implications. In 2006, the Diversity and Minority Affairs Committee of the American Society of Transplantation sponsored a symposium to examine disparities in transplantation for underserved and minority populations.29 Although the council noted that differences in socioeconomic status and barriers to access to care have an effect on disparities in outcomes, the council also stressed that the biologic mechanisms for outcome differences in diverse ethnic populations have been understudied and might provide novel and rational approaches for improvement in transplant outcomes. Thus, our findings that immunologic factors play a larger role in the development of GF in blacks and Hispanics argues that more research is needed that explores racial differences in the genetics and regulation of immune responses. For instance, prior data from kidney transplant recipients has suggested a more vigorous immune response in black patients partially related to the increased expression of costimulatory molecules that may contribute to enhanced T-cell activation.30 Furthermore, despite multiple observational studies confirming racial differences in pharmacogenetic profiles that influence tacrolimus metabolism, few clinical studies have incorporated this information in a prospective fashion nor have guidelines been issued that might guide transplant centers
how to systematically adopt this information in their clinical practice. To truly affect racial disparities in transplant outcomes, future research efforts must examine the biologic basis of outcome differences in order to well define alternative endpoints for clinical studies that may have greater applicability to diverse populations.29 In conclusion, we found race-related and ethnicity-related differences in the rate of incident GF and risk factor profiles for incident GF. Considering the inferior long-term survival in racial/ethnic minorities after HT and the worsening disparities in post-transplant outcomes,3 our study underscores the need for focused research and efforts attempting to eliminate these disparities and equalize post-transplant survival for all groups.
Disclosure statement None of the authors has a financial relationship with a commercial entity that has an interest in the subject of the presented manuscript or other conflicts of interest to disclose. This work was partly supported by funding from National Institutes of Health/National Heart, Lung and Blood Institute grant 1-U10-HL-110302 (J.B.), National Institutes of Health/National Institute of Nursing Research grant 5-P01-NR-011587-03 (J.B), and Health Resources and Services Administration contract 2342005-37011C. The content is the responsibility of the authors alone and does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the United States Government.
References 1. Stehlik J, Edwards LB, Kucheryavaya AY, et al. The Registry of the International Society for Heart and Lung Transplantation: 29th official adult heart transplant report—2012. J Heart Lung Transplant 2012;31:1052-64. 2. Allen JG, Weiss ES, Arnaoutakis GJ, et al. The impact of race on survival after heart transplantation: an analysis of more than 20,000 patients. Ann Thorac Surg 2010;89:1956-64. 3. Singh TP, Almond C, Givertz MM, Piercey G, Gauvreau K. Improved survival in heart transplant recipients in the United States: racial differences in era effect. Circ Heart Fail 2011;4:153-60. 4. Cohen O, De La Zerda D, Beygui RE, Hekmat D, Laks H. Ethnicity as a predictor of graft longevity and recipient mortality in heart transplantation. Transplant Proc 2007;39:3297-302. 5. Singh TP, Givertz MM, Semigran M, DeNofrio D, Costantino F, Gauvreau K. Socioeconomic position, ethnicity, and outcomes in heart transplant recipients. Am J Cardiol 2010;105:1024-9. 6. Morris AA, Cole RT, Veledar E, et al. Influence of race/ethnic differences in pre-transplantation panel reactive antibody on outcomes in heart transplant recipients. J Am Coll Cardiol 2013;62:2308-15. 7. Levin M. The occurrence of lung cancer in man. Acta Unio Int Contra Cancrum 1953;9:531-41. 8. Organ Procurement and Transplantation Network. Policies. Available at: http://optn.transplant.hrsa.gov/ContentDocuments/OPTN_Policies.pdf. Accessed August 15, 2014. 9. Young J, Naftel D, Bourge R, et al. Matching the heart donor and heart transplant recipient. Clues for successful expansion of the donor pool: a multivariable, multiinstitutional report. The Cardiac Transplant Research Database Group. J Heart Lung Transplant 1994;13:353-64. 10. Chen JM, Edwards NM. Management of high-risk donor. In: Edwards NM, Chen JM, Mazzeo PA, editors. Cardiac transplantation: the Columbia University Medical Center/New York-Presbyterian Hospital Manual. Totowa, NJ: Humana Press; 2004. p. 19-36.
Morris et al.
Association of Race/Ethnicity with Cardiac Graft Failure
11. Greenland S, Drescher K. Maximum likelihood estimation of the attributable fraction from logistic models. Biometrics 1993;49:865-72. 12. Kalogeropoulos A, Georgiopoulou V, Kritchevsky SB, et al. Epidemiology of incident heart failure in a contemporary elderly cohort: the health, aging, and body composition study. Arch Internal Med 2009;169:708-15. 13. Stehlik J, Edwards LB, Kucheryavaya AY, et al. The Registry of the International Society for Heart and Lung Transplantation: twentyeighth adult heart transplant report—2011. J Heart Lung Transplant 2011;30:1078-94. 14. Hong KN, Iribarne A, Worku B. Who Is the High-Risk Recipient? Predicting Mortality After Heart Transplant Using Pretransplant Donor and Recipient Risk Factors. The Annals of Thoracic Surgery 2011;92:520-7. 15. Naesens M, Kuypers DR, Sarwal M. Calcineurin inhibitor nephrotoxicity. Clin J Am Soc Nephrol 2009;4:481-508. 16. Khan UA, Williams SG, Fildes JE, Shaw SM. The pathophysiology of chronic graft failure in the cardiac transplant patient. Am J Transplant 2009;9:2211-6. 17. Mehra MR, Crespo-Leiro MG, Dipchand A, et al. International Society for Heart and Lung Transplantation working formulation of a standardized nomenclature for cardiac allograft vasculopathy—2010. J Heart Lung Transplant 2010;29:717-27. 18. Liu V, Bhattacharya J, Weill D, Hlatky MA. Persistent racial disparities in survival after heart transplantation. Circulation 2011;123:1642-9. 19. Kilic A, Weiss ES, Allen JG, et al. Simple score to assess the risk of rejection after orthotopic heart transplantation. Circulation 2012;125:3013-21. 20. Farmer SA, Grady KL, Wang E, McGee EC, Cotts WG, McCarthy PM. Demographic, psychosocial, and behavioral factors associated with survival after heart transplantation. Ann Thorac Surg 2013;95: 876-883.
831
21. Allen JG, Weiss ES, Arnaoutakis GJ, et al. Insurance and education predict long-term survival after orthotopic heart transplantation in the United States. J Heart Lung Transplant 2012;31:52-60. 22. Singh TP, Naftel DC, Addonizio L, et al. Association of race and socioeconomic position with outcomes in pediatric heart transplant recipients. Am J Transplant 2010;10:2116-23. 23. Oliva M, Singh TP, Gauvreau K, VanderPluym CJ, Bastardi HJ, Almond CS. Impact of medication non-adherence on survival after pediatric heart transplantation in the USA. J Heart Lung Transplant 2013;32:881-8. 24. Butkus DE, Meydrech EF, Raju SS. Racial differences in the survival of cadaveric renal allografts. N Engl J Med 1992;327:840-5. 25. Macphee IA, Fredericks S, Tai T, et al. Tacrolimus pharmacogenetics: polymorphisms associated with expression of cytochrome p4503A5 and P-glycoprotein correlate with dose requirement. Transplantation 2002;74:1486-9. 26. Staatz CE, Goodman LK, Tett SE. Effect of CYP3A and ABCB1 single nucleotide polymorphisms on the pharmacokinetics and pharmacodynamics of calcineurin inhibitors: part I. Clin Pharmacokinet 2010;49:141-75. 27. Kanter KR, Berg AM, Mahle WT, et al. Donor-recipient race mismatch and graft survival after pediatric heart transplantation. Ann Thorac Surg 2009;87:204-10. 28. Greenland S, Robins J. Conceptual problems in the definition and interpretation of attributable fractions. Am J Epidemiol 1988;128:1185-97. 29. Higgins RS, Fishman JA. Disparities in solid organ transplantation for ethnic minorities: facts and solutions. Am J Transplant 2006;6:2556-62. 30. Hutchings A, Purcell WM, Benfield MR. Increased costimulatory responses in African-American kidney allograft recipients. Transplantation 2001;71:692.