Abstract
Background
There is continued and significant debate regarding the salient etiologies associated with graft loss and racial disparities in kidney transplant (KTX) recipients.
Methods
This was a longitudinal cohort study of all adult KTX recipients, comparing patients with early graft loss (<5 yrs) to those with graft longevity (surviving graft with at least 5 yrs of follow-up) across racial cohorts (African-American (AA) and non-AA) to discern risk factors.
Results
524 patients were included, 55% AA, 151 with early graft loss (29%) and 373 with graft longevity (71%). Consistent within both races, early graft loss was significantly associated with disability income (adjusted odds-ratio [AOR] 2.2, 95% CI: 1.1-4.5), kidney donor risk index (AOR 3.2, 1.4-7.5), rehospitalization (AOR 2.1, 1.0-4.4) and acute rejection (AOR 4.4, 1.7-11.6) Unique risk factors in AAs included Medicare only insurance (AOR 8.0, 2.3-28) and BK infectio (AOR 5.6, 1.3-25). Unique protective factors in AAs included cardiovascular risk factor control: AAs with a mean systolic BP <150 mmHg had 80% lower risk of early graft loss (AOR 0.2, 0.1-0.7), while LDL <100 mg/dL (AOR 0.4, 0.2-0.8), triglycerides <150 mg/dL (AOR 0.4, 0.2-1.0) and HgbA1C <7% (AOR 0.2, 0.1-0.6) were also protective against early graft loss in AA, but no in non-AA recipients.
Conclusions
AA recipients have a number of unique risk factors for early graft loss, suggesting that controlling cardiovascular comorbidities may be an important mechanism to reduce racial disparities in kidney transplantation.
Keywords: Kidney transplantation, African-American, graft loss, cardiovascular disease
Introduction
Based on the most recent scientific registry of transplant recipients (SRTR) data, African-American (AA) kidney transplant recipients have a significantly higher risk of graft loss at five years post-transplant.1 Despite over 35 years of focused research endeavors into this disparity, little has changed with this racial inequality. In the landmark 1977 analysis exposing this, Opelz et al demonstrated a 10% absolute lower rate in three-year graft survival rates in AA recipients (25% vs. 35%, p<0.001).2,3 Thirty-five years later, it is stubbornly similar; the 2012 SRTR report demonstrated a 5-year absolute difference of 12% in graft survival rates between AA and non-AA patients.1 This inequality has often been attributed to immunologic risks leading to higher acute rejection rates,4-7 socioeconomic status (SES),8-9 medication non-adherence10,11 and more aggressive/poorly controlled cardiovascular (CV) risk factors.12-14
Despite a large number of studies focusing on the salient etiologies associated with the higher rates of graft loss in AA kidney transplant recipients, there continues to be significant debate into which areas to focus efforts to improve outcomes in this high-risk population.15-18 This is likely due to the contradicting and inconsistent studies published in this area of research.19-22 Large national analyses, using registry data, lack SES and clinical data to determine true causal relationships with a number of important risks; while smaller single-center experiences usually lack sample-size or long-term follow-up to draw meaningful conclusions.23 Additionally, to date, there have been limited studies analyzing the effects of SES,9-11,18 CV comorbidities and disease state management on racial disparities in transplantation.12-15 Our transplant program is high-volume (∼200 kidney transplants per year) and contains a significant number of AA recipients (∼55% of kidney cohort), thus having an ideal and unique set of characteristics to study racial disparities in a meaningful and clinical relevant manner. Therefore, the aim of this study was to use a large cohort of patients with detailed comprehensive data and long-term follow-up to determine if there are racial differences with the established factors associated with early graft loss in kidney transplant recipients.
Materials and Methods
Study Design and Patients
This was an IRB-approved longitudinal cohort study of renal recipients that underwent transplant between Jan 2005 and Dec 2012; 2005 was chosen as the start of this study because this is when our program began utilizing contemporary immunosuppression (tacrolimus and induction therapy) and flow cytometry in the histocompatibility lab. For the purposes of this study, risk and control cohorts were developed for comparison based on two primary factors: race and graft loss. Initially, groups were apportioned based on self-reported race: African-American versus non-African-American. Subsequently, the groups were established based on graft outcomes: the control cohort consisted of patients with ongoing graft function and at least five years of post-transplant follow-up (graft longevity cohort). The risk cohort consisted of patients with graft loss within five years post-transplant (early graft loss cohort). Patients with ongoing graft survival but follow-up of less than five-years and patients with graft loss after five years were excluded from this study. The end analysis compared four unique cohorts (patients with graft longevity vs. those with early graft loss separately between African-American and non-African-American patients). Thus, in the end, this comparison produced a strong juxtaposition in patient types in which to compare risk factors for graft loss and do so across racial cohorts. Additional exclusion criteria included age less than 18 years at the time of transplant, recipients of non-renal solid organ transplants, or those that were lost to follow-up with inadequate clinical data available for analysis. Data was collected at baseline, from time of kidney transplant, to death, loss to follow-up or end of the study (July 2013).
Outcome Measures
The primary outcome for this study was to determine the significant factors associated with early graft loss and compare these between African-American and non-African-American recipients. The secondary outcomes of this study were to use multivariate models to determine independent variables that significantly influence early graft loss and compare these models across race. The final goal of the study was to determine the correlation and predictive ability of these models on determining early graft loss and compare this between African-American and non-African-American patients.
Data Variables and Definitions
All data was collected in a retrospective longitudinal manner from electronic and paper medical records. Initially, detailed baseline recipient and donor sociodemographics and transplant characteristics were recorded from the past medical history taken at the time of evaluation and updated at the time of transplant. All documented clinical post-transplant events were captured, including acute rejections, graft failures, and deaths. Laboratory measurements, vital signs, and clinical assessments were recorded daily for the first week post-transplant, then weekly for one month, then at three and six months, and finally biannually thereafter. For analyses, values were aggregated for each year independently, and then averaged across the entire follow-up time period for each patient. Graft failure was defined as a return to chronic dialysis or death. Delayed graft function was defined as the need for dialysis within 7 days following transplantation. GFRs were estimated using the 4-variable MDRD equation. The Kidney Donor Risk Index (KDRI) is a risk index which combines a number of donor factors into an aggregate score. The KDRI objectively expresses the relative risk of the graft failure for a given donor compared to the median kidney donor from the previous year. CMV infection was defined as the presence of CMV viremia of at least 2,000 copies/mL or any CMV viremia with signs and symptoms consistent with infection. BK infection was defined as BK viremia of at least 2,000 copies/mL or biopsy proven BK nephropathy. Per institutional protocol, all PCRs were repeated for confirmation. Acute rejection was defined as biopsy proven and at least Banff grade of 1A. Antibody mediated rejection was biopsy proven with C4D+ staining, the presence of DSA and graft dysfunction.
Immunosuppression and Anti-Infective Prophylactic Regimens
The majority of patients received induction therapy with either thymoglobulin 1.5 mg/kg IV daily for 3 to 5 doses, daclizumab 1 mg/kg IV on day 0 and day 7 post-transplant, or basiliximab 20 mg IV on day 0 and day 4 post-transplant. Choice of induction therapy was based on protocols that utilized thymoglobulin for high immunologic risk patients (re-transplantation, CIT >24 hours, or PRA >20%), and an IL2 receptor antibody in most other patients. Maintenance immunosuppression consisted of tacrolimus, with dose adjustments made to maintain target trough whole blood concentrations between 8 and 12 ng/mL for weeks 1 through 6, 6 to 10 ng/mL for weeks 7 through month 12, and > 5 ng/mL after 1 year. In addition, all patients received mycophenolate mofetil 1 gram twice daily and prednisone titrated to 5 mg daily by day 60 post-transplant, with tapers below 5 mg occurring rarely. Patients were converted to mTORs for CNI intolerance (nephrotoxicity being the primary etiology) and patients were converted from mycophenolate to leflunomide for BK nephropathy.
Patients at high-risk of CMV (D+/R- or cytolytic induction therapy) received three to six months of valganciclovir prophylaxis depending on medication tolerability and affordability. Patients at low-risk of CMV (D-/R-) did not receive CMV prophylaxis, but received Herpes simplex prophylaxis with three months of acyclovir therapy. Moderate CMV risk patients either received three months of valganciclovir therapy or preemptive CMV PCR monitoring based on medication health insurance coverage and affordability. All patients received PJP prophylaxis with sulfamethoxazole/trimethoprim for three months and fungal prophylaxis with nystatin swish and swallow suspension for one month. BK viral infection monitoring was performed by BK PCR monitoring at a minimum of 1, 3, 6, 9, and 12 months post-transplant.
Statistical Analysis
For the initial univariate analysis, patients with early graft loss (graft loss within 5-years post-transplant) were compared across racial groups for all baseline and follow-up variables. Likewise, patients with graft longevity (graft survival with at least 5-years of post-transplant follow-up) were compared in a similar fashion. All variables that demonstrated statistically significant associations on univariate analysis were included in multivariate modeling using binary logistic regression (dependent variable of early graft loss). Additional variables that were controlled for in these models included gender, age, CV history and immunologic factors (HLA mismatches, PRA, warm and cold ischemic times). Multivariate model exploration was conducted using a stepwise backward conditional approach, with variables removed at each step that lacked statistical association (p>0.2); model results are reported as odds ratios and 95% confidence intervals. Variables included in the initial model were those that demonstrated significant differences within the univariate comparisons. Model performance was assessed by developing risk probabilities for each patient to determine positive and negative predictive value, while also placing these predictive values in receiver operating characteristic (ROC) curves, with output reported in graphical form and numerically as the area under the curve (AUC, C-statistic) with 95% confidence intervals. Statistical significance was based on a p-value of less than 0.05. All data was manually input into a spreadsheet (Excel, MS Office, version 2010, Microsoft Corporation, Seattle WA) with statistical analyses performed using SPSS [version 20, SPSS Inc. Chicago, IL.].
Results
Patients
Between January 2005 and December 2012, 1,508 adult kidney transplants were performed at our institution. Figure 1 displays the study cohort flowchart. Of the 1,508 transplants performed, 198 were excluded due to receiving extra renal transplants (liver, pancreas, heart or lung), 134 were excluded due to missing follow-up data and 652 were excluded due to lack of follow-up of at least five years or without early graft loss; leaving 524 patients in the final cohort. Of these patients, 286 (55%) were AA and 238 (45%) were non-AA. The AA cohort had 87 patients (30%) with early graft loss (risk cohort) and 199 patients (70%) in the control group (graft longevity), while the non-AA group had 64 patients (27%) with early graft loss and 174 patients (73%) with graft longevity. Mean follow-up for the entire cohort was 5.0±2.5 years, which was similar between racial cohorts (AA 4.9±2.5 yrs, non-AA: 5.0±2.5 yrs, p=0.648). Expectedly, the early graft loss cohort had significantly shorter follow-up compared to the graft longevity cohort (1.4±1.3 yrs vs. 6.4±0.8, p<0.001, respectively), which was consistent between racial groups (AA: 1.4±1.3 yrs vs. 6.4±0.8, p<0.001; non-AA: 1.3±1.3 yrs vs. 6.4±0.9, p<0.001).
Figure 1. Study flowchart displaying cohort designation and patients excluded from the analysis with reasons.
Factors Associated with Early Graft Loss in Univariate Analyses
Table 1 displays the baseline sociodemographics and transplant characteristics for patients with early graft loss versus those with graft longevity, comparing these between the AA and non-AA cohorts. For both AA and non-AA patients, three baseline factors were consistently more common in those with early graft loss, regardless of race. These included receiving disability income (OR 2.5-3.6, 95% CI 1.8-7.0), KDRI greater than 1.3 (OR 2.4-2.9, 95% CI 1.1-5.5) and donors with a history of diabetes (OR 11.2-12.3, 95% CI 2.4-59). AA patients had several baseline risk factors that were common in those with early graft loss (as compared to those with graft longevity), which was not demonstrated in non-AA patients; these included receiving Medicare only health insurance (OR 3.9, 95% CI 1.8-8.3), history of cardiac catheterization or CABG (OR 2.5, 95% CI 1.1-5.7), not receiving induction therapy (OR 29, 95% CI 3.6-226) and development of delayed graft function (OR 3.9, 95% CI 1.9-8.3). Receiving peritoneal dialysis prior to transplant was more common in AA patients with graft longevity, as compared to those with early graft loss (OR 0.3, 95% CI 0.1-0.8). In the non-AA patients, unique factors that were more common in those with early graft loss (as compared to those with graft longevity) included recipients of hemodialysis prior to transplant (OR 2.3, 95% CI 1.3-4.0) and donor history of hypertension (OR 3.8, 95% CI 1.7-8.2).
Table 1. Baseline donor and recipient characteristics compared between those with early graft loss versus graft longevity across racial cohorts.
| Baseline Characteristics | Graft Loss within 5 years of Transplant | Graft Survival of at least 5 years Post-Transplant | ||
|---|---|---|---|---|
|
| ||||
| African-American (n=87) | Non-African-American (n=64) | African-American (n=199) | Non-African-American (n=174) | |
|
| ||||
| Baseline Recipient Sociodemographics | ||||
|
| ||||
| Mean Age (yrs±SD) | 50±14 | 53±15 | 49±13 | 51±14 |
|
| ||||
| Female Gender | 39% | 36% | 43% | 36% |
|
| ||||
| Mean BMI (kg/m2±SD) | 28±6 | 28±6 | 28±5 | 27±5 |
|
| ||||
| Did Not Complete High School | 16% | 14% | 8% | 7% |
|
| ||||
| Health Insurance Medicare Only | 24% | 13% | 8% | 10% |
|
| ||||
| Cannot Read or Write | 6% | 0% | 2% | 0% |
|
| ||||
| Working at the Time of Transplant | 10% | 18% | 16% | 26% |
|
| ||||
| Only Income from Disability | 41% | 39% | 22% | 15% |
|
| ||||
| Primary Diagnosis Diabetes | 35% | 25% | 26% | 18% |
|
| ||||
| Primary Diagnosis Hypertension | 37% | 20% | 34% | 13% |
|
| ||||
| Past Medical History | ||||
| Hypertension | 92% | 90% | 94% | 89% |
| Diabetes | 40% | 34% | 34% | 23% |
| Smoker | 22% | 33% | 20% | 29% |
| Heart Disease | 23% | 23% | 15% | 18% |
| Hyperlipidemia | 36% | 47% | 43% | 52% |
| Stroke | 10% | 3% | 10% | 5% |
| Cardiac Catheterization or CABG | 16% | 20% | 7% | 13% |
| Acute Myocardial Infarction | 6% | 6% | 5% | 2% |
| Congestive Heart Failure | 3% | 3% | 3% | 3% |
| Peripheral Vascular Disease | 5% | 2% | 6% | 4% |
|
| ||||
| Pre-Transplant Dialysis | 91% | 75% | 94% | 65% |
| Peritoneal Dialysis | 5% | 13% | 15% | 23% |
| Hemodialysis | 86% | 63% | 79% | 42% |
|
| ||||
| Baseline Immunologic Characteristics | ||||
|
| ||||
| Mean HLA Mismatches (±SD) | 5±1 | 4±2 | 5±1 | 4±2 |
|
| ||||
| Mean Cold Ischemic Time (hrs±SD) | 18±8 | 16±10 | 19±10 | 14±11 |
|
| ||||
| Mean Warm Ischemic Time (min±SD) | 41±27 | 37±11 | 39±19 | 36±9 |
|
| ||||
| Re-Transplant | 14% | 17% | 7% | 13% |
|
| ||||
| Mean % Panel Reactive Antibody | 9±23 | 18±33 | 11±23 | 12±25 |
| >20% | 14% | 25% | 19% | 17% |
| >80% | 6% | 11% | 5% | 6% |
|
| ||||
| Baseline Donor Characteristics | ||||
|
| ||||
| Mean Donor KDRI (±SD) | 1.34±0.41 | 1.35±0.42 | 1.14±0.33 | 1.18±0.36 |
|
| ||||
| Donor Type | ||||
| Living Donor | 7% | 19% | 11% | 33% |
| Extended Criteria | 14% | 22% | 9% | 13% |
| Deceased after Cardiac Death | 0% | 0% | 1% | 2% |
|
| ||||
| Female Gender | 48% | 48% | 40% | 46% |
|
| ||||
| African-American | 32% | 19% | 34% | 14% |
|
| ||||
| Past Medical History | ||||
| Hypertension | 23% | 39% | 17% | 14% |
| Diabetes | 12% | 18% | 1% | 2% |
| Died due to Stroke | 44% | 45% | 30% | 32% |
| Hepatitis C Positive | 5% | 0% | 0% | 2% |
|
| ||||
| Peri-Operative Factors | ||||
|
| ||||
| Induction Therapy | ||||
| None/Not Documented | 13% | 5% | 0% | 2% |
| IL-2 Receptor Antibody | 52% | 52% | 62% | 65% |
| Cytolytic Therapy | 35% | 44% | 38% | 34% |
|
| ||||
| Delayed Graft Function | 23% | 9% | 7% | 5% |
Table 2 displays the immunosuppression, clinical outcomes and follow-up cardiometabolic indices for patients with early graft loss versus graft longevity, comparing these between the AA and non-AA cohorts. Three outcomes were significantly more common in those with early graft loss, regardless of race, including the development of rejection (OR 2.7-8.5, 95% CI 1.2-83), rehospitalization (OR 2.0-4.5, 95% CI 1.2-10.5) and BK viral infection (OR 2.5-3.4, 95% CI 1.0-8.8). Factors that were more common in early graft loss AA patients, but not in non-AA recipients were predominantly focused around CV risk factor control, including having a mean follow-up SBP <150 mmHg (OR 0.4, 95% CI 0.2-0.8), having a lower mean LDL (OR 0.1-0.7, 95% CI 0.01-0.9), having a lower mean triglyceride level (OR 0.3-0.4, 95% CI 0.1-0.9) and lower mean HgbA1Cs (OR 0.2-0.4, 95% CI 0.1-1.0). In non-AA patients, unique factors more common in those with early graft loss were primarily focused on immunosuppression dosing; non-AA patients with tacrolimus trough concentrations <8 ng/mL during the first month post-transplant had twice the risk of early graft loss (OR 2.1, 95% CI 1.0-4.7), while patients with mean mycophenolate doses less than 1000 mg/day after the first year post-transplant had nearly five times the risk of early graft loss (OR 4.6, 95% CI 2.1-10.3).
Table 2. Immunosuppression and clinical outcomes compared between those with early graft loss versus graft longevity across racial cohorts.
| Post-Transplant Outcomes | Graft Loss within 5 years of Transplant | Graft Survival of at least 5 years Post-Transplant | ||
|---|---|---|---|---|
|
| ||||
| African-American (n=87) | Non-African-American (n=64) | African-American (n=87) | Non-African-American (n=64) | |
|
| ||||
| Immunologic Characteristics | ||||
|
| ||||
| Mean Tacrolimus Trough Concentrations | ||||
| First Week | 7.7±2.2 | 8.6±2.6 | 7.4±2.6 | 9.0±2.5 |
| First Month | 8.4±1.7 | 8.7±2.0 | 8.5±1.8 | 9.5±1.7 |
| First Year | 8.3±1.5 | 8.5±1.8 | 8.5±1.4 | 8.9±1.3 |
| After First Year | 6.7±2.7 | 6.0±2.5 | 6.8±1.8 | 6.6±1.8 |
|
| ||||
| Mean Mycophenolate Doses | ||||
| First Month | 1,603±490 | 1,466±508 | 1,652±462 | 1,534±419 |
| Three Month | 1,498±573 | 1,271±598 | 1,581±511 | 1,465±478 |
| After First Year | 1,327±710 | 885±786 | 1,452±582 | 1,371±551 |
|
| ||||
| Acute Cellular Rejection | 35% | 17% | 16% | 6% |
| Within 3 months | 15% | 6% | 7% | 2% |
| Within 6 months | 17% | 8% | 8% | 2% |
| Within 1 year | 24% | 13% | 9% | 3% |
| Severe Rejection (Banff ≥1B) | 26% | 11% | 9% | 2% |
| Acute Antibody Mediated | 14% | 5% | 6% | 1% |
| Borderline | 12% | 20% | 15% | 13% |
|
| ||||
| Readmissions | ||||
| 30-Day | 28% | 23% | 11% | 6% |
| 90-Day | 41% | 36% | 18% | 12% |
| 1-Year | 58% | 45% | 30% | 24% |
| Any | 70% | 58% | 54% | 36% |
| Readmit for CV Event | 12% | 22% | 8% | 8% |
| Readmit for Infectious Event | 22% | 22% | 21% | 15% |
| Readmit for Surgical Issue | 16% | 9% | 10% | 10% |
|
| ||||
| Infections | ||||
| Cytomegalovirus | 15% | 6% | 12% | 8% |
| BK Virus | 10% | 9% | 7% | 6% |
| Any Significant Infection | 45% | 36% | 34% | 27% |
|
| ||||
| Convert to mTOR | 10% | 12% | 15% | 13% |
|
| ||||
| Cardiometabolic Characteristics | ||||
|
| ||||
| Mean Protein/Creatinine Ratio | 1.17±3.69 | 1.23±4.53 | 0.70±1.27 | 0.43±0.81 |
|
| ||||
| Mean Systolic Blood Pressure | 140±15 | 134±12 | 136±15 | 134±14 |
| Mean Diastolic Blood Pressure | 79±13 | 74±9 | 77±8 | 76±8 |
|
| ||||
| Mean Low Density Lipoproteins+ | 109±38 | 83±31 | 96±25 | 95±53 |
| Mean Triglycerides+ | 143±71 | 165±78 | 120±60 | 163±78 |
|
| ||||
| Mean Hemoglobin A1C+* | 7.8±2.5 | 7.1±1.8 | 7.5±1.6 | 6.8±1.2 |
|
| ||||
| Compelling Medication Use | ||||
| Anti-platelet Therapy | 41% | 34% | 53% | 44% |
| β-Blocker | 51% | 44% | 58% | 51% |
| ACE inhibitor or ARB | 39% | 36% | 54% | 46% |
| Statin Therapy | 43% | 34% | 53% | 49% |
| Other Anti-Lipid Therapy | 37% | 28% | 43% | 43% |
Values represent the mean during the entire post-transplant follow-up period
This is reported for only patients with a diagnosis of diabetes
Multivariate Analysis for Factors Associated with Early Graft Loss
Figure 2 displays the Forest plots of the final multivariate model results for the dependent variable of early graft loss, separated by recipient race. Models for both AA and non-AA groups controlled for gender, age, CV history, immunologic risks (HLA mismatches, PRA, warm and cold ischemic times) and donor criteria (KDRI). CV risk factor control (HgbA1C, TG, LDL and SBP) and two SES surrogates (income only from disability and Medicare only health insurance) continued to be unique factors significantly associated with early graft loss in AA patients, while in non-AA patients, immunosuppression continued to be unique factors associated with early graft loss. Both rejection and infection were significant risk factors in both cohorts, regardless of race.
Figure 2.
Forest plots of the final multivariate model results for the dependent variable of early graft loss. AA patients are in the top blue panel and non-AA patients are in the bottom red panel.
Figure 3 displays the ROC curves and C-statistics for the predictability of multivariate models for early graft loss, separated by race. A model containing four CV risk factor control metrics (LDL <100 mg/dL, TG <150 mg/dL, SBP <150 mmHg and HgbA1C <7%) was a significant predictor of early graft loss in AA patients (Figure 3A, C-statistic 0.69 (0.60-0.78), p<0.001), accounting for 15% of the variability associated with early graft loss; however, a model containing these same four CV indices was not associated with early graft loss in non-AA patients (Figure 3C, C-statistic 0.57 (0.46-0.67), p=0.211), accounting for 2% of the variability associated with early graft loss. The final models for both the AA and non-AA cohorts, which included the unique risk factors described in the above section, were strong predictors of early graft loss (Figures 3B and 3D, C-statistic 0.81-0.83 (0.74-0.90), p<0.001), accounting for more than one-third of the variability associated with this outcome (R2 >33%). In AA patients, the positive predictive value (PPV) of this model in discerning patients that will develop early graft loss was 78%, while the negative predictive value (NPV) in discriminating which patients will not have graft loss within five years post-transplant was 88%. In non-AA patients, the PPV was 60% and the NPV was 88%.
Figure 3.
ROC curves displaying the association between various post-transplant factors and early graft loss for both AA patients (Figures 3A and 3B) and non-AA patients (Figures 3C and 3D). Models containing four cardiovascular risk factor control metrics (LDL <100 mg/dL, TG <150 mg/dL, SBP <150 mmHg and HgbA1C <7%) was a significant predictor of early graft loss in AA patients (Figure 3A, C-statistic 0.69 (0.60-0.78), p<0.001), while a model containing these same four cardiovascular metrics was not associated with early graft loss in non-AA patients (Figure 3C, C-statistic 0.57 (0.46-0.67), p=0.211). The final models for both the AA and non-AA cohorts, which included the risk factors demonstrating statistical significance in Tables 1 and 2, were strong predictors of early graft loss (Figures 3B and 3D, C-statistic 0.81-0.83 (0.74-0.90), p<0.001).
Discussion
Despite years of focused research in racial disparities, AA kidney transplant recipients continue to experience a disproportionately higher rate of graft loss1 and the predominant factors driving this disparity are heavily debated. The reasons for this continued controversy are likely related to the complex interacting etiologies converging to create this disparity, coupled with numerous studies that provide contradicting findings.15-17 The results of the analysis presented in this paper provides detailed and comprehensive assessment of baseline and follow-up clinical data that cannot be captured in national registry studies.23 The results demonstrate that AA patients have unique baseline and follow-up characteristics that predispose them to graft loss. In particular, it appears that surrogates of baseline SES (Medicare only insurance and disability only income), coupled with CV risk factor control during the post-transplant follow-up period provide strong associations with graft loss. Other than disability income, these factors were unique to AA recipients, and did not appreciably influence early graft loss in the non-AA cohort. Taken in their entirety, these results provide evidence to potentially support interventional studies aimed at improving CV risk factor control as a means to diminish racial disparities in kidney transplantation.
The finding that the development of rejection was a consistent risk factor for early graft loss in both AA and non-AA patients is an expected result.24,25 Given the consistent predictability of rejection on graft loss across all racial groups, future interventions should focus on methods to reduce rejection to optimize graft outcomes in all patients, and given the additional unique risk factors in AA patients identified in this study, it is unlikely that solely focusing on reducing rejection will produce robust improvements in racial disparities in transplantation. It is interesting to note that in the univariate analysis, lack of induction therapy was a strong risk factor for early graft loss in AA recipients, while cytolytic induction was a risk factor for graft loss in non-AA patients. This data supports evidence from previous studies, suggesting AA patients may obtain more benefit from potent induction therapy.16 A recently completed prospective study from our institution demonstrated the AA patients likely benefited more than non-AA from antithymocyte globulin induction by reducing acute rejection rates, even within the context of a potent contemporary triple-drug maintenance immunosuppressant regimen of tacrolimus, mycophenolate and prednisone.26
The development of post-transplant infection was also a consistent risk factor for early graft loss, regardless of race. Given the abundance of previous studies demonstrating the risks of infection on graft loss, this is also an expected finding. Focusing efforts on reducing the impact of infection or BK virus on graft loss will likely improve outcomes, but given the consistent effect of this on graft loss across both AA and non-AA patients, this would not likely reduce racial disparities. 27,28
Previous studies have demonstrated similar results to the data presented in this analysis with regards to a number of risk factors particularly important in AA kidney transplant recipients. These include SES factors8-11 and post-transplant hypertension control.12 Although modifying SES is a difficult, improving post-transplant CV risk factor control is certainly obtainable.15 Chisholm et al demonstrated improved hypertension control in AA patients with interventions focused on medication adherence and cost.29 However, within the transplant population, there is paucity in literature demonstrating improved patient or graft survival through enhanced CV risk factor control. Certainly, given the strong body of literature within the general population30-32 and the fact that transplant patients usually have multiple CV risk factors, improving hypertension, dyslipidemia and diabetes management in transplant recipients should be expected to improve graft and patient survival.32,34 An important question, which the results of the data presented in this study cannot answer, is what is the best mechanism to optimize CV risk factor control in transplant patients, and in particular AA transplant recipients? It is unlikely that a single strategy to this complex issue will produce robust results. Rather, a multidimensional intervention that focuses on compelling medication prescribing and dosing optimization, coupled with improving adherence and healthy lifestyle choices leading to weight management and exercise regimens may be the best approach to obtain enduring results.32,35
There are several important limitations to this study that are worthy of discussion. First, this was a retrospective study; limitations with retrospective cohort studies include the potential for selection bias, recall misclassification and missing data. Data collection occurred through the use and validation of both electronic and paper medical records to minimize these potential limitations. An additional limitation to the retrospective study design includes lacking the ability to prospectively randomize and risk-stratify non-modifiable donor and recipient characteristics at the study onset; however, multivariate modeling was used to minimize this limitation. Although this study contained a large number of patients with significant follow-up, its single-center design did limit study power, and certainly separating groups by both graft loss and race also decreased the studies overall power. However, these limitations are somewhat offset by the fact that detailed baseline SES and medical histories were captured in this study. Additionally, thorough follow-up clinical data, including blood pressures, serum lipids and HgbA1Cs were collected and included in this analysis, which is lacking from previous analyses in this area of research. Baseline and follow-up data, such as education level, was included in this analysis, which would not be easily obtainable in a multicenter retrospective study or one that utilizes national registry data from the SRTR.23 Finally, it should be noted that limiting the timeframe of this study from beginning in 2005 may have limited the sample-size and reduce study power, it did produce a study cohort that is well-representative of contemporary kidney transplant recipients, including age, medical comorbidities, donor risk and immunosuppression regimens.
In conclusion, the results of this analysis demonstrate that in a contemporary cohort of kidney transplant recipients, acute allograft rejection and infection continue to be prominent risk factors associated with graft loss. Within AA recipients, SES and CV risk factor control are significant and unique factors associated with early graft loss. Future studies should focus interventions on modifying these factors as a mechanism to improve graft outcomes in this high-risk cohort of kidney transplant recipients.
Acknowledgments
Research reported in this publication was supported by the National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health under Award Numbers K23DK099440 and T35 DK007431.
Footnotes
Conflict of interest statement: The authors have no conflicts of interest to disclose as it relates to the content of this manuscript.
References
- 1.OPTN / SRTR 2010 Annual Data Report. Rockville, MD: Department of Health and Human Services, Health Resources and Services Administration, Healthcare Systems Bureau, Division of Transplantation; 2011. Organ Procurement and Transplantation Network (OPTN) and Scientific Registry of Transplant Recipients (SRTR) [Google Scholar]
- 2.Opelz G, Mickey MR, Terasaki PI. Influence of race on kidney transplant survival. Transplant Proc. 1977;9(1):137–142. [PubMed] [Google Scholar]
- 3.Terasaki PI, Opelz G, Mickey MR. Summary of kidney transplant data, 1977 – factors affecting graft outcome. Transplant Proc. 1978;10(2):417–421. [PubMed] [Google Scholar]
- 4.Ciancio G, Burke GW, Suzart K, Mattiazzi A, Vaidya A, Roth D, Kupin W, Rosen A, Johnson N, Miller J. The use of daclizumab, tacrolimus and mycophenolate mofetil in African-American and Hispanic first renal transplant recipients. Am J Transplant. 2003;3(8):1010–1016. doi: 10.1034/j.1600-6143.2003.00181.x. [DOI] [PubMed] [Google Scholar]
- 5.Neylan JF. Racial differences in renal transplantation after immunosuppression with tacrolimus versus cyclosporine. FK506 kidney transplant study group. Transplantation. 1998;65(4):515–523. doi: 10.1097/00007890-199802270-00011. [DOI] [PubMed] [Google Scholar]
- 6.Podder H, Podbielski J, Hussein I, Katz S, Buren C, Kahan BD. Sirolimus improves the two-year outcome of renal allografts in African-American patients. Transplant Int. 2001;14(3):135–142. doi: 10.1007/s001470100315. [DOI] [PubMed] [Google Scholar]
- 7.Weber M, Deng S, Arenas J, Aradhye S, Grossman R, Shaw L, Naji A, Barker C, Brayman KL. Decreased rejection episodes in African-American renal transplant recipients receiving mycophenolate mofetil/tacrolimus therapy. Transplant Proc. 1997;29(8):3669–70. doi: 10.1016/s0041-1345(97)01067-1. [DOI] [PubMed] [Google Scholar]
- 8.Butkus DE, Meydrech EF, Raju SS. Racial differences in the survival of cadaveric renal allografts. overriding effects of HLA matching and socioeconomic factors. N Engl J Med. 1992;327(12):840–845. doi: 10.1056/NEJM199209173271203. [DOI] [PubMed] [Google Scholar]
- 9.Curtis JJ. Kidney transplantation: Racial or socioeconomic disparities? Am J Kidney Dis. 1999;34(4):756–758. doi: 10.1016/S0272-6386(99)70404-X. [DOI] [PubMed] [Google Scholar]
- 10.Schweizer RT, Rovelli M, Palmeri D, Vossler E, Hull D, Bartus S. Noncompliance in organ transplant recipients. Transplantation. 1990;49(2):374–377. doi: 10.1097/00007890-199002000-00029. [DOI] [PubMed] [Google Scholar]
- 11.Kalil R, Heim-Duthoy K, Kasiske B. Patients with a low income have reduced renal allograft survival. Am J Kidney Dis. 1992;20(1):63. doi: 10.1016/s0272-6386(12)80318-0. [DOI] [PubMed] [Google Scholar]
- 12.Cosio FG, Dillon JJ, Falkenhain ME, Tesi RJ, Henry ML, Elkhammas EA, Davies EA, Bumgardner GL, Ferguson RM. Racial differences in renal allograft survival: The role of systemic hypertension. Kidney Int. 1995;47(4):1136–1141. doi: 10.1038/ki.1995.162. [DOI] [PubMed] [Google Scholar]
- 13.Cosio FG, Pesavento TE, Kim S, Osei K, Henry M, Ferguson RM. Patient survival after renal transplantation: IV. Impact of post-transplant diabetes. Kidney Int. 2002;62(4):1440–1446. doi: 10.1111/j.1523-1755.2002.kid582.x. [DOI] [PubMed] [Google Scholar]
- 14.Cosio FG, Hickson LJ, Griffin MD, Stegall MD, Kudva Y. Patient survival and cardiovascular risk after kidney transplantation: The challenge of diabetes. Am J Transplant. 2008;8(3):593–599. doi: 10.1111/j.1600-6143.2007.02101.x. [DOI] [PubMed] [Google Scholar]
- 15.Young CJ, Kew C. Health disparities in transplantation: Focus on the complexity and challenge of renal transplantation in African Americans. Med Clin N Am. 2005;89:1003–1031. doi: 10.1016/j.mcna.2005.05.002. [DOI] [PubMed] [Google Scholar]
- 16.Malat GE, Culkin C, Palya A, Ranganna K, Kumar MS. African American kidney transplantation survival: The ability of immunosuppression to balance the inherent pre- and post-transplant risk factors. Drugs. 2009;69(15):2045–2062. doi: 10.2165/11318570-000000000-00000. [DOI] [PubMed] [Google Scholar]
- 17.Young CJ, Gaston RS. African Americans and renal transplantation: Disproportionate need, limited access, and impaired outcomes. Am J Med Sci. 2002;323(2):94. doi: 10.1097/00000441-200202000-00007. [DOI] [PubMed] [Google Scholar]
- 18.Butkus DE, Meydrech EF, Raju SS. Racial differences in the survival of cadaveric renal allografts. Overriding effects of HLA matching and socioeconomic factors. N Engl J Med. 1992;327(12):840–845. doi: 10.1056/NEJM199209173271203. [DOI] [PubMed] [Google Scholar]
- 19.Navaneethan SD, Singh S. A systematic review of barriers in access to renal transplantation among African Americans in the United States. Clin Transplant. 2006;20(6):769–775. doi: 10.1111/j.1399-0012.2006.00568.x. [DOI] [PubMed] [Google Scholar]
- 20.Pallet N, Thervet E, Alberti C, Emal-Aglaé V, Bedrossian J, Martinez F, Roy C, Legendre C. Kidney transplant in black recipients: are African Europeans different from African Americans? Am J Transplant. 2005;5(11):2682–2687. doi: 10.1111/j.1600-6143.2005.01057.x. [DOI] [PubMed] [Google Scholar]
- 21.Yeates K, Wiebe N, Gill J, Sima C, Schaubel D, Holland D, Hemmelgarn B, Tonelli M. Similar outcomes among Black and White renal allograft recipients. J Am Soc Nephrol. 2009;20(1):172–179. doi: 10.1681/ASN.2007070820. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Chakkera HA, O'Hare AM, Johansen KL, Hynes D, Stroupe K, Colin PM, Chertow GM. Influence of race on kidney transplant outcomes within and outside the department of veterans affairs. J Am Soc Nephrol. 2005;16(1):269–277. doi: 10.1681/ASN.2004040333. [DOI] [PubMed] [Google Scholar]
- 23.Kaplan B, Schold J, Meier-Kriesche HU. Overview of large database analysis in renal transplantation. Am J Transplant. 2003;3:1052–56. doi: 10.1034/j.1600-6143.2003.00193.x. [DOI] [PubMed] [Google Scholar]
- 24.Schold JD, Srinivas TR, Braun WE, et al. The relative risk of overall graft loss and acute rejection among African American renal transplant recipients is attenuated with advancing age. Clin Tranpslant. 2011;25:721–30. doi: 10.1111/j.1399-0012.2010.01343.x. [DOI] [PubMed] [Google Scholar]
- 25.Sellares J, de Freitas DG, Mengel M, et al. Understanding the causes of kidney transplant failure: the dominant role of antibody-mediated rejection and nonadherence. Am J Transplant. 2012;12:388–99. doi: 10.1111/j.1600-6143.2011.03840.x. [DOI] [PubMed] [Google Scholar]
- 26.Pilch NA, Taber DJ, Moussa O, et al. Prospective randomized controlled trial of rabbit antithymocyte globulin with IL-2 receptor antagonist induction therapy in kidney transplantation. Ann Surg. 2014;259(5):888–93. doi: 10.1097/SLA.0000000000000496. [DOI] [PubMed] [Google Scholar]
- 27.Binet I, Volker N, Hirsch HH, et al. Polyomavirus disease under new immunosuppressive drugs: a cause of renal graft dysfunction and graft loss. Transplantation. 1999;67:918–22. doi: 10.1097/00007890-199903270-00022. [DOI] [PubMed] [Google Scholar]
- 28.Johnston O, Jaswal D, Gill JS, et al. Treatment of polyomavirus infection in kidney transplant recipients: a systematic review. Transplantation. 2010;89:1057–70. doi: 10.1097/TP.0b013e3181d0e15e. [DOI] [PubMed] [Google Scholar]
- 29.Chisholm MA, Mulloy LL, Jagadeesan M, Martin BV, DiPiro JT. Effect of clinical pharmacy services on the blood pressure of African-American renal transplant patients. Ethn Dis. 2002;12:392–7. [PubMed] [Google Scholar]
- 30.Brugts JJ, Yetgin T, Hoeks SE, et al. The benefits of statins in peiole without established cardiovascular disease with cardiovascular risk factors: meta-analysis of randomised controlled trials. Brit Med J. 2009;338:2376–83. doi: 10.1136/bmj.b2376. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Kelly TN, Bazzano LA, Fonseca VA, et al. Systematic review: glucose control and cardiovascular disease in type 2 diabetes. Ann Intern Med. 2009;151:394–403. doi: 10.7326/0003-4819-151-6-200909150-00137. [DOI] [PubMed] [Google Scholar]
- 32.Thompson AM, Hu T, Eshelbrenner CL, et al. Antihypertensive treatment and secondary prevention of cardiovascular disease events among person without hypertension. J Am Med Assoc. 2011;305:913–22. doi: 10.1001/jama.2011.250. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Carpenter MA, Weir MR, Adey DB, House AA, Bostom AG, Kusek JW. Inadequacy of cardiovascular risk factor management in chronic kidney transplantation–evidence from the FAVORIT study. Clin Transplant. 2012;26(4):E438–46. doi: 10.1111/j.1399-0012.2012.01676.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Pilmore HL, Skeans MA, Snyder JJ, Israni AK, Kasiske BL. Cardiovascular disease medications after renal transplantation: Results from the patient outcomes in renal transplantation study. Transplantation. 2011;91(5):542–551. doi: 10.1097/TP.0b013e31820437bd. [DOI] [PubMed] [Google Scholar]
- 35.Taber DJ, Pilch NA, Meadows HB, McGillicuddy JW, Bratton CF, Chavin KD, Baliga PK, Egede LE. The impact of cardiovascular disease and risk factor control on ethnic disparities in kidney transplant. J Cardiovasc Pharm T. 2013;18(3):243–50. doi: 10.1177/1074248412469298. [DOI] [PubMed] [Google Scholar]