Materials and Methods
Data Source
We utilized data from the Scientific Registry of Transplant Recipients (SRTR). The SRTR data system includes data on all donors, waitlisted candidates and transplant recipients in the United States, submitted by the members of the Organ Procurement and Transplantation Network (OPTN). The Health Resources and Services Administration (HRSA), U.S. Department of Health and Human Services provides oversight to the activities of the OPTN and SRTR contractors.
Study Design and Participants
Data submitted to the SRTR was examined to identify all kidney-alone transplant recipient pairs with a common deceased donor in the United States between 2000 and 2013. Exclusions were multi-organ transplants and recipients that were pediatric, hepatitis B surface antigen positive, had missing or unknown HIV or HCV serostatus or received a previous liver transplant. Two data sets were created. One data set (HCV mate) was created to analyze recipients with and without HCV wherein one kidney transplant recipient was hepatitis C status-seropositive (HCV+) and the mate kidney recipient was hepatitis C status-seronegative (HCV−). Additional exclusions were recipient HIV-seropositivity and donor hepatitis C seropositivity. The other data set (HIV mate) was created to analyze recipients with and without HIV wherein one kidney transplant recipient was HIV-seropositive (HIV+), regardless of hepatitis C status and the mate kidney recipient was both HIV-seronegative (HIV−) and hepatitis C-seronegative (HCV−).
Exposure and Outcome Classification and Assessment
HCV and HIV status are reported but not necessarily confirmed or assessed at the time of transplant for recipients. Paired donors were categorized using a donor identifier available in the registry.
Recipient outcomes were compared between paired donors by infection group. The infected group in the HCV mate cohort comprised HCV positive (HCV+) recipients. The noninfected group consisted of recipients who were HCV negative (HCV−). The infected group in the HIV mate cohort comprised recipients all of whom were HIV-positive (HIV+) and some of whom also had HCV (HCV+). The noninfected group consisted of recipients who had neither HIV nor HCV (HIV−/HCV−).
The primary outcome was death-censored graft survival (DCGS) following transplantation defined as the earliest of re-transplantation or return to dialysis. Secondary outcomes were (a) patient mortality from any cause following transplantation and (b) 1-year acute rejection (indicated by reported acute rejection or treatment for acute rejection in the follow-up forms).
Potential Confounders
Donor variables were not included in the Cox regression models based on the paired kidney study design. The following recipient factors were included in the models: age (continuous), sex, race (African-American, other), previous KTX, duration of maintenance dialysis prior to transplantation (<3, ≥3 years, missing), number of HLA-A, B and DR mismatches (HLA MM; ≤3, >3), peak panel reactive antibody (PRA) level (≤30%, >30%, missing), body mass index (BMI; ≤30, >30 kg/m, missing), cold ischemia time (CIT; ≤24, >24 h, missing), insurance status (private, other), comorbidity and year of transplantation. In the HIV mate cohort, recipient HCV infection status was also included as a potential confounder.
BMI was calculated as weight (kg)/height (m). Comorbidity was defined as history of at least one of the following conditions: angina, malignancy, peptic ulcer disease, drug-treated chronic obstructive pulmonary disease, cerebrovascular disease or peripheral vascular disease. The appropriate functional form of model covariates was determined by exploratory data analysis in unadjusted models and perceived impact on clinical meaningfulness.
Statistical Analysis
Univariate associations was examined using chi-square tests for categorical variables and t-tests for continuous variables whose distributions approximated normality. Survival distributions for mortality and graft failure were estimated using the Kaplan–Meier method. Survival curves were compared using the log-rank test. For DCGS and patient survival Cox proportional hazards models were fit to estimate hazard ratios and 95% confidence intervals (CIs) for infection groups after accounting for all potential confounders. Infection group as well as all other covariates were examined for adherence to the proportional hazard assumption. Ties in the failure time were handled using the Breslow method. The dependence of observations derived from the kidneys from the same donor was accounted for in Cox models with adjustment of the standard error of the hazard ratio (sandwich estimator). Time to outcome was defined as time from the date of transplant until date of outcome (death or graft failure), censored for loss to follow-up and end of study period (October 31, 2013).
Acute rejection within 1 year was defined as the presence of rejection regardless of treatment as coded in follow-up forms at 3, 6 or 12 months. The odds of acute rejection at 1 year were assessed using logistic regression with all confounders included in the model. When acute rejection data were missing or unknown (HCV mate, 11.1% of cases; HIV mate, 16.7% of cases); these cases were not included in the analysis. Due to the instability of both models, because of low sample sizes within missing categories, the final model excluded cases with missing PRA, BMI and dialysis time.
All statistical analyses were conducted using the SAS system version 9.2 (SAS Institute, Inc., Cary, NC). Statistical significance was identified by a p-value of less than 0.05 and all CIs also used a 95% threshold. All p-values were two-sided. This study was approved by the Institutional Review Board of the Albert Einstein College of Medicine.