INCIDENCE AND PREDICTORS OF MYOCARDIAL INFARCTION AFTER KIDNEY TRANSPLANTATION IN THE UNITED STATES

Abstract

O149* Aims: To describe the incidence and predictors of myocardial infarction after kidney transplantation and assess the impact of this complication on mortality. Methods: We used registry data collected by the United States Renal Data System (USRDS) to perform a retrospective investigation of post-transplant myocardial infarction (PTMI) among adult (≥18 years-old) first renal allograft recipients. Eligibility was defined as transplantation in 1995-2000 with Medicare as the primary payer. Patients with prior and/or simultaneous multi-organ transplants were excluded. PTMI events were ascertained from billing and death records, and the incidence was estimated by the product-limit method. We employed multivariate Cox’s proportional hazards analysis to obtain covariate-adjusted estimates of the risk of PTMI (hazard ratio, HR) associated with recipient, donor, and transplant-related factors. Observations were censored at the earliest of the following events: loss of Medicare, loss to follow-up, three years after the start of Medicare, or end of observation (December 31, 2000). We examined the impact of PTMI on mortality using Cox’s proportional hazards analysis with PTMI as a time-dependent covariate. 95% confidence intervals (CI) were used to measure the precision of all estimates. Results: We identified 35,847 eligible participants. The cumulative incidence of PTMI was 4.3% (CI 4.1-4.5%), 5.6% (CI 5.3-5.8%), and 11.1% (CI 10.7-11.5%) at 6, 12, and 36 months, respectively. Recipient factors associated with increased risk of PTMI included older age, diabetes as the primary cause of end-stage renal failure (HR 1.34, CI 1.19-1.50), and histories of angina (HR1.68, CI 1.51-1.88) and peripheral vascular disease (HR 1.28, CI 1.18-1.39). Myocardial infarction history strongly predicted PTMI (HR 4.16, CI 3.70-4.68) but risk decreased with time after transplant. Pre-transplant diabetes (related or unrelated to primary cause of renal failure) also predicted PTMI (HR 1.18, CI 1.04-1.33), and risk rose further over observation. Recipient factors that reduced the risk of PTMI included African American race (HR 0.78, CI 0.71-0.86), Hispanic ethnicity (HR 0.70, CI 0.62-0.80), and employment (HR 0.82, CI 0.75-0.90). Female recipients had lower risk compared with males at transplant (HR 0.74, CI 0.66-0.82), but protection diminished with time. Donor and transplant-related risk factors for PTMI included older donor age, deceased-donor source (HR 1.18, CI 1.05-1.31), and delayed graft function (HR 1.16, CI 1.07-1.27). The risk of PTMI changed after important post-transplant complications. In particular, risk was higher after a diagnosis of post-transplant diabetes (HR 1.60, CI 1.36-1.89) and markedly elevated after graft failure (HR 2.79, CI 2.42-3.20). The risk of PTMI was not significantly (P≥0.01) related to time from first dialysis to transplant, degree of HLA-matching, CMV sero-pairing, the type of maintenance immunosuppression started after transplant, or the use of induction therapy. In a second analysis designed to identify independent risk factors for mortality, PTMI was strongly predictive of death in a manner that declined with time after PTMI. Specifically, the hazard of death was 55.6 (CI 48.3-64.0) in the first week after PTMI, 9.9 (CI 8.2-11.9) in the next three weeks, and 2.7 (CI 2.4-3.0) thereafter. Conclusions: Potentially modifiable risk factors for PTMI include delayed graft function, post-transplant diabetes, and graft failure. Because PTMI is a potent independent predictor of death, reducing the risk of PTMI should improve outcomes after renal transplantation.