FP663COMPARATIVE LONGER-TERM OUTCOMES OF IRON SUCROSE VERSUS SODIUM FERRIC GLUCONATE IN U.S. PATIENTS INITIATING HEMODIALYSIS. A PSEUDO-EXPERIMENTAL STUDY

Abstract

Introduction and Aims: Little is known about any potential differences in hard, longer-term study outcomes from different intravenous iron formulations routinely used in hemodialysis (HD) patients. We exploited the natural experiment that occurs when facilities make formulary decisions and treat essentially all of their patients with a single intravenous iron agent. We compared outcomes of patients initiating HD in facilities that predominantly (i.e., in >95% of patients each month) used sodium ferric gluconate (FG) vs. iron sucrose (IS). Methods: Using the USRDS, we matched HD facilities using FG to similar ones using IS on geographic region and facility type (chain-affiliated vs. not; hospital-based vs. free-standing). Facilities were only eligible after one year of consistent iron formulation choice to ensure that they were accustomed to the use and administration of the respective formulation. We then identified all subsequent incident HD patients in these facilities and assigned them their facility iron formulation exposure regardless of whether and how often they received intravenous iron. Patients were followed for mortality from any, cardiovascular, or infectious cause. Among those who had Medicare as their primary payor on day 90, patients were followed for infectious and cardiovascular (stroke, myocardial infarction) hospitalizations as well as composite outcomes with the corresponding cause-specific deaths. Results: We matched each of 2015 FG facilities with one IS facility, in which 51603 patients (FG: 26692; IS: 24911) subsequently initiated HD. Of those, 24865 patients (FG: 12771; IS: 12094) had Medicare insurance on day 90. All 28 recorded patient characteristics (demographics, comorbidities, biometrics) were exquisitely balanced between groups (all standardized difference <5%) in the overall population and in the Medicare subset. Throughout follow-up, monthly percent use and mean dose of erythropoiesis-stimulating agents as well as mean hemoglobin concentrations were essentially identical between groups. During 49989 person-years of follow-up, 10381 deaths (3908 cardiovascular; 1209 infectious) occurred. Multivariable Cox models found similar mortality (HR=0.98; 95% CI: 0.93-1.03), cardiovascular mortality (HR=0.96; 95% CI: 0.89-1.03), and infectious mortality (HR=0.98; 95% CI: 0.86-1.13) of patients treated in FG and IS facilities. Among Medicare beneficiaries, 3346 experienced a cardiovascular endpoint over 22373 person-years of follow up. No differences between FG and IS facilities were observed (HR=1.01; 95% CI: 0.93-1.09). The composite infectious endpoint was observed in 12185 patients over 12902 person-years of follow-up. In multivariable Cox models, IS was associated with slightly, but significantly lower rate of fatal or non-fatal infection (HR=0.92; 95% CI: 0.88-0.96). Conclusions: Outcomes of patients initiating HD in facilities almost exclusively using FG versus IS experienced similar longer term outcomes. However, there was a small, but on a population level meaningful decrease in infectious hospitalizations and deaths in patients dialyzing in facilities predominantly using IS. Formulary decisions regarding routinely administered medications may serve as useful instruments to conduct longer-term comparisons of intermittently administered drugs in which standard methods for observational comparative effectiveness research are limited.