(MTPPI) in collaboration with Harvard School of Public Health (HSPH) to explore the causal relationship between epoetin dose and survival of hemodialysis patients receiving this therapy.

Recent public attention has been directed toward the risks associated with certain drugs that are in wide-spread use. As with most drugs after FDA approval, assessment of the risks and clinical benefits of these products becomes very complicated because the relationship between drug use and patient outcomes is often confounded in complex ways that can not be addressed by conventional statistical techniques. Recently, researchers at Harvard School of Public Health (HSPH) have developed advanced statistical methods (referred to as marginal structural models - MSM) that adjust for such complex confounding practice patterns. In this grant, "Epoetin Therapy and Survival of Hemodialysis Patients," MSM techniques will be applied to explore the causal relationship of epoetin therapy for treatment of anemia on survival for hemodialysis patients. The application of these methods will provide a state-of-the-art estimate of the effect of epoetin on survival.

Background. More than 20 million Americans have chronic renal disease and an equal number are at increased risk. Approximately 400,000 Americans progress to end-stage renal disease (ESRD) requiring routine dialysis for the rest of their lives unless they undergo a kidney transplant. The ESRD population is growing dramatically at an annual rate of 8%. According to the NIDDK, the mortality rate of 24% annually among this population remains 'unacceptably' high. A common occurrence among ESRD patients is anemia (below normal red blood cell count). To treat their anemia, over 90% of all hemodialysis patients receive epoetin (recombinant human erythropoietin, rHuEPO, epoetin alpha, EPO or EPOGEN(TM)) treatment at a cost of approximately $3 billion dollars/year. The relationship between epoetin and survival remains controversial. Currently, clinical practice guidelines recommend epoetin treatment based in part on a multitude of published studies that report a positive association between higher hematocrit and survival. However, some of these studies also suggest that hematocrit is both affected by prior epoetin therapy and affects future decisions about epoetin dosing. That is, hematocrit may confound the relation of epoetin and survival. It is doubtful therefore whether the published associations between hematocrit and survival based on conventional statistical techniques can be interpreted as a causal relationship. MSMs that adjust for complex confounding practice patterns - such as the one existing between hematocrit, epoetin, and survival - will help elucidate these relationships.

The Goal. A better understanding of the relationship between epoetin and survival will provide a basis for improving current clinical practice guidelines and may thereby decrease the mortality rate of ESRD patients. Study results will be disseminated to the public through peer-reviewed journals. Our goal is to insure that treatment decisions for this high-risk population are based on the best possible evidence to with the best possibilities for survival.