Supplementary Material for: Mortality and Cardiovascular Morbidity Associated with Haemoglobin Levels: A Pooled Analysis of Randomised Controlled Trials

<p><b><i>Background/Aims:</i></b> Several randomised controlled trials (RCTs) have raised concerns about potential harm associated with erythropoiesis-stimulating agents (ESAs) in chronic kidney disease patients, especially when haemoglobin (Hb) levels above 13 g/dl were targeted. We report the relationship between Hb levels and outcomes in the methoxy polyethylene glycol-epoetin beta RCT programme. <b><i>Methods:</i></b> We assessed the association between Hb and a composite end point, as well as its components [all-cause mortality, myocardial infarction (MI) or cerebrovascular events (CVE)], in multiple post hoc analyses of 9 prospective RCTs (3,405 chronic kidney disease patients). Mean Hb levels over time and deviation from target were analysed using a Cox regression model. Time-adjusted average Hb, deviation from target, the last Hb, Hb slope and within-patient Hb variability preceding an event were analysed using a time-dependent Cox model. Hazard ratios and 95% confidence intervals were calculated. <b><i>Results:</i></b> Average Hb <10 g/dl, decrease from stable baseline Hb >1 g/dl, last Hb <10 g/dl, Hb decline >1.5 g/dl/4 weeks and increased Hb variability were associated with a higher risk of the composite end point and all-cause mortality. An increased risk for CVE and MI was found with a last Hb <10 g/dl and with a decrease from baseline >1 g/dl in the preceding month. <b><i>Conclusion:</i></b> In multiple analyses from a large programme of prospective clinical trials of ESA treatment, risk of all-cause mortality and cardiovascular morbidity risk was consistently higher at Hb <10 g/dl and in patients whose Hb fell below target.