Proefschrift

CHAPTER 3.1 86 METHODS Data sources Detailed information on the methods including data sources, endpoint definitions and collection codes can be found in a previous study.18 Briefly, five HFrEF RCTs and two HF registries were included in this study. BEAUTIFUL and SHIFT were phase III ivabradine trials (n= 15732),19,20 FAIR-HF and CONFIRM were phase III and phase IV studies on intravenous iron supplementation (n=763)21,22 and PANTHEON was a phase II trial for neladenosone bialanate (n=427).23 For final analysis, aggregated data from both treatment and placebo arms of each RCT were pooled to represent one RCT population (n = 16 917). The Dutch CHECK-HF and SwedeHF registries enroll patients with clinicianjudged HF and detailed information on the methods can be found elsewhere.24,25 For the current analysis, only HFrEF patients, defined as those enrolled with left ventricular ejection fraction (LVEF) <40% were considered. To ensure consistency with CHECK-HF, only outpatients registered between 2000 to 2016 in SwedeHF (n=40 230) were included. Contrary to Dutch CHECK-HF, SwedeHF contains followup data, therefore any analysis of clinical outcomes was restricted to patients from SwedeHF. Ethics approvals were obtained by the original study investigators for the RCTs. CHECK-HF received approval for anonymized analysis of routine clinical data. In SwedeHF, patient consent to enrollment in the registry allows analysis of individual patient data. Eligibility criteria, study population, and outcomes The inclusion and exclusion criteria listed in the study protocols of the five RCTs were tabulated to identify common eligibility and ineligibility criteria (see Figure 1, Supplementary Table 1).18 These common criteria were applied to the SwedeHF and CHECK-HF dataset to identify subgroups of patients who would have been eligible for trial participation or not. Data were then presented by the following groups and additionally stratified by sex: RCT, RCT-eligible, and RCT-ineligible (Figure 1). The following clinical outcomes at one-year were assessed: all-cause mortality, CVmortality, and first HF hospitalization.

RkJQdWJsaXNoZXIy MjY0ODMw