Prognostication in heart failure across the ejection fraction spectrum : implications for clinical practice and trial design

Abstract

Background. The heterogeneity of heart failure (HF) across the ejection fraction (EF) spectrum presents a challenge when triaging for follow-up, implementing therapies, and designing trials. There is a preconceived notion that the prognosis of patients with HF and preserved EF (HFpEF) patients is dominated by old age and non-HF comorbidities, potentially leading to de-prioritization of cardiological follow-up, diagnostic work-up, and treatments. In HF with reduced EF (HFrEF), several treatments have been shown to improve prognosis, but implementation is lacking. One potential contributing factor is the notion that the heterogenous HFrEF population encountered in real-world practice is poorly represented in the landmark trials that have proven the efficacy of HFrEF treatments.

Aims. The overarching aim was to improve the understanding of prognostic profiles in HF across EF phenotypes, with implications for follow-up in clinical practice and trial design, by studying: • HF follow-up in specialty vs. primary care in HFrEF, HF with mildly reduced EF (HFmrEF), and HFpEF, in terms of use, associated patient profiles, and associated outcomes (study I). • Longitudinal changes in functional capacity (New York Heart Association [NYHA] class) in HFrEF, HFmrEF, and HFpEF in terms of incidence, predictors, and associated outcomes (study II). • The role of iron deficiency (ID) in HFrEF, HFmrEF, and HFpEF, in terms of screening, prevalence, iron need, predictors, and prognosis (study III). • The real-world eligibility for contemporary HF trials (VICTORIA [study IV] and GALACTIC-HF [study V]) in terms of patient characteristics and outcomes.

Materials. All studies used data from the Swedish HF Registry (SwedeHF), linked with Statistics Sweden to obtain socioeconomic variables and the Cause of Death Register to obtain time and cause of death. Additional baseline comorbidities and hospitalizations were obtained from the National Patient Register (studies I-II and studies IV-V) and Region Stockholm’s central data warehouse (VAL-database, study III). For study III, laboratory data including iron biomarkers and hemoglobin were obtained from the Stockholm CREAtinine Measurements project (SCREAM).

Study I: Patient profile and outcomes associated with follow-up in specialty vs. primary care in heart failure. In 75’518 patients from SwedeHF (53% HFrEF, 23% HFmrEF, 24% HFpEF), 64% received follow-up in specialty and 36% in primary care. Specialty care was less likely in females, in-patients, patients with lower socioeconomic status, higher EF, and comorbidities, but more likely in those already receiving HFrEF therapies. Specialty care was independently associated with lower risk of all-cause and cardiovascular death regardless of EF.

Study II: Trajectories in New York Heart Association functional class in heart failure across the ejection fraction spectrum. In 13’535 patients with ≥2 NYHA class assessments (median 1 year apart) recorded in SwedeHF, 69% of patients had stable, 17% improved, and 14% worsened NYHA class. Follow-up in specialty care predicted improving NYHA class, whereas older age, severe HF, and comorbidities predicted worsening. NYHA class trajectories were associated with mortality/morbidity independently of the 1st (baseline) NYHA class assessment, but this association disappeared when adjusting for the 2nd (follow-up) NYHA class assessment. Results were consistent across the EF spectrum.

Study III: Iron deficiency in heart failure: screening, prevalence, incidence and outcome. In 15’197 patients from Region Stockholm with available EF in SwedeHF and laboratory measurements in SCREAM between 2006-2018, iron screening improved since 2016 but remained <25% as of 2018. In 1’486 patients with iron biomarkers at baseline, the prevalence of ID was 55% (HFrEF 54%, HFmrEF 51%, HFpEF 61%). Iron need, i.e. the amount of iron estimated necessary to replete ID, was ≥1’500 mg in 72% of patients. ID was independently associated with higher risk for HF rehospitalizations and of CV death/HF rehospitalizations regardless of EF, but not with all-cause death, CV death, or first HF hospitalization.

Study IV: Eligibility for vericiguat in a real-world heart failure population according to trial, guideline and label criteria. In 23’573 patients with chronic HFrEF from SwedeHF, estimated eligibility for vericiguat according to trial, guidelines, and label criteria was 21%, 47%, and 47%, respectively. The requirement for recent HF hospitalization (within 6 months prior) was the criterion limiting eligibility the most in all scenarios (excluding 49% of the population). In all scenarios, eligibility was much greater in patients who were enrolled in an HF inpatient setting at baseline. Eligible patients were overall older, had more severe HF, greater comorbidity burden, and higher cardiovascular mortality/HF hospitalization rates compared with ineligible patients in all scenarios.

Study V: Eligibility for omecamtiv mecarbil in a real-world heart failure population. In 31’015 patients with HFrEF for ≥3 months registered in SwedeHF, we calculated eligibility for omecamtiv mecarbil by i) applying all the GALACTIC-HF selection criteria (trial scenario), and ii) applying only criteria deemed likely to determine use in clinical practice, i.e. NYHA class, history of worsening HF, Nterminal pro-B-type natriuretic peptides, blood pressure, and renal failure criteria (pragmatic scenario). Eligibility was 21% and 36% in the trial and pragmatic scenarios, respectively. The criterion that most limited eligibility was history of a recent worsening HF event. Eligibility was higher in those with markers of more severe HF. Eligible patients were characterized by more severe HF and higher CV event-rates in both scenarios.

Conclusions. The studies in this thesis suggest that: 1) there are disparities in current triage practices for specialist HF follow-up and iron deficiency screening; 2) iron deficiency is more common and equally prognostic in HFpEF as in HFrEF/HFmrEF; 3) favorable trajectories in functional class are rare in HF overall, and rarer in HFpEF/HFmrEF than HFrEF, and 4) NYHA class changes may be suboptimal for use as trial endpoints; 5) HF trials’ eligibility criteria successfully enrich for HF events also in a real-world population, but limit eligibility.