Biomarkers for eligibility and surrogate endpoints in heart failure trials

Sammanfattning: Background: In heart failure (HF) with reduced ejection fraction (HFrEF), randomized controlled trials have provided effective treatments, but prognosis still remains poor. HF with mid-range EF (HFmrEF) has no evidencebased therapy and represents a newly characterized and relevant population for future trials. Trials in HF with preserved EF (HFpEF) have failed to provide any effective treatment, with several concerns about their design. Aims: Overall aim is to provide evidence to improve trial design in HF, investigating the use of natriuretic peptides (NPs) as surrogate endpoint, and as eligibility criterion to foster the enrichment of trials for cardiovascular (CV) vs. non-CV events. Specific aims were: • to assess the associations between changes in NP [B-type NP (BNP) and N-Terminal pro-BNP (NTproBNP)] levels over time and prognosis in chronic HFpEF and HFmrEF (Study I) and in acute decompensated HFpEF (Study II); • to compare levels, the independent determinants of levels and the prognostic role of NT-proBNP across EF categories (Study III); • to evaluate the associations between NT-proBNP and CV and non-CV outcomes across EF categories and in specific subgroups, and the associations between HF therapies and outcomes according to NTproBNP levels (Study IV). Changes in NT-proBNP and prognosis in chronic HFpEF and HFmrEF: We studied 650 HFpEF/HFmrEF outpatients enrolled in the Swedish Heart Failure registry (SwedeHF) between 2000 and 2012, reporting serial NT-proBNP assessments. A reduction in NT-proBNP at the median time of 7 months from the first measurement was associated with a reduction of mortality/HF hospitalization risk by 54% in the overall population, by 51% in HFpEF and by 61% in HFmrEF. Changes in BNP/NTproBNP levels and prognosis in acute decompensated HFpEF: From the Karolinska-Rennes (KaRen) study, 361 patients with acute decompensated HFpEF and BNP/NTproBNP measurements at the baseline and at the 4-8 weeks follow-up visit were analyzed. Changes in NPs from baseline to follow-up visit were not significantly associated with the risk of mortality/HF hospitalization although a trend toward a reduction in risk following the reduction in levels was observed. Levels, predictors of levels and prognostic/discriminatory role of NT-proBNP across EF categories: We analyzed 9,847 outpatients with HFpEF (18%), HFmrEF (22%) or HFrEF (60%) with at least one NTproBNP assessment, enrolled in the SwedeHF between 2000 and 2012. NT-proBNP levels were significantly higher in HFrEF (2,288 pg/ml) vs. HFpEF (1.428 pg/ml) and HFmrEF (1,540 pg/ml). Across EF categories, there were several different independent determinants for NT-proBNP levels, with atrial fibrillation more important in HFmrEF and HFpEF, diabetes and hypertension in HFmrEF, and age and body mass in HFrEF and HFmrEF, whereas there were no differences for renal function, New York Heart Association class, heart rate and anemia. NT-proBNP >vs. ≤median was associated with increased risk of mortality and mortality/ hospitalization with hazard ratios significantly higher in HFmrEF and HFpEF vs. HFrEF. NT-proBNP had greater area under the curve for death/HF hospitalization in HFmrEF vs. HFpEF and HFrEF. NT-proBNP levels and risk of CV/non-CV events across EF categories: We studied 15,849 patients with HFpEF (23%), HFmrEF (21%) and HFrEF (56%) and at least one NT-proBNP assessment, enrolled in SwedeHF between 2000 and 2012. Increasing NT-proBNP levels were associated with a steeper increase in CV vs. non-CV event rates in HFpEF vs. HFmrEF vs. HFrEF. CV to non-CV event ratio increased together with the increase in NT-proBNP in HFpEF and HFrEF, but only in the lower range in HFmrEF. The association between HF treatments (angiotensin converting enzyme inhibitors or angiotensin receptor blockers and beta-blockers) and CV/non-CV events risk was consistent in NT-proBNP ≤ and >median. Conclusions: The association between NT-proBNP levels and prognosis across the EF spectrum, together with the association between reduction in NT-proBNP levels and improvement in prognosis in HFpEF, HFmrEF and HFrEF supports the use of NT-proBNP as surrogate endpoint in phase II trials in chronic HF. We did not observe any significant association between changes in BNP/NT-proBNP and prognosis in acute decompensated HFpEF. The observed relationship between NT-proBNP levels and CV and non-CV events supports the use of NTproBNP for eligibility and enrichment for CV events in HF trials, but the cut-off levels should consider the differences in comorbidities across the EF spectrum. Potential treatment response according to NT-proBNP levels deserves further investigation.