Classification and regression tree analysis for identification of long-term risk in patients with heart failure with preserved ejection fraction. final results of the Argentine ARFEY-Preser Registry

Introduction Heart failure with preserved ejection fraction (HFpEF) is a complex clinical syndrome characterized by a wide range of phenotypes and the presence of various comorbidities. Despite its prevalence, there remains a limited understanding of the current prognosis for patients affected by this condition. Objective To evaluate the incidence of long-term events as well as to identify prognostic predictors of mortality in HFpEF in Argentina. Material and Methods The ARFEY-Preser registry was a multicenter, prospective, and observational study conducted by Heart Failure and Pulmonary Hypertension Committee of the Argentine Federation of Cardiology, and included both outpatients and patients evaluated at the time of hospital discharge. Enrollment criteria required a LVEF of 50% or higher, as well as the presence of current or previous signs and symptoms of heart failure and documented evidence of structural heart disease. The primary endpoint was all-cause mortality by the end of the follow-up period. With the variables that showed a significant association with mortality, a classification and regression tree analysis (CART) was performed to categorize the risk. Results Between December 22 and May 24, 47 investigators included 638 patients. The mean age was 72.5±11.2 years (26-99), 62.1% female, and mean LVEF was 60.5±6.4% (50-84%). The mean follow-up was 21±5 months and was completed in 97% of the population. At 12 months, the incidence rates for key clinical outcomes were as follows: death 7.3%, heart failure hospitalization (HFH) 11.5%, and the combined endpoint of death or HFH 16%. Survival at 24 months was 90% and the variables associated with mortality in the univariate analysis were: obesity (OR=0.49, p=0.013), type 2 diabetes mellitus (T2DM) (OR=0.49, p=0.014), HFH in the previous year (HFH1a)(OR=3.3, p<0.0001), functional class 3-4 (FC34) (OR=2.1, p=0.006), anemia (OR=2.1, p=0.009), systolic pressure less than 130 mmHg (OR=1.8, p=0.046), sodium ≤135 mEq/l (HypoNa)(OR=2.9, p<0.0001), estimated pulmonary systolic pressure ≥40 mmHg (PH) (OR=2.5, p=0.002) and EQ5D questionnaire <60 (OR=2.0, p=0.013). The CART analysis (Graph 1) allowed the population to be classified into 4 risk groups: 1-Low (No HypoNa, no T2DM and no HFH1a; or no HypoNa, T2DM without PH); 2-Lower intermediate (HypoNa, without PH, or no HypoNa, T2DM, with PH)); 3-Upper intermediate (HypoNa, without T2DM, with HFH1a) and 4-High (HypoNa and PH), with a survival rate of 96% (HR=1.0); 91% (HR=2.6; 95%CI=1.0-6.9, p=0.051); 87% (HR=4.0; 95%CI=1.7-9.6; p=0.002) and 65% (HR=11.3; 95%CI=4.6-28.0; p<0.001), respectively. Conclusions The current course of HFpEF reflects a poor prognosis despite the availability of new therapeutic options. Beyond its clinical profile, parameters that reflect the severity of HF (hospitalizations, pulmonary hypertension and hyponatremia) emerge as straightforward effective markers for identifying patients at higher risk.Graph 1Survival curves