Prognosis of Heart Failure with Preserved Ejection Fraction in Primary Care by the H2FPEF Score

After 18 years of the classic pathophysiological characterization of what we currently call “heart failure (HF) with preserved ejection fraction (HFpEF)”, its diagnostic criteria remain questionable and evolving.1 Even with historically different diagnostic criteria, establishing its prevalence and assessing its prognosis is an increasing challenge in primary care. In 2016, the European Society of Cardiology (ESC) established the following as criteria for HFpEF: the presence of signs and or symptoms of HF; left ventricle ejection fraction (LVEF) greater than or equal to 50%; elevation of natriuretic peptides; and the presence of cardiac structural or functional alteration.2 In general, primary care physicians find it difficult to deal with patients with multiple comorbidities, signs and symptoms of HF and preserved LVEF and, in such cases, easily accessible tools that assist the physician in diagnosis and prognosis can be extremely useful. Recently, a scoring system called the H2FPEF score3 was proposed to estimate the diagnostic probability of HFpEF in patients assisted in a specialized HF unit. However, the prognostic utility of this score remains 666

unknown in primary care patients. The aim of the present study is to apply the H 2 FPEF score to patients in primary care and to verify its predictive power of outcomes.

Study Design
We carried out a longitudinal study, derived from the DIGITALIS study, whose design has been previously published 4 , which included 402 individuals consecutively, aged 45 years or over, enrolled in primary care, in a city of 400,000 inhabitants, in Rio de Janeiro State, Brazil. Initial data were collected from July 2011 to December 2012 and the revaluation took place between January and December 2017, that is, five years later.

Study Population
The selection of the primary care units and the population was performed through random sequence T h e T D E e x a m s we r e p e r f o r m e d b y t w o echocardiographists without previous knowledge of the results of the other tests using two devices: Cypress 20 (Acuson, Siemens, USA) and AU-3 Partner (Esaote, Italy). The examinations were performed according to the recommendations for quantification of chambers of the American Society of Echocardiography (ASE) and the European Association of Echocardiography (EAE) 5 .

Outcome Measures
The diagnosis of HFpEF was confirmed in individuals with signs and symptoms of HF, LVEF ≥50%, elevation of BNP, and presence of structural alteration or diastolic dysfunction 2 . This evaluation was performed by two independent cardiologists who were blind to the study.
The H 2 FPEF score uses six clinical and echocardiographic variables obtained in the evaluation of patients with HF symptoms. The variables were scored according to the strength of their respective association. The overall score of H 2 FPEF ranged from 0 to 9. The variables used and their score were: body mass index >30kg/m 2 (2 points); use of two or more drugs to treat hypertension (1 point); atrial fibrillation (3 points); PASP >35 mmHg (1 point); age >60 years (1 point); and high left ventricular filling pressures, E/e'> 9 (1 point) 3 ( Figure 2). After five years, the patients in this study were reassessed as to the occurrence of the composite outcome: death from any cause or hospitalization for cardiovascular disease, including decompensated HF, coronary artery disease, stroke and vascular diseases.

Statistical Analysis
Statistical analysis was performed with SPSS v 21.0 software (Chicago, Illinois, USA). Continuous variables were expressed as median and interquartile range, as none of them was positive for normality when tested using the Kolmogorov-Smirnov test. Categorical variables were expressed in absolute numbers and/or percentages. For comparison between groups (categorical variables), we used chi-square tests with continuity correction and Fisher's exact test when necessary. The Mann Whitney test was used to verify the existence Figure 2 -H 2 FPEF -score for each characteristic (maximum total of 9 points) 3 of differences between continuous variables. In all comparisons, bilateral tests were performed and p-values <0.05 were considered statistically significant.

Ethical Considerations
This study was conducted in accordance with the principles set out in the Declaration of Helsinki, revised in 2000 (Scotland 2000). The study protocol was approved by the Institution's Research Ethics Committee under the number 0077.0.258.000-10.

Results
Among the 402 subjects (mean age = 60.2±10.0 years, 71% women) involved in the study, HFpEF was diagnosed in 58 subjects (14.4%) and these patients with HFpEF had a score H 2 FPEF 2 (1)(2)(3)(4). All the parameters used in the score model were significantly different between patients with and without HFpEF, except for body mass index (BMI) and pulmonary artery systolic pressure (PASP). The anthropometric, clinical and laboratory characteristics are shown in Table 1. Table 2 shows the main echocardiographic parameters.
Among patients with a H 2 FPEF score ≥ 4, 30% had HFpEF and among patients with a score ≤ 4, 12% presented HFpEF. BNP values were higher in patients with HFpEF compared to those without HFpEF, regardless of the score obtained in the H 2 FPEF. In patients with HFpEF and score ≤3, a lower outcome rate (21%) was observed compared to patients with HFpEF and a score ≥4 (53%), showing that the higher the H 2 FPEF score, the greater the risk of death and/or hospitalization due to cardiovascular disease.
After a 5-year follow-up, 42 (10.4%) composite outcomes were observed, with 21% in patients with HFpEF and a score ≤4, and 53% in patients with HFpEF and a score ≥4. In patients without HFpEF, the rate of outcomes was 7% in patients with a score ≤4 and 6% in patients with a score ≥4 (Table 3).

Discussion
This study evaluates the H 2 FPEF score in patients in primary care. Our data show that patients with HFpEF and a H 2 FPEF score ≥4 are at increased risk of death from any cause or hospitalization for cardiovascular disease.
A secondary analysis of the TOPCAT study evaluated the association between the probability of HFpEF by the H 2 FPEF score system and the primary endpoint composed of cardiovascular death, aborted cardiac arrest or HF hospitalization in patients with HFpEF using spironolactone or placebo. The high probability of HFpEF according to the H 2 FPEF score was associated with worsening renal function, elevated natriuretic peptide values, increased left ventricle mass and left atrium (LA) size 6 .
When comparing the patients with and without HFpEF according to the H 2 FPEF score, we observed that in those with a score ≤4 there were significant differences between worsening renal function, increased left ventricle (LV) mass and LA volume index (LAV-I). However, in individuals with a score ≥4, the difference was only observed in relation to LAV-I.
Although natriuretic peptides are part of the diagnosis of HFpEF in the ESC guideline, their values did not contribute the score. NT-proBNP values were missed in 24% of the patients enrolled in the elaboration of the score because some cardiologists did not request the test during clinical evaluation 3 . Our data show that BNP was higher in patients with HFpEF compared to individuals without HFpEF, regardless of the H 2 FPEF score.
The Meta-Analysis Global Group in Chronic Heart Failure (MAGGIC) 7 , a risk score composed of 13 clinical variables, evaluated patients with HF across the spectrum of LVEF. Among the 13 variables used in the MAGGIC score, three were common to the H 2 FPEF score (age, BMI and hypertension). The MAGGIC score did not use echocardiographic parameters and, like H 2 FPEF, natriuretic peptides values were not considered. Of the patients defined with HFpEF (n = 407), followed during 3.6 ± 1.8 years, 28% died, 32% were hospitalized for HF and 55% had cardiovascular hospitalization and/or death. When compared to our assessment of the H 2 FPEF score, we observed similar values in relation to the composite outcome observed in the MAGGIC score (55%) and in H 2 FPEF, with a score ≥ 4 (53%). The H 2 FPEF score, like the MAGGIC, is a simple score, but with a smaller number of variables, and shows to be useful for morbidity and mortality risk stratification in HFpEF.
The study has limitations related to a single center design, the relative small sample size, with female predominance, and the score was applied retrospectively. Multicenter studies with large populations are needed to confirm our data.