Data were retrieved from a national multicenter, prospective registry (REDINSCOR II), which includes input from as many as 20 Spanish hospitals of varying complexity. Patients in the registry were older than 18 years and were hospitalized in the cardiology unit for at least 24hours; the main reason for admission was presentation of symptoms compatible with acute HF, both de novo and decompensated, and a chest X-ray indicating pulmonary congestion. The REDINSCOR II registry did not include any participants from the previous REDINSCOR study, which examined patients with chronic HF recruited during cardiology consultations.3

Exclusion criteria included ST-elevation acute coronary syndrome, end-stage disease with a life expectancy < 1 year, and any condition likely to preclude follow-up. HF was diagnosed by the patient's cardiologist in accordance with current HF guidelines.

The registry was compiled consecutively between October 2013 and December 2014 and included a total of 1831 patients admitted with acute HF. Patients were included in the analysis only if they had undergone an echocardiography exam to determine LVEF on admission (1084 patients) or in the preceding 6 months (336), resulting in a study population of 1420 patients (77.6% of the total).

The baseline characteristics of patients excluded due to lack of echocardiography-determined LVEF (411 patients; 22%) are shown in Table 1 of the Supplementary Material.

Follow-up consisted of vital-status and events assessment by telephone interview at 3 months, 6 months, and 1 year after inclusion in the registry. There was minimal loss of follow-up (30 patients; 1.6%), and patients lost to follow-up were not included in the statistical analysis. The study was approved by the ethics committees at the participating hospitals, and all patients gave written informed consent.

Participants were assigned to 1 of 3 groups according to LVEF: HF with reduced EF (HFrEF) for LVEF < 40%; HFmrEF for LVEF between 40% and 49%; and HF with preserved EF (HFpEF) for LVEF ≥ 50%.

The objective of this study was to analyze the REDINSCOR II registry for the clinical, therapeutic, and prognostic characteristics of patients with HFmrEF and compare them with the 2 classic HF categories (HFrEF and HFpEF).

Quantitive variables are expressed as mean ± standard deviation (SD) and qualitative variables are expressed as frequency (percentage). Qualitative variables were compared using the chi-square test or the Fisher exact test, and quantitative variables were compared by ANOVA or the Student t test.

Propensity-score matching was used to balance groups and minimize the bias arising from an observational study of the effect of LVEF category. Propensity-score matching balances the baseline group characteristics for a set of defined variables, allowing analysis of the effect of an intervention or factor. Here, we used 1:1 nearest-neighbor matching without replacement, with a caliper width of 0.2 standard deviations in the propensity score. Between-group balance was assessed by comparing the means of continuous and binary variables using “standardized difference”,4 which is not influenced by sample size and allows comparison of relative balance between variables with different units. In our case, we expected standardized differences < 0.2.

Propensity scores were matched using the MatchIt statistical package in R. The following independent variables were evaluated: age, sex, HF history, diabetes mellitus, hypertension, chronic obstructive pulmonary disease, obstructive sleep apnea hypopnea syndrome, stroke, peripheral vascular disease, cardiac frequency, systolic blood pressure, ischemic etiology, hemoglobin, glomerular filtration rate (CKD-EPI),5 N-terminal pro-brain natriuretic peptide > 1500 ng/L, Charlson age-comorbidity index,6 Barthel frailty index,7 Pfeiffer cognitive impairment test,8 and pharmacological treatment (angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, beta-blockers, and aldosterone antagonists). This analysis generated 2 matched groups: 212 patients with LVEF 40% to 49% matched against 212 with LVEF <40%, and 208 patients with LVEF 40% to 49% matched against 208 with LVEF ≥ 50%.

Kaplan-Meier survival plots were first generated for the full cohort, with comparison by the log-rank test. Survival was then analyzed in the matched cohorts. A Cox regression model was constructed for all-cause death, with LVEF group as the explanatory variable (%). The incidence of readmission for HF was analyzed by drawing the accumulated incidence curves, and between-group differences were analyzed by Gray's test9; the risk effect was determined using a regression model devised by Fine and Gray for competing risks.10,11

The proportional hazard assumption was evaluated by the Shoenfeld residuals test.

Missing data were imputed using the MICE package in R (Multivariate Imputation by Chained Equations). Only 1 imputation was carried out because the percentage of missing data was < 5%. The only exception was plasma N-terminal pro-brain natriuretic peptide, which was classified into 3 categories: < 1500 ng/L, ≥ 1500 ng/L, and “data unavailable” if there were missing data. This was done because the occurrence of missing data for this variable could be related to the event.

Data were analyzed with the statistical packages SPSS 22 and R 3.2. Differences were considered statistically significant at P < .05.

Baseline characteristics on admission of the 1420 patients in the study population are summarized in Table 1.

The mean age of the analyzed study population was 71.8 ± 12.1 years. Of these, 884 (62.3%) were men. The most frequent risk factor was hypertension, affecting 1083 patients (76.6%); 664 patients (47%) had diabetes. More than half the patients (805; 57.2%) had a previous HF diagnosis. The most frequent HF etiology was ischemic heart disease, affecting 491 patients (39.1%). There was also a high frequency of atrial fibrillation on admission, affecting 540 patients (39.4%).

Among all included patients, 583 (41%) had LVEF <40% (HFrEF), 227 (16%) had LVEF 40% to 49% (HFmrEF), and 610 (43%) had LVEF ≥ 50% (HFpEF).

The 3 groups differed significantly in age, sex, hypertension history, and ischemic etiology, and in clinical variables on admission such as the presence of atrial fibrillation, hemoglobin, and N-terminal pro-brain natriuretic peptide. Overall, the HFmrEF group was more similar to the HFpER group in age, hypertension prevalence, and the presence of atrial fibrillation, whereas this group was more similar to the HFrEF group in the predominance of men and the high prevalance of ischemic etiology.

The 3 groups also differed significantly in their capacity for daily living activities measured on the Barthel scale and in their cognitive state measured on the Pfeiffer scale. On these measures, the HFmrEF group tended to more closely resemble the HFpEF group. However, the 3 groups showed no differences in the association between age and comorbidity measured by the Charlson index.

The 3 groups differed significantly in the main treatments prescribed at discharge (Table 1). The HFmrEF group had a high prescription rate for treatments considered class I treatments for HFrEF: more than 70% of HFmrEF patients were prescribed angiotensin-converting enzyme inhibitors (72.4%) or angiotensin receptor blockers and beta-blockers (71.8%), and almost half (45%) were prescribed aldosterone antagonists.

Among the hospitalized HF patients included in the analysis, the number of in-hospital deaths was 53 (3.9%): 21 of these patients had HFrEF (3.9% mortality), 9 had HFmrEF (4.1% mortality), and 23 had HFpEF (3.9% mortality) (P = .972) (Table 2). In-hosptial mortality did not differ between the 3 groups (HFrEF vs HFpEF: odds ratio [OR], 0.96; 95% confidence interval [95%CI], 0.53-1.76; P = .896; HFmrEF vs HFpEF: OR, 1.06; 95%CI, 0.48-2.31; P = .891). At 1-month follow-up, the HFmrEF group showed a tendency toward higher mortality (5.3% mortality in HFrEF, 8.4% in HFrEF, and 6.1% in HFpEF); however, the difference was not statistically significant (P = .265) (HFrEF vs HFpEF: OR, 0.87; 95%CI, 0.53-1.42; P = .578; HfmrEF vs HFpEF: OR, 1.42; 95%CI, 0.80-2.52; P = .237).

Similarly, there were no significant between-group differences in 6-month mortality (P = .37) (HFrEF vs HFpEF: OR, 1.00; 95%CI, 0.73-1.38; P = .993; HFmrEF vs HFpEF: OR, 1.31; 95%CI, 0.87-1.96; P = .196) or at 1 year (P = .278) (HFrEF vs HFpEF: OR, 1.04; 95%CI, 0.78-1.39; P = .810; HFmrEF vs HFpEF: OR, 1.33; 95%CI, 0.93-1.92; P = .122). Overall mortality was 15% at 6 months and 20.4% at 1 year. Kaplan-Meier survival curves at 1-year follow-up are shown for the 3 LVEF categories in Figure 1.

Figure 1.

One-year Kaplan-Meier survival curves for the 3 left ventricular ejection fraction groups in the full study population. HFmrEF, heart failure with mid-range ejection fraction; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; LVEF, left ventricular ejection fraction.

(0.09MB).

The HRrEF, HFmrEF, and HFpEF groups showed no significant differences in the distribution of cause of death. In all 3 groups, the most frequent cause of death was refractory HF, followed by death due to noncardiovascular causes (Table 3).

The overall readmission rate for HF was 9.6% at 1 month, 24.2% at 6 months, and 30.1% at 1 year (Table 2).

No significant differences were found between the 3 groups at any stage of follow-up.

The propensity score matching analysis produced 2 sets of matched patients: 212 patients with LVEF 40% to 49% vs 212 with LVEF < 40%, and 208 patients with LVEF 40% to 49% vs 208 with LVEF ≥ 50%. The balance of variable distribution between appropriately matched patients is shown in Table 4A and Table 4B.

Twelve-month survival in the matched cohorts was analyzed by Cox regression, using LVEF group as the explanatory variable. This analysis identified no significant differences in all-cause mortality between the HFmrEF and HFrEF groups (hazard ratio [HR], 0.77; 95%CI, 0.53-1.13; P = .19) or between the HFmrEF and HFpEF groups (HR, .03; 95%CI, 0.69-1.54; P = .87). Kaplan-Meier 1-year survival curves for the matched cohorts are shown in Figures 2A and 2B.

Figure 2.

One-year Kaplan-Meier survival curves for the matched cohorts. A: LVEF 40% to 49% vs LVEF <40%. B: LVEF 40% to 49% vs LVEF ≥ 50%. LVEF, left ventricular ejection fraction.

(0.14MB).

Matched groups similarly showed no significant differences in readmission for HF (Figures 3A and 3B) or in the distribution of cause of death (Table 2A and Table 2B of the Supplementary Material). However, the HFrEF group showed a tendency toward higher rates for readmission and death due to HF progression.

Figure 3.

Cumulative incidence curves for HF readmission in the matched cohorts. A, LVEF 40% to 49% vs LVEF <40%. B, LVEF 40% to 49% vs LVEF ≥ 50%. HF, heart failure; LVEF, left ventricular ejection fraction.

(0.15MB).

The REDINSCOR II registry includes patients admitted to Spanish cardiology services with decompensated or de novo HF and who were monitored prospectively for 1 year. Our results are therefore limited to this patient category, which is underrepresented in the literature because most registries include outpatients with stable chronic HF or a mix of acute and chronic HF patients.

A notable number of patients (411; 22.4%) were excluded from the analysis because LVEF was not determined on admission or in the preceding 6 months; however, on comparison, this group showed no significant difference from the analyzed study population in terms of mortality or HF readmission.

A further point is that patients were assigned to an HF category based on LVEF recorded during hosptitalization, a period of clinical instability. LVEF progression during follow-up is thus unknown, and information is therefore lacking on patients who may have undergone dynamic changes in LVEF during this period, possibly recovering LVEF and changing category from HFrEF to HFmrEF or HFpEF. Most LVEF values (76.3%) were recorded during hospitalization, and LVEF values were reported by the investigators (cardiologists) at each participating center, without any cross-checking to control for measurement variability; it is thus possible that some patients might have been assigned to a different LVEF group than similar patients at another center, which might limit our results to some degree. The evaluated treatments are those prescribed at hospital discharge, and we are therefore also unable to confirm whether the therapeutic strategy was maintained throughout the follow-up period, a factor that could influence prognosis.

There may also be limitations related to the classification of cause of death, especially out-of-hospital death; nonetheless, this is the first study to describe the specific cause of death in HFmrEF patients, showing no significant differences from the other analyzed groups, either in the full study population or in the matched cohorts.

Although we found no statistically significant between-groups differences for the clinical parameters analyzed, our results may be somewhat limited by a lack of sufficient statistical power in these comparisons. Moreover, although propensity score matching is more robust than classic regression analysis, it has its weaknesses, such as the inability to correct for some unmeasured confounding factors, resulting in a degree of residual confounding.