ESC HEART FAILURE ORIGINAL ARTICLE ESC Heart Failure 2022; 9: 3804–3813 Published online 2 August 2022 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/ehf2.14098 Quality of life in patients with heart failure and improved ejection fraction: one-year changes and prognosis Elisabet Zamora1,2,3, Beatriz González1, Josep Lupón1,2,3, Andrea Borrellas1, Mar Domingo1, Evelyn Santiago-Vacas1, Germán Cediel1, Pau Codina1,2, Carmen Rivas1, Ana Pulido1, Eva Crespo1, Patricia Velayos1, Violeta Diaz1 and Antoni Bayes-Genis1,2,3* 1 Heart Failure Clinic and Cardiology Service, University Hospital Germans Trias i Pujol, Barcelona, Spain; 2Department of Medicine, Universitat Autonoma de Barcelona, Barcelona, Spain; and 3CIBERCV, Instituto de Salud Carlos III, Madrid, Spain Abstract Aims The criteria for patients with heart failure (HF) and improved ejection fraction (HFimpEF) are a baseline left ventricular ejection fraction (LVEF) ≤40%, a ≥10-point increase from baseline LVEF, and a second LVEF measurement >40%. We aimed to (i) assess patients with HF and reduced LVEF (HFrEF) at baseline and compare quality of life (QoL) changes between those that fulfilled and those that did not fulfil the HFimpEF criteria 1 year later and (ii) assess the prognostic role of QoL in patients with HFimpEF. Methods We reviewed data from a prospective registry of real-world outpatients with HF that were assessed for LVEF and QoL at a first visit to the HF clinic and 1 year later. QoL was evaluated with the Minnesota Living with Heart Failure Question- naire (MLWHFQ). The primary prognostic endpoint was the composite of all-cause death or HF hospitalization. Results Baseline and 1-year LVEF and MLWFQ scores were available for 1040 patients with an initial LVEF ≤40% (mean age, 65.2 ± 11.7 years; 75.9% men). The main aetiology was ischaemic heart disease (52.9%), and patients were mostly in New York heart Association Classes II (71.1%) and III (21.6%). At baseline, the mean LVEF was 28.5% ± 7.3, and the mean MLWHFQ score was 30.2 ± 19.5. After 1 year, the mean LVEF increased to 38.0% ± 12.2, and the MLWHFQ scores improved to 17.4 ± 16.0. In 361 patients that fulfilled the HFimpEF criteria (34.7%), significant improvements were observed in both LVEF (from 28.7% ± 6.6 to 50.9% ± 7.6, P < 0.001) and QoL (from 32.9 ± 20.6 to 16.9 ± 16.0, P < 0.001). Patients that did not fulfil the HFimpEF criteria also showed significant improvements in LVEF (from 28.4% ± 7.6 to 31.1% ± 7.9, P < 0.001) and QoL (from 28.7 ± 18.8 to 17.6 ± 15.9, P < 0.001). However, the QoL improvement was significantly higher in the HFimpEF group ( 16.0 ± 23.8 vs. 11.1 ± 20.3, P = 0.001), despite the worse mean baseline MLWHFQ score, compared with the non-HFimpEF group (P = 0.001). The 1-year QoL was similar between groups (P = 0.50). The 1-year MLWHFQ score was independently as- sociated with outcomes; the hazard ratio for the composite endpoint was 1.02 (95% CI: 1.01–1.03, P = 0.006). In contrast, the QoL improvement (with a cut-off ≥5 points) was not independently associated with the composite outcome. Conclusions Patients with HFrEF showed improved QoL after 1 year, regardless of whether they met the HFimpEF criteria. The similar 1-year QoL perception between groups suggested that factors other than LVEF influenced QoL perception. The 1-year QoL was superior to the QoL change from baseline for predicting prognosis in patients with HFimpEF. Keywords Heart failure; Quality of life; Left ventricular ejection fraction; Heart failure with improved ejection fraction; Outcomes Received: 13 April 2022; Revised: 16 June 2022; Accepted: 20 July 2022 *Correspondence to: Antoni Bayes-Genis, Heart Institute, Hospital Universitari Germans Trias i Pujol, Carretera del Canyet, s/n, 08916 Badalona, Barcelona, Spain. Tel: +34 934978915; Fax: +34 934978939. Email: abayesgenis@gmail.com © 2022 The Authors. ESC Heart Failure published by John Wiley & Sons Ltd on behalf of European Society of Cardiology. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. Quality of life in patients with heart failure and improved ejection fraction: one-year changes and prognosis 3805 Introduction or other hospital departments. The criteria for referral to the HF clinic were HF diagnosed according to the European Heart failure (HF) is a chronic condition with signs and symp- Society of Cardiology guidelines, regardless of aetiology, and toms that affect the patient’s quality of life (QoL).1 From the at least one HF hospitalization and/or reduced LVEF.31,32 All very start of implementing QoL questionnaires in HF, QoL was patients were examined regularly during follow-up visits at reported to be related to many factors, including age, sex, the HF clinic, according to their clinical needs. All patients New York Heart Association (NYHA) functional class, hospital- were treated according to a unified protocol. Follow-up visits izations, fragility, and different new HF treatments, among included a minimum of one visit with a nurse every 3 months others.2–10 However, studies have provided inconsistent and and one visit with a physician (cardiologist, internist, or family controversial results about the association between QoL physician) every 6 months. Patients also attended optional and left ventricular ejection fraction (LVEF).2,3,6,11–13 The visits with specialists in geriatrics, psychiatry, and 2021 universal definition of HF specifically describes the rehabilitation,31,32 and in recent years, visits with a nephrolo- criteria for patients with HF that have shown improved LVEF gist and endocrinologist were included. (HFimpEF) as follows: HF with a baseline LVEF ≤40%, a ≥10- point increase from baseline LVEF, and a second measure- ment of LVEF >40%.14 Improvement or recovery in LVEF Outcomes has been associated with better clinical outcomes, including HF-related hospitalizations and survival.15–23 However, the The primary study endpoint was the change in QoL related to influence of LVEF improvement (or LVEF recovery) on QoL, an improvement in LVEF. Secondary clinical outcome end- although reported in a small number of studies,6,22–24 has points included the composite endpoint of all-cause death not been completely established. Indeed, it has been widely or HF-related hospitalization, all-cause death alone, and the reported that QoL could predict outcomes in HF.25–30 More- number of subsequent HF-related hospitalizations. Fatal over, it remains to be defined whether the QoL or QoL events were identified from patient health records (including changes actually influence the outcomes in patients with records from hospital wards, the emergency room, and gen- HFimpEF. eral practitioners) or by contacting relatives. Data were veri- The present study included patients with HF and reduced fied with the Catalan and Spanish Health Systems databases LVEF (HFrEF) at a first visit to an outpatient HF clinic. The and the Spanish National Death Registry. Adjudication of Minnesota Living with Heart Failure Questionnaire events was performed by an ad hoc committee (JL, MdeA, (MLWHFQ) was administered to assess QoL at the first visit BG, and MD), and discrepancies were resolved by two inde- and at the 1-year follow-up. We aimed to determine (i) pendent researchers (PM and GC). Hospitalizations were whether patients that fulfilled the HFimpEF criteria after identified from the clinical records of patients with HF, from 1 year experienced greater QoL improvement than patients hospital ward records, and from the electronic Catalan his- that did not fulfil the HFimpEF criteria after 1 year and (ii) tory records. whether QoL had prognostic value for outcomes in patients All patients provided written informed consent, during the with HFimpEF. baseline visit, for the use of their clinical data for research purposes. The study was performed in compliance with the laws protecting personal data, in accordance with the inter- national guidelines on clinical investigations from the World Methods Medical Association’s Declaration of Helsinki. The local ethics committee approved the study. Study population We reviewed a prospective registry of real-world outpatients HF type definition with HF to identify consecutive, ambulatory patients admit- ted to a structured multidisciplinary HF clinic at a university LVEF was assessed at the first visit and at the 1-year follow- hospital between August 2001 and August 2021. All patients up, according to the recommendations of the American Echo- were evaluated to determine LVEF and QoL with the cardiography Society guidelines. LVEF was measured from MLWHFQ at their first visit to the HF clinic and at a 1-year fol- apical two-chamber and four-chamber views with Simpson’s low-up. The study inclusion criteria were a diagnosis of HFrEF method. Patients were first classified by their ventricular at baseline, a prospectively scheduled second LVEF measure- function, according to the 2021 universal definition of HF.14 ment, and two QoL assessments with the MLWHFQ, one at Only patients with LVEF ≤40% at the first visit were included baseline and one at 1 year. Patients were referred to the HF in the present study. After 1 year, echocardiography was per- clinic mostly by the cardiology or internal medicine depart- formed, and patients were reclassified into two groups: (i) ment, and to a lesser extent, by the emergency department HFimpEF: HF with a baseline LVEF ≤40%, ≥10-point increase ESC Heart Failure 2022; 9: 3804–3813 DOI: 10.1002/ehf2.14098 3806 E. Zamora et al. from baseline LVEF, and a second measurement of LVEF latter analyses, an out-of-hospital death due to HF was con- >40%; (ii) non-HFimpEF: patients who did not fulfil such sidered an additional event. Statistical analyses were per- criteria (permanent HFrEF, HFmrEF without 10-point increase formed with SPSS 24 (SPSS Inc., Chicago, Illinois) and Stata. in LVEF). A two-sided P < 0.05 was considered significant. QoL assessment QoL was measured with an HF-specific QoL questionnaire, the Results MLWHFQ.1 The Spanish version of this questionnaire has been widely used3 and was prospectively validated.33 The We retrieved data from August 2001 to August 2021 and MLWHFQ consisted of 21 questions that evaluated the im- identified 1040 patients with both baseline and 1-year LVEF pact of HF on physical, psychological, and social aspects of and MLWFQ scores and an initial LVEF ≤40%. Table 1 shows the patient’s life. Answers ranged from 0 (no limitation) to the baseline demographic and clinical characteristics of pa- 5 (maximal limitation); thus, the global scores ranged from tients and the treatments administered during follow-up. 0 to 105, and higher scores reflected a worse QoL. A The mean age was 65.2 ± 11.7 years, 75.9% of the patients 5-point change was considered the minimally important dif- were men, the main aetiology was ischaemic heart disease ference (https://qol.thoracic.org/sections/instruments/ko/ (52.9%), and most patients were in New York heart Associa- pages/mlwhfq.html); thus, in addition to the magnitude of tion Classes II (71.1%) and III (21.6%). Most patients received change between the two assessments (baseline and 1 year), currently recommended HF treatments. Table S1 shows the we categorized patients according to whether they showed treatments administered at baseline, at 1 year, and during a 5-point improvement in the score (i.e., a 5-point reduction follow-up. in the MLWHFQ score was considered an improved QoL). When necessary, an HF-specialized nurse assisted patients in completing the questionnaire.3 The level of assistance Changes in LVEF and MLWHFQ depended on the patient’s reading and writing capabilities. When a patient had difficulty completing the questionnaire, The mean baseline LVEF was 28.5% ± 7.3, and the mean base- the nurse read the MLWHFQ aloud and completed each line MLWHFQ score was 30.2 ± 19.5. At 1 year, the mean LVEF question, based on the patient’s oral answer. In all cases, increased to 38.0% ± 12.2, and the mean MLWHFQ score im- the nurses attempted to ensure that they did not alter the re- proved to 17.4 ± 16.0 (LVEF, Figure 1; MLWHFQ, Figure 2). sponse of the patient in any way, but simply intervened for The HFimpEF criteria were fulfilled by 361 patients guidance or assistance; they never acted in an interested (34.7%). These patients showed significant, marked improve- way that could compromise the patient’s independence. ments in both LVEF (from 28.7% ± 6.6 to 50.9% ± 7.6, P < 0.001) and QoL (from 32.9 ± 20.6 to 16.9 ± 16.0 points, P < 0.001). Patients that did not fulfil the HFimpEF criteria Statistical analysis also showed improvements in LVEF (from 28.4% ± 7.6 to 31.1% ± 7.9, P < 0.001) and QoL (from 28.7 ± 18.8 to Categorical variables are expressed as percentages. Continu- 17.6 ± 15.9, P < 0.001). The improvement in QoL was signif- ous variables are expressed as the mean [standard deviation icantly higher in the HFimpEF group ( 16.0 ± 23.8 vs. (SD)] or median [interquartile range: Q1–Q3 (IQR)], according 11.1 ± 20.3 points, P = 0.001), because the baseline to whether the data were normally or non-normally distrib- MLWHFQ score was worse in the HFimpEF group uted. Normally distributed data were assessed with normal (P = 0.001) than in the non-HFimpEF. However, at 1 year, Q-to-Q plots. Differences between HF types were assessed the QoL scores were similar between groups (P = 0.50). with the chi-squared test, Student’s t-test, or Mann–Whitney When patients were categorized into those with and with- U test, as appropriate. Univariable and multivariable binomial out at least a 5-point improvement in the MLWHFQ score logistic regression was performed to assess which variables (improved QoL and non-improved QoL, respectively), QoL im- were associated with QoL improvement, and results are provement was observed significantly more frequently in the expressed as the odds ratio (OR) or hazard ratio (HR) and HFimpEF group (67.9%) than in the non-HFimpEF group 95% confidence interval (95% CI). (61.4%, P = 0.04). Indeed, in univariable logistic regression, The prognostic role of QoL in HFimpEF was assessed with HFimpEF was significantly associated with an improved QoL Cox regression analyses. Univariable and multivariable analy- (OR: 1.33, 95% CI: 1.01–1.74, P = 0.04). However, when other ses were performed. Multivariable models included covari- variables classically associated with QoL, such as age, sex, ates with P values < 0.10 in the univariable analyses. Recur- NYHA functional class, and the number of HF-related hospi- rent HF-related hospitalizations were analysed with binomial talizations in the previous year were added to a multivariable negative regression (univariable and multivariable), and re- model, HFimpEF did not remain independently associated sults are expressed as the incidence rate ratio (IRR). For the with an improved QoL (Table 2). ESC Heart Failure 2022; 9: 3804–3813 DOI: 10.1002/ehf2.14098 Quality of life in patients with heart failure and improved ejection fraction: one-year changes and prognosis 3807 Table 1 Baseline clinical characteristics of patients with HFrEF that either showed improved EF (HFimpEF) or no EF improvement (no HFimpEF) HFimpEF Non-HFimpEF Characteristic N = 361 N = 679 P value Age, years 63.4 ± 12.1 66.2 ± 11.3 <0.001 Male sex, n (%) 255 (70.6) 535 (78.8) 0.003 Aetiology, n (%) <0.001 Ischaemic HD 114 (31.6) 437 (64.4) Dilated CM 96 (26.6) 101 (14.9) Hypertensive CM 26 (7.2) 33 (4.9) Alcoholic CM 50 (13.9) 21 (3.1) Drug-induced CM 16 (4.4) 17 (2.5) Valvular disease 31 (8.6) 35 (5.2) Other 28 (7.8) 35 (5.2) a LVEF, % 28.2 ± 6.6 28.4 ± 7.6 0.55 HF duration, months 2 (1–12) 12 (2–60) <0.001 NYHA class, n (%) 0.001 I 29 (8.0) 41 (6.0) II 273 (75.6) 467 (68.8) III 58 (16.1) 166 (24.4) IV 1 (0.3) 5 (0.7) Diabetes, n (%) 144 (39.9) 276 (40.6) 0.81 Hypertension, n (%) 226 (62.6) 405 (59.6) 0.35 COPD, n (%) 51 (14.1) 123 (18.1) 0.81 Atrial fibrillation/flutter 63 (17.5) 112 (16.5) 0.10 a Treatments FU , n (%) ACEI or ARB 331 (91.7) 580 (85.4) 0.37 Beta-blockers 342 (94.7) 640 (94.4) 0.75 ARM 271 (75.1) 501 (73.8) 0.65 Loop diuretics 334 (92.5) 639 (94.1) 0.32 ARNI 64 (17.7) 118 (17.4) 0.89 Digoxin 145 (40.2) 312 (45.9) 0.07 Ivabradine 122 (33.8) 157 (23.1) <0.001 CRT 40 (11.1) 138 (20.3) <0.001 ICD 39 (10.8) 184 (27.1) <0.001 ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; ARNI, angiotensin receptor neprilysin inhibitor; CM, cardiomyopathy; COPD, chronic obstructive pulmonary disease; CRT, cardiac resynchronization therapy; EF, ejection fraction; HF, heart failure; HFrEF, heart failure with reduced ejection fraction; ICD, implantable cardiac defibrillator; LVEF, left ventricular ejection fraction; MRA, mineralocorticoid receptor antagonist; NYHA, New York Heart Association. a At any moment during follow-up. Relationship between QoL and outcomes plot representation of different potential QoL assessments and their associations with all-cause death and the composite Among the 361 patients with HFimpEF, during a mean endpoint of all-cause death or HF hospitalization. follow-up of 6.7 ± 4.6 years, 152 patients died [50.7% from A QoL improvement was not associated with either non-cardiovascular causes (Table S2)], 86 patients experi- all-cause death (HR: 0.80, 95% CI: 0.57–1.11) or recurrent enced 166 HF-related hospitalizations (Table S3), and 175 pa- HF-related hospitalizations (IRR: 0.90, IQR: 0.43–1.86; tients experienced the composite endpoint of all-cause death P = 0.77). In contrast, the 1-year MLWHFQ score was signifi- or HF-related hospitalization. cantly associated with all-cause death in both the univariable Analyses of the relationships between an improved QoL and multivariable analyses (Table 4). However, its association and HFimpEF outcomes showed that a QoL improvement with recurrent HF-related hospitalizations was only signifi- was significantly associated with the composite endpoint cant in the univariable analysis (Table S5). of all-cause death or HF-related hospitalization in the univariable analysis (HR: 0.73, 95% CI: 0.54–0.999; P = 0.049), but this association did not remain significant in a multivariable analysis, when age, sex, NYHA functional Discussion class, ischaemic aetiology, and the number of HF admissions in the previous year were included in the model (Table S4). The two main findings in this study were: (i) Patients with In contrast, the 1-year MLWHFQ score showed a significant HFimpEF showed greater improvement in QoL than those association with the composite endpoint, both in the with non-HFimpEF, although the groups showed similar univariable analysis and in a very comprehensive multivari- one-year MLWHFQ scores; and (ii) QoL was significantly asso- able Cox regression analysis (Table 3). Figure 3 shows a forest ciated with outcomes in HFimpEF, which was not previously ESC Heart Failure 2022; 9: 3804–3813 DOI: 10.1002/ehf2.14098 3808 E. Zamora et al. Figure 1 Violin plots of baseline and one-year LVEF measurements. Each Figure 2 Violin plots of baseline and one-year MLWHFQ scores. The vio- violin plot illustrates the kernel probability density (i.e. the width of the lin plot illustrates the kernel probability density (i.e. the width of the shaded area represents the proportion of the data located there). Inside shaded area represents the proportion of the data located there). Inside the violin plots, boxplots indicate the median and interquartile range; the the violin plots, boxplots indicate the median and interquartile range; the whiskers indicate 1.5 times the interquartile range. Green violin plots rep- whiskers indicate 1.5 times the interquartile range. Blue violin plots rep- resent the baseline LVEFs. Orange violin plots represent the 1-year LVEFs. resent the baseline scores. Red violin plots represent the 1-year scores. described, to the best of our knowledge. As a prognostic fac- relationship.3,11–13 Nevertheless, QoL affects all patients with tor, the 1-year QoL score was stronger to the change in QoL HF; thus, we expected to find that the perceived QoL im- from baseline. QoL is enormously important to patients with proved more among patients with HFimpEF than in those HF; among the chronic diseases, HF has one of the largest ef- with non-HFimpEF. In a small sample of 35 patients with fects on QOL.14,34 Indeed, in some patients with advanced HFrEF that showed improved LVEFs to 50% or more, diseases, relief from symptoms35 and QoL36 were reported Wohlfahrt et al.24 showed that recovery of systolic function to be even more important than life expectancy. was associated with HF-associated QoL improvements, and The prevalence of HFimpEF depends strongly on the co- for each 10% increase in LVEF, the Kansas City Cardiomyopa- hort characteristics and the definition and cut-offs used. We thy Questionnaire score improved by a mean (SD) of 4.8 ± 1.6 found that 34.7% of our patients with non-HFimpEF evolved points (P = 0.003). In addition, DeVore et al.22 very recently to HFimpEF. In the meta-analysis performed by He et al.,37 reported QoL improvements related to a ≥10% increase in the pooled prevalence of HFimpEF was only 22.64%, although LVEF in 635 patients out of 2092 with initial HFrEF. Among it ranged from 10 to 52%. patients with HFimpEF, the Kansas City Cardiomyopathy The first major result of this study was that both HFimpEF Questionnaire overall summary score changed by a mean of and non-HFimpEF groups showed significant improvements 7.6 points (range: 6.0–9.2), compared with 3.5 points (range: in the perception of QoL. In previous studies, we25 and 2.3–4.8) in those with non-HFimpEF (P < 0.001). Moreover, others5,38 showed that QoL improved during specialized HF the statistical difference between groups persisted after management. In the present study, the improved QoL from adjusting for clinical variables, such as age, baseline LVEF, baseline HFrEF in the entire cohort during the first year was blood pressure, serum creatinine, and the baseline score. In likely to be due to therapy optimization and the structured addition, the difference between groups remained signifi- educational and monitoring programme carried out by cant, when the model was adjusted for other variables, such HF-specialized nurses, who performed all the follow-ups ev- as sex, history of HF, history of coronary artery disease, and ery 3 months. Moreover, improvements in depressive history of diabetes mellitus. However, the latter adjustment symptoms39 may have influenced the improvement in QoL. decreased the effect estimate to 2.98. In the present study, However, the association between QoL and LVEF has been we also observed greater improvement in QoL in the controversial, because many authors did not find any HFimpEF group, compared with the non-HFimpEF group, ESC Heart Failure 2022; 9: 3804–3813 DOI: 10.1002/ehf2.14098 Quality of life in patients with heart failure and improved ejection fraction: one-year changes and prognosis 3809 Table 2 Characteristics associated with QoL improvement a QoL improvement Univariable Multivariable Characteristic OR 95% CI P value OR 95% CI P value Age 1.00 0.99–1.01 0.60 1.01 1.00–1.02 0.22 Female sex 1.07 0.79–1.44 0.18 0.74 0.77–1.43 0.74 NYHA improvement 1.66 1.18–2.33 0.76 1.68 1.19–2.37 0.003 HF hospitalizations in previous year 0.74 0.56–0.97 0.03 0.51 0.37–0.70 <0.001 b HF duration 0.90 0.85–0.94 <0.001 0.75 0.57–0.98 0.04 HFimpEF 1.33 1.01–1.74 0.04 1.24 0.85–1.52 0.39 HF, heart failure; HFimpEF, heart failure with improved ejection fraction; mo, months; NYHA, New York Heart Association; QoL, quality of life. Improvement in the MLWHF score by ≥5 points. a b Months, Log(2) transformed. Table 3 Characteristics associated with the composite primary endpoint of all-cause death or HF-related hospitalization in patients with HFimpEF Composite of all-cause death or HF hospitalization Univariable Multivariable Characteristic HR 95% CI P value HR 95% CI P value Age 1.06 1.04–1.07 <0.001 1.05 1.03–1.07 <0.001 Female sex 1.10 0.79–1.51 0.58 0.77 0.54–1.10 0.15 One-year NYHA class 2.25 1.62–3.12 <0.001 1.47 0.97–2.22 0.07 HF hospitalizations in previous year 4.04 2.75–5.91 <0.001 4.61 3.10–6.86 <0.001 Ischaemic aetiology 1.71 1.26–2.32 <0.001 1.53 1.10–2.12 0.01 Diabetes 1.60 1.19–2.16 0.002 1.43 1.04–1.98 0.03 COPD 1.45 0.98–2.15 0.06 1.13 0.74–1.74 0.57 ACEI or ARB FU 0.34 0.19–0.59 <0.001 0.44 0.25–0.80 0.007 Beta-blockers FU 0.33 0.19–0.56 <0.001 0.39 0.21–0.73 0.003 ARNI FU 0.55 0.33–0.94 0.03 0.70 0.39–1.26 0.24 Ivabradine FU 0.69 0.48–0.97 0.03 0.88 0.61–1.28 0.51 One-year MLWHFQ score 1.02 1.01–1.02 0.001 1.02 1.01–1.03 0.006 ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; ARNI, angiotensin receptor neprilysin inhibitor; FU, follow-up; HF, heart failure; HFimpEF, heart failure with improved ejection fraction; MLWHFQ, Minnesota Living with Heart Failure Ques- tionnaire; NYHA, New York Heart Association. but the difference was due to a lower baseline QoL in the treatments, etc.2–10 Indeed, the physical dimension is HFimpEF group. At 1 year, the MLWHFQ scores were quite also very important in the QoL assessment; it is not surpris- similar between the HFimpEF and non-HFimpEF groups, ing that QoL was reported to be worse in patients with which indicated a similar perception of QoL at the end of more co-morbidities or higher NYHA functional classes.26 the study period. Previous studies revealed that an improvement in the NYHA The issue of how to interpret the differences in QoL im- functional class translated into a favourable impact on provement is debatable. On one hand, it could be simply a QoL.2,8 In the present study, we found that the perceived matter of chance that the patients with HFimpEF had a worse QoL improvement in patients with HFimpEF was mainly re- baseline QoL than patients with non-HFimpEF. However, our lated to the number of HF-related hospitalizations experi- results showed that, compared with the non-HFimpEF group, enced in the previous year and with the NYHA functional the HFimpEF group had a shorter HF duration, a larger pro- class. portion was classified as NYHA I–II, and the HF aetiology Our second study aim represented a novelty, to our was distributed differently; all these factors could have influ- knowledge, and it is probably the most important finding enced our results. On the other hand, if the HFimpEF group of our results. Previous studies have shown that patients had improved to the same extent as the non-HFimpEF group, with HFimpEF or ‘recovered’ LVEF had a better prognosis15– 23 at 1 year, the perception of QoL would have been worse. In- ; but that was not the objective of the present study. stead, at the end of the study, the groups reported similar Moreover, QoL was previously associated with outcomes levels of QoL. in patients with HF.25–30 However, no study had investi- QoL is a subjective assessment affected by multiple fac- gated how QoL might influence the prognosis of patients tors, such as age, sex, previous hospitalizations, diabetes, with HFimpEF. Based on our results, we concluded that ESC Heart Failure 2022; 9: 3804–3813 DOI: 10.1002/ehf2.14098 3810 E. Zamora et al. Figure 3 Forest plots show associations between different QoL assessments and either all-cause death or the composite end-point of all-cause death or HF hospitalization. (A) QoL assessed as the delta change between baseline and 1-year MLWHFQ scores (per 1%). (B) QoL improvement assessed as the continuous change between baseline and 1-year MLWHFQ scores (per 1 point). (C) QoL assessed as a significant categorical improvement in MLWHFQ scores (the minimal significant improvement was 5 points). (D) QoL assessed as the continuous 1-year MLWHFQ score (per 1 point). Blue = all-cause death; red = the primary composite endpoint of all-cause death or HF-related hospitalization. Note that the scale on the x-axis is not the same for all plots. Table 4 Characteristics associated with all-cause death in patients with HFimpEF All-cause death Univariable Multivariable Characteristic HR 95% CI P value HR 95% CI P value Age 1.07 1.05–1.08 <0.001 1.05 1.03–1.07 <0.001 Female sex 1.09 0.77–1.55 0.61 0.75 0.51–1.10 0.14 One-year NYHA class 2.08 1.46–2.95 <0.001 1.34 0.86–2.07 0.19 HF hospitalizations in previous year 2.03 1.36–3.03 0.001 1.59 1.05–2.40 0.03 a HF duration 1.06 0.99–1.13 0.09 1.08 1.00–1.17 0.04 Ischaemic aetiology 1.96 1.42–2.72 <0.001 1.80 1.24–2.61 0.002 Diabetes 1.54 1.12–2.13 0.008 1.27 0.90–1.78 0.18 ACEI or ARB FU 0.32 0.17–0.58 <0.001 0.37 0.20–0.70 0.002 Beta-blockers FU 0.25 0.14–0.45 <0.001 0.22 0.11–0.40 <0.001 ARNI FU 0.40 0.20–0.78 0.007 0.48 0.22–0.40 0.07 Ivabradine 0.64 0.43–0.94 0.02 0.90 0.59–1.36 0.61 One-year MLWHFQ score 1.01 1.00–1.02 0.01 1.01 1.00–1.03 0.03 ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; ARNI, angiotensin receptor neprilysin inhibitor; FU, follow-up; HF, heart failure; HFimpEF, heart failure with improved ejection fraction; MLWHFQ, Minnesota Living with Heart Failure Ques- tionnaire; mo, months; NYHA, New York Heart Association. a Months, Log(2) transformed. ESC Heart Failure 2022; 9: 3804–3813 DOI: 10.1002/ehf2.14098 Quality of life in patients with heart failure and improved ejection fraction: one-year changes and prognosis 3811 the QoL at the 1-year follow-up was significantly associated ment. Therefore, our study population comprised mainly with outcomes in patients with HFimpEF. By definition, relatively young men with HF of ischaemic aetiology and re- these patients had significantly improved their LVEFs. More- duced LVEF. Consequently, the results we obtained might over, remarkably, QoL was prognostically important, inde- not necessarily be extrapolated to a community-based HF pendent of other strong prognostic factors, like age, previ- population. Moreover, it remains to be determined whether ous hospital admissions, ischaemic aetiology, diabetes, the data for this study might be generalizable to a larger NYHA functional class, and treatments. Indeed, the percep- population of mainly older women with less systolic tion of QoL at 1 year was independently associated with dysfunction. outcomes, whereas the change in QoL from baseline was In conclusion, our results showed that QoL improved in pa- not. A similar result was observed in a secondary analysis tients treated for HF, regardless of whether they achieved of the TOPCAT and HF-ACTION studies.29 Those studies con- HFimpEF in 1 year. The QoL perception at 1 year was similar ducted serial QoL assessments in patients with in both groups—HFimpEF and non-HFimpEF—which sug- non-HFimpEF and HFrEF, respectively. They measured the gested that factors other than LVEF had influenced the QoL current, prior, and change in the Kansas City Cardiomyopa- perception. Remarkably, QoL at 1 year was found to be an in- thy Questionnaire score and found that the most recent as- dependent prognostic factor, contrary to the QoL change sessment provided the most important information about from baseline. the risks of subsequent clinical events. Thus, from the prog- nostic point of view, the future outcome was related to how the QoL was perceived by patients at one specific mo- Acknowledgements ment, independent of how they felt previously. In this sense, the perception of QoL might be considered a prog- We would like to thank all nurses of the Heart Failure Unit nostic biomarker in this subgroup of patients. Extensive that assisted in this study for their efforts in data collection studies have shown that improvements in LVEF had prog- and for their invaluable work in the Unit. Also, we thank Dr nostic implications on ‘hard’ endpoints, like death or Marta de Antonio and Dr Pedro Moliner for the revision of HF-related hospitalizations. However, increasingly, QoL has events and also for their invaluable work in the Unit during been considered an important endpoint for both patient years. well-being and its association with outcomes. In this study, we showed that QoL was related to outcomes beyond the improvement in LVEF; thus, the measurement of QoL per- ception is of clinical interest and important, particularly in Conflict of interest patients with improved cardiac function. None declared. Limitations This study had some limitations. First, LVEF was measured Funding with quantitative transthoracic echocardiography. However, None. LV function and volumes might have been assessed more precisely with 3D echocardiography or cardiac MRI. Second, similar to previous studies, we only included patients that had both baseline and 1-year echocardiography data avail- Supporting information able for analysis. Third, the QoL is a subjective measure that is difficult to measure on a group level.40 However, we imple- Additional supporting information may be found online in the mented a valid approach for highlighting the importance of Supporting Information section at the end of the article. the prognostic value of MLWHFQ scores. Another potential limitation was the lack of data on sodium/glucose co-trans- Table S1. Treatments. porter 2 inhibitor treatments. However, those data were of Table S2. Causes of death in HFimpEF patients. limited use in the study, and thus, they were not included Table S3. Number of HF-related hospitalizations in HFimpEF in the analyses; consequently, we could not ascertain patients. whether the use of these inhibitors could have influenced Table S4. Cox regression for the composite primary end-point the results. Finally, this study was conducted in a single cen- of all-cause death or HF-related hospitalization in HFimpEF tre; our population was a general population with HF, treated patients using QoL-improved as the variable of interest. at a specific, multidisciplinary HF unit in a tertiary hospital; Table S5. Binomial negative regression for recurrent HF and most patients were referred from the cardiology depart- hospitalizations. ESC Heart Failure 2022; 9: 3804–3813 DOI: 10.1002/ehf2.14098 3812 E. Zamora et al. References 1. Moradi M, Daneshi F, Behzadmehr R, FA, Ponikowski P, Sabatine MS, Anand failure in the community: Prevalence, Rafiemanesh H, Bouya S, Raeisi M. IS, Bělohlávek J, Böhm M, Chiang CE, prognosis, and predictors of the com- Quality of life of chronic heart failure Chopra VK, de Boer RA, Desai AS, Diez plete clinical recovery with return of patients: A systematic review and meta- M, Drozdz J, Dukát A, Ge J, Howlett JG, cardiac size and function to normal in analysis. Heart Fail Rev 2020; 25: Katova T, Kitakaze M, Ljungman CEA, patients undergoing optimal therapy. J 993–1006. Merkely B, Nicolau JC, O’Meara E, Petrie Card Fail 2004; 10: 250–257. 2. Quittan M, Wiesinger GF, Crevenna R, MC, Vinh PN, Schou M, Tereshchenko S, 17. Merlo M, Pyxaras SA, Pinamonti B, Nuhr MJ, Posch M, Hülsman M, Müller Verma S, Held C, DeMets DL, Docherty Barbati G, Di Lenarda A, Sinagra G. Prev- D, Pacher R, Fialka-Moser V. Cross-cul- KF, Jhund PS, Bengtsson O, Sjöstrand alence and prognostic significance of left tural adaptation of the Minnesota living M, Langkilde AM, DAPA-HF Trial Com- ventricular reverse remodeling in dilated with heart failure questionnaire for mittees and Investigators. Dapagliflozin cardiomyopathy receiving tailored medi- German-speaking patients. J Rehabil in patients with heart failure and re- cal treatment. J Am Coll Cardiol 2011; Med 2001; 33: 182–186. duced ejection fraction. N Engl J Med 57: 1468–1476. 3. Parajón T, Lupón J, González B, Urrutia 2019; 381: 1995–2008. 18. Kalogeropoulos AP, Fonarow GC, A, Altimir S, Coll R, Prats M, Valle V. 11. Riegel B, Moser DK, Glaser D, Carlson B, Georgiopoulou V, Burkman G, Use of the Minnesota living with heart Deaton C, Armola R, Sethares K, Siwamogsatham S, Patel A, Li S, failure quality of life questionnaire in Shively M, Evangelista L, Albert N. The Papadimitriou L, Butler J. Characteris- Spain. Rev Esp Cardiol 2004; 57: Minnesota living with heart failure ques- tics and outcomes of adult outpatients 155–160. tionnaire: Sensitivity to differences and with heart failure and improved or re- 4. Gastelurrutia P, Lupón J, Altimir S, de responsiveness to intervention intensity covered ejection fraction. JAMA Cardiol Antonio M, González B, Cabanes R, in a clinical population. Nurs Res 2002; 2016; 1: 510–518. Cano L, Urrutia A, Domingo M, Zamora 51: 209–218. 19. De Groote P, Fertin M, Duva Pentiah A, E, Díez C, Coll R, Bayes-Genis A. Effect 12. Austin BA, Wang Y, Smith GL, Vaccarine Goéminne C, Lamblin N, Bauters C. of fragility on quality of life in patients V, Krumholz HM, McNamara RL. Systolic Long-term functional and clinical with heart failure. Am J Cardiol 2013; function as a predictor of mortality and follow-up of patients with heart failure 112: 1785–1789. quality of life in long-term survivors with with recovered left ventricular ejection 5. Hole T, Grundtvig M, Gullestad L, heart failure. Clin Cardiol 2008; 31: fraction after β-blocker therapy. Circ Flønæs B, Westheim A. Improved quality 119–124. Heart Fail 2014; 7: 434–439. of life in Norwegian heart failure pa- 13. Juenger J, Schellberg D, Kraemer S, 20. Savarese G, Vedin O, D’Amario D, Uijl A, tients after follow-up in outpatient heart Haunstetter A, Zugck C, Herzog W, Dahlström U, Rosano G, Lam CSP, Lund failure clinics: Results from the Haass M. Health related quality of life LH. Prevalence and prognostic implica- Norwegian heart failure registry. Eur J in patients with congestive heart failure: tions of longitudinal ejection fraction Heart Fail 2010; 12: 1247–1252. Comparison with other chronic diseases change in heart failure. JACC Heart Fail 6. Joyce E, Chung C, Badloe S, Odutayo K, and relation to functional variables. 2019; 7: 306–317. Desai A, Givertz MM, Nohria A, Heart 2002; 87: 235–241. 21. Strange G, Playford D, Scalia GM, Lakdawala NK, Stewart GC, Young M, 14. Bozkurt B, Coats AJS, Tsutsui H, Celermajer DS, Prior D, Codde J, Chan Weintraub J, Stevenson LW, Lewis EF. Abdelhamid CM, Adamopoulos S, Albert YK, Bulsara MK, Stewart S, NEDA Inves- Variable contribution of heart failure to N, Anker SD, Atherton J, Böhm M, tigators. Change in ejection fraction and quality of life in ambulatory heart failure Butler J, Drazner MH, Michael Felker long-term mortality in adults referred with reduced, better, or preserved ejec- G, Filippatos G, Fiuzat M, Fonarow GC, for echocardiography. Eur J Heart Fail tion fraction. JACC Heart Fail 2016; 4: Gomez-Mesa JE, Heidenreich P, 2021; 23: 555–563. 184–193. Imamura T, Jankowska EA, Januzzi J, 22. DeVore AD, Hellkamp AS, Thomas L, Al- 7. Lewis EF, Claggett BL, McMurray JJV, Khazanie P, Kinugawa K, Lam CSP, bert NM, Butler J, Patterson JH, Spertus Packer M, Lefkowitz MP, Rouleau JL, Matsue Y, Metra M, Ohtani T, Francesco JA, Williams FB, Shen X, Hernandez AF, Liu J, Shi VC, Zile MR, Desai AS, Piepoli M, Ponikowski P, Rosano GMC, Fonarow GC. The association of im- Solomon SD, Swedberg K. Health-re- Sakata Y, Seferović P, Starling RC, provement in left ventricular ejection lated quality of life outcomes in PARA- Teerlink JR, Vardeny O, Yamamoto K, fraction with outcomes in patients with DIGM-HF. Circ Heart Fail 2017; 10: Yancy C, Zhang J, Zieroth S. Universal heart failure with reduced ejection frac- e003430. definition and classification of heart fail- tion: Data from CHAMP-HF. Eur J Heart 8. Ekman I, Chassany O, Komajda M, Böhm ure: A report of the Heart Failure Society Fail 2022; 24: 762–770. M, Borer JS, Ford I, Tavazzi L, Swedberg of America, Heart Failure Association of 23. Wang K, Youngson E, Bakal JA, Thomas K. Heart rate reduction with ivabradine the European Society of Cardiology, J, McAlister FA, Oudit GY. Cardiac re- and health related quality of life in pa- Japanese Heart Failure Society and Writ- verse remodelling and health status in tients with chronic heart failure: Results ing Committee of the Universal Defini- patients with chronic heart failure. ESC from the SHIFT study. Eur Heart J 2011; tion of Heart Failure: Endorsed by the Heart Fail 2021; 8: 3106–3118. 32: 2395–2404. Canadian Heart Failure Society, Heart 24. Wohlfahrt P, Nativi-Nicolau J, Zhang M, 9. Ponikowski P, van Veldhuisen DJ, Failure Association of India, Cardiac So- Selzman CH, Greene T, Conte J, Biber Comin-Colet J, Ertl G, Komajda M, ciety of Australia and New Zealand, JE, Hess R, Mondesir FL, Wever-Pinzon Mareev V, McDonagh T, Parkhomenko and Chinese Heart Failure Association. O, Drakos SG, Gilbert EM, Kemeyou L, A, Tavazzi L, Levesque V, Mori C, Eur J Heart Fail 2021; 23: 352–380. LaSalle B, Steinberg BA, Shah RU, Fang Roubert B, Filippatos G, Ruschitzka F, 15. Lupón J, Díez-López C, de Antonio M, JC, Spertus JA, Stehlik J. Quality of life Anker SD, CONFIRM-HF Investigators. Domingo M, Zamora E, Moliner P, in patients with heart failure with recov- Beneficial effects of long-term intrave- González B, Santesmases J, Troya MI, ered ejection fraction. JAMA Cardiol nous iron therapy with ferric Bayés-Genís A. Recovered heart failure 2021; 6: 957–962. carboxymaltose in patients with symp- with reduced ejection fraction and out- 25. Lupón J, Gastelurrutia P, de Antonio M, tomatic heart failure and iron deficiency. comes: A prospective study. Eur J Heart González B, Cano L, Cabanes R, Urrutia Eur Heart J 2015; 36: 657–668. Fail 2017; 19: 1615–1623. A, Díez C, Coll R, Altimir S, Bayes-Genis 10. McMurray JJV, Solomon SD, Inzucchi 16. Cioffi G, Stefenelli C, Tarantini L, A. Quality of life monitoring in ambula- SE, Køber L, Kosiborod MN, Martinez Opasich C. Chronic left ventricular tory heart failure patients: Temporal ESC Heart Failure 2022; 9: 3804–3813 DOI: 10.1002/ehf2.14098 Quality of life in patients with heart failure and improved ejection fraction: one-year changes and prognosis 3813 changes and prognostic value. Eur J 30. Alta F, Briançon S, Guillemin F, Jullière 36. Lewis EF, Johnson PA, Johnson W, Heart Fail 2013; 15: 103–109. Y, Mertès PM, Villemot JP, Zannad F, EP- Collins C, Griffin L, Stevenson LW. Pref- 26. Gastelurrutia P, Lupón J, Moliner P, ICAL investigators. Self-rating of quality erences for quality of life or survival Yang X, Cediel G, de Antonio M, of live provides additional prognostic in- expressed by patients with heart failure. Domingo M, Altimir S, González B, formation in heart failure. Insights into J Heart Lung Transplant 2001; 20: Rodríguez M, Rivas C, Díaz V, Fung E, the EPICAL study. Eur J Heart Fail 1016–1024. Zamora E, Santesmases J, Núñez J, 2002; 4: 337–343. 37. He Y, Ling Y, Guo W, Li Q, Yu S, Huang Woo J, Bayes-Genis A. Comorbidities, 31. Zamora E, Lupón J, Vila J, Urrutia A, de H, Zhang R, Gong Z, Liu J, Mo L, Yi S, fragility, and quality of life in heart fail- Antonio M, Sanz H, Grau M, Ara J, Lai D, Yao Y, Liu J, Chen J, Liu Y, Chen ure patients with midrange ejection Bayés-Genís A. Estimated glomerular fil- S. Prevalence and prognosis of fraction. Mayo Clin Proc Innov Qual Out- tration rate and prognosis in heart fail- HFimpEF developed from patients with comes 2018; 2: 176–185. ure: Value of the modification of diet in heart failure with reduced ejection frac- 27. Konstam V, Salem D, Pouleur H, Kostis J, renal disease Study-4, chronic kidney tion: Systematic review and meta-anal- Gorkin L, Shumaker S, Mottard I, Woods disease epidemiology collaboration, and ysis. Front Cardiovasc Med 2021; 8: P, Konstam MA, Yusuf S. Baseline quality Cockroft-Gault formulas. J Am Coll 757596. of life as a predictor of mortality and Cardiol 2012; 59: 1709–1715. 38. Konstam V, Gregory D, Chen J, hospitalisation in 5,025 patients with 32. Lupón J, Gavidia-Bovadilla G, Ferrer E, Weintraub A, Patel A, Levine D, Venesy congestive heart failure. Studies of the de Antonio M, Perera-Lluna A, López- D, Perry K, Delano C, Konstam MA. SOLVD investigators. Am J Cardiol Ayerbe J, Domingo M, Núñez J, Zamora Health-related quality of life in a multi- 1996; 78: 890–895. E, Moliner P, Díaz-Ruata P, Santesmases center randomized controlled compari- 28. Sepehrvand N, Savu A, Spertus JA, J, Bayés-Genís A. Dynamic trajectories son of telephonic disease management Dyck JRB, Anderson T, Howlett J, of left ventricular ejection fraction in and automated home monitoring in pa- Paterson I, Oudit GY, Kaul P, McAlister heart failure. J Am Coll Cardiol 2018; tients recently hospitalized with heart FA, Ezekowitz JA, Alberta HI. Change 72: 591–601. failure: SPAN-CHF II trial. J Card Fail of health-related quality of life over 33. Garin O, Soriano N, Ribera A, Ferrer M, 2011; 17: 151–157. time and its association with Pont A, Alonso J, Permanyer G, Grupo 39. Diez-Quevedo C, Lupón J, González B, patient outcomes in patients with heart IC-QoL. Validation of the Spanish ver- Urrutia A, Cano L, Cabanes R, Altimir failure. J Am Heart Assoc 2020; 9: sion of the Minnesota living with heart S, Coll R, Pascual T, de Antonio M, e017278. failure questionnaire. Rev Esp Cardiol Bayes-Genis A. Depression, antidepres- 29. Pokharel Y, Khariton Y, Tang Y, Nassif 2008; 61: 251–259. sants, and long-term mortality in heart ME, Chan PS, Arnold SV, Jones PG, 34. McMurray JJ, Petrie MC, Murdoch DR, failure. Int J Cardiol 2013; 167: Spertus JA. Association of Serial Kansas Davie AP. Clinical epidemiology of heart 1217–1225. City cardiomyopathy questionnaire as- failure: Public and private health bur- 40. Aaronson KD, Schwartz JS, Chen TM, sessments with death and hospitaliza- den. Eur Heart J 1998; 19: 9–12. Wong KL, Goin JE, Mancini DM. Devel- tion in patients with heart failure with 35. Stanek E, Oates M, McGhan F, Denofrio opment and prospective validation of a preserved and reduced ejection fraction: D, Loh E. Preferences for treatment out- clinical index to predict survival in am- A secondary analysis of 2 randomized comes in patients with heart failure: bulatory patient referred for cardiac clinical trials. JAMA Cardiol 2017; 2: Symptoms versus survival. J Card Fail transplant evaluation. Circulation 1997; 1315–1321. 2000; 3: 225–232. 95: 2660–2667. ESC Heart Failure 2022; 9: 3804–3813 DOI: 10.1002/ehf2.14098