Beyond initial recovery: Heart failure with transient vs sustained improvement in left ventricular ejection fraction

Abstract

BACKGROUND

There is no available data about the trajectory of heart failure (HF) with improved ejection fraction (EF) and patient clinical outcomes in Qatar.

AIM

To explore the difference in characteristics and outcomes between patients with transient and sustained improvement in left ventricular ejection fraction (LVEF) and to determine the independent predictors for sustained improvement in LVEF.

METHODS

This is a retrospective cohort study that was conducted at the advanced HF clinic of a tertiary care hospital in Qatar between January 2017 and December 2018. This study included adult patients with improved LVEF and had at least three echocardiographic studies. The patients were divided into two groups: HF with transient improvement in EF (HFtimpEF) and HF with sustained improvement in EF (HFsimpEF).

RESULTS

A total of 175 patients with HF and improved EF were included. Among them 136 (77.7%) patients showed sustained improvement in LVEF. The remaining patients with HFtimpEF were predominantly males [37 (94.9%) vs 101 (74.3%), P = 0.005] with a higher incidence of ischemic cardiomyopathy [32 (82.1%) vs 68 (50.4%), P = 0.002], dyslipidemia [24 (61.5%) vs 54 (39.7%), P = 0.03], and hypertension [34 (87.2%) vs 93 (68.4%), P = 0.03] than those with HFsimpEF. The latter experienced significantly lower rates of hospitalization [39 (28.7%) vs 20 (51.3%), P = 0.01] and diagnosis of new cardiovascular conditions during the follow-up (e.g., acute coronary syndrome, stroke, decompensated HF, and atrial fibrillation) [14 (10.3%) vs 10 (25.6%), P = 0.03] without a difference in emergency department visits or in-hospital death. Sustained improvement in LVEF was positively associated with being female [adjusted odds ratio (aOR) = 6.8, 95% confidence interval (CI): 1.4-32.3, P = 0.02], having non-ischemic etiology of HF (aOR = 3.1, 95%CI: 1.03-9.3, P = 0.04), and using a mineralocorticoid receptor antagonist (aOR = 7.0, 95%CI: 1.50-31.8, P = 0.01).

CONCLUSION

Patients with HFsimpEF experienced significantly lower rates of hospitalization and diagnosis of new cardiovascular conditions than patients with HFtimpEF. Sustained improvement in LVEF was positively associated with being a female, having non-ischemic etiology of HF, and using a mineralocorticoid receptor antagonist.

Key Words: Asia; Cardiomyopathy; Improved ejection fraction; Middle East; Qatar

Core Tip: Our study reported observations from a population with heart failure in Qatar, a Middle Eastern country that is usually underrepresented in major clinical trials. This study was the first in the Middle East to characterize the clinical features and outcomes of patients with heart failure who demonstrate either sustained or temporary improvement in left ventricular ejection fraction.

INTRODUCTION

Heart failure (HF) remains one of the key causes of cardiovascular mortality[1]. The current guidelines propose stratifying patients in phenotypes based on left ventricular ejection fraction (LVEF) to provide individualized care[2]. Guideline-directed medical therapy in HF with reduced EF (HFrEF) has resulted in improvements in cardiac function and left ventricular reverse remodeling[3,4]. Several studies have shown that these patients with HF and improved EF (HFimpEF) have a distinct HF phenotype and may carry favorable clinical outcomes[5-8].

Yet this improvement is not always sustained; reverse remodeling is often accompanied by detrimental neurohormonal changes that can precipitate fluctuations in EF[9]. This can lead to variations in EF over time and highlights the importance of longitudinal assessment of EF to further sub-stratify patients with HFimpEF into sustained improvement and transient improvement. This distinction has led to the need to further subclassify HFimpEF as HF with sustained improvement in EF (HFsimpEF) or HF with transient improvement in EF (HFtimpEF).

The factors that determine whether patients with HFrEF go on to become HFsimpEF or HFtimpEF has been an area of interest yet poorly studied. There is also a need to assess the long-term outcomes and prognosis of these patients. There is a lack of available data on the trajectory of HFimpEF and patient clinical outcomes in the Middle East. Hence, this study was conducted to explore several aspects related to patients with HFimpEF and their sustained improvement in LVEF. The hypothesis was that patients with sustained improvement in LVEF would have better clinical outcomes compared with those with transient improvement building on preliminary findings from previous studies exploring outcomes of patients with HFimpEF, who showed a generally favorable prognosis[10]. The main objective of this study was to explore the difference in characteristics and outcomes between patients with transient and sustained improvement in LVEF and to determine the independent predictors for sustained improvement in LVEF.

MATERIALS AND METHODS

This was a retrospective cohort study that was conducted at the advanced HF clinic in a tertiary care hospital in Qatar between January 1, 2017 and December 31, 2018. The study was approved by the Institutional Review Board of the Heart Hospital and the Medical Research Center, approval No. MRC-01-20-139. An informed consent was waived due to the retrospective nature of the study. The study was in line with the principles of the Declaration of Helsinki, Good Clinical Practice and the laws and regulations of the Ministry of Public Health in Qatar. This study included all adult patients (≥ 18 years) who were initially diagnosed with HFrEF, defined as LVEF < 40.0% at baseline, who demonstrated an initial improvement in LVEF and had at least three echocardiographic studies from their initial visit to the clinic.

The study population was selected from our previously published study that described the characteristics of patients with HFrEF who had improvement in their LVEF and analyzed the independent predictors of LVEF improvement. Our previous study screened all patients who visited the advanced HF clinic during the study period and included adult patients aged 18 years or older with a diagnosis of HFrEF and who had two echocardiographic examinations performed at least 6 months apart[11].

For the current study, all patients had a first (or baseline) echocardiogram (Echo-1) where the diagnosis of HFrEF was established and showed improvement in LVEF of at least 1.0% on the second echocardiogram (Echo-2) performed at least six months after Echo-1. Hence, Echo-2 would identify the HFimpEF cohort. The specific cutoff values for the improvement in LVEF have not been established yet by the international guidelines[12], and patients with improvement of any magnitude were included. A third echocardiogram (Echo-3) at least 12 months after from Echo-2 was included to distinguish patients with transient or sustained improvement in LVEF. An increase in LVEF by at least 1.0% in both the Echo-2 and the Echo-3 compared to the Echo-1 was defined as sustained improvement in EF (i.e. HFsimpEF). An initial improvement in LVEF of at least 1.0% in the Echo-2 followed by any reworsening of LVEF or lack of further improvement in the Echo-3 was defined as having transient improvement in EF (i.e. HFtimpEF). Thus, the patients were divided into two groups: HFtimpEF and HFsimpEF.

Furthermore, the improvement in LVEF was considered large if the increase or change in LVEF (∆ LVEF) was 10.0% or more and modest if ∆ LVEF was 1.0%-9.0%[13]. A subgroup analysis for the patients with an initial large LVEF improvement (i.e. the difference between Echo-2 and Echo-1 of ≥ 10.0%) was conducted to compare the patients with HFtimpEF and HFsimpEF.

LVEF was measured using the Simpson biplane method, unless the Simpson method was not possible. This study excluded patients with specific cardiomyopathies (e.g., hypertrophic, infiltrative, restrictive, stress-induced, or chemotherapy-induced), primary right-sided disease, congenital heart disease as the etiology of HF, cardiac transplant, left ventricular assist device, or primary valvular disease.

RESULTS

Baseline characteristics

Within the specified study duration, 582 patients with HFrEF underwent at least two echocardiograms. Of them, 175 patients (30.1%) showed any improvement in LVEF (i.e. ≥ 1.0%) between Echo-1 and Echo-2 and who had a second follow-up echocardiogram (i.e. Echo-3) performed at least 12 months after Echo-2. Among them, 136 patients (77.7%) showed sustained improvement in LVEF on Echo-3 while the remaining had reworsening of the LVEF (i.e. had transient improvement).

Patients with HFtimpEF were predominantly males (94.9% vs 74.3%, P = 0.005) with ischemic cardiomyopathy (82.1% vs 50.4%, P = 0.002). There were no differences between the groups in terms of age, body mass index, or duration of HF diagnosis. Patients with HFtimpEF had a higher incidence of dyslipidemia (61.5% vs 39.7%, P = 0.03) and hypertension (87.2% vs 68.4, P = 0.03) than those with HFsimpEF.

Table 1 presents details of baseline characteristics. At baseline, there were no differences between the groups in terms of blood pressure measurements, heart rate values, and the presence of left bundle branch block.

Table 1 Baseline characteristics of the included patients.

Variable	Transient improvement (n = 39)	Sustained improvement (n = 136)	P value
Age (years), mean ± SD	58.15 ± 11.69	54.51 ± 13.52	0.13
BMI (kg/m2), mean ± SD	28.36 ± 5.67	30.49 ± 7.20	0.09
Male	37 (94.9)	101 (74.3)	0.005
Female	2 (5.1)	35 (25.7)	-
HF etiology
Ischemic	32 (82.1)	68 (50.4)	0.002
Dilated	5 (12.8)	57 (42.2)
Others	2 (5.1)	10 (7.4)
Duration of HF diagnosis, mean ± SD	7.05 ± 3.81	5.79 ± 4.0	0.08
Comorbidities
Dyslipidemia	24 (61.5)	54 (39.7)	0.03
Hypertension	34 (87.2)	93 (68.4)	0.03
Diabetes	28 (71.8)	90 (66.2)	0.57
Atrial fibrillation	8 (20.5)	25 (18.4)	0.82
Ventricular arrhythmia	5 (12.8)	18 (13.2)	1.00
CVA/TIA	5 (12.8)	16 (11.8)	0.79
PAD	2 (5.1)	8 (5.9)	1.00
Renal impairment	17 (43.6)	35 (25.7)	0.05
Liver disease	2 (5.1)	5 (3.7)	0.65
Anemia	10 (25.6)	38 (27.9)	0.84
Sleep apnea	0 (0)	9 (6.6)	0.21
Chronic lung disease	6 (15.4)	21 (15.4)	1.00
Cancer	0 (0)	7 (5.1)	0.35
Dementia	2 (5.1)	3 (2.2)	0.31
Depression	0 (0)	6 (4.4)	0.34
Hypothyroidism	2 (5.1)	11 (8.1)	0.74
Hyperthyroidism	0 (0)	4 (2.9)	0.58

Data are presented as n (%). BMI: Body mass index; CVA: Cerebrovascular accident; HF: Heart failure; PAD: Peripheral artery disease; SD: Standard deviation; TIA: Transient ischemic attack.

Similarly, there was no difference in the previous parameter measurements between the groups at follow-up except for the heart rate, which was significantly lower in patients with HFsimpEF [72.84 ± 12.90 vs 79.41 ± 15.45 beats per minute (bpm), P = 0.02] as shown in Table 2. Blood test parameters did not differ between the two groups at baseline and follow-up except for the estimated glomerular filtration rate, which was significantly higher in the sustained improvement group at both baseline (79.68 ± 33.27 vs 66.47 ± 27.79 mL/minute, P = 0.03) and follow-up (70.35 ± 29.24 vs 59.89 ± 28.49 mL/minute, P = 0.05) (Table 3).

Table 2 Vital signs and electrocardiographic measures of the included patients.

Variable	Transient improvement (n = 39)	Sustained improvement (n = 136)	P value
Baseline (Echo-1), mean ± SD
Systolic BP (mmHg)	124.38 ± 22.41	125.30 ± 20.1 (n = 134)	0.81
Diastolic BP (mmHg)	72.64 ± 12.36	75.97 ± 12.99 (n = 134)	0.16
Heart rate (bpm)	86.43 ± 18.02 (n = 37)	89.69 ± 19.12 (n = 132)	0.36
Achieved target BP	26 (66.7)	87 (64.9)	1.00
Sinus rhythm	28 (90.3)	103 (86.6)	0.77
LBBB	7 (14.1)	19 (17.1)	0.42
QRS (ms), mean ± SD	109.45 ± 26.29 (n = 29)	109.43 ± 44.04 (n = 129)	1.0
QTc (ms), mean ± SD	468.86 ± 43.48 (n = 29)	462.84 ± 52.19 (n = 109)	0.57
Second follow-up (Echo-3), mean ± SD
Systolic BP (mmHg)	127.87 ± 21.43	133.47 ± 98.18 (n = 129)	0.73
Diastolic BP (mmHg)	73.85 ± 13.62	73.81 ± 11.49 (n = 129)	0.99
Heart rate (bpm)	79.41 ± 15.45	72.84 ± 12.90 (n = 128)	0.02
Achieved target BP	26 (66.7)	90 (69.8)	0.67
Sinus rhythm	33 (89.2)	111 (88.8)	1.00
LBBB	7 (19.4)	21 (17.4)	0.81
QRS (ms), mean ± SD	112.00 ± 26.87 (n = 37)	110.07 ± 26.34 (n = 120)	0.70
QTc (ms), mean ± SD	455.08 ± 31.57 (n = 37)	448.91 ± 42.92 (n = 120)	0.35

Data are presented as n (%). Echo-1: First (or baseline) echocardiogram; BP: Blood pressure; bpm: Beats per minute; LBBB: Left bundle branch block; QTc: Corrected QT interval; QRS: QRS complex duration; Echo-3: Third echocardiogram; SD: Standard deviation.

Table 3 Laboratory blood tests of the included patients.

Variable	Transient improvement (n = 39)	Sustained improvement (n = 136)	P value
Baseline (Echo-1)
NT-pro-BNP (pg/mL)	5106.63 ± 6035.19 (n = 27)	4730.48 ± 6792.82 (n = 93)	0.80
Hemoglobin (g/dL)	13.16 ± 2.22 (n = 38)	14.20 ± 12.89 (n = 132)	0.62
Urea (mmol/L)	8.16 ± 4.53 (n = 38)	7.15 ± 4.17 (n = 133)	0.20
Creatinine (µmol/L)	120.03 ± 56.87 (n = 38)	113.55 ± 114.84 (n = 132)	0.74
eGFR (mL/min)	66.47 ± 27.79 (n = 38)	79.68 ± 33.27 (n = 132)	0.03
HbA1c (%)	8.38 ± 2.34 (n = 26)	7.54 ± 2.19 (n = 102)	0.09
LDL-C (mmol/L)	2.69 ± 0.92 (n = 35)	2.57 ± 1.18 (n = 120)	0.56
Iron (µmol/L)	8.87 ± 3.69 (n = 10)	10.43 ± 6.49 (n = 37)	0.47
TSAT (%)	16.77 ± 5.42 (n = 10)	20.37 ± 15.74 (n = 35)	0.48
Ferritin (ug/L)	325.75 ± 347.95 (n = 8)	208.50 ± 277.71 (n = 30)	0.32
TSH (pmol/L)	1.76 ± 1.51 (n = 22)	2.38 ± 2.18 (n = 92)	0.21
T4 (mIU/L)	13.51 ± 1.70 (n = 23)	14.87 ± 3.58 (n = 87)	0.08
Second follow-up (Echo-3)
NT-pro-BNP (pg/mL)	7672.15 ± 13459.32 (n = 27)	3184.52 ± 6564.79 (n = 81)	0.11
Hemoglobin (g/dL)	12.23 ± 2.25 (n = 37)	12.87 ± 2.08 (n = 120)	0.11
Urea (mmol/L)	10.28 ± 7.27 (n = 38)	8.50 ± 8.94 (n = 125)	0.26
Creatinine (µmol/L)	142.84 ± 88.49 (n = 38)	127.25 ± 126.01 (n = 128)	0.48
eGFR (mL/min)	59.89 ± 28.49 (n = 38)	70.35 ± 29.24 (n = 127)	0.05
HbA1c (%)	7.66 ± 2.04 (n = 33)	7.90 ± 4.88 (n = 103)	0.79
LDL-C (mmol/L)	1.81 ± 0.99 (n = 33)	2.03 ± 0.93 (n = 99)	0.25
Iron (µmol/L)	9.64 ± 4.33 (n = 18)	11.23 ± 6.24 (n = 41)	0.33
TSAT (%)	17.44 ± 8.69 (n = 18)	23.48 ± 16.53 (n = 41)	0.15
Ferritin (ug/L)	276.17 ± 489.76 (n = 14)	244.41 ± 429.58 (n = 30)	0.83
TSH (pmol/L)	2.43 ± 1.45 (n = 22)	3.22 ± 3.12 (n = 83)	0.26
T4 (mIU/L)	15.73 ± 2.23 (n = 19)	15.05 ± 3.86 (n = 73)	0.46

Data are presented as mean ± SD. Echo-1: First (or baseline) echocardiogram; eGFR: Estimated glomerular filtration rate; HbA1c: Glycated hemoglobin; Echo-3: Third echocardiogram; LDL-C: Low-density lipoprotein cholesterol; NT-proBNP: N-terminal pro-brain natriuretic peptide; T4: Thyroxine hormone; TSAT: Transferrin saturation; TSH: Thyroid-stimulating hormone.

Echocardiographic parameters

Upon initial diagnosis, mean LVEF was significantly lower in the sustained improvement group (26.77 ± 7.05% vs 30.56 ± 6.95%, P = 0.003) with no difference in other echocardiographic parameters. At follow-up, LVEF became significantly higher in the sustained improvement group (40.18 ± 0.12 vs 28.46 ± 6.59%, P = 0.001) with significant improvement in other parameters such as lower left atrium volume index (34.30 ± 11.26 vs 41.41 ± 12.92 mL/m², P = 0.003), left ventricular end-systolic diameter (4.50 ± 0.95 vs 4.99 ± 0.75 cm, P = 0.004), and right ventricular systolic pressure (RVSP; 33.10 ± 11.82 vs 38.78 ± 14.61 mmHg, P = 0.04). The initial improvement in LVEF (between Echo-1 and Echo-2) at almost 20 months of follow-up did not differ in magnitude between the groups (Echo-2 - Echo-1: ∆ LVEF: 9.54 ± 7.69% vs 8.05 ± 6.53%, P = 0.27). After another 20 months of follow-up, the overall improvement in LVEF differed significantly between the groups (Echo-3 - Echo-1: ∆ LVEF: 13.43% ± 9.09% vs 2.10% ± 5.05%, P = 0.001) (Table 4).

Table 4 Echocardiographic parameters of the included patients.

Variable	Transient improvement (n = 39)	Sustained improvement (n = 136)	P value
Baseline (Echo-1)
LVEF (%)	30.56 ± 6.95	26.77 ± 7.05	0.003
Average E/e’	15.01 ± 5.88 (n = 20)	14.52 ± 5.78 (n = 67)	0.74
LA volume index (mL/m2)	35.97 ± 12.51 (n = 25)	37.09 ± 11.35 (n = 82)	0.68
LVEDD (cm)	5.76 ± 0.79 (n = 37)	6.70 ± 4.17 (n = 130)	0.18
LVESD (cm)	4.76 ± 0.84 (n = 34)	5.08 ± 0.89 (n = 115)	0.80
RVSP (mmHg)	33.46 ± 11.06	37.35 ± 12.89 (n = 130)	0.13
Second follow-up (Echo-3)
LVEF (%)	28.46 ± 6.59	40.18 ± 0.12	0.001
Average E/e’	13.12 ± 5.46 (n = 19)	12.08 ± 5.86 (n = 77)	0.49
LA volume index (mL/m2)	41.41 ± 12.92 (n = 31)	34.30 ± 11.26 (n = 114)	0.003
LVEDD (cm)	6.06 ± 0.62 (n = 38)	5.81 ± 0.89 (n = 134)	0.05
LVESD (cm)	4.99 ± 0.75 (n = 37)	4.50 ± 0.95 (n = 131)	0.004
RVSP (mmHg)	38.78 ± 14.61 (n = 37)	33.10 ± 11.82 (n = 129)	0.04
LVEF (Echo-2)	38.59 ± 7.57	36.33 ± 7.69	0.11
∆ LVEF (Echo-2 - Echo-1) (%)	8.05 ± 6.53	9.54 ± 7.69	0.27
Duration between Echo-1 and Echo-2 (months)	19.82 ± 26.61	17.24 ± 21.15	0.53
∆ LVEF (Echo-3 - Echo-1) (%)	-2.10 ± 5.05	13.43 ± 9.09	0.001
Duration between Echo-2 and Echo-3 (months)	20.49 ± 13.40	21.79 ± 23.39	0.74

Data are presented as mean ± SD. Echo-1: First (or baseline) echocardiogram; Echo-2: Second echocardiographic examination at first follow-up; Echo-3: Third echocardiographic; LA: Left atrium/atrial; LVEDD: Left ventricular end-diastolic diameter; LVEF: Left ventricular ejection fraction; LVESD: Left ventricular ejection fraction; RVSP: Right ventricular systolic pressure.

Medications

At the time of diagnosis, more than half of the patients in both groups were on beta-blockers and renin-angiotensin-aldosterone system inhibitors without a difference between the two groups. Significantly more patients in the sustained improvement group were on mineralocorticoid receptor antagonists (MRA; 21.3% vs 5.1%, P = 0.02). At the second follow-up (Echo 3), the use of guideline-directed medical therapy for HF increased in both groups without a difference between them (Table 5).

Table 5 Medications at baseline and at second follow-up.

Variable	Transient improvement (n = 39)	Sustained improvement (n = 136)	P value
Baseline (Echo-1)
Beta-blockers	19 (48.7)	78 (57.4)	0.37
ACEI/ARB	25 (64.1)	81 (59.6)	0.71
MRA	2 (5.1)	29 (21.3)	0.02
Diuretics	13 (33.3)	62 (45.6)	0.20
HDZ/ISDN	1 (2.6)	4 (2.9)	1.00
Ivabradine	0 (0)	6 (4.4)	0.34
Digoxin	0 (0)	9 (6.6)	0.21
Antiplatelet therapy	24 (61.5)	78 (57.4)	0.71
LLA	24 (61.5)	72 (52.9)	0.38
Iron (IV)	7 (17.9)	17 (12.5)	0.43
Inotropes (hospital)	4 (10.3)	23 (16.9)	0.45
Second follow-up (Echo-3)
Beta-blockers	36 (92.3)	126 (95.5)	0.43
ACEI/ARB	30 (76.9)	110 (80.9)	0.58
MRA	13 (33.3)	54 (40.9)	0.46
Diuretics	31 (79.5)	99 (75.0)	0.67
HDZ/ISDN	4 (10.3)	10 (7.6)	0.53
Ivabradine	3 (7.7)	11 (8.3)	1.00
Digoxin	2 (5.1)	7 (5.3)	1.00
Antiplatelet therapy	34 (87.2)	97 (73.5)	0.09
Lipid-lowering agents	36 (92.3)	106 (80.3)	0.08
Iron (IV)	4 (10.3)	5 (3.7)	0.12
Inotropes (hospital)	3 (7.7)	3 (2.2)	0.13

Echo-1: First (or baseline) echocardiogram; Echo-3: Third echocardiographic; ACEI: Angiotensin-converting enzyme inhibitors; ARB: Angiotensin receptor blockers; HDZ: Hydralazine; ISDN: Isosorbide dinitrate; IV: Intravenous; MRA: Mineralocorticoid receptor antagonists.

Clinical outcomes

Patients with sustained improvement in LVEF demonstrated a significantly lower rate of all-cause hospitalization compared with those with transient improvement [39 (28.7%) vs 20 (51.3%) patients, P = 0.01]. Among all the hospitalizations, 28 were specifically due to HF with 16 occurring in the sustained improvement group and 12 in the transient improvement group. The incidence of developing new cardiovascular conditions during follow-up including acute coronary syndrome, stroke, decompensated HF, and atrial fibrillation, was also lower in the sustained improvement group [14 (10.3%) vs 10 (25.6%) patients, P = 0.03]. However, there was no statistically significant difference in the rates of emergency department visits [38 (27.9%) vs 16 (41.0%) patients, P = 0.12] or in-hospital deaths [6 (4.4%) vs 1 (2.6%) patient(s), P = 1.00] between the groups.

Multivariate nominal logistic regression analysis was performed to identify factors associated with sustained improvement in LVEF. Female sex [adjusted odds ratio (aOR) = 6.8, 95% confidence interval (CI): 1.4-32.3, P = 0.02], non-ischemic etiology of HF (aOR = 3.1, 95%CI: 1.03-9.3, P = 0.04), and the use of MRA (aOR = 7.0, 95%CI: 1.50-31.8, P = 0.01) were significantly associated with sustained improvement of LVEF. Other factors such as age, hyperlipidemia, and hypertension did not show a significant association with sustained improvement (Table 6).

Table 6 Multivariate nominal logistic regression analysis.

Variable	Adjusted OR (95%CI)	P value
Age in years	0.99 (0.95-1.0)	0.42
Female	6.8 (1.4-32.3)	0.02
HF etiology
Ischemic	1 (reference)	-
Dilated	3.1 (1.03-9.3)	0.04
Others	1.7 (0.30-9.3)	0.56
Hyperlipidemia	0.50 (0.21-1.11)	0.09
Hypertension	0.40 (0.13-1.16)	0.09
MRA use	7.0 (1.50-31.8)	0.01

CI: Confidence interval; OR: Odds ratio; HF: Heart failure; MRA: Mineralocorticoid receptor antagonists.

Subgroup analysis

A subgroup analysis was performed to identify the characteristics and outcomes of patients with HFimpEF and a large improvement in LVEF between the baseline (Echo-1) and first follow-up echocardiogram (Echo-2). Seventy-one (40.6%) patients had a large improvement in LVEF (i.e. ∆ LVEF ≥ 10.0%) at Echo-2. Of them, 14 (19.3%) and 57 (80.3%) patients showed transient and sustained improvement in LVEF, respectively. Patients with transient improvement were significantly more dyslipidemic (71.4% vs 35.1%, P = 0.014) and hypertensive (92.9% vs 61.4%, P = 0.024) with significantly lower heart rates (79.92 ± 19.74 vs 94.18 ± 21.11 bpm, P = 0.030) and RVSP values (29.24 ± 7.29 vs 38.77 ± 12.87 mmHg, P = 0.010) at baseline. The heart rate was significantly lower (81.14 ± 15.37 vs 70.85 ± 11.77 bpm, P = 0.032) in patients with sustained improvement at the second follow-up.

Guideline-directed medications for HF did not differ between the sustained and transient improvement subgroups. The clinical outcome findings of the subgroups were in line with the overall study findings. Patients with sustained improvement in LVEF experienced a significantly lower incidence of new cardiovascular conditions (28.6% vs 7.0%, P = 0.022) without a difference in the emergency department visits (35.7% vs 29.8%, P = 0.750) or in-hospital deaths (7.1% vs 5.3%, P = 1.000). However, there was a trend towards lower hospitalizations in those with sustained improvement (50.0% vs 22.8%, P = 0.054).

DISCUSSION

This study explored the independent predictors for sustained improvement in LVEF at the advanced HF clinic in Qatar. One-quarter of our study population had reworsening in LVEF despite initial improvement. These patients with HFtimpEF were predominantly males with ischemic cardiomyopathy and higher incidence of hypertension and dyslipidemia compared with patients with HFsimpEF. The latter experienced significantly lower rates of hospitalization and diagnosis of new cardiovascular conditions during the follow-up. Sustained improvement of LVEF was positively associated with being a female, having a non-ischemic etiology of HF, and using MRA (Figure 1).

Figure 1 Characteristics and outcomes of patients with improved ejection fraction. AHFC: Advanced heart failure clinic; aOR: Adjusted odds ratio; Dec: December; ED: Emergency department; Echo 1: First (or baseline) echocardiogram; Echo 2: Second echocardiogram; Echo 3: Third echocardiogram; HF: Heart failure; ICM: Ischemic cardiomyopathy; LVEF: Left ventricular ejection fraction; LVEF1/3: LVEF at first (or baseline) echocardiogram/third echocardiogram; MRA: Mineralocorticoid receptor antagonists; NICM: Non-ischemic cardiomyopathy.

Although the current diagnosis and management of HF depends on LVEF, it is important to consider that LVEF may vary over time. Many patients may erroneously be managed in the long-term as HFimpEF based on the initial improvement seen on their follow-up echocardiogram when in fact their LVEF may have later reworsened. This misclassification impacts their predicted mortality and hospitalizations. One study found that one-quarter of all patients in their cohort, similar to ours, had subsequent deterioration in their LVEF despite initial improvement[14]. This finding likely underpredicts the long-term course of LVEF owing to the lack of routine follow-up echocardiograms conducted in clinical practice once the patient is labelled to have recovered LVEF.

Currently, there is no consensus in the international guidelines on the exact cutoffs for LVEF improvement to define HFimpEF[12]. Due to the small sample size, we included all patients who demonstrated any (rather than large) improvement in LVEF on their first follow-up echocardiogram (Echo-2). However, our study found that the average initial improvement of LVEF was approximately 10.0% (∆ LVEF 8.1%-9.5%). Furthermore, the findings from subgroup analysis of patients who had an initial large improvement in LVEF of 10.0% or more were not different from the findings of the overall population.

Patients with HFsimpEF made up 77.7% of the total cohort in our study. Similarly, in a study that enrolled patients with dilated cardiomyopathy (n = 188), 46.0% of them showed an initial improvement in their LVEF (from 26.0 ± 7.0% to 48.0 ± 10.0%). At 36 months of follow-up after Echo-2, most patients had Echo-3 (n = 183), which also showed sustained EF improvement in 70.0% of patients[15]. In contrast the reported frequency of HFtimpEF varies between 11.0% and 38.0% in prior studies[16]. This variation is thought to be due to different definitions of LVEF reworsening and variations in the timing and length of the follow-up echocardiograms. This led to increased rates of reported reworsening as the follow-up period was extended.

In our study patients with HFsimpEF showed concomitant improvement in echocardiographic parameters such as left atrium volume index, left ventricular end-systolic diameter, and RVSP despite having an initially lower LVEF than those with HFtimpEF. Similar findings have been observed in a study by Blechman et al[15]. In our study patients with HFsimpEF had a lower mean heart rate than those with HFtimpEF at follow-up, highlighting the continued use of beta-blockers despite initial improvement in patients with HFrEF. However, another study found no significant association between continued beta blocker therapy and sustained improvement in LVEF[17]. Hence, further studies are needed to confirm the long-term benefit of beta-blockers in HFsimpEF, especially to determine whether the benefits exist beyond just the lowering of heart rate.

Understanding the factors that lead to long-term sustained vs transient improvement followed by deterioration in LVEF is crucial for individualizing patient care. By recognizing the characteristics of such patients in each group, we can improve risk stratification and identify the vulnerable patients who may benefit from more aggressive treatment approaches. Currently, there is substantial evidence on the characteristics and outcomes of patients with HFimpEF. It is also known that patients with recovered EF have an improved prognosis after recovery[6,7,18]. However, it is not an entirely benign outlook. Despite experiencing improvement in their LVEF, these patients tend to have increased rates of all-cause hospitalizations and hospitalizations for HF[8]. Hence, HFimpEF should not be considered recovered. However, data is lacking about whether it is due to reworsening of the LVEF or whether there is a difference in the characteristics and clinical outcomes between those with HFtimpEF and those with HFsimpEF.

Our patients with HFsimpEF had favorable outcomes in terms of hospitalization rates and diagnosis of new cardiovascular diseases without an impact on mortality. In contrary other studies demonstrated a significant impact on mortality. Park et al[19] demonstrated that LVEF can be a predictor of mortality and that a declining LVEF was associated with a two-fold higher mortality. Sustained improvement of LVEF was positively associated with being female, having non-ischemic etiology of HF, and using MRA in our study.

In a recent study by McElderry et al[17], who recruited 7070 patients with HFimpEF, White race (OR = 1.31, 95%CI: 1.17-1.48) and continued use of renin-angiotensin-aldosterone system inhibitors or angiotensin receptor-neprilysin inhibitor (ARNi) (OR = 1.13, 95%CI: 1.03-1.25) were associated with increased odds of maintaining an improved LVEF. Male sex (OR = 0.84, 95%CI: 0.76-0.93), use of loop diuretics (OR = 0.79, 95%CI: 0.72-0.87), atrial fibrillation (OR = 0.85, 95%CI: 0.77-0.94), and history of myocardial infarction (OR = 0.76, 95%CI: 0.67-0.85) were correlated with a decline in LVEF over time in patients with HFimpEF[17].

Another study (n = 183) concluded that sustained improvement in LVEF was associated with a favorable long-term prognosis and reported the factors that were associated with the sustained improvement in patients with dilated cardiomyopathy. Pregnancy or chemotherapy-associated cardiomyopathy were associated with sustained improvement, but family history of dilated cardiomyopathy and long-standing disease had worse prognoses. Independent predictors of sustained improvement included pregnancy-associated disease, shorter disease duration, left ventricular hypertrophy by echocardiogram, and baseline LVEF of ≤ 25.0%. There was no difference in terms of medical therapy between the groups (HFsimpEF and HFtimpEF). However, beta-blocker dose was significantly higher in those with HFsimpEF. Interestingly, to predict sustained improvement in LVEF, the investigators established a score that assigned one point to each of the following variables: (1) Short disease duration (< 3.0 years) and no familial cardiomyopathy; (2) Baseline LVEF of ≤ 25.0%; (3) Pregnancy-associated presentation; and (4) Left ventricular wall thickness of ≥ 12.0 mm. A score of ≥ 3.0 reliably predicted sustained improvement in LVEF in 91.0% of patients[15]. It is important to recognize in our study that patients with pregnancy-associated cardiomyopathy were excluded. Yet most of our patients exhibited sustained improvement underscoring the importance of treatment-related factors in LVEF improvement.

To the best of our knowledge, only a few studies have investigated the factors associated with sustained improvement in LVEF despite the accumulating studies examining the characteristics of patients with HFimpEF[17]. Enzan et al[20] concluded that beta-blocker therapy prevented the deterioration of LVEF (i.e. decrease in LVEF ≥ 10.0%) in patients with recovered dilated cardiomyopathy (i.e. mean LVEF was 49.3 ± 8.2%) at the 2-year follow-up (OR = 0.77, 95%CI: 0.63-0.95; P = 0.013). Chen et al[21] prospectively determined the safety of withdrawing spironolactone in patients with idiopathic dilated cardiomyopathy with improved LVEF (i.e. mean 44.0%-47.0%). Withdrawing spironolactone increased the likelihood of relapse (i.e. reduction in LVEF < 10.0%) within 12 months (relative risk = 4.31, 95%CI: 1.67-11.11; P < 0.001).

Halliday et al[22] in their pilot study concluded that tapering medications gradually can lead to relapses in patients with recovered dilated cardiomyopathy. Additionally, recent research identified echocardiographic parameters, namely global longitudinal strain (GLS), that could predict sustained improvement in LVEF. Adamo et al[23] reported that a normal GLS was an indicator of maintaining a stable LVEF after its recovery and abnormal GLS was a predictor of a decrease in LVEF during follow-up. However, the authors defined a reduced LVEF as < 50.0%.

The results of our study should be interpreted in the context of the following potential limitations. Firstly, the retrospective observational design of the study may have introduced confounding by unmeasured variables, potentially affecting the reported associations between baseline characteristics and improvement in LVEF. Secondly, we defined improvement and sustained improvement in LVEF as 1.0% or greater on follow-up echocardiograms regardless of a specific cutoff for the final LVEF. Hence, interpreting the results must account for the inter-echocardiographer variability in the reporting of LVEF on each follow-up and must recognize the limitation of echocardiography as an imaging modality (i.e. it may be unable to recognize small improvements in LVEF of 1.0%). This has the potential to misclassify some patients between the two groups. Thirdly, the possibility of survival bias or lead-time bias cannot be fully excluded considering that patients who did not have a second follow-up echocardiogram (Echo-3) were excluded. Fourth, our study did not evaluate subjective measures of improvement to compare quality-of-life differences between HFsimpEF and HFtimpEF patients as such data were not available. It is also important to note that our study period concluded in 2018 when only a few of our patients were on an ARNi or a sodium-glucose cotransporter-2 inhibitors. This may result in potential underestimation of patients with long-term sustained improvement in LVEF considering their proven benefits in patients with HFrEF. Regarding clinical outcomes, our study investigated only the in-hospital death rates. Therefore, potential differences in the overall mortality rates between the groups remains to be elucidated. Finally, we did not evaluate other factors such as adherence to dietary recommendations, medication compliance, alcohol consumption, or smoking cessation, all of which can impact improvement in LVEF[24].

Despite these limitations, the key strength of our study was the inclusion of all patients from the largest tertiary care center for cardiovascular disease in Qatar, a country with a predominantly South East Asian and Middle Eastern/North African populations who are underrepresented in most HF studies. Other peripheral hospitals in Qatar share the same electronic medical records, hence the potential to miss a clinical outcome was reduced. Our study offers further baseline data for long-term longitudinal assessment of HFrEF patients with regular follow-up echocardiograms to prevent lack of recognition of HFtimpEF.

Further prospective studies are needed for the validation of the associated factors of improvement mentioned in this study to allow the creation of a risk stratification model and to identify patients at risk of reworsening LVEF to intensify their management strategy. In an era marked by extensive research demonstrating the benefits of ARNi and sodium-glucose cotransporter-2 in patients with HF, our study reported favorable outcomes despite the majority of patients not receiving these novel therapies. This highlights the need for future studies focusing on patients with improved LVEF who are treated with contemporary, guideline-directed HF therapies.

CONCLUSION

Patients with HFsimpEF experienced significantly lower rates of hospitalization and diagnosis of new cardiovascular conditions than patients with HFtimpEF. Sustained improvement in LVEF was positively associated with being female, having non-ischemic etiology of HF, and using an MRA.