Between January 2016 and December 2021, the HF registry of Chi Mei Medical Center enrolled 165 patients who had symptomatic HFrEF and New York Heart Association (NYHA) functional class $\ge$II, and were receiving a sequential combination of ARNI and SGLT2i for HF management. Participants who lacked baseline and follow-up laboratory data, follow-up echocardiographic parameters, regular follow-up, or used ARNI or SGLT2i for less than one month were excluded from the study. Ninety-five and 70 patients received the ARNI-first and SGLT2i-first strategies, respectively, by physician’s choice. Data on general demographics, baseline hemodynamic condition, etiologies of HF, comorbidities, baseline and follow-up serum creatinine, follow-up N-terminal pro-B-type natriuretic peptide (NT-ProBNP), echocardiographic parameters, and clinical outcomes were compared between the ARNI and SGLT2i-first strategy groups. Echocardiographic examination was recommended every 3–6 months for patients with HFrEF. Baseline serum creatinine levels were based on data obtained within 1 month before using ARNI or SGLT2i.

Echocardiographic parameters, including left ventricular ejection fraction (LVEF), left atrial dimension (LAD), and left ventricular end-systolic and end-diastolic volume (LVESV and LVEDV), were measured using GE Vivid 9 (GE Healthcare Chicago, IL, USA), or Philips IE33, or Philips EPIQ 7 (Philips Healthcare, Amsterdam, Netherlands). LVEF, LAD, LVESV, and LVEDV were quantified using the M mode and corrected using the two-dimensional guided biplane Simpson’s method of disc measurements by using echocardiography. Echocardiographic examination was recommended every 3–6 months for patients with HFrEF thereafter in the absence of clinical events and the onset of HF.

Symptomatic HFrEF was defined as left ventricular ejection fraction 40% and experienced symptoms of HF such as shortness of breath, swollen ankles, and fatigue, as well as signs of HF including increased jugular venous pressure, crackling sounds in the lungs, and swelling in the extremities [9]. HF hospitalization was defined as the occurrence of HF events of NYHA functional classes II–IV in the absence of other alternative diagnoses. Symptoms of HF were defined as the need for medical treatment with NYHA functional class II–IV symptoms. Cardiovascular (CV) mortality was defined as sudden death related to arrhythmias, HF, and myocardial infarction. All-cause mortality was defined as death due to any cause.

The study endpoints were hospitalization for HF, CV mortality, and all-cause mortality during one-year follow-up period.

Data were presented as the mean $\pm$ standard deviation, median and interquartile range, or numbers (percentages). Median with the interquartile range presented if data were not normally distributed. The clinical characteristics of the two groups were compared using Student’s t-test or Mann-Whitney U test for continuous variables and the Chi-square test for categorical variables. Statistical significance was defined as p 0.05. All analyses were performed using statistical software (SPSS for Windows, Version 22, IBM Corp., Armonk, NY, USA).

In the ARNI-first group, the mean age was 60 $\pm$ 14.7 years, and the prevalence in men was 74.7%. In the SGLT2i-first group, the mean age was 60 $\pm$ 11.6 years, and the prevalence in men was 67.1% (Table 1). A shorter interval of combination with another medication was noted in the ARNI-first group (ARNI-first vs. SGLT2i-first; 74 [49–100] days vs. 112 [86–138]; p = 0.044). A higher prevalence of adding another medication within 1 month was noted in the ARNI-first group (ARNI-first vs. SGLT2i-first; 60.0% vs. 32.9%; p 0.001). The average age and prevalence in men did not differ between the two groups. Hemodynamic conditions (systolic blood pressure and heart rate), etiologies of HF, comorbidities, baseline renal function, and NT-proBNP, and medication were not significantly different between the two groups. A higher mean dose of ARNI presented in the ARNI-first group (ARNI-first vs. SGLT2i-first; 199.5 $\pm$ 125.6 mg vs. 147.5 $\pm$ 102.7 mg; p = 0.005).

The baseline and follow-up left ventricular performances are listed in Table 2. At the baseline phase, the mean LVEF, LAD, LVESV, LVEDV, E/A, mitral E/E’, lateral E/E’, and RV S’ did not differ between the two groups. A numerically higher prevalence of mitral regurgitation $\ge$III was noted in the ARNI-first group. In the follow-up phase, the mean LVEF, LAD, LVESV, LVEDV, E/A, mitral E/E’, lateral E/E’, and RV S’ did not differ between the two groups. There was no significant difference in the change in LVEF, LAD, LVESV, and LVEDV between the ARNI-first group and SGLT2i-first groups.

The incidence of HF hospitalization, CV mortality, and all-cause mortality were higher in the ARNI-first group; however, they did not achieve a significant difference when the ARNI-first group was compared with the SGLT2i-first group (Table 3).

Follow-up serum creatinine levels and estimated glomerular filtration rate (eGFR) did not differ between the two groups (Table 4). The decrease in eGFR was not significantly higher in the SGLT2i-first group. A non-significant trend of lower NT-proBNP levels was noted in the SGLT2i-first group (ARNI-first vs. SGLT2i-first; 1383 [319–2507] pg/mL vs. 570 [206–1314] pg/mL; p = 0.055). A significantly higher discontinuation rate of diuretic agents was noted in the SGLT2i-first group (ARNI-first vs. SGLT2i-first; 6.8% vs. 17.5%; p = 0.039).

Fig. 1A showed non-significant trend of improving EF $>$5% when early combination ($\le$14D) compared to late combination ($>$14D) ($\le$14D vs. $>$14D; 80.0% vs. 64.8%; p = 0.060). The decreasing LVEDV $>$ 10% did not differ between $\le$14D and $>$14D groups (Fig. 1B), but the decreasing LVESV $>$10% showed significant higher in the early combination ($\le$14D) group ($\le$14D vs. $>$14D; 72.0% vs. 54.4%; p = 0.042) (Fig. 1C).

Fig. 1.

The improving ejection fraction, decreasing left ventricular end-diastolic volume, and decreasing left ventricular end-systolic volume when early combination ($\mathrm{\le }$14D) compared to late combination ($\mathrm{>}$14D). (A) A non-significant trend of improving ejection fraction $>$5% presented when early combination ($\le$14D) compared to late combination ($>$14D) ($\le$14D vs. $>$14D; 80.0% vs. 64.8%; p = 0.060). (B) The decreasing left ventricular end-diastolic volume $>$10% did not differ between early combination ($\le$14D) and late combination ($>$14D) groups ($\le$14D vs. $>$14D; 52.0% vs. 41.1%; p = 0.221). (C) The decreasing left ventricular end-systolic volume $>$10% showed significant higher when early combination ($\le$14D) compared to late combination ($>$14D) group ($\le$14D vs. $>$14D; 72.0% vs. 54.4%; p = 0.042). Abbreviation: EF, ejection fraction; LVEDV, left ventricular end-systolic volume; LVESV, left ventricular end-systolic volume; ARNI, angiotensin receptor neprilysin inhibitor; SGLT2i, sodium-glucose cotransporter 2 inhibitors; D, day.

In patients with HFrEF, several factors may affect renal function, including cardiac pump failure related to lower blood supply and increased congestion, chronic activation of the adrenergic system and renin-angiotensin-aldosterone system (RAAS), and direct and indirect effects of anti-HF medication [11, 12]. Approximately 30% of the patients with HF also have chronic kidney disease, and worsening renal function occurs in approximately 25% of the cases, with a deleterious effect on outcomes [13]. Additionally, renal function may also be affected by age, and comorbidities, including atherosclerosis, diabetes mellitus, hypertension, and renal dysfunction, are associated with poor outcomes [14]. Acute worsening renal function after starting RAAS inhibitors occurs in approximately 13% of the patients with HF, more so in the older hypertensive patients, following exposure to iodinated contrast, during acute illness, heart failure exacerbation, hypotension, and dehydration [15, 16, 17]. According to recent studies on SGLT2i for HF, a lower risk of kidney failure and slower progression of renal function were observed when the SGLT2i group was compared with the control group [18, 19]. In our study, no patients needed emerge hemodialysis and experienced renal death. The progression of renal function did not differ between ARNI and SGLT2i-first strategy groups. Therefore, sequential combination of ARNI and SGLT2i may be feasible for the patients with symptomatic HFrEF.

Among the patients with HFrEF, a low proportion of patients received the target dose (or received at least 50% of the target dose), partly due to hypotension [20]. In the PIONEER-HF study, ARNI could safely be initiated in patients with SBP $>$100 mmHg for 6 hours, without using vasodilators or increasing the dose of intravenous diuretics in the preceding 6 hours, and no use of inotropes in the preceding 24 hours [21]. When the SBP is 100–120 mmHg, ARNI should be initiated at a low dose to prevent hypotension [20]. SGLT2i does exhibit a modest blood pressure lowering effect and approximately 1–4 mmHg in patients with HF, and the volume status of such a population on diuretics should be reevaluated owing to the risk of over-diuresis and hypotension [22]. The risk of hypotension and acute worsening of renal function may allow physicians to choose one medication first. The expert also suggested that SGLT2i could be initiated on at a minimum dose of RAAS inhibitors and $\beta$-blockers without systemic hypotension [20]. In our study, a higher prevalence of discontinuing diuretic agents and trend toward lower serum NT-proBNP levels were observed in patients with HFrEF and SGLT2i-first strategies. Better effect of decongestion was expected in the SGLT2i-first group. The mechanisms of improving decongestion of SGLT2i are related to reduction of interstitial fluid volume, natriuresis and osmotic diuresis, and neurohumoral and inflammation attenuation [23].

In a retrospective study of diabetic patients with HFrEF, those with the initiation of sacubitril/valsartan presented a more prominent improvement in LV performance than those with the initiation of SGLT2i, and the combination of ARNI and SGT2i showed significant improvement in cardiac function and prognosis in this population [24]. Another real-world study also found that treatment with a combination of SGLT2i and ARNI was associated with a lower risk of composite HF hospitalization or all-cause mortality and was well tolerated [25]. The subgroups with SGLT2i presented better results than those without SGLT2i [26]. A meta-analysis concluded that SGLT2i and ARNI demonstrated similar effects (indirect comparison hazard ratio 0.93, 95% confidence interval 0.82–1.06, p-value = 0.28), while the combination of SGLT2i and ARNI resulted in a better cardiovascular protective effect [25]. In our study, the sequence of ARNI and SGLT2i did not affect reverse LV remodeling and renal function progression and did not present with acute kidney injury. The SGLT2i-first strategy may provide better decongestion. Early combination may provide better LV remodeling effect.

This study had several limitations. First, it was a retrospective, nonrandomized study with limited size, and we could not rule out selective bias. Second, the timing of adding another medication based on the judgement of physicians and a shorter interval was noted in the ARNI-first group. Third, echocardiographic examination was performed every 3–6 months for patients with HFrEF if no new events. However, we still provide important information for clinical practice regarding the sequence of ARNI and SGLT2i implementation in patients with symptomatic HFrEF.