Abstract
Objective: Implantable cardioverter defibrillators (ICDs) are the standard treatment for patients with reduced left ventricular ejection fraction (LVEF ≤35%) to reduce the risk of sudden cardiac death. Loop diuretics can cause electrolyte imbalances, leading to an increased incidence of ICD shocks. Sodium-glucose cotransporter-2 inhibitors (SGLT2is) have shown cardiovascular benefits in patients with heart failure (HF), but their effects on ventricular arrhythmias and ICD shocks, particularly in patients receiving different doses of loop diuretics, are not fully understood. This study evaluated the effects of furosemide dose and SGLT2i use on ICD shocks in HF patients with reduced left ventricular ejection fraction (HFrEF). Materials and Methods: HFrEF patients using oral furosemide and undergoing ICD implantation in our clinic were followed for 12 months to monitor ICD shocks for ventricular arrhythmias. They were grouped according to daily oral furosemide dose and SGLT2i use. Results: Out of 175 patients, the use of high-dose furosemide (>80 mg/day) was significantly higher in the ICD shock group compared to the non-shock group (38.8% vs. 16.7%, p = 0.001), while the use of SGLT2i was lower (19.4% vs. 45.4%, p < 0.001). ICD shocks occurred in 67.6% of patients on high-dose furosemide without SGLT2i and 30.0% with SGLT2i (p < 0.001). Multivariate analysis identified the absence of SGLT2i as an independent predictor of ICD shocks. Conclusions: SGLT2i was associated with reduced ventricular arrhythmias and ICD shocks in HF patients, even when high doses of furosemide were used. The absence of SGLT2i in HF treatment was an independent predictor of ICD shocks.
Clinical and demographic characteristics of heart failure patients using oral furosemide were compared between those with and without shocks from implantable defibrillators.
High-dose oral furosemide (>80 mg/day) was associated with an increased risk of defibrillator shocks.
Patients using SGLT2 inhibitors with furosemide experienced fewer shocks compared to those not using SGLT2 inhibitors.
Introduction
Heart failure (HF) patients are at increased risk of cardiac mortality due to ventricular tachycardia and ventricular fibrillation (VT/VF). This condition is characterised by complex molecular and structural changes in the heart, including hypertrophy and fibrosis. This remodelling process is closely associated with an increased susceptibility to arrhythmias and can exacerbate the pro-arrhythmic effects of factors such as hypokalemia [1, 2]. Implantable cardioverter defibrillators (ICD) are considered the standard of treatment protocol for patients with a reduced left ventricular ejection fraction (LVEF ≤35%), aiming to reduce the risk of sudden cardiac death [3, 4]. In HF, electrolyte imbalances such as hypokalemia and hypomagnesaemia are associated with VT/VF development over time and increase the likelihood of ICD shocks [1, 2]. ICD shocks not only have a negative impact on quality of life in HF but have also been associated with increased mortality in subsequent analyses [5]. A previous study by Sweeney et al. [6] reported an increased risk of mortality associated with the termination of ventricular arrhythmia episodes with ICD shocks. Similarly, results from the Sudden Cardiac Death in Heart Failure Trial showed that a single appropriate shock increased the risk of death by a factor of five [7]. Therefore, treatment that reduces the burden of ventricular arrhythmias should be continued after ICD implantation.
Loop diuretics and thiazides play a pivotal role in preventing congestion in HF treatment, and they can also cause a dose-dependent decrease in potassium levels [8]. Hypokalemia, defined as serum potassium (K+) levels below 3.5 mEq, has been associated in numerous studies with an increased risk of fatal arrhythmias such as VT/VF in HF patients with serum K+ <4 mEq [9, 10]. This situation creates a paradox: intensive diuretic therapy, which is essential to prevent congestion in HF patients, simultaneously leads to electrolyte imbalances and the development of ventricular arrhythmias.
Sodium-glucose cotransporter-2 inhibitors (SGLT2is) have been associated with a reduction in cardiovascular mortality, but their impact on the adverse effects of other HF treatments is not fully understood [3, 11]. HF medications, particularly diuretics, can sometimes adversely affect metabolic parameters and cause ICD shocks [12]. However, the effect of SGLT2i on the development of ventricular arrhythmias and subsequent ICD shocks remains unclear, particularly in patients receiving furosemide therapy. We aimed to evaluate the effects of furosemide dose and SGLT2i use on ventricular arrhythmias and ICD shocks in HF patients with reduced LVEF (HFrEF) and thus understand the cardiovascular benefits of SGLT2i.
Materials and Methods
Study Design and Population
In this retrospective cohort study, patients with HFrEF who underwent ICD implantation for primary or secondary prophylaxis in the cardiology department of our hospital were evaluated. The study had the approval of the Institutional Review Board. Patients were followed at the Pacemaker Control Outpatient Clinic for 12 months after ICD implantation. Data on ICD-detected arrhythmias and shock therapies were carefully collected by evaluating electrogram (EGM) data.
Inclusion criteria were patients over 18 years of age, ICD implantation at least 12 months prior, oral furosemide use, and LVEF ≤35%. Patients on SGLT2i were excluded if they had discontinued treatment. Patients whose furosemide dose changed during follow-up, type 1 diabetes mellitus, eGFR <25 mL/min/1.73 m2 or on dialysis, with electrolyte imbalance due to diarrhoea or vomiting, receiving arrhythmia therapy such as amiodarone, mexiletine or with a history of VT ablation, hypertrophic cardiomyopathy, right ventricular cardiomyopathy, congenital heart disease, percutaneous or surgical coronary revascularisation within the previous 12 months were excluded.
Data Collection
Parameters such as demographic and baseline characteristics, device type, ICD therapies, laboratory and echocardiographic data concurrent with pacemaker controls, and mortality information were carefully collected from the hospital registry system and patient files. If mortality data were not available, they were obtained by contacting the registration system of the Ministry of Health.
Endpoints
The primary endpoint of the study was the occurrence of clinically significant ventricular arrhythmias requiring an ICD shock during this period; the secondary endpoint was death from cardiovascular (CV) causes.
ICD Programming
Only treatments that resulted in an appropriate shock for VT or VF were included in this study. In our department, the programming of the ICD algorithms was based on established guidelines. The event was recorded without intervention if the heart rate was between 130 and 171 bpm (zone 1). For heart rates above 171 bpm (zone 2), the ICD initially performed anti-tachycardia pacing with three bursts and ramps. If the arrhythmia persisted, a defibrillator shock was delivered. The device was set to deliver an immediate shock at heart rates above 207 bpm (zone 3) (shown in Fig. 1). For secondary prevention, we programmed the ICDs according to the patient’s previous VT-VF cycle length.
Device electrograms (EGMs). a Detection of VT onset and anti-tachycardia pacing (ATP) are shown. b VT episode accelerated to VF. c Defibrillation of the device with 34.0 j. d Return to basal rhythm.
Device electrograms (EGMs). a Detection of VT onset and anti-tachycardia pacing (ATP) are shown. b VT episode accelerated to VF. c Defibrillation of the device with 34.0 j. d Return to basal rhythm.
Statistical Analysis
We analysed the normal distribution of continuous variables using the Kolmogorov-Smirnov and Shapiro-Wilk tests. Descriptive statistics were presented as mean ± standard deviation (SD) for normally distributed continuous variables and median (interquartile range, Q1–Q3) for non-normally distributed variables. Student’s t test and Mann-Whitney U test were used to compare data between the groups with and without ICD shock, as appropriate. Categorical variables were compared using the chi-squared test. At baseline, patients were divided into low-dose (10–80 mg furosemide daily) and high-dose (81–240 mg furosemide daily) diuretic groups. Each group was further subdivided according to SGLT2i use. A total of four groups were analysed. Differences in patient characteristics based on diuretic dose and SGLT2i use were assessed using Fisher’s exact test for categorical variables, Mann-Whitney U test for continuous variables, or Kruskal-Wallis test for comparisons between the four groups. Spearman correlation analysis was used to assess the relationship between furosemide dose and blood electrolyte levels. Logistic regression models were constructed to predict the occurrence of ICD shocks. Univariate regression analysis was performed first, and variables with p < 0.10 in the analyses, together with gender, were included in the multivariate analysis using the Enter method. An ROC curve analysis was used to determine the optimal furosemide cutoff dose for CV mortality. We used the Youden index to identify the optimal cutoff point. Statistical analyses and visualisations were performed using GraphPad Prism 10 software, with a significance level of p < 0.05.
Results
We screened 231 patients who underwent ICD implantation at our clinic between January 1, 2019, and April 31, 2023. Among them, 16 patients were not receiving diuretic therapy; 16 patients were using antiarrhythmic drugs such as amiodarone and mexiletine; 12 patients underwent ICD implantation not related to HF; 7 patients experienced ICD dysfunction due to factors such as lead fracture or sensing defects; 5 patients experienced inappropriate ICD shocks and were excluded from the study. As a result, 175 patients were included in the final analysis. The study flowchart is shown in Figure 2.
The baseline clinical characteristics of the patients classified according to the occurrence of an ICD shock are summarised in Table 1. There were no significant differences between the groups in terms of gender, body mass index, hypertension, diabetes mellitus, hyperlipidaemia, or cerebrovascular disease. Patients who experienced an ICD shock during follow-up were more likely to be older and had a higher prevalence of ischaemic aetiology (Table 1). The ICD shock group had a significantly lower median LVEF than those in the non-shock group (30.0 [25.0–33.0] vs. 34.0 [28.0–37.0], respectively, p < 0.001). Regarding medication use, patients in the non-shock group were more likely to receive guideline-directed medical therapy (GDMT) consisting of a β-blocker, a mineralocorticoid receptor antagonist (MRA), and either an angiotensin-converting enzyme inhibitor, an angiotensin receptor blocker, or an angiotensin receptor neprilysin inhibitor compared to the shock group (78.7% vs. 65.7%, p = 0.045). The shock group had a significantly higher rate of high-dose furosemide use than the non-shock group (38.8% vs. 16.7%, p = 0.001), while they had a lower rate of SGLT2i use (19.4% vs. 45.4%, p < 0.001). There was a significant negative correlation between the dose of furosemide and the levels of potassium, calcium and magnesium in patients' blood samples (p < 0.001, for each; shown in Fig. 3).
Comparison of variables between with and without ICD shock groups
Variables . | Without shock group (n = 108) . | Shock group (n = 67) . | p value . |
---|---|---|---|
Age, years | 52.4±14.4 | 57.0±15.2 | 0.047 |
Male, n (%) | 64 (59.3) | 39 (58.2) | 0.891 |
Body mass index, kg/m2 | 27.7 [24.2–31.4] | 26.8 [24.9–30.6] | 0.667 |
Dyslipidemia, n (%) | 65 (60.2) | 39 (58.2) | 0.796 |
Hypertension, n (%) | 56 (51.9) | 42 (62.7) | 0.160 |
Diabetes mellitus, n (%) | 36 (33.3) | 28 (41.8) | 0.259 |
CVD, n (%) | 16 (14.8) | 8 (11.9) | 0.591 |
Furosemide dose, mg/day | 40.0 [40.0–80.0] | 80.0 [40.0–120.0] | 0.014 |
Baseline ECG | |||
Sinus, n (%) | 54 (50.0) | 30 (44.8) | 0.746 |
AF, n (%) | 21 (19.4) | 13 (19.4) | |
Pace rhythm, n (%) | 33 (30.6) | 24 (35.8) | |
QRS duration, ms | 120.0 [110.0–130.0] | 125.0 [110.0–140.0] | 0.244 |
Echocardiographic characteristics | |||
EDD, mm | 62.0 [56.2–68.0] | 62.0 [57.0–70.0] | 0.710 |
LA, mm | 47.0 [43.2–50.0] | 46.0 [42.0–51.0] | 0.332 |
LVEF, % | 34.0 [28.0–37.0] | 30.0 [25.0–33.0] | 0.001 |
ICD type | |||
VVI, n (%) | 78 (72.2) | 42 (62.7) | 0.220 |
DDD, n (%) | 20 (18.5) | 13 (19.4) | |
CRT-D, n (%) | 10 (9.3) | 12 (17.9) | |
Reason for ICD implantation | |||
Secondary prevention, n (%) | 23 (21.3) | 19 (28.4) | 0.288 |
Primary prevention, n (%) | 85 (78.7) | 48 (71.6) | |
Medication | |||
Statin, n (%) | 50 (46.3) | 31 (46.3) | 0.997 |
Beta blocker, n (%)a | 106 (98.1) | 67 (100.0) | 0.263 |
ACEi/ARB/ARNI, n (%) | 65/24/5 (87.0) | 38/11/3 (77.6) | 0.103 |
MRA, n (%) | 91 (84.3) | 47 (70.1) | 0.026 |
CCB, n (%) | 14 (13.0) | 18 (26.9) | 0.021 |
Digoxin, n (%) | 28 (25.9) | 17 (25.4) | 0.935 |
Ivabradine, n (%) | 13 (12.0) | 4 (6.0) | 0.188 |
Thiazide, n (%) | 33 (30.6) | 22 (32.8) | 0.752 |
SGLT2i, n (%) | 49 (45.4) | 13 (19.4) | <0.001 |
GDMT usec | |||
One of GDMT, n (%) | 2 (1.9) | 6 (9.0) | 0.045 |
Two of GDMT, n (%) | 21 (19.4) | 17 (25.4) | |
Three of GDMT, n (%) | 85 (78.7) | 44 (65.7) | |
HF aetiology | |||
Non-ischaemic CMP, n (%) | 41 (38.0) | 14 (20.9) | 0.018 |
Ischaemic CMP, n (%) | 67 (62.0) | 53 (79.1) | |
Furosemide dose group | |||
Low-dose, ≤80 mg/day, n (%) | 90 (83.3) | 41 (61.2) | 0.001 |
High-dose, >80 mg/day, n (%) | 18 (16.7) | 26 (38.8) | |
NYHA class | |||
NYHA class I, n (%) | 64 (59.3) | 2 (3.0) | <0.001 |
NYHA class II, n (%) | 41 (38.0) | 46 (68.7) | |
NYHA class III, n (%) | 3 (2.8) | 19 (28.4) | |
NT-proBNP, pg/mL | 354.5 [181.2–790.7] | 846.0 [499.0–2535.0] | <0.001 |
eGFR, mL/min/1.73 m2 | 78.2±25.3 | 70.4±27.4 | 0.058 |
Urea, mg/dL | 45.0 [31.0–61.7] | 44.0 [34.0–53.0] | 0.639 |
Creatinine, mg/dL | 2.1 [1.8–2.2] | 1.0 [0.8–1.1] | 0.676 |
Sodium, mEq/L | 138.0 [136.0–141.0] | 138.0 [135.0–141.0] | 0.604 |
Potassium, mEq/L | 4.4±0.4 | 3.9±0.6 | <0.001 |
Magnesium, mEq/L | 2.1 [1.8–2.2] | 1.9 [1.6–2.1] | 0.001 |
Adjusted calcium, mEq/Lb | 9.2±0.6 | 8.6±0.9 | <0.001 |
Albumin, mg/dL | 4.3 [3.8–4.6] | 4.3 [3.9–4.5] | 0.704 |
Clinical outcomes | |||
Hospitalisation due to cardiac decompensation, n (%) | 9 (8.3) | 23 (34.4) | <0.001 |
Cardiovascular mortality, n (%) | 3 (2.8) | 15 (22.4) | <0.001 |
Variables . | Without shock group (n = 108) . | Shock group (n = 67) . | p value . |
---|---|---|---|
Age, years | 52.4±14.4 | 57.0±15.2 | 0.047 |
Male, n (%) | 64 (59.3) | 39 (58.2) | 0.891 |
Body mass index, kg/m2 | 27.7 [24.2–31.4] | 26.8 [24.9–30.6] | 0.667 |
Dyslipidemia, n (%) | 65 (60.2) | 39 (58.2) | 0.796 |
Hypertension, n (%) | 56 (51.9) | 42 (62.7) | 0.160 |
Diabetes mellitus, n (%) | 36 (33.3) | 28 (41.8) | 0.259 |
CVD, n (%) | 16 (14.8) | 8 (11.9) | 0.591 |
Furosemide dose, mg/day | 40.0 [40.0–80.0] | 80.0 [40.0–120.0] | 0.014 |
Baseline ECG | |||
Sinus, n (%) | 54 (50.0) | 30 (44.8) | 0.746 |
AF, n (%) | 21 (19.4) | 13 (19.4) | |
Pace rhythm, n (%) | 33 (30.6) | 24 (35.8) | |
QRS duration, ms | 120.0 [110.0–130.0] | 125.0 [110.0–140.0] | 0.244 |
Echocardiographic characteristics | |||
EDD, mm | 62.0 [56.2–68.0] | 62.0 [57.0–70.0] | 0.710 |
LA, mm | 47.0 [43.2–50.0] | 46.0 [42.0–51.0] | 0.332 |
LVEF, % | 34.0 [28.0–37.0] | 30.0 [25.0–33.0] | 0.001 |
ICD type | |||
VVI, n (%) | 78 (72.2) | 42 (62.7) | 0.220 |
DDD, n (%) | 20 (18.5) | 13 (19.4) | |
CRT-D, n (%) | 10 (9.3) | 12 (17.9) | |
Reason for ICD implantation | |||
Secondary prevention, n (%) | 23 (21.3) | 19 (28.4) | 0.288 |
Primary prevention, n (%) | 85 (78.7) | 48 (71.6) | |
Medication | |||
Statin, n (%) | 50 (46.3) | 31 (46.3) | 0.997 |
Beta blocker, n (%)a | 106 (98.1) | 67 (100.0) | 0.263 |
ACEi/ARB/ARNI, n (%) | 65/24/5 (87.0) | 38/11/3 (77.6) | 0.103 |
MRA, n (%) | 91 (84.3) | 47 (70.1) | 0.026 |
CCB, n (%) | 14 (13.0) | 18 (26.9) | 0.021 |
Digoxin, n (%) | 28 (25.9) | 17 (25.4) | 0.935 |
Ivabradine, n (%) | 13 (12.0) | 4 (6.0) | 0.188 |
Thiazide, n (%) | 33 (30.6) | 22 (32.8) | 0.752 |
SGLT2i, n (%) | 49 (45.4) | 13 (19.4) | <0.001 |
GDMT usec | |||
One of GDMT, n (%) | 2 (1.9) | 6 (9.0) | 0.045 |
Two of GDMT, n (%) | 21 (19.4) | 17 (25.4) | |
Three of GDMT, n (%) | 85 (78.7) | 44 (65.7) | |
HF aetiology | |||
Non-ischaemic CMP, n (%) | 41 (38.0) | 14 (20.9) | 0.018 |
Ischaemic CMP, n (%) | 67 (62.0) | 53 (79.1) | |
Furosemide dose group | |||
Low-dose, ≤80 mg/day, n (%) | 90 (83.3) | 41 (61.2) | 0.001 |
High-dose, >80 mg/day, n (%) | 18 (16.7) | 26 (38.8) | |
NYHA class | |||
NYHA class I, n (%) | 64 (59.3) | 2 (3.0) | <0.001 |
NYHA class II, n (%) | 41 (38.0) | 46 (68.7) | |
NYHA class III, n (%) | 3 (2.8) | 19 (28.4) | |
NT-proBNP, pg/mL | 354.5 [181.2–790.7] | 846.0 [499.0–2535.0] | <0.001 |
eGFR, mL/min/1.73 m2 | 78.2±25.3 | 70.4±27.4 | 0.058 |
Urea, mg/dL | 45.0 [31.0–61.7] | 44.0 [34.0–53.0] | 0.639 |
Creatinine, mg/dL | 2.1 [1.8–2.2] | 1.0 [0.8–1.1] | 0.676 |
Sodium, mEq/L | 138.0 [136.0–141.0] | 138.0 [135.0–141.0] | 0.604 |
Potassium, mEq/L | 4.4±0.4 | 3.9±0.6 | <0.001 |
Magnesium, mEq/L | 2.1 [1.8–2.2] | 1.9 [1.6–2.1] | 0.001 |
Adjusted calcium, mEq/Lb | 9.2±0.6 | 8.6±0.9 | <0.001 |
Albumin, mg/dL | 4.3 [3.8–4.6] | 4.3 [3.9–4.5] | 0.704 |
Clinical outcomes | |||
Hospitalisation due to cardiac decompensation, n (%) | 9 (8.3) | 23 (34.4) | <0.001 |
Cardiovascular mortality, n (%) | 3 (2.8) | 15 (22.4) | <0.001 |
Median values are given with interquartile ranges (the 25th and 75th percentiles).
ACEi, angiotensin-converting enzyme inhibitor; AF, atrial fibrillation; ARB, angiotensin receptor blocker; ARNI, angiotensin receptor neprilysin inhibitor; CCB, calcium channel blocker; CMP, cardiomyopathy; CVD, cerebrovascular disease; ECG, electrocardiogram; EDD, end diastolic diameter; eGFR, estimated glomerular filtration rate; GDMT, guideline-directed medical therapy; HF, heart failure; ICD, implantable cardioverter defibrillator; LA, left atrium; LVEF, left ventricular ejection fraction; MRA, mineralocorticoid receptor antagonist; NT-proBNP, N-terminal pro-B-type natriuretic peptide; NYHA, New York Heart Association; SGLT2i, sodium-glucose cotransporter-2 inhibitor.
aFisher’s exact test was used.
bAdjusted calcium values are corrected according to albumin levels.
cOne of GDMT; β-blocker or ACEi/ARB/ARNI, two of GDMT; β-blocker and ACEi/ARB/ARNI, three of GDMT; β-blocker, ACEi/ARB/ARNI, and MRA.
Correlation between furosemide dose and serum levels of potassium, calcium, and magnesium.
Correlation between furosemide dose and serum levels of potassium, calcium, and magnesium.
The comparison of clinical parameters according to the use of SGLT2i with different doses of furosemide in patients with ICD revealed several significant findings. The groups include low-dose furosemide with SGLT2i (LowF/SGLT2i(+)), high-dose furosemide with SGLT2i (HighF/SGLT2i(+)), low-dose furosemide without SGLT2i (LowF/SGLT2i(−)), and high-dose furosemide without SGLT2i (HighF/SGLT2i(−)). The comparison shows notable differences between these groups with respect to ICD shocks and cardiovascular mortality. The incidence of ICD shocks varied significantly between groups, with rates of 19.2%, 30.0%, 39.2%, and 67.6% for the LowF/SGLT2i(+), HighF/SGLT2i(+), LowF/SGLT2i(−), and HighF/SGLT2i(−) groups, respectively (p < 0.001). In addition, cardiovascular mortality was 3.8%, 0%, 8.9%, and 26.5% in the respective groups (p = 0.004; Table 2). The analysis also showed significant differences in serum potassium, calcium, and magnesium levels between the four groups. Pairwise comparisons between groups showed that lower serum levels of potassium, calcium, and magnesium were associated with higher doses of furosemide, but not with SGLT2i use (p < 0.001, p < 0.001, p = 0.004, and p < 0.001 for electrolytes, respectively; Table 2).
Comparison of clinical parameters according to SGLT2i use with different furosemide doses in patients with ICD
Variables . | LowF/SGLT2i(+) (n = 52) . | HighF/SGLT2i(+) (n = 10) . | LowF/SGLT2i(−) (n = 79) . | HighF/SGLT2i(−) (n = 34) . | p value . |
---|---|---|---|---|---|
ICD shock, n (%) | 10 (19.2) | 3 (30.0) | 31 (39.2) | 23 (67.6) | <0.001 |
CV mortality, n (%) | 2 (3.8) | 0 (0) | 7 (8.9) | 9 (26.5) | 0.004 |
Potassium, mEq/La | 4.3 [3.9–4.7] | 3.6 [3.4–4.1] | 4.3 [4.0–4.7] | 3.6 [3.3–4.4] | <0.001 |
Calcium, mEq/La | 9.3 [8.6–9.7] | 8.7 [7.9–8.8] | 9.3 [8.7–9.6] | 8.25 [7.8–8.9] | <0.001 |
Magnesium, mEq/La | 2.0 [1.8–2.2] | 1.8 [1.4–2.0] | 2.0 [1.9–2.2] | 1.75 [1.5–2.1] | 0.004 |
Sodium, mEq/La | 139.0 [136.0–141.0] | 136.5 [134.5–140.5] | 139.0 [136.0–141.0] | 135.5 [133.0–138.0] | <0.001 |
GDMTb, n (%) | |||||
One of GDMT | 2 (3.8) | 0 (0) | 6 (7.6) | 0 (0) | 0.246 |
Two of GDMT | 11 (21.2) | 0 (0) | 20 (25.3) | 7 (20.6) | |
Three of GDMT | 39 (75.0) | 10 (100) | 53 (67.1) | 27 (79.4) |
Variables . | LowF/SGLT2i(+) (n = 52) . | HighF/SGLT2i(+) (n = 10) . | LowF/SGLT2i(−) (n = 79) . | HighF/SGLT2i(−) (n = 34) . | p value . |
---|---|---|---|---|---|
ICD shock, n (%) | 10 (19.2) | 3 (30.0) | 31 (39.2) | 23 (67.6) | <0.001 |
CV mortality, n (%) | 2 (3.8) | 0 (0) | 7 (8.9) | 9 (26.5) | 0.004 |
Potassium, mEq/La | 4.3 [3.9–4.7] | 3.6 [3.4–4.1] | 4.3 [4.0–4.7] | 3.6 [3.3–4.4] | <0.001 |
Calcium, mEq/La | 9.3 [8.6–9.7] | 8.7 [7.9–8.8] | 9.3 [8.7–9.6] | 8.25 [7.8–8.9] | <0.001 |
Magnesium, mEq/La | 2.0 [1.8–2.2] | 1.8 [1.4–2.0] | 2.0 [1.9–2.2] | 1.75 [1.5–2.1] | 0.004 |
Sodium, mEq/La | 139.0 [136.0–141.0] | 136.5 [134.5–140.5] | 139.0 [136.0–141.0] | 135.5 [133.0–138.0] | <0.001 |
GDMTb, n (%) | |||||
One of GDMT | 2 (3.8) | 0 (0) | 6 (7.6) | 0 (0) | 0.246 |
Two of GDMT | 11 (21.2) | 0 (0) | 20 (25.3) | 7 (20.6) | |
Three of GDMT | 39 (75.0) | 10 (100) | 53 (67.1) | 27 (79.4) |
For potassium, comparisons between groups, Mann-Whitney U test results: LowF/SGLT2i(+) versus HighF/SGLT2i(+), p = 0.009; LowF/SGLT2i(+) versus LowF/SGLT2i(−), p = 0.856; LowF/SGLT2i(+) versus HighF/SGLT2i(−), p = 0.001; HighF/SGLT2i(+) versus LowF/SGLT2i(−), p = 0.004; HighF/SGLT2i(+) versus HighF/SGLT2i(−), p = 0.772; LowF/SGLT2i(−) versus HighF/SGLT2i(−), p < 0.001.
For calcium, comparisons between groups, Mann-Whitney U test results: LowF/SGLT2i(+) versus HighF/SGLT2i(+), p = 0.002; LowF/SGLT2i(+) versus LowF/SGLT2i(−), p = 0.457; LowF/SGLT2i(+) versus HighF/SGLT2i(−), p < 0.001; HighF/SGLT2i(+) versus LowF/SGLT2i(−), p = 0.002; HighF/SGLT2i(+) versus HighF/SGLT2i(−), p = 0.591; LowF/SGLT2i(−) versus HighF/SGLT2i(−), p < 0.001.
For magnesium, comparisons between groups, Mann-Whitney U test results: LowF/SGLT2i(+) versus HighF/SGLT2i(+), p = 0.049; LowF/SGLT2i(+) versus LowF/SGLT2i(−), p = 0.910; LowF/SGLT2i(+) versus HighF/SGLT2i(−), p = 0.002; HighF/SGLT2i(+) versus LowF/SGLT2i(−), p = 0.075; HighF/SGLT2i(+) versus HighF/SGLT2i(−), p = 0.923; LowF/SGLT2i(−) versus HighF/SGLT2i(−), p = 0.003.
For sodium, comparisons between groups, Mann-Whitney U test results: LowF/SGLT2i(+) versus HighF/SGLT2i(+), p = 0.141; LowF/SGLT2i(+) versus LowF/SGLT2i(−), p = 0.664; LowF/SGLT2i(+) versus HighF/SGLT2i(−), p < 0.001; HighF/SGLT2i(+) versus LowF/SGLT2i(−), p = 0.201; HighF/SGLT2i(+) versus HighF/SGLT2i(−), p = 0.249; LowF/SGLT2i(−) versus HighF/SGLT2i(−), p < 0.001.
CV, cardiovascular; GDMT, guideline-directed medical therapy; HighF/SGLT2i(+), high-dose furosemide with SGLT2 inhibitor; HighF/SGLT2i(−), high-dose furosemide without SGLT2 inhibitor; ICD, implantable cardioverter defibrillator; LowF/SGLT2i(+), low-dose furosemide with SGLT2 inhibitor; LowF/SGLT2i(−), low-dose furosemide without SGLT2 inhibitor.
aKruskal-Wallis H test was conducted.
bOne of GDMT; β-blocker or ACEi/ARB/ARNI, Two of GDMT; β-blocker and ACEi/ARB/ARNI, Three of GDMT; β-blocker, ACEi/ARB/ARNI, and MRA.
Multivariate logistic regression analysis revealed significant associations between clinical variables and the incidence of ICD shocks. Patients not receiving SGLT2i were significantly more likely to experience an ICD shock than those receiving SGLT2i (adjusted odds ratio [aOR]: 3.401, 95% confidence interval [CI]: 1.359–8.516, p = 0.009). Higher furosemide doses were not statistically significant predictors of ICD shocks after adjustment for other variables (aOR: 1.235, 95% CI: 0.446–3.421, p = 0.685). The absence of MRA was significantly associated with increased odds of ICD shocks (aOR: 4.171, 95% CI: 1.467–11.862, p = 0.007). Additionally, lower serum potassium levels, hospitalisation for cardiac decompensation, and older age were significantly associated with increased odds of ICD shocks (Table 3).
Multivariate logistic regression analysis of variables associated with ICD shock incidence
Variables . | OR (95% CI) . | p value . | aOR (95% CI)a . | p value . |
---|---|---|---|---|
SGLT2i (ref: yes) | 3.450 (1.689–7.047) | 0.001 | 3.401 (1.359–8.516) | 0.009 |
Furosemide dose (ref: low dose) | 3.171 (1.566–6.419) | 0.001 | 1.235 (0.446–3.421) | 0.685 |
MRA (ref: yes) | 2.278 (1.091–4.756) | 0.028 | 4.171 (1.467–11.862) | 0.007 |
LVEF | 0.957 (0.920–0.996) | 0.032 | 0.847 (0.928–1.022) | 0.281 |
Potassium | 0.213 (0.112–0.460) | <0.001 | 0.189 (0.079–0.451) | <0.001 |
Calcium | 0.449 (0.299–0.674) | <0.001 | 0.577 (0.302–1.104) | 0.097 |
Magnesium | 0.173 (0.062–0.483) | 0.001 | 0.602 (0.127–2.855) | 0.523 |
Hospitalisation due to cardiac decompensation (ref: none) | 5.750 (2.461–13.433) | <0.001 | 3.826 (1.295–11.302) | 0.015 |
QRS duration | 1.012 (0.966–1.029) | 0.147 | ||
Age | 1.021 (1.000–1.043) | 0.049 | 1.038 (1.007–1.071) | 0.016 |
Gender (ref: male) | 1.044 (0.562–1.939) | 0.891 | 0.971 (0.419–2.246) | 0.944 |
HF aetiology (ref: dilated CMP) | 2.317 (1.144–4.691) | 0.020 | 2.063 (0.849–5.012) | 0.110 |
Thiazide (ref: none) | 1.111 (0.578–2.137) | 0.752 |
Variables . | OR (95% CI) . | p value . | aOR (95% CI)a . | p value . |
---|---|---|---|---|
SGLT2i (ref: yes) | 3.450 (1.689–7.047) | 0.001 | 3.401 (1.359–8.516) | 0.009 |
Furosemide dose (ref: low dose) | 3.171 (1.566–6.419) | 0.001 | 1.235 (0.446–3.421) | 0.685 |
MRA (ref: yes) | 2.278 (1.091–4.756) | 0.028 | 4.171 (1.467–11.862) | 0.007 |
LVEF | 0.957 (0.920–0.996) | 0.032 | 0.847 (0.928–1.022) | 0.281 |
Potassium | 0.213 (0.112–0.460) | <0.001 | 0.189 (0.079–0.451) | <0.001 |
Calcium | 0.449 (0.299–0.674) | <0.001 | 0.577 (0.302–1.104) | 0.097 |
Magnesium | 0.173 (0.062–0.483) | 0.001 | 0.602 (0.127–2.855) | 0.523 |
Hospitalisation due to cardiac decompensation (ref: none) | 5.750 (2.461–13.433) | <0.001 | 3.826 (1.295–11.302) | 0.015 |
QRS duration | 1.012 (0.966–1.029) | 0.147 | ||
Age | 1.021 (1.000–1.043) | 0.049 | 1.038 (1.007–1.071) | 0.016 |
Gender (ref: male) | 1.044 (0.562–1.939) | 0.891 | 0.971 (0.419–2.246) | 0.944 |
HF aetiology (ref: dilated CMP) | 2.317 (1.144–4.691) | 0.020 | 2.063 (0.849–5.012) | 0.110 |
Thiazide (ref: none) | 1.111 (0.578–2.137) | 0.752 |
R2 value: 0.437, −2 log likelihood: 165.087.
aOR, adjusted odds ratio; CI, confidence ınterval; CMP, cardiomyopathy; HF, heart failure; ICD, implantable cardioverter defibrillator; LVEF, left ventricular ejection fraction; MRA, mineralocorticoid receptor antagonists; OR, odds ratio; ref, reference; SGLT2i, sodium-glucose cotransporter-2 inhibitors.
aSignificant variables from univariate analyses plus gender were included in the model (Enter method was used).
The ROC curve for cardiovascular mortality is shown in Figure 4. The largest area under the ROC curve (0.64) for CV death was obtained with a threshold of 100 mg furosemide, yielding a sensitivity of 50.0% and a specificity of 78.3% (p = 0.045).
Discussion
Our findings showed that the use of high-dose furosemide (>80 mg/day) was associated with an increased risk of ventricular arrhythmias and subsequent ICD shocks in HF patients. However, concomitant use of SGLT2i with high- and low-dose furosemide is associated with reduced ICD shocks.
The management of fluid overload in HF is essential to maintain clinical stability. Because there is no accurate way to measure excess fluid in the body, loop diuretics are often used based on clinical assessment, with doses adjusted until symptoms are relieved. However, if the dose becomes too high, the ongoing neurohormonal activation caused by the overuse of loop diuretics can be harmful [13]. Patients with HFrEF are particularly susceptible to complications associated with high-dose diuretic therapy and are at increased risk of ventricular arrhythmias [12]. There is pathophysiological evidence to support the association between diuretic therapy and the increased incidence of fatal arrhythmias due to electrolyte depletion. Studies have shown that hypokalemia and hypomagnesaemia predispose to ventricular ectopic activity [1, 8], with this relationship being particularly significant in HF. Non-potassium-sparing diuretics decrease serum and total body levels of potassium and magnesium dose-dependent [1, 14], with significant hypokalemia occurring. Indeed, our study showed a negative correlation between furosemide dose and serum potassium and magnesium levels, with hypokalemia emerging as an independent predictor of ICD shock.
The definition of “high-dose” furosemide in the literature supports our study. Some studies have defined a daily dose of furosemide between 80 and 160 mg as high dose in patients with HF [15, 16]. In a study by Kapelios et al. [17], the use of high-dose furosemide (>80 mg/day) independently predicted adverse outcomes such as death and lethal arrhythmias in HF patients during 3 years of follow-up, and the frequency of hypokalemic episodes was higher in patients receiving high-dose furosemide. Similarly, in our study, higher rates of ICD shocks were observed in patient groups using >80 mg/day of furosemide, and significantly lower potassium levels were detected in groups using high-dose furosemide. Given the significant number of patients receiving high-dose furosemide therapy, it is imperative to identify the patient characteristics that may delineate a high-risk group for fatal arrhythmias and ICD shocks and to identify adjunctive therapies that may mitigate this risk. Classical diuretics are known to induce abnormal serum potassium or calcium levels, which can adversely affect cardiac repolarisation and precipitate life-threatening arrhythmias [18]. However, the expanding indications for SGLT2i highlight the broadening therapeutic horizon of these drugs. They have shown efficacy in reducing the risk of cardiovascular death and hospitalisation not only in patients with HFrEF but also in those with preserved LVEF [3]. Unlike classical diuretics, which can lower serum potassium levels and potentially cause arrhythmias, SGLT2i have a neutral effect on potassium levels [19]. In addition, SGLT2i have been shown not to prolong the QT interval during cardiac repolarisation [20]. Our study supports these findings, showing a significant reduction in ICD shocks in patient groups using SGLT2i compared to those using low-dose or high-dose furosemide alone.
In our study, multivariate regression analysis identified the absence of SGLT2i treatment as an independent predictor of ICD shocks, and this effect persisted after adjustment for baseline factors. This finding contrasts with a previous study of dapagliflozin, which found no significant difference in ICD shocks during 6 months of follow-up after initiation of dapagliflozin treatment [21]. Notably, this study did not provide information on diuretic use between the groups with and without ICD shocks, nor did it find a significant difference in potassium levels between the two groups [21].
The components of GDMT, including β-blockers and neurohormonal blockers targeting the renin-angiotensin-aldosterone pathway, have been well documented to reduce mortality related to pump failure, as well as the incidence of ventricular arrhythmias and sudden cardiac death [3, 22‒26]. A meta-analysis of 11,302 patients with symptomatic HFrEF showed that those treated with MRAs had a 23% lower risk of sudden cardiac death compared with placebo [27]. In our study, one of the independent predictors of ICD shock was the absence of MRA on treatment. Moreover, when we analysed ICD shocks and other clinical characteristics across the four groups divided by furosemide dose and SGLT2i use, we found no significant differences in the use of triple GDMT (p = 0.246). This lack of difference allowed us to more clearly evaluate the effects of furosemide dose and SGLT2i use on ICD shocks.
Cardiac decompensation triggers ventricular arrhythmias. In our study, one of the independent predictors of ICD shocks during the 12-month follow-up period was hospitalisation for cardiac decompensation. Similarly, a study by Lampropoulou et al. [28] investigating the triggers for appropriate ICD shocks analysed 710 patients from a prospective ICD registry and found that the most common trigger was cardiac decompensation, accounting for 39% of cases.
Although we analysed the effect of furosemide dose and SGLT2i use on ICD shocks, CV-related deaths during follow-up were significantly higher in the high-dose furosemide group without SGLT2i. There is a large body of literature suggesting that SGLT2i reduce CV mortality in HF regardless of diabetes status and have been included in GDMT in current guidelines [3, 26].
We also performed ROC curve analysis to examine the association between furosemide dose and CV mortality. Our results showed that a daily dose of 100 mg was the most accurate predictor of CV death, which we defined as high-dose furosemide. Patients who require higher doses of diuretics to maintain stability tend to have higher levels of congestion and mortality [29]. While loop diuretics can effectively relieve persistent congestion in patients with signs and symptoms of decompensated HF [26], they are often continued at high doses even after congestion resolves. This may result from inadequate regular assessments or difficulty in detecting excessive fluid volumes. Consequently, overuse of loop diuretics may lead to adverse outcomes, such as increased mortality from ventricular arrhythmias or other causes.
Limitations
Despite its comprehensive approach, this study has several limitations. The retrospective design and the relatively small sample size limit the generalisability of the results. The inclusion of patients with specific characteristics, such as the use of furosemide, may also affect the applicability of the results. Variability in ICD programming and treatment algorithms between centres and clinicians may also influence the results. A multicentre study design would have added value by taking these variations into account.
Conclusion
SGLT2i was associated with reduced ventricular arrhythmias and ICD shocks in HFrEF patients receiving low-dose and high-dose furosemide. The absence of SGLT2i in HF therapy emerged as an independent predictor of ICD shocks. There is limited information in the literature on the preventive effect of SGLT2i against potential arrhythmic side effects of other drugs used to treat HF. Future research could strengthen our findings by investigating the effect of SGLT2i on arrhythmogenic substrates.
Statement of Ethics
The study was designed by the principles of the Declaration of Helsinki and the principles of Good Clinical Practice and did not violate the ethical rules of research involving human subjects. All participants provided their informed consent to participate in the study. Approval for the study was obtained from the Bioethics Committee of Zonguldak Bulent Ecevit University (No. 2024/09, Date: May 08, 2024). All participants gave their consent to participate. All authors consent to publication.
Conflict of Interest Statement
Authors declare no conflict of interest.
Funding Sources
The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.
Author Contributions
Ilke Erbay and Ahmet Avci: study design; Ilke Erbay, Naile Eris Gudul, Ugur Kokturk, and Meltem Kandazoglu: data collection; Ilke Erbay and Pelin Aladag: statistical analysis; Ilke Erbay, Naile Eris Gudul, Ugur Kokturk, and Meltem Kandazoglu: data interpretation; Ilke Erbay, Naile Eris Gudul, Ugur Kokturk, and Pelin Aladag: manuscript preparation; and Ilke Erbay and Ahmet Avci: literature search.
Data Availability Statement
The datasets generated during and/or analysed during the current study are available from the corresponding author upon reasonable request.