Holter-detected arrhythmias and associated risk factors in hospitalized elderly patients with diagnosed heart failure

Abstract

Objective

This study aimed to investigate the prevalence, characteristics, and associated clinical factors of arrhythmias—particularly atrial fibrillation (AF) and high-grade premature ventricular contractions (PVCs)—as detected by 24-hour Holter ECG monitoring in hospitalized elderly patients with diagnosed heart failure with ejection fraction ≤ 50%.

Methods

A cross-sectional study was conducted between May 2023 to March 2024 at the Cardiovascular Institute, 108 Central Military Hospital, Hanoi, Vietnam. A total of 72 hospitalized patients aged ≥ 65 years with confirmed heart failure (LVEF < 50%) were enrolled. Clinical characteristics, laboratory data, echocardiographic parameters, and 24-hour Holter monitoring results were collected. Multivariable logistic regression with stepwise forward selection (p-entry < 0.2) was used to identify predictors of AF and high-grade PVCs (Lown grade III–V).

Results

The mean age was 73.7 ± 8.9 years, and 66.7% were male. AF was detected in 19.4% of patients, while a high PVC burden was common, with a median (25th -75th ) of 1955 (54-38514) and 58.3% falling into Lown grades III–V. Multivariable analysis identified increased body mass index (OR = 1.71, 95% CI: 1.18–2.48, p = 0.005) and elevated systolic pulmonary artery pressure (sPAP) (OR = 1.06, 95% CI: 1.00–1.13, p = 0.049) as independent predictors of AF. For high-grade PVCs, significant predictors included elevated sPAP (OR = 1.06, 95% CI: 1.01–1.12, p = 0.024), lower creatinine (OR = 0.99, 95% CI: 0.98–1.00, p = 0.014).

Conclusion

AF and high-grade PVCs are common in hospitalized elderly patients with diagnosed heart failure and are associated with distinct clinical and laboratory parameters. Elevated sPAP was a shared predictor of both arrhythmias. These findings support the use of integrated echocardiographic and Holter monitoring for risk stratification and management in this population.

Introduction

Heart failure with reduced ejection fraction (HFrEF) is increasingly prevalent in the elderly and is frequently accompanied by cardiac arrhythmias, which contribute to adverse outcomes including stroke, hospitalization, and sudden cardiac death [1]. Among these, atrial fibrillation (AF) and premature ventricular contractions (PVCs) are especially common and has been closely associated with heightened risks of stroke, hospitalization, and sudden cardiac death (SCD) [2]. Early detection and characterization of these arrhythmias are essential for optimizing risk stratification and management in this vulnerable population [3].

Although standard 12-lead electrocardiograms (ECG) are routinely used, they often miss intermittent arrhythmias [4]. In contrast, 24-hour Holter ECG monitoring offers continuous rhythm assessment, enhancing the detection of asymptomatic or paroxysmal events, such as AF and high-grade PVCs [5]. Several studies have shown that Holter monitoring significantly improves arrhythmia detection in patients with HFrEF compared to standard ECG [6]. Furthermore, it proves valuable in capturing transient arrhythmias such as paroxysmal atrial fibrillation (PAF), which are often implicated in cryptogenic strokes and are undetectable through shorter-term monitoring [7].

Despite these advantages, the prevalence and risk factors of arrhythmias detected by Holter monitoring remain underexplored in elderly patients with newly diagnosed HFrEF—a subgroup characterized by advanced age, multiple comorbidities, and unique pathophysiological features. Understanding the arrhythmic profile in this group may inform early interventions and prevent complications [8, 9]. 

Recognizing arrhythmic patterns via Holter monitoring has substantial implications for clinical management. It guides pharmacological therapy adjustments, such as initiating anticoagulation for AF or optimizing beta-blocker use in response to ventricular ectopy [10]. Moreover, the identification of high-risk ventricular arrhythmias may warrant the use of device-based therapies like implantable cardioverter-defibrillators (ICDs) or cardiac resynchronization therapy (CRT), which have demonstrated survival benefits in appropriately selected HFrEF patients [11, 12].

This study aimed to investigate the prevalence, types, and associated clinical factors of arrhythmias, particularly AF and high-grade PVCs, as detected by 24-hour Holter ECG monitoring in hospitalized elderly patients with newly diagnosed HFrEF.

Materials and methods

Study design and population

This was a cross-sectional study conducted at the Cardiovascular Institute, 108 Central Military Hospital, Hanoi, Vietnam, from May 2023 to March 2024. The study included 72 elderly patients aged ≥ 65 years with a newly confirmed diagnosis of heart failure with reduced ejection fraction (HFrEF), defined as a left ventricular ejection fraction (LVEF) < 40%, in accordance with the latest AHA/ACC/HFSA 2022 guideline [13]. Patients with LVEF 41–49% were classified as having heart failure with mildly reduced ejection fraction (HFmrEF), but were included in the analysis for exploratory purposes. A formal sample size calculation was not performed prior to study initiation due to the exploratory and descriptive nature of this cross-sectional design [9]. Instead, all eligible hospitalized patients meeting the inclusion criteria during the study period were consecutively recruited to maximize representativeness. Exclusion criteria included patients with prior implantable cardioverter-defibrillators (ICDs), cardiac resynchronization therapy (CRT), or those with advanced-stage malignancies to ensure a homogenous study population. All patients provided written informed consent prior to enrollment.

Data collection procedure

Eligible participants were consecutively recruited during their hospitalization. Clinical and demographic data were extracted from medical records and structured interviews, including age, sex, comorbidities (e.g., hypertension, diabetes, coronary artery disease), smoking and alcohol use, and presenting symptoms. All patients were in sinus rhythm at the time of Holter initiation, as confirmed by 12-lead ECG. Functional status was assessed using the New York Heart Association (NYHA) classification. Blood pressure, weight, height, and other anthropometric measurements were obtained on admission. All patients underwent transthoracic echocardiography and 24-hour Holter ECG monitoring as part of standard clinical evaluation during hospitalization.

Data measurements

All data was collected when patient stablized and before hospital’s discharge. Echocardiographic measurements were performed using standard protocols, including assessments of left atrial diameter (LA), aortic root diameter (A0), left ventricular end-diastolic (Dd) and end-systolic diameters (Ds), ejection fraction (EF), and pulmonary artery systolic pressure (sPAP). Arrhythmias such as atrial fibrillation (AF) and ventricular arrhythmias were identified by cardiologists using 12-lead ECG and/or confirmed by 24-hour Holter monitoring. Holter ECG was recorded according to our hospital protocol using Digitrak XT device (Philips Electronics, USA). Holter parameters included mean heart rate, number of supraventricular episodes, premature ventricular complexes (PVCs), QT interval, corrected QT interval (QTc), RR interval, and QRS duration. PVCs were categorized using the Lown classification. Laboratory measurements included glucose, creatinine, cholesterol, LDL-C, HDL-C, triglycerides, and BNP, collected from fasting venous blood samples.

Statistical analysis

Descriptive statistics were used to summarize patient characteristics, with categorical variables presented as frequencies and percentages, and continuous variables reported as means and standard deviations (SD). Group comparisons were performed using chi-square tests for categorical variables and t-tests or ANOVA for continuous variables, as appropriate. Multivariable logistic regression analyses with stepwise forward selection (entry p < 0.2) were conducted to identify independent predictors of atrial fibrillation and high-grade PVCs (Lown grade III–V). Results were expressed as odds ratios (ORs) with 95% confidence intervals (CIs). A two-sided p-value < 0.05 was considered statistically significant. All analyses were performed using Stata version 16.0 (StataCorp, College Station, TX, USA).

Results

Table 1 summarizes the background characteristics of the 72 patients included in the study. The mean age was 73.7 years (SD = 8.9), and the average BMI was 22.3 kg/m² (SD = 3.3). Males accounted for 66.7% of the sample. The most prevalent comorbidity was hypertension (61.1%), followed by diabetes (30.6%), coronary artery disease (18.1%), stroke (9.7%), cardiomyopathy (9.7%), gout (6.9%), dyslipidemia (5.6%), and valvular heart disease (2.8%). Regarding health behaviors, 11.1% of patients reported smoking, and only 1.4% reported alcohol use. Dyspnea was the most common reason for hospital admission (76.4%), followed by fatigue (37.5%), chest pain (23.6%), edema (13.9%), and syncope (5.6%). In terms of NYHA functional class, most patients were in class III (41.7%), with fewer in class II (23.6%), class I (19.4%), and class IV (15.3%). Mean values for key clinical parameters included systolic and diastolic blood pressure (128.1 ± 20.4 mmHg and 77.8 ± 12.5 mmHg, respectively), glucose (7.1 ± 3.1 mmol/L), creatinine (122.9 ± 59.9 µmol/L), cholesterol (4.8 ± 1.2 mmol/L), triglycerides (2.1 ± 0.9 mmol/L), HDL-C (1.1 ± 0.4 mmol/L), LDL-C (2.9 ± 1.1 mmol/L), BNP (8161.1 ± 11006.0 pg/mL). The most prescribed medication for heart failure is SGLT2i (72.2%) followed by ACEi/ARB/ARNI group (59.7%) and betablockers used in 33.3% population.

Table 1.

Clinical characteristics of patients

Variable	Category	Value
Sex, Male, n(%)		48 (66.7%)
Comorbidities, n(%)	Hypertension	44 (61.1%)
	Stroke	7 (9.7%)
	Coronary artery dis.	13 (18.1%)
	Cardiomyopathy	7 (9.7%)
	Valvular	2 (2.8%)
	Gout	5 (6.9%)
	Diabetes	22 (30.6%)
	Dyslipidemia	4 (5.6%)
Behaviors, n(%)	Smoking	8 (11.1%)
	Alcohol use	1 (1.4%)
Reasons for admission, n(%)	Dyspnea	55 (76.4%)
	Fatigue	27 (37.5%)
	Edema	10 (13.9%)
	Chest pain	17 (23.6%)
	Syncope	4 (5.6%)
NYHA, n(%)	I	14 (19.4%)
	II	17 (23.6%)
	III	30 (41.7%)
	IV	11 (15.3%)
Age (years), Mean (SD)		73.7 (8.9)
BMI (kg/m2), Mean (SD)		22.3 (3.3)
Systolic blood pressure (mmHg), Mean (SD)		128.1 (20.4)
Diastolic blood pressure (mmHg), Mean (SD)		77.8 (12.5)
Glucose (mmol/L), Mean (SD)		7.1 (3.1)
Creatinine (µmol/L), Mean (SD)		122.9 (59.9)
Cholesterol (mmol/L), Mean (SD)		4.8 (1.2)
Triglyceride (mmol/L), Mean (SD)		2.1 (0.9)
HDL-C (mmol/L), Mean (SD)		1.1 (0.4)
LDL-C (mmol/L), Mean (SD)		2.9 (1.1)
BNP (pg/mL), Mean (SD)		8161.1 (11006.0)
Diuretic, n(%)		31(43.1)
SGLT2i, n(%)		52(72.2)
ACEi/ARB/ARNI, n(%)		43(59.7)
Spironolacton, n(%)		19(26.3)
Beta blockers, n(%)		24(33.3)

Abbrev: HDL-C high-density lipoprotein cholesterol, LDL-C low-density lipoprotein cholesterol, BNP brain natriuretic peptide, SGLT2i sodium glucose cotransporter 2 inhibitor, ACEi angiotensin converting enzyme inhibitor, ARB angiotensin receptor blocker, ARNI angiotensin receptor neprilysin inhibitor

Table 2 presents the echocardiographic and arrhythmia characteristics of the 72 patients included in the study. The mean left atrial diameter was 38.9 mm (SD = 6.6), aortic root diameter was 30.2 mm (SD = 4.0), and left ventricular end-diastolic and end-systolic diameters were 54.7 mm (SD = 10.6) and 44.8 mm (SD = 10.3), respectively. The mean ejection fraction was reduced at 33.3% (SD = 10.8), and the mean pulmonary artery systolic pressure was 35.8 mmHg (SD = 12.0). Regional wall motion abnormalities were observed in 45 patients (62.5%), while pericardial and pleural effusion were present in 4 (5.6%) and 15 (20.8%) patients, respectively. Atrial fibrillation was detected in 14 patients (19.4%). Notably, no cases of non-sustained ventricular tachycardia (NSVT) or supraventricular tachycardia (SVT) were recorded in this cohort.

Table 2.

Echocardiographic and arrhythmia characteristics of patients

Parameter	Mean (SD)
Left atrial diameter (LA, mm)	38.9 (6.6)
Aortic root diameter (A0, mm)	30.2 (4.0)
LV end-diastolic diameter (Dd, mm)	54.7 (10.6)
LV end-systolic diameter (Ds, mm)	44.8 (10.3)
Ejection fraction (EF, %)	33.3 (10.8)
Pulmonary artery systolic pressure (mmHg)	35.8 (12.0)
	n (%)
Regional wall motion abnormality	45 (62.5%)
Pericardial effusion	4 (5.6%)
Pleural effusion	15 (20.8%)
Atrial fibrillation (AF)	14 (19.4%)
Non-sustained ventricular tachycardia (NSVT)	0 (0.0%)
Supraventricular tachycardia (SVT)	0 (0.0%)

Abbrev: LA left atrium, A0 aortic root, Dd end-diastolic diameter, Ds end-systolic diameter, EF ejection fraction, ALĐMP pulmonary artery systolic pressure, NSVT non-sustained ventricular tachycardia, SVT supraventricular tachycardia

Table 3 summarizes key Holter monitoring parameters among 72 patients. The mean Holter-derived average heart rate was 81.96 bpm, with a mean ventricular rate of 91.63 bpm. The average number of supraventricular episodes recorded was 862.65, while premature ventricular complexes (PVCs) were notably frequent, with a median (25th-75th) of 1955 (54-38514). The minimum and maximum heart rates observed were 55.46 bpm and 133.39 bpm, respectively. QT and corrected QT (QTc) intervals averaged 392.47 ms and 476.41 ms, with a mean RR interval of 696.64 ms and QRS duration of 99.63 ms. Regarding Lown classification of PVCs, the majority of patients fell into higher-risk categories: Grade IVa (38.8%) and IVb (18.1%), while only 27.8% were in Grade I.

Table 3.

Summary of holter-related characteristics (n = 72)

Group	Mean (SD)
Holter average heart rate (bpm)	81.96 (14.67)
Ventricular rate (bpm)	91.63 (23.06)
No. of supraventricular episodes	862.65 (2114.14)
No. of PVCs, mean (25th −75th)	1955 (54-38514)
Min heart rate (bpm)	55.46 (12.09)
Max heart rate (bpm)	133.39 (31.08)
QT interval (ms)	392.47 (46.76)
RR interval (ms)	696.64 (175.85)
QTc interval (ms)	476.41 (47.77)
QRS duration (ms)	99.63 (23.74)
Lown grade, n(%)
I	20 (27.8%)
II	9 (12.5%)
III	1 (1.4%)
IVa	28 (38.8%)
IVb	13 (18.1%)
V	1 (1.4%)

Abbrev: bpm  beats per minute, ms milliseconds, QTc corrected QT interval, PVC premature ventricular complex

Table 4 presents the results of multivariable logistic regression models with stepwise forward selection strategy to assess clinical and laboratory factors associated with atrial fibrillation (AF) and frequent premature ventricular contractions (PVCs) classified as Lown grade III– V. In the AF model (n = 53, LR χ²(5) = 14.76, p = 0.0115, Pseudo R² = 0.2411), increased body mass index (OR = 1.71, 95% CI: 1.18–2.48, p = 0.005) and higher systolic pulmonary artery pressure (sPAP) (OR = 1.06, 95% CI: 1.00–1.13, p = 0.049) were statistically significant. For Lown grade III–V PVCs (n = 72, LR χ²(7) = 22.99, p = 0.0017, Pseudo R² = 0.2368), elevated sPAP (OR = 1.06, 95% CI: 1.01–1.12, p = 0.024), lower creatinine (OR = 0.99, 95% CI: 0.98–1.00, p = 0.014), and higher sodium levels (OR = 1.24, 95% CI: 1.04–1.47, p = 0.014) were significant predictors.

Table 4.

Stepwise multivariable logistic regression analysis of factors associated with atrial fibrillation and high-grade pvcs (Lown grade III–V)

Variables	Atrial Fibrillation		PVC Lown Grade III–V
	OR	95%CI	OR	95%CI
BMI (per 1 kg/m²)	1.71*	1.18–2.48	–	–
Systolic Pulmonary Artery Pressure (per 1 mmHg)	1.06*	1.00–1.13	1.06*	1.01–1.12
Age (per year)	1.12	1.00–1.27	1.06	0.99–1.15
LDL-C (per 1 mmol/L)	1.74	0.84–3.60	–	–
Glucose (per 1 mmol/L)	1.21	0.92–1.60	1.30	0.99–1.70
Creatinine (per 1 µmol/L)	–	–	0.99*	0.98–1.00
Hypertension (presence vs. absence)	–	–	3.33	0.96–11.58

*p < 0.05

Abbrev: OR  Odds Ratio, CI  Confidence Interval, BMI Body Mass Index, LDL-C  Low-Density Lipoprotein Cholesterol, PVC Premature Ventricular Contraction,sPAP Systolic Pulmonary Artery Pressure

Discussion

This study provides important insights into the clinical characteristics, arrhythmic burden, and associated risk factors among hospitalized elderly patients with newly diagnosed heart failure, using a combination of clinical, echocardiographic, and Holter monitoring data. Notably, both AF and PVCs, classified by Lown grade III–V, were relatively prevalent and linked to distinct clinical and laboratory parameters. The study used a threshold of LVEF <50% to define systolic dysfunction, which does not fully align with the latest guideline classifications. According to AHA/ACC/HFSA 2022, only LVEF <40% qualifies as HFrEF, while values from 41–49% fall under HFmrEF. This discrepancy should be considered when interpreting the findings.

In our study, echocardiographic evaluation revealed significant cardiac structural abnormalities among elderly patients with HF, including a reduced left ventricular ejection fraction (mean EF = 33.3%) and a high prevalence of regional wall motion abnormalities (62.5%). These findings are consistent with previous literature, where the majority of elderly HF patients were classified as having HFrEF. For instance, Mani Teja K et al. (2022) reported that 79.2% of elderly HF patients had LVEF < 50% [14]. The mean left atrial (LA) diameter in our cohort was 38.9 mm, suggesting LA enlargement, which is recognized as a marker of atrial remodeling and a risk factor for AF [15, 16]. The prevalence of AF in our sample was 19.4%, which is slightly lower than the 22–44% range reported in large HF registries [17, 18], but still clinically relevant. The observed elevated pulmonary artery systolic pressure (mean = 35.8 mmHg) may further contribute to the atrial pressure overload and arrhythmogenesis. Notably, no cases of non-sustained ventricular tachycardia (NSVT) or supraventricular tachycardia (SVT) were recorded, which contrasts with a prior study suggesting higher rates of complex arrhythmias in HF patients [19]. However, this discrepancy may be attributed to variations in patient selection, HF severity, or detection methods. Overall, our echocardiographic findings support the importance of comprehensive cardiac imaging in the elderly HF population, as it provides crucial information on structural remodeling, hemodynamic burden, and arrhythmic risk.

Holter monitoring revealed substantial arrhythmic burden in this cohort, with a high number of PVCs (median (25^th−75^th)= 1955 (54-38514)) and a majority of patients falling into high-risk Lown grades IVa (38.8%) and IVb (18.1%). Although the Lown grading system was originally developed for risk stratification in patients with ischemic heart disease [20], it has also been employed in broader heart failure populations in contemporary studies [21]. In our cohort, only 18.1% of patients had documented coronary artery disease; thus, applying the Lown classification to patients with non-ischemic etiologies may have limitations. However, it still provides a practical framework for characterizing the burden of ventricular ectopy and has been shown to correlate with prognosis in various heart failure subtypes. This adaptation should be interpreted with caution, and further validation in non-ischemic cohorts is warranted. These findings align with prior reports that Holter monitoring offers superior sensitivity in detecting transient arrhythmias compared to surface ECG [6]. In that study, Holter monitoring identified AF in nearly 25% of elderly HF patients—higher than typical detection rates via standard ECG—supporting its role in comprehensive rhythm surveillance. Our data also reinforce the association between HF and a high burden of ventricular arrhythmias. The mean QTc interval in our sample was prolonged at 476.41 ms, and QRS durations were widened, both of which are known markers of increased sudden cardiac death risk [22, 23]. These results highlight the role of Holter monitoring not only for arrhythmia detection but also for risk stratification in elderly HF patients.

Multivariable regression analysis identified several clinical factors independently associated with AF and high-grade PVCs. For AF, increased BMI and elevated systolic pulmonary artery pressure (sPAP) emerged as significant predictors. These findings are supported by prior studies indicating that obesity contributes to atrial structural remodeling and increased atrial pressure, facilitating AF development [24, 25]. Elevated sPAP, reflecting pulmonary hypertension, has also been linked to atrial pressure overload and AF incidence in HF patients [26]. Although age, LDL-C, and glucose levels showed trends toward significance, they did not reach statistical thresholds in our cohort. For high-grade PVCs (Lown grade III–V), elevated sPAP, lower creatinine, and higher sodium levels were significant predictors. The association between sPAP and PVCs likely reflects the underlying hemodynamic stress and myocardial strain in HF, while the inverse relationship with creatinine was unexpected and warrants further investigation—possibly reflecting confounding or survivor bias. The positive association between sodium and PVCs aligns with literature suggesting that electrolyte imbalances may influence ventricular excitability. Overall, these findings emphasize the multifactorial nature of arrhythmogenesis in HF and the need for individualized risk assessment [27–29].

The findings of this study have important clinical implications for the management of elderly patients with heart failure. The high prevalence of arrhythmias—particularly atrial fibrillation and high-grade premature ventricular contractions—underlines the necessity of routine rhythm surveillance, especially using extended monitoring such as Holter ECG. Identifying independent predictors such as increased BMI, elevated systolic pulmonary artery pressure, and altered electrolyte levels enables clinicians to better stratify arrhythmic risk and tailor monitoring and treatment strategies accordingly. These results also support integrating echocardiographic and laboratory parameters into comprehensive risk models to improve early detection and prevention of adverse cardiac events.

Limitations

This study has several limitations that warrant consideration. Fistly, the sample size was relatively small (n = 72), which may have limited the statistical power to detect certain associations and may reduce the generalizability of findings. Second, the cross-sectional design precludes any conclusions about causality or temporal relationships between risk factors and arrhythmia development. Third, the absence of NSVT and SVT cases in this cohort may reflect selection bias or underdetection, potentially affecting the completeness of arrhythmic profiling. Additionally, the study population was drawn from a single center, and its demographic and clinical characteristics may not reflect those of broader or more diverse populations. Another important limitation is the duration of Holter monitoring. Although 24-hour Holter ECG is a widely used tool for arrhythmia detection, its sensitivity remains limited—especially for paroxysmal or intermittent arrhythmias such as atrial fibrillation. Recent evidence suggests that prolonged monitoring over 7 to 14 days significantly increases the detection rate of clinically relevant arrhythmias. Therefore, our reliance on 24-hour Holter monitoring may have underestimated the true arrhythmic burden in this population and restricts the generalizability and strength of the findings. Future studies employing extended monitoring durations are warranted to more comprehensively capture arrhythmic events in elderly heart failure patients.

Conclusion

In conclusion, this study highlights that higher sPAP is a shared predictor of AF and high-grade PVCs in newly diagnosed heart failure patients. Other contributing factors include BMI for AF and creatinine for PVCs. These findings emphasize the value of integrated clinical and diagnostic assessments to identify patients at elevated arrhythmic risk and guide tailored management strategies.

Authors’ contributions

Do Chien provided idea, wrote manuscript, prepared figures Luyen Kien collected data and literature searchAll authors reviewd the manuscript.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

The study protocol was reviewed and approved by the Institutional Review Board of 108 Central Military Hospital (Approval Number: 2150/CN-HDDD). All participants provided written informed consent prior to participation. The study was conducted in accordance with the principles of the Declaration of Helsinki. 

Competing interests

The authors declare no competing interests.