Abnormal electrocardiogram and poor prognosis in heart failure with preserved ejection fraction

Abstract

Purpose

Electrocardiogram (ECG) is generally performed in patients with heart failure with preserved ejection fraction (HFpEF), but the prognostic value of abnormal ECG is not fully understood. We aim to explore the prognostic value of abnormal ECG at baseline in HFpEF using data from the TOPCAT trial.

Methods

A total of 1736 patients from TOPCAT-Americas were included and divided into normal versus abnormal ECG groups. Survival analyses were performed for the following outcomes: the primary endpoint [a composite of cardiovascular death, heart failure (HF) hospitalization, and aborted cardiac arrest], all-cause death, cardiovascular death, and HF hospitalization.

Results

Abnormal ECG was significantly associated with higher risks of the primary endpoint [hazard ratio (HR): 1.480, P = 0.001] and HF hospitalization (HR: 1.400, P = 0.015), and borderline significantly with cardiovascular death (HR: 1.453, P = 0.052) in patients with HFpEF after multivariate adjustment. As for specific ECG abnormalities, bundle branch block was associated with the primary endpoint (HR: 1.278, P = 0.020) and HF hospitalization (HR: 1.333, P = 0.016), whereas atrial fibrillation/flutter was associated with all-cause death (HR: 1.345, P = 0.051) and cardiovascular death (HR: 1.570, P = 0.023), but ventricular paced rhythm, pathological Q waves, and left ventricular hypertrophy were not of prognostic significance. Besides, other unspecific abnormalities together were associated with the primary endpoint (HR: 1.213, P = 0.032).

Conclusion

Abnormal ECG at baseline could be associated with poor prognosis in patients with HFpEF. Physicians are encouraged to pay more attention to HFpEF patients who present an abnormal ECG instead of ignoring those obscure abnormalities.

Key messages

What is already known on this topic

Electrocardiogram (ECG) is a basic and easily accessible examination for patients with heart failure with preserved ejection fraction (HFpEF). Some findings from ECG such as frontal QRS-T angle, QTc interval, and the Cornell product have been shown to be associated with the prognosis of HFpEF but these results are from studies with relatively small sample sizes.

What this study adds

Using data from TOPCAT-Americas, this study found that an overall estimation of abnormal ECG significantly predicted poor prognosis in patients with HFpEF. As for specific abnormalities in ECG, bundle branch block mainly predicted heart failure hospitalization and atrial fibrillation mainly predicted death.

How this study might affect research, practice, or policy

This study reminds physicians to pay more attention to HFpEF patients who present an abnormal ECG.

Introduction

An electrocardiogram (ECG) assessment is essential for heart failure (HF) diagnosis. A normal ECG makes the diagnosis of HF unlikely, whereas an abnormal ECG increases the likelihood of a diagnosis of HF and also may guide therapy [1]. In patients with an established HF diagnosis, abnormality of ECG also has a prognostic significance. In patients with acute HF (AHF), ECG findings of prolonged QRS and junctional rhythm independently predict both in-hospital and long-term mortality [2], and ST-depression predicts 30-day mortality [3]. In chronic systolic HF patients, low ECG voltage predicts disease severity as well as 1-year adverse outcomes [4].

HF with preserved ejection fraction (HFpEF) has been regarded as a distinct category of HF and attracting more and more attention in recent years, because of its high mortality and hospitalization burden and the lack of effective treatment [5]. Therefore, risk prediction is important in HFpEF to identify patients at high risk of poor prognosis. Some ECG parameters are of prognostic significance in HFpEF, including frontal QRS-T angle [6], QTc interval [7], ST-segment and T wave polarity changes in lead aVR [8], and the Cornell product [(R_aVL + S_V3) × QRS duration] [9]. However, the numbers of patients in these studies are relatively small, and the prognostic value of other ECG abnormalities, as well as the overall evaluation of ECG (normal vs. abnormal), have not been explored in HFpEF.

In the present study, we aim to explore the prognostic value of abnormal ECG and its specific abnormalities in HFpEF patients using data from the TOPCAT trial.

Materials and methods

Study population

The Treatment of Preserved Cardiac Function Heart Failure with an Aldosterone Antagonist Trial (TOPCAT) was a phase III, international, multicenter, randomized, double-blind, placebo-control trial to explore the efficacy of spironolactone among patients with HFpEF. The detailed rationale and design and main results of the TOPCAT trial have been published previously [10, 11]. Briefly, symptomatic HF patients aged ≥50 years with a life expectancy of ≥3 years were recruited if they had left ventricular ejection fraction (LVEF) ≥45% and documented evidence of either elevated natriuretic peptide level within 60 days or hospitalization for HF within 12 months before randomization. Patients with uncontrolled systolic blood pressure, renal failure, or serum potassium level ≥ 5.0 mmol/L were excluded. Ethical approval was obtained from local institutional review boards, and all patients provided written consent forms. In the present study, only patients from the Americas (n = 1767) were included due to concerns about data validation from Russia and Georgia [12]. Besides, 25 patients without ECG data, 1 patient without follow-up data, 1 patient without baseline blood pressure, 3 patients without New York Heart Association (NYHA) class, and 1 patient without a medication record were excluded, leaving 1736 patients for analysis. The present study was approved by the Medical Ethics Commission of the First Affiliated Hospital of Sun Yat-sen University.

Electrocardiogram assessment

Each subject’s most recent ECG was obtained at baseline screening and sent to the ECG Core Lab at Brigham and Women’s Hospital for interpretation. An overall evaluation of the ECG (normal or abnormal) was documented. If an ECG was judged to be abnormal, the specific abnormality was further determined: bundle branch block (BBB), ventricular paced rhythm, pathological Q waves, left ventricular hypertrophy, atrial fibrillation/flutter (AF), or other.

Outcomes of interest

The primary endpoint was a composite of cardiovascular death, HF hospitalization, and aborted cardiac arrest. Other outcomes of interest were all-cause death, cardiovascular death, and HF hospitalization.

Statistical analysis

Continuous variables were presented as mean ± standard deviation (SD), and between-group differences were tested by Student’s t-test. Categorical variables were presented as number and percentage, and between-group differences were tested by the chi-square test or Fisher’s exact test, as appropriate. Missing data in all variables did not exceed 5%, and no imputation was performed. Kaplan–Meier curves were plotted to present the survival differences in outcomes of interest between patients with normal and abnormal ECG as well as specific ECG abnormalities intuitively, and the log-rank test was used to test the significance of differences. Survival analyses were performed using the Cox proportional hazards model; results were presented as hazard ratio (HR) and 95% confidence interval (CI). Variables included for multivariate adjustment were age, gender, treatment arm (i.e. with spironolactone treatment or not), body mass index, systolic blood pressure, NYHA class, LVEF, history of myocardial infarction, AF, and diabetes mellitus, which were selected based on their significant impact on outcomes and between-group baseline differences. C-statistics of multivariate models were reported, and integrated discrimination improvement (IDI) was calculated at 5-year follow-up between the model with the independent variable and confounders for adjustment versus the model with confounders for adjustment only. Statistical analyses were performed using R 4.2.0 (packages survminer, survival, survIDINRI). A two-tailed P < .05 was considered to be statistically significant.

Results

Baseline characteristics

Among 1736 patients included in the present study, only 416 of them had normal ECG at baseline, while 1320 of them had abnormal ECG. As shown in Table 1, compared with patients with normal ECG, those with abnormal ECG were older (72.20 vs. 69.48 years old, P < .001) and more often male (53.1% vs. 39.4%, P < .001). As for medical history, a higher proportion of patients with abnormal ECG had a myocardial infarction (21.6% vs. 16.3%, P = .025) and an AF history (49.9% vs. 18.0%, P < .001), but a lower proportion of them had diabetes mellitus history (43.1% vs. 50.7%, P = .008). Besides, patients with abnormal ECG had significantly lower systolic blood pressure (126.69 vs. 129.89 mm Hg, P < .001) and body mass index (33.55 vs. 34.70 kg/m², P = .012), as well as LVEF (57.74% vs. 59.59%, P < .001) and serum potassium (4.17 vs. 4.22 mmol/L, P = .044), but a higher proportion of them was in NYHA class III or IV (37.7% vs. 27.9%, P < .001) and under beta-blockers (80.5% vs. 73.3%, P = .002) and diuretics (90.5% vs. 85.3%, P = .004) treatment. Of note, baseline HF symptoms (dyspnea on exertion, dyspnea at rest, orthopnea, rales, and edema) were comparable between groups.

Abnormal electrocardiogram and outcomes

During a median follow-up of 2.93 years, all of the incidence rates of the outcomes of interest were higher in patients with abnormal ECG at baseline (Table 2). Kaplan–Meier curves illustrated that patients with abnormal ECG had significantly higher risks of the primary endpoint (P < .001), all-cause death (P = .001), cardiovascular death (P = .003), and HF hospitalization (P = .001) (Fig. 1). Similar results were found in unadjusted survival analyses (Table 2). However, after adjusting for potential confounders, abnormal ECG was not associated with the risk of all-cause death (HR: 1.248, 95% CI: 0.945–1.649, P = .118) and associated with cardiovascular death with a borderline significance (HR: 1.453, 95% CI: 0.997–2.118, P = .052), but was still significantly associated with higher risks of the primary endpoint (HR: 1.480, 95% CI: 1.165–1.881, P = .001) and HF hospitalization (HR: 1.400, 95% CI: 1.068–1.835, P = .015), with moderate C-statistics (0.636 for primary endpoint, 0.632 for cardiovascular death, and 0.653 for HF hospitalization) (Table 2). In terms of the incremental value of abnormal ECG in predicting risks of outcomes, IDIs showed a good incremental discriminating value of abnormal ECG (8.4% for the primary endpoint, 10.0% for cardiovascular death, and 7.2% for HF hospitalization) (Table 2).

Specific electrocardiogram abnormalities and outcomes

As for specific ECG abnormalities, there were 335 patients with BBB, 162 with ventricular paced rhythm, 104 with pathological Q waves, 167 with left ventricular hypertrophy, 445 with AF, and 755 with other abnormalities (Table 3). Ventricular paced rhythm, pathological Q waves, and left ventricular hypertrophy on ECG were not associated with any outcomes (all P > .05). As shown in Kaplan–Meier curves (Supplemental Figs 1–6) and unadjusted models, BBB was significantly associated with higher risks of the primary endpoint (HR: 1.413, 95% CI: 1.155–1.729, P < .001), all-cause death (HR: 1.373, 95% CI: 1.087–1.735, P = .008), and HF hospitalization (HR: 1.462, 95% CI: 1.163–1.836, P = .001); AF was significantly associated with higher risks of the primary endpoint (HR: 1.213, 95% CI: 1.002–1.468, P = .047), all-cause death (HR: 1.383, 95% CI: 1.113–1.717, P = 0.003), and cardiovascular death (HR: 1.528, 95% CI: 1.156–2.020, P = 0.003); and other abnormalities were significantly associated with higher risks of the primary endpoint (HR: 1.292, 95% CI: 1.087–1.536, P = .004) and HF hospitalization (HR: 1.239, 95% CI: 1.017–1.509, P = .034). After multivariate adjustment, BBB was significantly associated with higher risks of the primary endpoint (HR: 1.278, 95% CI: 1.040–1.571, P = .020) with a C-statistic of 0.634 and IDI of 7.4% and HF hospitalization (HR: 1.314, 95% CI: 1.033–1.671, P = .026) with a C-statistic of 0.653 and IDI of 6.3%; AF was significantly associated with a higher risk of cardiovascular death (HR: 1.570, 95% CI: 1.063–2.319, P = .023) with a C-statistic of 0.633 and IDI of 11.4% and borderline significantly associated with a higher risk of all-cause death (HR: 1.345, 95% CI: 0.999–1.811, P = .051) with a C-statistic of 0.646 and IDI of 7.8%; and other abnormalities were significantly associated with a higher risk of the primary endpoint (HR: 1.213, 95% CI: 1.017–1.446, P = .032) with a C-statistic of 0.636 and IDI of 9.0% (Table 3).

Discussion

Results of this study revealed that an abnormal ECG significantly predicted poor prognosis (primary endpoint, cardiovascular death, and HF hospitalization) in patients with HFpEF after adjusting for potential confounders, with moderate C-statistics and good incremental discriminating value. As for specific ECG abnormalities, BBB mainly predicted HF hospitalization and AF mainly predicted death, whereas ventricular paced rhythm, pathological Q waves, and left ventricular hypertrophy were not of prognostic significance.

ECG has been a examination in cardiology for centuries. In the field of HF, ECG has been recommended in guidelines for the diagnosis and primary evaluation of HF [1, 13]. Additionally, in the newly proposed “HFA–PEFF diagnostic algorithm” for HFpEF, ECG is also recommended to be performed in the first step when screening breathless patients for the diagnosis of HFpEF [14]. As for prognostic value, some specific ECG parameters, such as frontal QRS-T angle, QTc interval, the Cornell product, and QRS duration, have been reported to predict outcomes in patients with HFpEF [6–9, 15]. However, these are just a little part of the information from an ECG. Usually, BBB is unintentionally found in ECG examination but is significantly associated with poor outcomes. In primary care patients, BBB predicts risks of future HF and cardiovascular death [16]. In patients hospitalized with HF, BBB was reported to be associated with all-cause mortality [17], whereas another study found that left BBB was associated with death and HF hospitalization and right BBB did not have an independent prognostic value [18]. There is no study exploring BBB in HFpEF patients so far. Results of the present study showed that BBB in HFpEF patients was mainly associated with HF hospitalization but not death. However, patients in the present study were in a chronic pattern, so whether BBB predicts death in hospitalized HFpEF patients needs further investigation. Besides, the data we acquired did not show specific left or right BBB, so the comparison of left and right BBB could not be done. AF along with HFpEF are regarded as vicious twins because they share similar risk factors and often coexist [19]. The presence of AF was found to have a prognostic value on long-term mortality in AHF patients, even though the prognostic value attenuated after multivariate adjustment [2]. Our study consistently found that AF on ECG predicted death outcomes in HFpEF patients, and the predictive value was still of statistical significance after multivariate adjustment.

Of note, unspecific ECG abnormalities other than BBB, AF, ventricular paced rhythm, pathological Q waves, and left ventricular hypertrophy were also of prognostic significance. Any abnormality on an ECG means an electrophysiological change of the heart, and in the setting of HFpEF, this may implicate a pathophysiological change and could be associated with a worse heart and subsequently poor prognosis. When a physician receives a patient’s ECG, their primary impression of it will be “normal or not.” However, if there is only a “small” or “common” abnormality on an ECG, physicians may ignore it. Our results suggested that any abnormality in an ECG might be associated with poor prognosis in HFpEF patients, so physicians should pay more attention to HFpEF patients if they are presenting an abnormal ECG.

The present study has several limitations. First, left or right BBB was not specified in the present study. As their prognostic values might differ, this limits the depth of this study. Second, patients had chronic HFpEF in the present study, but ECG might alter greatly when acute decompensation occurs, yet we did not have data about that. Third, ventricular paced rhythm, pathological Q waves, and left ventricular hypertrophy were not of prognostic significance in the present study. However, the number of these abnormalities was rather small, which makes it harder to reach a statistical significance.

Conclusion

An abnormal ECG presented at baseline significantly predicts worse outcomes in patients with HFpEF, especially the abnormalities of BBB and AF, so it is important for physicians to pay more attention to HFpEF patients who present an abnormal ECG.

Acknowledgements

We would like to thank the TOPCAT investigators and participants, as well as the National Heart, Lung, and Blood Institute for summarizing and sharing the trial data.

Conflict of interest statement: None declared.

Funding

This study is funded by Natural Science Foundation of China (No. 82200415).

Data availability

Data used in the present study are from the TOPCAT trial, which can be acquired via a reasonable request to BioLINCC (https://biolincc.nhlbi.nih.gov/).

Contributorship statement

T.H.: analysis and interpretation of data for the work and drafting the manuscript; C.L.: acquisition of data for the work and reviewing the manuscript; W.L.: conception and design of the work and reviewing the manuscript.

References