Predictive value of QT interval correction formulas for all-cause mortality in patients with left bundle branch block

Abstract

Background

QT prolongation is recognized as a prognostic marker associated with poor clinical outcomes and is frequently observed in patients with left bundle branch block (LBBB). However, the identification of QT prolongation in LBBB patients remains challenging. Although multiple QT correction formulas have been specifically developed for LBBB, the prognostic significance and predictive value of these corrected QT intervals in LBBB patients have not been conclusively established.

Methods and Results

This retrospective cohort study consecutively enrolled LBBB patients from 2011 to 2021. The exclusion criteria comprised patients with persistent atrial fibrillation/flutter, atrioventricular block, right ventricular pacing, atypical LBBB morphology and malignant tumor. Comprehensive data collection included electrocardiographic parameters, biochemical profiles, and clinical characteristics for all-cause mortality analysis. Five established QT correction (QTc) formulas (Rautaharju, Yankelson, Wang, Bogossian, and Colunga) were applied to calculate QTc intervals in LBBB patients. During a median follow-up time of 5.1 ± 2.9 years, the overall mortality rate was 21.50% (106/493), with differential rates observed between cardiology (18.32%, 74/404) and non-cardiology (35.96%, 32/89) subgroups. Multivariable regression analysis identified age, left ventricular ejection fraction, diabetes mellitus, and coronary artery disease as significant predictors of all-cause mortality following adjustment. Cardiac resynchronization therapy demonstrated consistent protective effects across all five analytical models constructed by each QTc formula. Furthermore, prolonged QTc interval calculated using all 5 formulas was significantly associated with increased mortality risk in multivariable analyses. Among the correction formulas, the Rautaharju formula exhibited superior predictive performance for all-cause mortality (AUC = 0.812, cut-off = 416.59 ms).

Conclusion

In this retrospective cohort, patients with LBBB demonstrate substantially elevated mortality rates, particularly in non-cardiology populations, necessitating heightened clinical attention. While our findings demonstrate a robust association between QTc prolongation and all-cause mortality. Among the formulas evaluated, our study identifies the Rautaharju formula as the preferred tool for risk stratification in LBBB patients.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12872-026-05661-8.

Introduction

QT prolongation represents a frequent clinical finding in hospitalized patients [1–2] and has been well-established as a significant prognostic marker for poor clinical outcomes [3]. The measurement of QT correction (QTc) intervals is fundamentally influenced by heart rate (HR) and QRS complex morphology. While numerous correction formulas, most notably the Bazett formula [4], have been extensively used for HR-adjusted QTc calculations, these conventional methods exhibit a critical limitation in their failure to consider QRS morphological characteristics into their computational algorithms.

Left bundle branch block (LBBB) is commonly observed electrocardiographic abnormality, is diagnostically identified by the presence of wide notches or indistinct R waves across multiple electrocardiography (ECG) leads [5]. This conduction abnormality is inherently linked to QT interval prolongation, largely due to the characteristic widening of QRS complexes [6]. Although numerous QTc correction formulas have been systematically developed and validated for precise QTc interval estimation in patients with LBBB [7–11], their prognostic utility and clinical predictive value remain inadequately defined. The present study aims to comprehensively assess the prognostic significance of QTc intervals derived from each formula in relation to all-cause mortality among patients with confirmed LBBB.

Method

Study patients

This retrospective cohort study consecutively enrolled patients diagnosed with LBBB from the First Affiliated Hospital of Dalian Medical University from 2011 to 2021 through systematic screening of electronic medical records. The diagnosis of true LBBB was strictly based on Strauss criteria [12], requiring fulfillment of the following electrocardiographic characteristics: (1) QRS duration ≥ 140 ms in male patients or ≥ 130 ms in female patients; (2) Presence of QS or rS patterns in leads V₁ and V₂, accompanied by r wave amplitude < 0.1 mV and q wave amplitude < 0.1 mV in lead aVL; and (3) Demonstration of notched QRS complexes in at least two of the following leads: I, aVL, V₁, V₂, V₅, and V₆.

The exclusion criteria were rigorously defined as follows: (1) Patients with persistent atrial flutter or fibrillation; (2) Cases demonstrating LBBB patterns exclusively in escape rhythms, atrioventricular block, or during right ventricular pacing; (3) Patients exhibiting atypical LBBB morphologies, including but not limited to intraventricular conduction delay (IVCD); and (4) Individuals with malignant neoplasms or end-stage systemic diseases.

ECG evaluation

Standard 12-lead-ECG was obtained immediately at a speed of 25 mm/s after hospital admission in standardized ward. Electrocardiographic parameters, including HR, PR interval, QRS and QT duration, were systematically extracted from the Muse System (GE Co, Ltd, 7.1.1). The longest QT interval was manually measured from leads II, V₂ and V₅, consistent with our previously established methodology [13]. QTc intervals were calculated using the Bazett formula as the primary correction method, with the exception of the Rautaharju formula, which incorporates intrinsic HR correction. The following QTc formulas have been specifically developed and validated for LBBB populations [7–11]:

1. Rautaharju [7]: QT_R=QT_LBBB–155*(RR_LBBB–1)–0.93*(QRS_LBBB–139)+k, (k is –22 ms for male, and –34 ms for female)

2. Yankelson [8]: QT_Y=QT_LBBB–QRS_LBBB+k (k is 95 ms for male, and 88 ms for female)

3. Wang [9]: QT_W=QT_LBBB–(0.86*QRS_LBBB–71)

4. Bogossian [10]: QT_B=QT_LBBB–48.5%*QR_SLBBB

5. Colunga [11]: QT_Co=QT_LBBB−(1.058*QRS−108.397)−(0.0496*JT)−11. 038

Clinical evaluation and follow-up

Comprehensive clinical data, including detailed physical examination findings, standardized laboratory test results and drugs, were routinely collected during patient evaluation. Transthoracic echocardiography was performed using the Vivid 7 Ultrasound system (GE Vingmed Ultrasound, Horten, Norway) in a dedicated echocardiography room. Left ventricular end-diastolic diameter (LVEDD) measurements were obtained with patients in the standard left lateral decubitus position during resting conditions. Left ventricular ejection fraction (LVEF) was quantitatively assessed using the Simpson method. All statistical analyses were restricted to patients with an initial diagnosis of LBBB. Survival status was ascertained through systematic review of municipal medical insurance records, with additional verification via telephone follow-up check.

Statistic analysis

Continuous variables were expressed as mean ± standard deviation. Parametric continuous variables with normal distribution were analyzed using independent Student’s t-tests, while non-parametric variables were evaluated with Mann-Whitney u tests. Categorical variables were presented as percentages and compared using Pearson’s χ² tests, with a predetermined significance threshold of p < 0.05. Multivariate Cox proportional hazards regression models were employed to identify significant risk factors of all-cause mortality, with results expressed as hazard ratios (HR) and corresponding 95% confidence intervals (95%CI). The diagnostic performance of predictive variables was assessed through receiver operating characteristic (ROC) curve analysis, with optimal cut-off values determined by maximizing the Youden index. Survival probabilities were estimated using Kaplan-Meier analysis, with between-group comparisons performed by log-rank tests. All statistical analyses were conducted using R Studio software (version 3.6.0, R Foundation for Statistical Computing).

Results

Basic characteristics of patients

Among 1,140 initially screened LBBB patients who met the inclusion criteria, 493 were ultimately included in the final analysis cohort. The study population had a mean age of 68.1 ± 12.1 years, with female participants representing 46.45% (229/493) in the cohort. The mean LVEF was 44.93 ± 14.22%, with 31.85% (157/493) of patients demonstrating severely reduced cardiac function (LVEF < 35%). The QTc intervals calculated using five distinct formulas were as follows: Rautaharju (QTcR) = 425.89 ± 39.94 ms; Yankelson (QTcY) = 424.29 ± 42.01 ms; Wang (QTcW) = 424.80 ± 41.31 ms; Bogossian (QTcB) = 409.03 ± 40.10 ms; and Colunga (QTcCo) = 404.79 ± 40.43 ms.

Significant differences in clinical characteristics were observed between survivors and non-survivors. The mortality group demonstrated significantly higher mean age, longer QTc intervals of all correction formulas, greater prevalence of abnormal biochemical parameters, diuretic use and clinical comorbidities compared to the survival group (Table 1). The distribution of LBBB diagnoses varied substantially across clinical departments. Cardiology department accounted for the majority of cases (81.95%), followed by the intensive care unit (ICU) (6.09%) and neurology department (2.23%) (Table 2).

Table 1.

Clinical characteristics of patients between survivors and non-survivors

Variables	Alive (n = 387)	Dead (n = 106)	P value
Follow-up (years,‾x ± s)	5.7 ± 2.8	3.3 ± 2.1	< 0.001
Population
Age (years old,‾x ± s)	66.6 ± 12.0	73.5 ± 10.7	< 0.001
Female (n, %)	183 (47.29%)	46 (43.40%)	0.477
ECG data (‾x ± s )
HR (bpm)	74.95 ± 17.11	72.59 ± 11.30	0.398
QRS (ms)	151.96 ± 17.66	150.41 ± 14.62	0.431
QT (ms)	443.06 ± 50.93	461.68 ± 57.17	0.007
JT (ms)	291.11 ± 52.64	311.27 ± 53.11	0.001
QTc (ms)	487.82 ± 39.72	502.44 ± 39.97	0.014
QTcR (ms)	422.17 ± 38.33	440.54 ± 40.04	< 0.001
QTcY (ms)	420.60 ± 41.00	438.68 ± 40.97	0.001
QTcW (ms)	421.31 ± 40.23	438.92 ± 40.19	0.002
QTcB (ms)	405.82 ± 38.62	422.71 ± 40.33	0.003
QTcCo (ms)	401.32 ± 39.63	418.50 ± 38.60	0.001
Biochemical test (n, %)
Anemia	25 (6.46%)	19 (17.92%)	< 0.001
Hypokalemia#	29 (7.49%)	12 (11.32%)	0.206
Hypocalcemia#	34 (8.79%)	17 (16.04%)	0.030
UCG data (‾x ± s )
LVEDD (mm)	53.24 ± 10.66	55.54 ± 10.71	0.023
LVEF (%)	46.76 ± 13.80	38.44 ± 13.86	< 0.001
Clinical Disease (n, %)
Hypertension	215 (55.56%)	67 (63.21%)	0.158
Diabetes	99 (25.58%)	47 (44.34%)	< 0.001
CAD	153 (39.53%)	72 (67.92%)	< 0.001
CKD	19 (4.91%)	10 (9.43%)	0.079
Neurological disorder*	12 (3.10%)	6 (5.66%)	0.341
Heart failure	146 (37.73%)	73 (68.87%)	< 0.001
CRT/CRT-D	65 (16.80%)	7 (6.60%)	0.008
Drugs (n, %)
Diuretic	161 (41.60%)	75 (70.75%)	< 0.001
Class III antiarrhythmic drugs	24 (6.20%)	8 (7.55%)	0.656
Chemotherapeutic drugs	1 (0.26%)	0 (0.00%)	1.000
Quinolone antibiotics	17 (4.39%)	4 (3.77%)	0.993

CKD Chronic kidney disease, CAD Coronary artery disease, CRT/CRT-D Cardiac resynchronization therapy/CRT-defibrillator

# Hypokalemia: serum K⁺ <3.50mmol/L. Hypocalcemia: serum Ca²⁺ <2.10mmol/L

* Neurological disorder: cerebral infarction and hemorrhage

Table 2.

Departments distribution of LBBB patients

Diagnosed department	Number (%)
Cardiology	404 (81.95%)
ICU	30 (6.09%)
Neurology	11 (2.23%)
Emergency	8 (1.62%)
Endocrinology	5 (1.01%)
Respiratory	4 (0.81%)
Opthalmology	4 (0.81%)
General surgery	4 (0.81%)
Urinary surgery	4 (0.81%)
Nephrology	2 (0.41%)
Gastroenterology	2 (0.41%)
Hematology	2 (0.41%)
Acute abdomen	2 (0.41%)
Joint surgery	2 (0.41%)
Rheumatology and immunology	1 (0.20%)
Vascular department	1 (0.20%)
Neurology surgery	1 (0.20%)
Other*	6 (1.22%)

ICU Intensive care unit

*Other: Special needs ward and gerontology are included

Risk factors for QT prolongation in LBBB patients

The univariate and multivariate analysis of clinical factors associated with all-cause mortality was presented in Table 3; Fig. 1 respectively. After adjustment for potential confounders, prolonged QTc intervals calculated by five correction formulas, age, LVEF, diabetes mellitus (DM) and coronary artery disease (CAD) emerged as significant prognostic factors for mortality. Consistent with clinical expectations, cardiac resynchronization therapy with or without defibrillator capability (CRT/CRT-D) demonstrated significant protective effects in all five analytical models.

Table 3.

Univariate regression model for different risk factors

Variables	HR (95%CI)	P value
QTcR	1.011 (1.006–1.017)	< 0.001
QTcY	1.011 (1.006–1.016)	< 0.001
QTcW	1.011 (1.006–1.016)	< 0.001
QTcB	1.011 (1.006–1.016)	< 0.001
QTcCo	1.011 (1.006–1.017)	< 0.001
Age	1.055 (1.035–1.075)	< 0.001
Female	0.823 (0.560–1.208)	0.320
LVEF	0.958 (0.945–0.971)	< 0.001
CRT/CRT-D	0.453 (0.211–0.976)	0.043
Hypertension	1.281 (0.863–1.901)	0.219
Diabetes	2.053 (1.399–3.012)	< 0.001
CAD	2.602 (1.730–3.913)	< 0.001
CKD	2.056 (1.071–3.946)	0.030
Neurological disorder*	1.948 (0.854–4.444)	0.113
Anemia	2.669 (1.623–4.388)	< 0.001
Hypokalemia	1.871 (1.025–3.418)	0.041
Hypocalcemia	1.907 (1.134–3.206)	0.015
Diuretic	3.240 (2.130–4.929)	< 0.001
Class III antiarrhythmic drugs	1.422 (0.691–2.926)	0.338
Chemotherapeutic drugs	-	-
Quinolone antibiotic	0.971 (0.356–2.648)	0.954

CKD Chronic kidney disease, CAD Coronary artery disease, CRT/CRT-D Cardiac resynchronization therapy/CRT-defibrillator

* Neurological disorder: cerebral infarction and hemorrhage

Fig. 1.

Multivariate Cox regression forest map based on 5 different QT formula models

The prediction accuracy of QT correction formula

The predictive accuracy of each formula for all-cause mortality is summarized in Table 4. The Rautaharju formula demonstrated superior diagnostic performance with the largest areas under curves (AUC) with a sensitivity of 73.00% and specificity of 78.00% (Fig. 2). Based on the established QTc prolongation threshold (≥ 416 ms) made by Rautaharju formula, LBBB patients were stratified into two distinct groups: prolonged QT (LQT) and non-LQT groups. Kaplan-Meier analysis revealed significantly reduced survival rates in the LQT group compared to the normal QT group (72.76% vs. 86.70%, p = 0.001) (Fig. 3a). This mortality pattern remained consistent across age-stratified subgroups (Fig. 3b, c) and gender-specific analyses (Fig. 3d, e).

Table 4.

Prediction accuracy of QTc prolongation by 5 formulas

AUC	Rautaharju	Yankelson	Wang	Bogossian	Colunga
	0.812	0.809	0.810	0.809	0.810
Sensitivity (%)	73.00	69.00	75.00	68.00	77.00
Specificity (%)	78.00	81.20	75.60	82.20	74.30
Cut-off (ms)	416.59	420.40	414.20	404.73	390.14
C-index	0.787	0.786	0.786	0.785	0.787

Fig. 2.

ROC curve of Rautaharju formula model on predicting all cause mortality

Fig. 3.

Kaplan-Meier survival analysis comparing LBBB patients with and without LQT using the Rautaharju formula revealed significant mortality differences. a In the overall cohort, the LQT group demonstrated significantly reduced survival rates compared to the non-LQT group (72.76% vs. 86.70%, p = 0.001). b, c This survival disparity was particularly pronounced in elderly subgroups (≥ 65 years old), with 63.44% of LQT patients vs. 86.99% in non-LQT patients (p < 0.001), and those aged ≥ 75 years old exhibiting 49.50% survival compared to 84.37% in controls (p < 0.001). d, e Gender-stratified analysis consistently demonstrated poorer survival outcomes in the LQT group, with male patients showing 69.57% survival vs. 85.71% in non-LQT males (p = 0.004), and female patients demonstrating 75.66% survival compared to 88.31% in non-LQT females (p = 0.049)

Subgroup analyses

Recognizing the clinical complexity and potential confounding factors in patients with severe heart failure (LVEF ≤ 35%), we further conducted interaction analysis and LVEF subgroup analyses based on Rautaharju-corrected QTc intervals (Fig. 4). A statistically significant interaction was identified between LVEF and formula-corrected QTc intervals (Table 5). LVEF distribution among different subgroups is presented in Table 6. Multivariable regression analysis revealed that in the LVEF > 35% subgroup, QTc prolongation, advanced age, and diuretic use were significantly associated with all-cause mortality, while CRT/CRT-D demonstrated a protective effect. In contrast, within the LVEF≤35% subgroup, only hypokalemia showed significant association with mortality risk, with no other factors including QTc reaching statistical significance.

Fig. 4.

Multivariable regression analysis of Rautaharju-corrected QTc intervals in LVEF > 35% (A) and LVEF ≤ 35% (B) subgroups

Table 5.

Interaction Analysis between LVEF and QTc

Model	HR (95%CI)	P value
QTcR*LVEF	1.003 (1.002–1.003)	< 0.001
QTcY*LVEF	1.002 (1.002–1.003)	< 0.001
QTcW*LVEF	1.002 (1.002–1.003)	< 0.001
QTcB*LVEF	1.002 (1.002–1.003)	< 0.001
QTcCo*LVEF	1.003 (1.002–1.003)	< 0.001

Table 6.

Distribution of LVEF across QTc subgroup

Variable	QTc<416ms (n = 199)	QTc≥416ms (n = 294)	P value
LVEF ≤ 35%	53 (26.63%)	93 (31.63%)	0.266
LVEF	46.52 ± 14.32	43.86 ± 14.08	0.046

Discussion

Main finding

This study demonstrates a significant association between QTc intervals of 5 correction formulas and all-cause mortality in LBBB patients. We provide the first comprehensive validation of these QTc formulas, revealing comparable predictive accuracy as evidenced by similar AUC and C-index values across all formulas. However, the Rautaharju formula is recommended for clinical application due to its unique development through large-scale epidemiological studies utilizing general population cohort data, with an optimal QTcR cut-off value of 416.59 ms for identifying the high-risk patients. Notably, given the presence of comorbidities, stress conditions, or disease exacerbations in the study population, the prolonged QTc interval may reflect the overall disease burden or progression severity rather than solely indicating electrical instability. Consequently, its clinical significance leans more toward risk stratification than serving as an isolated electrophysiological marker.

Additionally, long-term follow-up revealed substantial mortality disparities between cardiology (18.32%, 74/404) and non-cardiology (35.96%, 32/89) departments, highlighting the critical need for improved ventricular synchrony assessment and implementation of comprehensive management strategies, particularly in non-cardiology settings.

QT correction fomulas

Bogossian et al. conducted a seminal study employing ventricular pacing at both the right ventricular apex and outflow tract to simulate LBBB, systematically comparing QRS, QT and JT intervals during sinus rhythm (SR) and paced conditions [10]. Wang et al. subsequently developed a QTc formula through right ventricular apical pacing validation in 22 patients with intermittent LBBB [9]. However, subsequent investigations including Wang formula have identified potential overcorrection of QT intervals using the Bogossian formula [14]. This concern was substantiated by comparative analysis demonstrating significant QTc differences between wide QRS and narrow QRS (441 ± 34 ms vs. 424 ± 31 ms) patients (p = 0.004) in Wang’s intermittent LBBB cohort. This is consistent with our findings of lower QTc values obtained using the Bogossian formula compared to the Wang method. Notably, these pacing-based studies utilized temporary LBBB induction through right ventricular apical pacing, which fundamentally differs from pathological LBBB in action potential characteristics and electrophysiological properties.

In contrast, the Yankelson and Colunga formulas were derived from LBBB patients following transcutaneous aortic valve replacement (TAVR) [8, 11]. These methodologies were predicated on the empirical assumption of constant JTc intervals before and after LBBB development. However, this assumption may be physiologically untenable given the well-documented cardiac memory effects on T-wave morphology [15], which likely accounts for the observed QT interval variations before and after acute LBBB onset in our study.

The rationale for preferentially recommending the Rautaharju formula is threefold. First, its development was based on a large-scale general population cohort comprising 11,739 patients with IVCD, including 342 confirmed LBBB cases [7], ensuring robust statistical power and broad clinical applicability. Second, unlike the other four formulas, the Rautaharju formula incorporates a linear function of RR interval to effectively eliminate HR dependence. Third, while the Bogossian, Wang, Yankelson, and Colunga formulas were primarily derived from short-term LBBB simulations (through pacing or TAVR) with limited sample with the lack of validation in real-world chronic LBBB populations. Importantly, the electrophysiological mechanisms of QT interval changes in chronic LBBB fundamentally differ from those in acute LBBB simulations. These critical distinctions provide a compelling explanation for our validation outcomes.

Other clinical factors

Multivariate regression analysis identified age, LVEF, DM, CAD, and CRT/CRT-D as significant predictors of patients' survival. These findings are consistent with established literature, as age, DM and CAD have been consistently associated with poor clinical prognosis across multiple studies [16–21]. Notably, current guidelines recommend CRT/CRT-D as a Class I indication for LBBB patients with LVEF ≤ 35%, given its demonstrated benefits in restoring physiological ventricular activation and improving heart failure outcomes [22]. Consistent with clinical expectations, CRT/CRT-D emerged as a significant protective factor in our multivariate analysis. Moreover, we conducted serial transthoracic echocardiographic records in patients receiving CRT/CRT-D during the follow-up period. The results demonstrated a statistically significant improvement in LVEF, increasing from a baseline value of 31.00 ± 11.47% to 38.76 ± 14.85% post-procedure (p < 0.001). This substantial enhancement in systolic function suggests that the protective effects of CRT/CRT-D may be primarily mediated through improvement in heart failure status. The observed benefits of CRT/CRT-D suggest that intra- and inter-ventricular dyssynchrony may represent critical pathophysiological mechanisms underlying heart failure progression in LBBB patients, although our study lacked detailed etiological stratification of heart failure causes (e.g., dilated, ischemic cardiomyopathy). In our cohort, the mean LVEF was 44.93 ± 14.22%, with 14.60% (72/493) meeting criteria for CRT/CRT-D implantation. Our study identifies low CRT/CRT-D utilization rates, particularly in non-cardiology patients (10.1%) and non-survivors (6.6%). In addition to limited patient acceptance in earlier periods, this finding reveals insufficient awareness and suboptimal management of LBBB within non-specialist clinical settings. Notably, our analysis revealed no significant gender-based differences in outcomes, despite established evidence of hormonal influences on QT interval prolongation in females [23]. This apparent discrepancy may be attributable to age-related attenuation of gender differences in QT interval duration, particularly in elderly populations.

Medication use is a significant factor influencing the QT interval. In this study, we systematically collected data on medications known to affect the QT interval among the observed subjects. The identified medications included diuretics (n = 236), quinolone antibiotics (n = 21), amiodarone (n = 32), and chemotherapeutic drugs (doxorubicin, n = 1). Regression analysis revealed no significant association in the overall sample. However, a significant association was observed in the LVEF > 35% subgroup, which disappeared in the LVEF ≤ 35% subgroup. Possible explanations for this finding include: (1) In patients with LVEF > 35%, more rapidly progressing heart failure may lead to systemic congestion and extensive diuretic use, which could indicate a poorer prognosis. (2) In patients with severe heart failure, multiple comorbidities, electrolyte imbalances, and diuretic resistance are more common, and interactions among these confounding factors may offset the adverse effects of diuretics. Amiodarone, quinolone and doxorubicin are considered to have minimal impact on the study outcomes due to the lack of statistically significant differences or their limited usage.

Notably, the significant association between hypokalemia and all-cause mortality observed in the LVEF ≤ 35% subgroup can be explained by the fact that severe diuretic resistance is common in such patients, and hypokalemia in this population largely results from chronic insufficient intake due to gastrointestinal congestion. Consequently, hypokalemia often presents as a persistent concomitant condition in these individuals. In the multivariate regression analysis, the interplay among these confounders may have counteracted other factors’ effects.

Limitation

The pathophysiological interplay between LQT and coexisting risk factors is complex and multifactorial [2, 18, 24–26]. The retrospective observational nature of this study limits causal inference. While we adjusted for multiple clinical confounders using standardized protocols, unmeasured confounding may persist. Furthermore, aspects such as patient selection and data collection in a retrospective design could introduce bias, despite our use of standardized protocols to minimize inconsistencies. Therefore, our findings should be interpreted as highlighting important associations rather than establishing definitive cause-and-effect relationships. Prospective studies are warranted to validate these observations and explore underlying mechanisms. Nevertheless, we believe that drug-induced acquired QT interval prolongation in hospitalized patients can generally be promptly identified and corrected (e.g., hypokalemia), suggesting that the potential for this underlying bias to have a significant long-term impact is limited. Moreover, the lack of specific causes of death for these patients implies that some individuals’ deaths may not have been cardiac origin or had a relatively weaker association with QT interval prolongation. The implementation of rigorous data collection protocols, comprehensive follow-up procedures, and standardized QT measurement methodologies by experienced investigators substantially enhances the reliability and validity of our findings, thereby partially mitigating these limitations.

Conclusion

Patients with LBBB demonstrate significantly elevated mortality rates, particularly in non-cardiology populations, warranting heightened clinical attention. While all investigated QT correction formulas exhibit predictive capacity for mortality risk assessment in LBBB patients, the Rautaharju formula demonstrates superior clinical reliability and is therefore recommended for routine application in the future.

Abbreviations

LBBB

Left bundle branch block

ECG

Electrocardiographic

QTc

QT correction

LVEF

Left ventricular ejection fraction

DM

Diabetes mellitus

CAD

Coronary artery disease

CKD

Chronic kidney disease

CRT/CRT-D

Cardiac resynchronization therapy / CRT-defibrillator

HR

Heart rate

IVCD

Intraventricular conduction delay

LVEDD

Left ventricular end-diastolic diameter

95%CI

95% confidence intervals

ROC

Receiver operating characteristic

ICU

Intensive care unit

AUC

Areas under curves

LQT

Prolonged QT

TAVR

Transcutaneous aortic valve replacement

SR

Sinus rhythm

Authors’ contributions

Rongqian Xu and Haolin Song collected and structured the dataset, wrote the manuscripte. Rongfeng Zhang and Xiaopeng Dong provided statistical support. Yingxue Dong, Xiaomeng Yin and Lianjun Gao provided the approval of administrative permissions for data acquisition and follow-up. Yunlong Xia provided the funding support and quality assurance. Yiheng Yang responsible study design, funding, and manuscript revision. All authors read and approved the final manuscript.

Funding

This work was supported by Natural Science Foundation of Liaoning Province (Grant number 2024-MS-154) and National Key Research and Development Program of China (Grant number 2022YFC2405002).

Data availability

The original data utilized in this study were derived from the real clinical records of patients and have been submitted to the scientific research management department for archiving. In accordance with the regulations by the ethics committee and the information center, such data are generally prohibited from being uploaded to external networks. If necessary, please contact the corresponding author or the first author, we will provide the original data through appropriate and compliant means. This approach ensures adherence to institutional policies while facilitating data accessibility for verification or further academic investigation when necessary.

Declarations

Ethics approval and consent to participate

Given that the study utilized non-interventional, anonymized clinical data and did not involve any biological samples, the informed consent could be waived according to the guidelines by the ethics committee of the First Affiliated Hospital of Dalian Medical University. Study protocol adheres to the principles of Helsinki Declaration, and has been approved by the Ethics committee of First affiliated hospital of Dalian Medical University (PJ-KS-KY-2023-558).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.