Abstract
Objective: The predictive value of different single electrocardiographic left ventricular hypertrophy (ECG-LVH) criteria and their combined indices for long-term all-cause and cardiovascular mortality in Chinese community-dwelling populations remains understudied.
Methods: 6016 participants aged ≥60 years from a Chinese community-based cohort free of prior cardiovascular events at baseline were followed up for 9.39 years (median). LVH was diagnosed using five individual criteria: the traditional Sokolow-Lyon, Cornell Voltage, Cornell Product, and RavL criteria and the recently proposed Peguero-Lo Presti (PLP) criterion. Two composite indices were constructed either with any of the four traditional criteria, or any of the four traditional criteria and the PLP criterion. Cox regression models were used to analyze the associations of these individual and composite indices with all-cause and cardiovascular mortality.
Results: Three individual ECG-LVH criteria (Cornell Voltage, Cornell Product and PLP) and two composite indices were associated with both all-cause mortality (HR range: 1.25–1.81, all P < 0.05) and cardiovascular mortality (HR range: 1.50–2.27, all P < 0.05). The novel PLP and composite indices yielded a significant net reclassification improvement (NRI) (all-cause mortality, NRI = 0.1116–0.1176; cardiovascular mortality, NRI = 0.1675–0.1948; all P < 0.010), with a mild improvement in discrimination (C-index change, 0.0146–0.0171) and non-significant integrated discrimination improvement (IDI).
Conclusion: The PLP and composite ECG-LVH criteria were associated with long-term mortality and provided modest incremental risk reclassification. External validation, structural cardiac assessment and evaluation of calibration and clinical net benefit are required before these approaches can be recommended for community screening.
Keywords: Cardiovascular risk, Mortality, Electrocardiography, Left ventricular hypertrophy, Peguero-Lo Presti
1. Introduction
Cardiovascular diseases (CVDs) are the leading cause of mortality among older adults in China, and the risk of CVDs increases progressively as left ventricular (LV) mass rises [1], [2]. Left ventricular hypertrophy (LVH) is an adaptive response of the heart to increased hemodynamic load, which gradually develops with age and is one of the most common adverse manifestations of the aging heart [3]. Community-dwelling older adults are particularly vulnerable to adverse outcomes due to high prevalence of chronic diseases and underutilization of preventive healthcare [4].
Echocardiography (ECHO) is an important non-invasive method for detecting LVH, however, due to its high requirements for operators, its availability is far behind electrocardiography (ECG). In Chinese community settings, ECG is a highly accessible, cost-effective diagnostic tool with minimal technical barriers, making it suitable for large-scale risk assessment among older adults. Extensive studies have demonstrated that ECG-LVH is a strong independent risk factor for cardiovascular events and mortality [5], [6], [7], [8]. The association persists significant even after adjusting for ECHO LVH [9], [10], which suggests that adverse electrical remodeling itself provides incremental prognostic value beyond structural remodeling assessed by ECHO.
Among the various ECG criteria for diagnosing LVH [11], the low sensitivity of any single criteria is a well-known limitation. For instance, the Sokolow-Lyon criterion detected only 6.9% of echocardiographic LVH in the Framingham Heart Study [12]. Previous studies have demonstrated that combining two or more ECG criteria improves diagnostic sensitivity at the cost of reduced specificity [13], [14], [15], while also enhancing cardiovascular risk stratification, supporting the adoption of composite ECG indices [6], [8], [16], [17], [18]. However, most studies investigating the prognostic value of composite indicators only focus on the classic Cornell Product and Sokolow-Lyon Voltage criteria. A novel indicator, the Peguero-Lo Presti (PLP) voltage index, which exhibits enhanced sensitivity and better predictive value for mortality, demonstrates higher accuracy than traditional criteria [19]. In Caucasian populations, PLP-diagnosed LVH correlates with increased all-cause and cardiovascular mortality [19], [20]. Nevertheless, while traditional ECG indicators have been associated with mortality in Chinese community-based populations [5], [18], [21], evidence regarding the prognostic value of the PLP criterion, either alone or in combination with conventional ECG-LVH criteria, for long-term cardiovascular and all-cause mortality in this population remains limited [6], [19]. In particular, whether incorporating the PLP criterion into composite indices enhances risk stratification beyond conventional criteria alone has not been systematically investigated, despite the high prevalence of hypertension and the expanding community-based cardiovascular screening across China.
Therefore, the present study utilized follow-up data from a community health examination cohort in Minhang District, Shanghai, with a median follow-up of 9.39 years. This study was designed to compare the long-term predictive performance of four classic ECG-LVH criteria (Sokolow-Lyon, Cornell Voltage, Cornell Product, and RaVL) and the novel PLP criterion, and to examine whether composite ECG-LVH indices offer improved mortality risk discrimination relative to individual criteria.
2. Methods
2.1. Study population
We established the cohort in 2012 among community-dwelling older adults (aged ≥60 years) attending the government-funded annual health examination program in Xinzhuang Community, Minhang District, Shanghai. Participants who provided written informed consent at the study site were enrolled. Follow-up surveys were conducted in 2016, 2018, and 2021. Information regarding participants' history of tobacco and alcohol consumption, history and treatment of hypertension and diabetes mellitus were collected by questionnaires. Body mass index (BMI) was calculated by dividing the weight in kilograms by the square of the height in meters. Blood pressure and heart rate were measured 3 times with 1-min interval using electronic sphygmomanometer by the medical staff. Participants had undergone a 12‑lead standard ECG. Digital ECG data were recorded using a GE MAC 800 electrocardiograph (GE, Milwaukee, Wisconsin) at 10 mm/mV calibration and a speed of 25 mm/s. All ECG tracings were initially inspected visually for technical errors and inadequate quality before being automatically processed using GE 12-SL Marquette version 2001.
During the analysis, we excluded participants who met any of the following criteria: (1) those with incomplete clinical data (including medical history and medication use), on-site blood pressure records or ECG results; (2) individuals diagnosed with arrhythmias including premature beats, atrial fibrillation, left/right bundle branch block and atrioventricular block, as confirmed by ECG or prior clinical diagnosis; (3) those with a history of cardiovascular events documented in the electronic health record (EHR) system. Finally, a total of 6016 individuals were left for analyses. The participant enrollment flowchart is presented in Fig. 1.
Fig. 1.
Participant enrollment flowchart.
2.2. Judgement of the end-point
Mortality data were ascertained on 31 December 2021 from participants' EHR and validated against documents provided by the Shanghai Centers for Disease Control and Prevention. The endpoints considered in the present analysis were all-cause and cardiovascular mortalities, which included mortality from cerebral hemorrhage [International Classification of Diseases, Tenth Revision (ICD-10) codes I60–62 and I69.1], cerebral infarction (ICD-10 codes I63 and I69.3), undetermined cerebral disorders (ICD-10 codes I64 and I67), and cardiac disorders (ICD-10 codes I07, I09–13, I20-I25, I33, I35, I38, I42, and I48-I51). All-cause mortality was defined as the primary outcome, and cardiovascular mortality as the secondary outcome.
2.3. Statistical methods
Analyses were performed with R statistics (version 4.2.2, The R foundation for Statistical Computing, Vienna, Austria). For continuous variables, normality was assessed via the Shapiro-Wilk test and homogeneity of variances via Levene's test. Categorical variables were reported as frequency and percentage, while continuous variables were reported as mean ± standard deviation. Differences in continuous variables were tested using ANOVA, whereas differences in categorical variables were tested using the chi-squared test or Fisher's exact test.
Four ECG-LVH diagnostic voltage criteria recommended by ESC/ESH guideline [22], including Sokolow-Lyon Voltage, Cornell Voltage, Cornell Product, RavL Voltage, as well as the recently proposed PLP Voltage criterion [19] were selected for present study. In addition, two composite ECG-LVH diagnostic indicators were generated: (1) the composite of the four ESH guideline-recommended voltage criteria, and (2) the composite of these traditional criteria combined with the PLP criterion. For both composite indicators, a participant was classified as LVH-positive if any one of the constituent criteria was met, and as LVH-negative only if all constituent criteria were negative. The cutoffs or definitions of all diagnostic indicators are presented in Table 1.
Table 1.
Definitions of ECG-LVH criteria.
| Criteria | Description | Cut-off |
|---|---|---|
| Individual ECG-LVH Criteria | ||
| Sokolow-Lyon Voltage | SV1 + RV5 or SV1 + RV6 | ≥ 3.5 mV |
| Cornell Voltage | SV3 + RavL | Men>2.8 mV; Women>2.0 mV |
| Cornell Product | Men (SV3 + RavL) * QRS-duration Women (SV3 + RavL +0.8) * QRS-duration |
>244 mV·ms |
| RavL Voltage | RavL | ≥ 1.1 mV |
| Peguero-Lo Presti | Deepest S wave in any single lead + SV4 | Men ≥2.8 mV；Women ≥2.3 mV |
| Composite ECG-LVH Criteria | Description | |
| Any of ESH 4 criteria | Any of ESH 4 criteria + /All of ESH 4 criteria - | |
| Any of ESH 4 criteria combined Peguero-Lo Presti | ≥1 of 5 criteria + / All 5 criteria - | |
The associations between different ECG-LVH criteria with all-cause and cardiovascular mortality were examined in separate Cox proportional hazards models. To avoid overfitting, covariates were pre-specified based on clinical importance and prior cardiovascular risk evidence, and included age, sex, BMI, smoking, drinking, history of hypertension and diabetes, antihypertensive treatment, and systolic and diastolic blood pressure. The proportional hazards assumption for all Cox models was assessed using the Schoenfeld residual test. The global test and individual covariate tests yielded P > 0.10 for all models, indicating no significant violation of the proportional hazards assumption. To address competing risk of non-cardiovascular death, Fine-Gray subdistribution hazard regression was applied as a sensitivity analysis for cardiovascular mortality. Additional sensitivity analyses included further adjustment for fasting glucose, low-density lipoprotein cholesterol, and high-density lipoprotein cholesterol in participants with available laboratory data.
In addition, three statistical indices were calculated to assess the potential utility of incorporating different ECG-LVH criteria for predicting all-cause and cardiovascular mortality separately. The concordance index (Harrell's C statistic) and integrated discrimination improvement (IDI) were used to evaluate model discriminative performance, and the net reclassification improvement (NRI) index was adopted to quantify reclassification improvement. For clarity, two forms of statistical significance are reported. Cox regression significance describes associations between ECG-LVH criteria and mortality, whereas incremental performance significance denotes changes in C-index, IDI and NRI after adding ECG-LVH indices to the baseline model.
3. Results
A total of 6016 participants (55.0% women, and 49.6% hypertensive patients) were included in final analysis. Table 2 shows the Clinical characteristics and ECG-LVH indexes of the study participants. At baseline, in terms of clinical characteristics, male participants were older, had higher diastolic blood pressure, and higher proportions of tobacco and alcohol consumption; otherwise, there were no significant differences in systolic blood pressure, heart rate, or the prevalence and treatment rates of hypertension and diabetes between the two gender groups. For the individual ECG-LVH diagnostic indices, the prevalence of ECG-LVH varied substantially across different criteria (range: 1.9% to 14.6%), with the PLP showing the highest prevalence. For the composite diagnostic indices, the incorporation of the PLP criterion increased the positive rate to 20.4%. Across all indices except the Sokolow-Lyon voltage criterion, the prevalence of ECG-LVH was higher in females than in males.
Table 2.
Clinical characteristics and ECG indexes of the study subjects.
| Overall (n = 6016) | Male (n = 2705) | Female (n = 3311) | p | |
|---|---|---|---|---|
| Clinical characteristics | ||||
| Age, y | 67.51 (6.15) | 68.03 (6.06) | 67.09 (6.19) | <0.001 |
| Body mass index, Kg/m2 | 24.67 (6.49) | 24.74 (7.54) | 24.62 (5.50) | 0.492 |
| Systolic blood pressure, mmHg | 131.98 (15.78) | 131.83 (15.28) | 132.10 (16.17) | 0.503 |
| Diastolic blood pressure, mmHg | 76.66 (8.39) | 77.59 (8.12) | 75.89 (8.52) | <0.001 |
| Heart rate, bpm | 75.89 (8.05) | 75.76 (8.11) | 75.99 (8.01) | 0.278 |
| Current smoking, n(%) | 380 (6.3) | 376 (13.9) | 4 (0.1) | <0.001 |
| Current drinking, n(%) | 198 (3.3) | 194 (7.2) | 4 (0.1) | <0.001 |
| Hypertension, n(%) | 2984 (49.6) | 1358 (50.2) | 1626 (49.1) | 0.389 |
| Antihypertensive treatment, n(%) | 2648 (44.0) | 1198 (44.3) | 1450 (43.8) | 0.720 |
| Diabetes mellitus, n(%) | 893 (14.8) | 422 (15.6) | 471 (14.2) | 0.145 |
| All-cause mortality, n(%) | 674 (11.2) | 407(15.0) | 267(8.1) | <0.001 |
| Caridovascular mortality, n(%) | 218 (3.6) | 131 (4.8) | 87 (2.6) | <0.001 |
| ECG variables, mean (SD) | ||||
| QRSdur, ms | 82.86 (12.97) | 87.61 (13.99) | 78.98 (10.59) | <0.001 |
| Sokolow-Lyon Voltage, mV | 2.27 (0.71) | 2.34 (0.75) | 2.22 (0.68) | <0.001 |
| Cornell Voltage, mV | 1.31 (0.55) | 1.36 (0.56) | 1.26 (0.53) | <0.001 |
| Cornell Product, mV·ms | 143.96 (61.19) | 120.11 (57.25) | 163.45 (57.29) | <0.001 |
| RavL Voltage, mV | 0.42 (0.27) | 0.40 (0.28) | 0.43 (0.26) | <0.001 |
| Peguero-Lo Presti, mV | 1.80 (0.72) | 1.93 (0.77) | 1.70 (0.67) | <0.001 |
| ECG-LVH by different criteria, n (%) | ||||
| Sokolow-Lyon Voltage | 323 (5.4) | 190 (7.0) | 133 (4.0) | <0.001 |
| Cornell Voltage | 264 (4.4) | 21 (0.8) | 243 (7.3) | <0.001 |
| Cornell Product | 239 (4.0) | 57 (2.1) | 182 (5.5) | <0.001 |
| RavL Voltage | 117 (1.9) | 50 (1.8) | 67 (2.0) | 0.692 |
| Peguero-Lo Presti | 878 (14.6) | 360 (13.3) | 518 (15.6) | 0.012 |
| Any of ESH 4 criteria | 675 (11.2) | 269 (9.9) | 406 (12.3) | 0.005 |
| Any of ESH 4 criteria | ||||
| combined Peguero-Lo Presti | 1229 (20.4) | 538 (19.9) | 691 (20.9) | 0.365 |
Data are expressed as mean (SD) for continuous variables and number (%) for categorical variables. ESH, European Society of Hypertension.
During a median follow-up of 9.39 years, 674 all-cause mortality events and 218 cardiovascular mortality events occurred, with the mortality rate of male participants being significantly higher than that of female participants. The unadjusted survival probability plots for all-cause (Fig. S1) and cardiovascular (Fig. S2) mortality were presented in supplementary materials.
Fig. 2 presents a forest plot of the all-cause and cardiovascular mortality risks associated with LVH diagnosed by different criteria, after adjusting for age, sex, BMI, smoking, drinking, history of hypertension and diabetes mellitus, antihypertensive treatment, and systolic and diastolic blood pressure. Among the individual indices, the ECG-LVH diagnosed by the 3 criteria (Cornell Voltage, Cornell Product and PLP) showed good predictive value for both all-cause mortality and cardiovascular mortality. For the prediction of all-cause mortality, the hazard ratios (HRs) for Cornell Voltage, Cornell Product and PLP were 1.61 (95% CI: 1.17–2.22, P = 0.004), 1.81 (95% CI: 1.34–2.46, P < 0.001), 1.34 (95% CI: 1.10–1.62, P = 0.003), while for cardiovascular mortality, HR values were 2.02 (95% CI: 1.23–3.33, P = 0.006), 2.27 (95% CI: 1.40–3.68, P < 0.001), 1.59 (95% CI: 1.15–2.19, P = 0.005) for the 3 indices, respectively. The RaVL criterion identified only 117 participants, and its hazard ratio should be interpreted cautiously due to limited statistical power. Of the composite criteria, both the combination of ESH 4 criteria and the addition of the PLP criterion remained statistically significant after adjustments for all-cause mortality and cardiovascular mortality. The competing-risk sensitivity analysis showed all ECG-LVH criteria had consistent risk directions in Fine-Gray versus Cox models, with mild HR fluctuations for most indicators, whereas Cornell Voltage lost statistical significance in the Fine-Gray model (Supplementary Table 2). In the 4711 participants (78.3%) with available lipid and glucose data, additional adjustment for these metabolic factors yielded materially unchanged results (all hazard ratio changes <4%, Supplementary Table 3).
Fig. 2.
Risk of all-cause mortality and cardiovascular mortality by different ECG-LVH criteria. Models were adjusted for age, sex, BMI, smoking, drinking, history of hypertension and diabetes mellitus, antihypertensive treatment, and systolic/diastolic blood pressure.
Furthermore, we assessed whether incorporating ECG-LVH indices into the conventional model would improve prognostic performance, as quantified by the C-index, IDI, and NRI. As shown in Table 3, the addition of composite criteria yielded a modest but significant net reclassification improvement for both outcomes, while improvements in the Harrell's C-index did not reach statistical significance for all-cause mortality. Although the addition of the PLP criterion resulted in only a marginal increment in the Harrell's C-index of the Cox model (All-cause mortality: 0.7436 to 0.7607, P = 0.0368; Cardiovascular mortality: 0.8035 to 0.8200, P < 0.001), we still observed a statistically significant difference.
Table 3.
Model discrimination and reclassification statistics for ECG-LVH indexes and mortality.
| Models | NRI(95%CI) | P value | IDI(95%CI) | P value | Harrell's C Statistic | P value for difference with base model |
|---|---|---|---|---|---|---|
| all-cause mortality | ||||||
| Base model | / | / | / | / | 0.7436 | / |
| + Sokolow-Lyon Voltage | 0.0493 (0.0079, 0.0927) | 0.034 | 0.0002 (−0.0005, 0.0009) | 0.602 | 0.7588 | 0.4708 |
| + Cornell Voltage | 0.0264 (−0.0464, 0.1024) | 0.474 | 0.0015 (−0.0004, 0.0034) | 0.148 | 0.7603 | 0.1004 |
| + Cornell Product | 0.0613 (0.0070, 0.1140) | 0.03 | 0.003 (0.0004, 0.0056) | 0.024 | 0.7604 | 0.0951 |
| + RavL Voltage | 0.0288 (−0.0002,0.0625) | 0.06 | 0.0006 (−0.0003,0.0017) | 0.238 | 0.7482 | 0.8627 |
| + Peguero-Lo Presti | 0.1116 (0.0508, 0.1760) | <0.001 | 0.0014 (−0.0004, 0.0034) | 0.118 | 0.7607 | 0.0368 |
| + Any of ESH 4 criteria | 0.1122 (0.0536, 0.1733) | <0.001 | 0.0012 (−0.0005, 0.0029) | 0.172 | 0.7601 | 0.1067 |
| + Any of ESH 4 criteria Combined Peguero-Lo Presti |
0.1176 (0.0510, 0.1874) | <0.001 | 0.0011 (−0.0003, 0.0028) | 0.132 | 0.7600 | 0.1263 |
| cardiovascular mortality | ||||||
| Base model | / | / | / | / | 0.8035 | / |
| + Sokolow-Lyon Voltage | 0.0728 (−0.0027, 0.1558) | 0.064 | 0.0004 (−0.0014, 0.0023) | 0.754 | 0.8154 | 0.4802 |
| + Cornell Voltage | −0.0724 (−0.1851, 0.0420) | 0.216 | 0.0007 (−0.0018, 0.0035) | 0.592 | 0.8179 | 0.0745 |
| + Cornell Product | 0.0572 (−0.0367, 0.1540) | 0.226 | 0.0037 (−0.0003, 0.0086) | 0.08 | 0.8177 | 0.1157 |
| + RavL Voltage | 0.0344 (0.0084,0.0719) | 0.012 | 0.0004 (−0.0008,0.0027) | 0.442 | 0.8161 | 0.1667 |
| + Peguero-Lo Presti | 0.1724 (0.0592, 0.2851) | 0.004 | 0.0004 (−0.0027, 0.0037) | 0.734 | 0.8200 | <0.001 |
| + Any of ESH 4 criteria | 0.1675 (0.0633, 0.2773) | <0.001 | 0.0018 (−0.0013, 0.0052) | 0.29 | 0.8181 | 0.0098 |
| + Any of ESH 4 criteria Combined Peguero-Lo Presti |
0.1948 (0.0796, 0.3249) | 0.002 | 0.0012 (−0.0016, 0.0044) | 0.45 | 0.8188 | 0.0046 |
Base model includes sex, age, BMI, smoking, drinking, history of hypertension and diabetes, antihypertensive treatment, systolic/diastolic blood pressure.
4. Discussion
The present prospective cohort study evaluated the predictive value of four traditional ESH-recommended ECG-LVH criteria, the PLP criterion, and their derived composite criteria for 10-year all-cause and cardiovascular mortality in a Chinese community-based elderly population. The key findings indicated that certain single and composite ECG-LVH criteria independently predicted long-term mortality. As evaluated by NRI and C-index, the PLP criterion and composite indices yielded incremental prognostic value over traditional ECG-LVH indices. These results provide valuable insights into the prognostic value of ECG-LVH criteria in Chinese community-dwelling adults.
4.1. Association between individual ECG-LVH criteria and long-term mortality
A core finding of our study is that two of the traditional ESH-recommended ECG-LVH criteria (Cornell Voltage and Cornell Product) remained significantly associated with 10-year all-cause and cardiovascular mortality after multivariable adjustment. This observation aligns with prior evidence indicating that ECG-detected LVH reflects subclinical cardiac impairment with broad prognostic relevance [5], [6], [7], [8], [9], [10], [23], [24], [25], supported by aggregated findings from meta-analyses and targeted cohort studies across diverse populations. For example, general population meta-analyses [24], [25] (over 100,000 participants) report consistent associations between traditional criteria and mortality outcomes (pooled RR = 1.30–1.87 for all-cause mortality; pooled RR = 1.38–1.66 for cardiovascular mortality).
Notably, our findings are fully consistent with most previous epidemiological studies in Chinese community populations [5], [9], [18]. The 9.5-year follow-up study [18] of 4530 Taiwanese community-dwelling elderly, which systematically evaluated 9 common ECG abnormalities and identified only ECG-LVH diagnosed by modified Cornell Product criterion as the only index consistently associated with both all-cause (HR = 1.39, 95% CI: 1.16–1.67, P = 0.0003) and cardiovascular mortality (HR = 2.37, 95% CI: 1.48–3.79, P = 0.0003), while in our study these two HR values of modified Cornell criteria were 1.81 (95% CI: 1.34–2.26) and 2.27 (95% CI: 1.40–3.68), respectively.
Regarding the prognostic predictive value of the PLP index [19], Afify et al.'s [20]analysis of 7825 U.S. nationally representative participants showed that ECG-LVH diagnosed by PLP criterion was independently associated with a 29% increased risk of all-cause mortality (HR = 1.29, 95% CI: 1.16–1.44) and 53% increased risk of cardiovascular mortality (HR = 1.53, 95% CI: 1.31–1.80). Our findings in Chinese community adults closely mirror this pattern, Peguero-Lo ECG-LVH predicted a 34% higher risk of all-cause mortality (HR = 1.34, 95% CI: 1.10–1.62) and a 59% higher risk of cardiovascular mortality (HR = 1.59, 95% CI: 1.15–2.19). This cross-population consistency observed between Afify et al.'s U.S. multiethnic cohort and our East Asian elderly sample further supports stable long-term mortality predictive performance of the novel PLP ECG-LVH index across diverse populations.
Importantly, a significant independent association in Cox regression does not guarantee meaningful improvements in risk discrimination when ECG-LVH criteria are added to multivariable baseline models, as baseline covariates may capture overlapping risk information. These statistical distinctions are illustrated in the comparative performance of the novel PLP criterion versus conventional ECG-LVH indices. Findings from Afify et al. [20] indicated no statistically significant difference in the long-term mortality prognostic value between the PLP criterion and other traditional ECG-LVH diagnostic indices, whereas our results showed that the PLP criterion exhibited slightly stronger associations with both all-cause and cardiovascular mortality, with modest increments in NRI and C-index, though IDI was non-significant.
It is also important to note that despite PLP criterion exhibits a modest diagnostic accuracy for ECHO LVH across Asian subgroups, while Sun et al. [26] reported lower diagnostic accuracy for echocardiographic LVH in Chinese adults (AUC = 0.623–0.689 vs. CV's 0.681–0.721), and Narita et al. [27] validated a modified Asian-specific variant in Japanese populations, our data confirm that PLP's predictive performance remains preserved. This discrepancy highlights that ECG-LVH criteria derive prognostic value from capturing functional and electrophysiological abnormalities, not just anatomical LVH, which is also supported by van Kleef et al.'s cohort study [8] focusing on individuals with vascular disease.
4.2. Composite ECG-LVH indices and long-term mortality: encompassing predictive utility
Previous studies have demonstrated that combination of two or more ECG-LVH criteria could significantly increase sensitivity for the detection of LVH and coexistence of these criteria further improved risk prediction for future events and mortality, even across diverse populations [8], [13], [23]. Tanaka et al. [23] reported that a composite of traditional criteria had a higher population attributable risk (11.6%) than individualcriteria (3.0–9.7%), which means that individualcriteria may miss hidden risk. While van Kleef et al. [8]and Okin et al. [13]similarly showed that multi-criterion approaches stratified highest risk (HR = 1.96–3.00) in vascular disease and hypertensive cohorts. In our study, the “any-positive” rule identified 20.4% of the cohort as ECG-LVH positive, nearly twice the maximum positivity rate for individual criteria. Higher case detection is an inherent sensitivity-specificity tradeoff of the any-positive composite rule. Though more high-risk subjects could be captured, the specificity and false-positive rate of this combined definition cannot be quantified without echocardiographic confirmation [10], leaving its net clinical impact to be clarified in future research. Even so, its higher positivity rate identifies a greater proportion of participants with ECG-LVH, who exhibited elevated mortality risk during long-term follow-up.
Compared with conventional individualECG indicators, the integration of PLP and composite criteria yielded a modest improvement in NRI, which may suggest potential advantages in reclassification performance. Composite criteria may capture complementary pathophysiological signals linked to LVH: specifically, Cornell Voltage and Cornell Product reflect septal and endocardial remodeling [28], Sokolow-Lyon criterion captures precordial voltage alterations [11], and PLP combines voltage with QRS duration [19]. This multi-signal approach could potentially offer more robust risk stratification across populations with varying left ventricular geometry or ECG voltage norms, which may aid the identification of individuals with elevated mortality risk within community risk assessment frameworks. However, this interpretation remains speculative, as our study did not include structural imaging or biomarkers. Future studies integrating multimodal cardiac imaging would be needed to test this hypothesis. It should be noted that the significant NRI improvements were accompanied by modest gains in discrimination and non-significant IDI, suggesting cautious extrapolation of the observed incremental predictive performance.
The study had several limitations. First, we did not perform echocardiographic validation for ECG-LVH indices. The absence of echocardiographic data precludes validation of anatomical LVH but does not compromise the assessment of the prognostic value of electrocardiographic criteria in relation to mortality outcomes. Routine echocardiography is limited by cost and technical constraints in community cohorts, and targeted local validation is recommended in future studies. Second, despite a median follow-up of 9.39 years, all ECG and clinical data were collected only at baseline without serial repeated measurements. Changes in antihypertensive therapy during follow-up may alter ECG-LVH progression and its prognostic links to mortality [29], [30], and longitudinal LVH trajectories (incident, persistent or regressive LVH) carry distinct predictive values beyond single baseline ECG testing [31]. Future analyses using time-varying Cox models could better account for these dynamic shifts and refine prognostic estimates in Chinese community residents. Third, participants with arrhythmia or conduction disorders were excluded to guarantee valid ECG-LVH analysis, which may reduce external validity for elderly populations prone to high cardiovascular risk. Most affected patients receive standardized clinical care with intensive cardiac surveillance. Fourth, important predictors such as renal function, specific medication classes, socioeconomic status, and frailty were not measured, and residual confounding from these unmeasured variables cannot be excluded. Our findings may not generalize to younger populations, other ethnic groups, or clinical populations with established cardiovascular disease. Additionally, we did not conduct model calibration, decision curve analysis (DCA), or net benefit assessment, leaving the sensitivity–specificity trade-off and clinical impact of our composite criteria unquantified. The lack of these analyses represents an important methodological limitation, as such approaches are essential for evaluating the clinical applicability and utility of composite ECG-LVH indices for risk stratification.
To our knowledge, our work is the first cohort study to assess the mortality risk of the novel PLP criterion and its composition with traditional criteria in this specific population. As such, our research partially fills an important gap in ethnicity-specific evidence for Chinese individuals. Notably, as shown in Table 3, while the addition of traditional ECG-LVH criteria to the base model did not reach statistical significance, the clinical implications of these findings warrant further investigation.
In conclusion, individual ECG-LVH criteria, particularly the PLP criterion, independently predicted long-term mortality in community-dwelling Chinese older adults, consistent with cumulative cross-population evidence. Cornell Voltage and Cornell Product criteria exhibited significant prognostic value in our Chinese cohort, aligning with prior epidemiological studies and supporting the ECG-LVH-mortality association for these two indices. Composite indices provided modest incremental risk reclassification by mitigating single-criterion limitations, though improvements in discrimination were small. External validation, structural cardiac assessment, and evaluation of calibration and clinical net benefit are required before these approaches can be considered for community screening.
CRediT authorship contribution statement
Jing Ma: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Ling Chen: Writing – original draft, Data curation. Hongmei Zhang: Data curation. Qin Wang: Investigation. Yan Wang: Visualization, Formal analysis. Jiguang Wang: Writing – review & editing, Supervision, Resources, Project administration, Funding acquisition, Conceptualization. Dingliang Zhu: Supervision, Conceptualization.
Funding
The study investigators were financially supported by grants from the National Natural Science Foundation of China (grants 82070435, 82270469 and 82370426), and National Health Commission (grant 2022YFC3601302), Beijing, China, and from the Shanghai Commissions of Science and Technology (grant 19DZ2340200), and Health(grant 2024ZZ1028 and a special grant for “leading academics” 2022LJ022), Shanghai, China, and the Key Discipline (2023–2025) of Public Health in Minhang District (MGWXK2023–06).
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Footnotes
Supplementary data to this article can be found online at https://doi.org/10.1016/j.ijcha.2026.102016.
Appendix A. Supplementary data
Supplementary material
References
- 1.Zhou M., et al. Mortality, morbidity, and risk factors in China and its provinces, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet. 2019;394(10204):1145–1158. doi: 10.1016/S0140-6736(19)30427-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Gardin J.M., et al. Sex, age, and disease affect echocardiographic left ventricular mass and systolic function in the free-living elderly. The Cardiovascular Health Study. Circulation. 1995;91(6):1739–1748. doi: 10.1161/01.cir.91.6.1739. [DOI] [PubMed] [Google Scholar]
- 3.Varagic J., Susic D., Frohlich E. Heart, aging, and hypertension. Curr. Opin. Cardiol. 2001;16(6):336–341. doi: 10.1097/00001573-200111000-00004. [DOI] [PubMed] [Google Scholar]
- 4.Clegg A., et al. Frailty in elderly people. Lancet. 2013;381(9868):752–762. doi: 10.1016/S0140-6736(12)62167-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Chen R., et al. Electrocardiographic left ventricular hypertrophy and mortality in an oldest-old hypertensive Chinese population. Clin. Interv. Aging. 2019;14:1657–1662. doi: 10.2147/CIA.S218106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Porthan K., et al. ECG left ventricular hypertrophy as a risk predictor of sudden cardiac death. Int. J. Cardiol. 2019;276:125–129. doi: 10.1016/j.ijcard.2018.09.104. [DOI] [PubMed] [Google Scholar]
- 7.Salles G.F., et al. Prognostic impact of baseline and serial changes in electrocardiographic left ventricular hypertrophy in resistant hypertension. Am. Heart J. 2010;159(5):833–840. doi: 10.1016/j.ahj.2010.02.012. [DOI] [PubMed] [Google Scholar]
- 8.van Kleef M., et al. Four ECG left ventricular hypertrophy criteria and the risk of cardiovascular events and mortality in patients with vascular disease. J. Hypertens. 2018;36(9):1865–1873. doi: 10.1097/HJH.0000000000001785. [DOI] [PubMed] [Google Scholar]
- 9.Du Z., et al. Complementary value of ECG and echocardiographic left ventricular hypertrophy for prediction of adverse outcomes in the general population. J. Hypertens. 2021;39(3):548–555. doi: 10.1097/HJH.0000000000002652. [DOI] [PubMed] [Google Scholar]
- 10.O’Neal W.T., et al. Electrocardiographic and echocardiographic left ventricular hypertrophy in the prediction of stroke in the elderly. J. Stroke Cerebrovasc. Dis. 2015;24(9):1991–1997. doi: 10.1016/j.jstrokecerebrovasdis.2015.04.044. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hancock E.W., et al. AHA/ACCF/HRS recommendations for the standardization and interpretation of the electrocardiogram: part V: electrocardiogram changes associated with cardiac chamber hypertrophy: a scientific statement from the American Heart Association Electrocardiography and Arrhythmias Committee, Council on Clinical Cardiology; the American College of Cardiology Foundation; and the Heart Rhythm Society. Endorsed by the International Society for Computerized Electrocardiology. J. Am. Coll. Cardiol. 2009;53(11):992–1002. doi: 10.1016/j.jacc.2008.12.015. [DOI] [PubMed] [Google Scholar]
- 12.Levy D., et al. Determinants of sensitivity and specificity of electrocardiographic criteria for left ventricular hypertrophy. Circulation. 1990;81(3):815–820. doi: 10.1161/01.cir.81.3.815. [DOI] [PubMed] [Google Scholar]
- 13.Okin P.M., et al. Combining ECG criteria for left ventricular hypertrophy improves risk prediction in patients with hypertension. J. Am. Heart Assoc. 2017;6(11) doi: 10.1161/JAHA.117.007564. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Jiang X., et al. Electrocardiographic criteria for the diagnosis of abnormal hypertensive cardiac phenotypes. J Clin Hypertens (Greenwich) 2019;21(3):372–378. doi: 10.1111/jch.13486. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Calderón A., et al. Detection of left ventricular hypertrophy by different electrocardiographic criteria in clinical practice. Findings from the Sara study. Clin. Exp. Hypertens. 2010;32(3):145–153. doi: 10.3109/10641960903254455. [DOI] [PubMed] [Google Scholar]
- 16.Wang D., et al. Performance of Electrocardiographic Criteria for Echocardiographically Diagnosed Left Ventricular Hypertrophy in Chinese Hypertensive Patients. Am. J. Hypertens. 2020;33(9):831–836. doi: 10.1093/ajh/hpaa083. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Hsieh B.P., Pham M.X., Froelicher V.F. Prognostic value of electrocardiographic criteria for left ventricular hypertrophy. Am. Heart J. 2005;150(1):161–167. doi: 10.1016/j.ahj.2004.08.041. [DOI] [PubMed] [Google Scholar]
- 18.Lu T.P., et al. Develop and Apply Electrocardiography-Based Risk Score to Identify Community-Based Elderly Individuals at High-Risk of Mortality. Front. Cardiovasc. Med. 2021;8 doi: 10.3389/fcvm.2021.738061. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Peguero J.G., et al. Electrocardiographic Criteria for the Diagnosis of Left Ventricular Hypertrophy. J. Am. Coll. Cardiol. 2017;69(13):1694–1703. doi: 10.1016/j.jacc.2017.01.037. [DOI] [PubMed] [Google Scholar]
- 20.Afify H.M.A., et al. Peguero Electrocardiographic Left Ventricular Hypertrophy Criteria and Risk of Mortality. Front. Cardiovasc. Med. 2018;5:75. doi: 10.3389/fcvm.2018.00075. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Zhao Y., et al. Electrocardiographic-left ventricular hypertrophy and incident stroke among Chinese hypertensive adults. J. Hum. Hypertens. 2020;34(4):286–292. doi: 10.1038/s41371-018-0155-x. [DOI] [PubMed] [Google Scholar]
- 22.Mancia G., et al. ESH guidelines for the management of arterial hypertension the task force for the management of arterial hypertension of the European Society of Hypertension: endorsed by the International Society of Hypertension (ISH) and the European Renal Association (ERA) J. Hypertens. 2023;41(12):1874–2071. doi: 10.1097/HJH.0000000000003480. [DOI] [PubMed] [Google Scholar]
- 23.Tanaka K., et al. Prognostic value of electrocardiographic left ventricular hypertrophy on cardiovascular risk in a non-hypertensive community-based population. Am. J. Hypertens. 2018;31(8):895–901. doi: 10.1093/ajh/hpy055. [DOI] [PubMed] [Google Scholar]
- 24.You Z., et al. Predictive value of electrocardiographic left ventricular hypertrophy in the general population: a meta-analysis. J. Electrocardiol. 2020;62:14–19. doi: 10.1016/j.jelectrocard.2020.07.001. [DOI] [PubMed] [Google Scholar]
- 25.Zhang H., Hu L., Wei X. Prognostic value of left ventricular hypertrophy in hypertensive patients: a meta-analysis of electrocardiographic studies. J Clin Hypertens (Greenwich) 2020;22(2):254–260. doi: 10.1111/jch.13795. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Sun G.Z., et al. Assessment of novel Peguero-Lo Presti electrocardiographic left ventricular hypertrophy criteria in a large Asian population: newer may not be better. Can. J. Cardiol. 2018;34(9):1153–1157. doi: 10.1016/j.cjca.2018.05.013. [DOI] [PubMed] [Google Scholar]
- 27.Narita M., et al. Novel electrocardiographic criteria for the diagnosis of left ventricular hypertrophy in the Japanese general population. Int. Heart J. 2019;60(3):679–687. doi: 10.1536/ihj.18-511. [DOI] [PubMed] [Google Scholar]
- 28.Monzo L., et al. Electrocardiographic voltage criteria in patients with hypertrophic cardiomyopathy. J. Cardiovasc. Med. (Hagerstown) 2020;21(9):696–703. doi: 10.2459/JCM.0000000000001034. [DOI] [PubMed] [Google Scholar]
- 29.Brooks J.E., Soliman E.Z., Upadhya B. Is left ventricular hypertrophy a valid therapeutic target? Curr. Hypertens. Rep. 2019;21(6):47. doi: 10.1007/s11906-019-0952-9. [DOI] [PubMed] [Google Scholar]
- 30.Okin P.M., et al. Regression of electrocardiographic left ventricular hypertrophy during antihypertensive treatment and the prediction of major cardiovascular events. JAMA. 2004;292(19):2343–2349. doi: 10.1001/jama.292.19.2343. [DOI] [PubMed] [Google Scholar]
- 31.Iribarren C., et al. Cohort study of ECG left ventricular hypertrophy trajectories: ethnic disparities, associations with cardiovascular outcomes, and clinical utility. J. Am. Heart Assoc. 2017;6(10) doi: 10.1161/JAHA.116.004954. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary material