Skip to main content
NCBI home page
Annals of Noninvasive Electrocardiology logoLink to Annals of Noninvasive Electrocardiology
. 2019 Feb 5;24(4):e12635. doi: 10.1111/anec.12635

Diagnostic performance of electrocardiographic criteria for left ventricular hypertrophy among various body mass index groups compared to diagnosis by cardiac magnetic resonance imaging

Veerawat Nomsawadi 1, Rungroj Krittayaphong 2,✉
PMCID: PMC6931489  PMID: 30719815

Abstract

Background

Presence of left ventricular hypertrophy (LVH) increases the risk for cardiovascular event. Many electrocardiographic (ECG) criteria can be used to diagnose LVH; however, high body mass index (BMI) may reduce ECG amplitudes.The aim of this study was to investigate the diagnostic performance of ECG criteria for diagnosing LVH among various BMI groups compared to diagnosis by cardiac magnetic resonance (CMR) imaging.

Methods

Patients who were referred for CMR were enrolled. CMR and ECG were performed on the same day. Left ventricular function, volume, and mass were calculated from CMR. Standard ECG criteria were measured, including: Cornell voltage, Cornell product, Romhilt‐Estes point score system, Sokolow‐Lyon index, and Sokolow‐Lyon‐Rappaport index. Diagnostic performance of each ECG criterion was calculated and analyzed in the following four BMI groups: underweight (<18.5 kg/m2), normal (18.5–22.9 kg/m2), overweight (23–24.9 kg/m2), and obese (≥25 kg/m2).

Results

Of the 1,882 patients that were included, 67 were underweight, 459 were normal weight, 434 were overweight, and 922 were obese. LVH was diagnosed in 34 (50.7%) underweight, 144 (31.4%) normal weight, 100 (23.0%) overweight, and 181 (19.6%) obese patients. Overall specificity of ECG was high (0.89–0.95), and overall sensitivity was low (0.25–0.37). The specificity of each ECG criterion was similar among BMI groups; however, the sensitivity of ECG criteria demonstrated a decreasing trend in the higher BMI groups.

Conclusion

All ECG criteria demonstrated relatively high specificity and relatively low sensitivity. Although the specificity across groups remained similar, higher BMI was found to be associated with decreased sensitivity.

Keywords: body mass index, diagnostic performance, electrocardiography, left ventricular hypertrophy, magnetic resonance imaging

1. INTRODUCTION

Left ventricular hypertrophy (LVH) is a common consequence (Dzau & Braunwald, 1991) of uncontrolled hypertension, and it contributes to increased cardiovascular events in both hypertensive patients (Jissho et al., 2010; Vakili, Okin, & Devereux, 2001) and in general population (Bikkina et al., 1994). The risk of developing a major adverse cardiac event increases commensurately with increases in left ventricular (LV) mass (Krittayaphong et al., 2009). LVH reflects increased left ventricular mass, which contributes to increased myocardial oxygen demand, and which may cause inadequate blood supply to the myocardium and may also cause myocardial ischemia or infarction (Aguilar et al., 2004). Treatments that reduce LV mass are associated with lower morbidity and mortality (Koren, Ulin, Koren, Laragh, & Devereux, 2002; Verdecchia et al., 2003, 1998).

The most commonly used tool for diagnosis of LVH is 12‐lead electrocardiography (ECG), which is simple, low‐cost, and widely available (Estes & Jackson, 2009). Many ECG criteria have demonstrated varying levels of accuracy relative to their ability to diagnose LVH (Krittayaphong et al., 2013; Schlegel et al., 2010; Xie & Wang, 2010), including Cornell voltage (Casale, Devereux, Alonso, Campo, & Kligfield, 1987; Casale et al., 1985), Cornell product (Ishikawa et al., 2009), Romhilt‐Estes point score system (Romhilt et al., 1969; Romhilt & Estes, 1968), Sokolow‐Lyon index (Sokolow & Lyon, 1949), and Sokolow‐Lyon‐Rappaport index (Levy et al., 1990). Echocardiography is a commonly used comparative method due to its wide availability; however, echocardiography has significant inter‐observer variability (Dai, Ayres, Harrist, Bricker, & Labarthe, 1999). Cardiac magnetic resonance (CMR) imaging is a gold standard investigation for assessing left ventricular volume, left ventricular mass, and left ventricular ejection fraction (LVEF) due to its high image resolution, its three‐dimensional image acquisition, and the fact that it has less intra‐ and inter‐observer variability compared to echocardiogram (Buchner et al., 2009; Mor‐Avi et al., 2004; Myerson, Bellenger, & Pennell, 2002).

Various ECG criteria contain R, S, and/or QRS wave amplitudes as components of their criteria (Casale et al., 1985; Levy et al., 1990; Molloy, Okin, Devereux, & Kligfield, 1992; Romhilt & Estes, 1968; Sokolow & Lyon, 1949). However, there are many factors that influence ECG amplitudes, including chest wall thickness. Higher body mass index (BMI) was shown to be associated with lower ECG amplitudes in patients with similar LV mass (Nasir, Rubal, Jones, & Shah, 2012). Prior study showed the sensitivity of Cornell voltage product to be significantly lower in obese normotensive patients than in non‐obese normotensive patients (Norman & Levy, 1996). However, data relating to the performance of ECG criteria for diagnosing LVH among four BMI groups is scarce. Accordingly, the aim of this study was to investigate the diagnostic performance of ECG criteria for diagnosing LVH among various BMI groups compared to diagnosis by CMR imaging.

2. METHODS

2.1. Study population

This study included patients >18 years of age that underwent both ECG and CMR on the same day at Faculty of Medicine Siriraj Hospital, Mahidol University, during the 2005–2009 study period. Siriraj Hospital is a 2,300‐bed national tertiary referral hospital that is located in Bangkok, Thailand. Patients unable to complete the CMR examination or with known contraindication for CMR, such as pacemaker, internal defibrillator, or intracranial clip, were excluded. Patients with unstable clinical conditions, claustrophobia, conduction abnormalities (e.g., Wolff‐Parkinson‐White syndrome), complete left bundle branch block, or right bundle branch block were also excluded. One hundred and eighty‐four healthy volunteers that underwent CMR comprised the control group for this study. Control group CMR results were used to establish the diagnosis threshold for LVH by CMR. The protocol for this study was approved by the Siriraj Institutional Review Board (SIRB), and written informed consent was obtained from all study participants.

2.2. CMR protocol

Cardiac magnetic resonance for left ventricular function, volume, and mass was performed using a Philips Gyroscan NT 1.5 T MRI Scanner (Philips Medical Systems, Best, the Netherlands). Cine images and spin echo images were performed using a steady‐state free‐precession (SSFP) technique. Images were acquired in 4‐chamber, vertical long‐axis, horizontal long‐axis, and multiple‐slice short‐axis series. Functional images were developed according to the following parameters: echo time/repetition time/number of excitations = 1.8/3.7/2; 256 × 240 matrix; 390 × 312 mm field of view; 8 mm slice thickness; 1.52 × 1.21 reconstruction pixel; and, 70 degree flip angle.

2.3. Analysis of CMR images

Cardiac magnetic resonance images were analyzed on a View Forum Workstation (Philips Medical Systems). Left ventricular volume was measured and calculated in end‐systolic (left ventricular end‐systolic volume, LVESV) and end‐diastolic phase (left ventricular end‐diastolic volume, LVEDV). Left ventricular mass (LVMASS) and LVEF were also calculated. Indices of LVEDV, LVESV, and LVMASS (LVEDVI, LVESVI, and LVMASSI) were calculated by index to body surface area.

Patients were considered to have LVH when their LVMASSI was above 95% of the LVMASSI (Solberg, 1987). Similarly, we used the 95th percentile for LVMASSI in the healthy population as the cut‐off value for diagnosis of LVH by CMR (Krittayaphong, Saiviroonporn, Boonyasirinant, Nakyen, & Kangkagate, 2004).

2.4. Analysis of 12‐lead ECG results

Twelve‐lead ECG was performed on the same day prior to CMR examination. ECG output data was recorded at a paper speed of 25 mm/s and 10 mm/mV with the patient resting in the supine position and breathing calmly. ECG results were interpreted by experienced investigators who were unaware of the patient's clinical information and CMR results. The following ECG criteria were used in this study: (a) Cornell voltage (Casale et al., 1987, 1985); (b) Cornell product (Ishikawa et al., 2009); (c) Romhilt‐Estes point score system of at least five points (Romhilt et al., 1969; Romhilt & Estes, 1968); (d) Sokolow‐Lyon index (Sokolow & Lyon, 1949); and (e) Sokolow‐Lyon‐Rappaport index (Levy et al., 1990). The criteria for the diagnosis of LVH for each of the aforementioned ECG criteria are presented in Supporting Information Table S1.

2.5. Statistical analysis

SPSS Statistics version 20 (SPSS, Inc., Chicago, IL, USA) was used to perform all data analyses. Continuous data are reported as mean ± standard deviation (SD), and categorical data are reported as number and percentage. Continuous data were compared using independent t‐test or analysis of variance (ANOVA), and categorical data were compared using chi‐square test. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were calculated. Patients were classified into four groups according to BMI: underweight (BMI: <18.5 kg/m2), normal weight (BMI: 18.5–22.9 kg/m2) group, overweight (BMI: 23.0–24.9 kg/m2) and obese (BMI: ≥25 kg/m2) group. Chi‐square test for trend was used to evaluate sensitivity and specificity trends among different BMI groups when using the same ECG criteria. Non‐parametric test for related samples was used to compare the sensitivity and specificity of each ECG criterion in the same BMI group. A p‐value < 0.05 was considered to be statistically significant.

3. RESULTS

3.1. Demographic data

Two thousand two hundred and twenty‐six (2,226) patients were referred for CMR during 2005–2009. After the exclusion of 344 patients, the remaining 1,882 patients were enrolled in this study. The baseline demographic and clinical characteristics of each BMI group are given in Table 1. Sixty‐seven (3.6%) patients were in the underweight group, 459 (24.4%) patients were in the normal weight group, 434 (23.1%) patients were in the overweight group, and 922 (49.0%) patients were in the obese group. LVH was diagnosed in 459 patients (24.4%), including 34 underweight patients (50.7%), 144 normal weight patients (31.4%), 100 overweight patients (23.0%), and 181 obese patients (19.6%). The higher BMI groups had a significantly higher prevalence of diabetes mellitus and hypertension. The lower BMI groups had a significantly higher prevalence of history of myocardial infarction and abnormal wall motion on CMR.

Table 1.

Baseline demographic and clinical characteristics

Characteristics Total (n = 1,882) Underweight (n = 67) Normal (n = 459) Overweight (n = 434) Obese (n = 922) p‐Value
Male gender 994 (52.8%) 36 (53.7%) 247 (53.8%) 249 (57.4%) 462 (50.1%) 0.088
Age (years) 64.5 ± 11.3 66.2 ± 16.1 65.2 ± 12.0 64.5 ± 11.2 64.1 ± 10.6 <0.001
Weight (kg) 65.2 ± 12.3 45.3 ± 4.5 54.7 ± 6.5 62.5 ± 6.3 73.1 ± 10.9 <0.001
Height (cm) 160.4 ± 8.4 161.2 ± 7.1 160.5 ± 8.4 161.1 ± 8.0 160.0 ± 8.7 0.123
BSA (m2) 1.7 ± 0.2 1.4 ± 0.1 1.6 ± 0.1 1.7 ± 0.1 1.8 ± 0.2 <0.001
BMI (kg/m2) 25.3 ± 4.1 17.4 ± 0.9 21.2 ± 1.2 24.0 ± 0.6 28.5 ± 3.1 <0.001
SBP (mmHg) 137.8 ± 23.8 144.0 ± 30.8 138.0 ± 25.9 138.3 ± 22.7 137.0 ± 22.8 <0.001
DBP (mmHg) 75.7 ± 13.3 75.8 ± 15.6 75.5 ± 13.3 76.9 ± 12.5 75.3 ± 13.5 0.265
Current smoker 374 (19.9%) 17 (25.4%) 101 (22.0%) 80 (18.4%) 176 (19.1%) 0.320
Hypercholesterolemia 1,206 (64.1%) 42 (62.7%) 276 (60.1%) 284 (65.4%) 604 (65.5%) 0.228
Diabetes mellitus 663 (35.2%) 18 (26.9%) 130 (28.3%) 139 (32.0%) 376 (40.8%) <0.001
Hypertension 1,177 (62.5%) 35 (52.2%) 246 (53.6%) 258 (59.4%) 638 (69.2%) <0.001
History of MI 272 (14.5%) 16 (23.9%) 76 (16.6%) 65 (15.0%) 115 (12.5%) 0.024
History of angina 92 (4.9%) 4 (6.0%) 21 (4.6%) 24 (5.5%) 43 (4.7%) 0.861
History of PCI 182 (9.7%) 10 (14.9%) 42 (9.2%) 41 (9.4%) 89 (9.7%) 0.515
History of CABG 116 (6.2%) 5 (7.5%) 29 (6.3%) 30 (6.9%) 52 (5.6%) 0.784
History of DOE 863 (45.9%) 37 (55.2%) 214 (46.6%) 181 (41.7%) 431 (46.7%) 0.123
Medications
Beta blocker 944 (50.2%) 34 (50.7%) 220 (47.9%) 221 (50.9%) 469 (50.9%) 0.751
CCB 424 (22.5%) 14 (20.9%) 102 (22.2%) 85 (19.6%) 223 (24.2%) 0.292
Nitrate 679 (36.1%) 33 (49.3%) 162 (35.3%) 154 (35.5%) 330 (35.8%) 0.153
ASA 1,174 (62.4%) 40 (59.7%) 278 (60.6%) 264 (60.8%) 592 (64.2%) 0.456
Clopidogrel 68 (3.6%) 6 (9.0%) 12 (2.6%) 23 (5.3%) 27 (2.9%) 0.009
ACEI 612 (32.5%) 23 (34.3%) 145 (31.6%) 120 (27.6%) 324 (35.1%) 0.049
ARB 163 (8.7%) 6 (9.0%) 26 (5.7%) 41 (9.4%) 90 (9.8%) 0.073
Statins 1,057 (56.2%) 40 (59.7%) 222 (48.4%) 250 (57.6%) 545 (59.1%) 0.002
CMR variables
LVH 459 (24.4%) 34 (50.7%) 144 (31.4%) 100 (23.0%) 181 (19.6%) <0.001
LVEF (%) 60.5 ± 18.5 49.0 ± 19.5 56.6 ± 20.0 59.9 ± 18.8 63.5 ± 16.8 <0.001
Abnormal wall motion 686 (36.5%) 40 (59.7%) 197 (42.9%) 158 (36.4%) 291 (31.6%) <0.001
LVEDVI (ml/m2) 75.7 ± 35.3 99.8 ± 48.1 84.8 ± 41.3 76.7 ± 37.7 69.0 ± 26.9 <0.001
LVESVI (ml/m2) 35.5 ± 35.9 56.5 ± 49.2 44.1 ± 41.4 36.4 ± 37.8 29.2 ± 28.6 <0.001
LVMASSI (g/m2) 55.6 ± 20.8 65.9 ± 25.5 58.4 ± 22.8 55.5 ± 20.8 53.5 ± 18.9 <0.001
Open in a new tab

Data presented as mean ± standard deviation or number and percentage.

A p‐value < 0.05 indicates statistical significance (bold and italic values).

ACEI: angiotensin converting enzyme inhibitor; ARB: angiotensin receptor blocker; ASA: acetylsalicylic acid; BMI: body mass index; BSA: body surface area; CABG: coronary artery bypass graft; CCB: calcium channel blocker; DBP: diastolic blood pressure; DOE: dyspnea on exertion; LVEDVI: left ventricular end‐diastolic volume index; LVEF: left ventricular ejection fraction; LVESVI: left ventricular end‐systolic volume index; LVH: left ventricular hypertrophy; LVMASSI: left ventricular mass index; MI: myocardial infarction; PCI: percutaneous coronary intervention; SBP: systolic blood pressure.

3.2. Diagnostic performance of ECG criteria for LVH diagnosis

Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of each ECG criterion are shown in Table 2. Overall sensitivity and specificity of each ECG criterion is presented in Figure 1. The overall specificity of ECG for diagnosing LVH was higher than the overall sensitivity. The three parameters with the highest sensitivity were Cornell product, Romhilt‐Estes point score system, and Sokolow‐Lyon‐Rappaport index—all of which were greater than 0.360 (p > 0.05). The three parameters with the highest specificity were Cornell voltage, Cornell product, and Sokolow‐Lyon index—all of which were greater than 0.90 (p > 0.05).

Table 2.

Diagnostic performance of five evaluated ECG criteria

Criteria BMI group Sensitivity Specificity PPV NPV Accuracy
Cornell voltage Overall 0.29 (0.25–0.33) 0.94 (0.93–0.96) 0.62 (0.56–0.69) 0.80 (0.78–0.82) 0.78 (0.77–0.80)
Cornell voltage Underweight 0.41 (0.26–0.58) 0.97 (0.85–0.99) 0.93 (0.70–0.99) 0.62 (0.48–0.74) 0.69 (0.57–0.79)
Cornell voltage Normal 0.33 (0.26–0.41) 0.94 (0.91–0.96) 0.73 (0.61–0.82) 0.76 (0.71–0.80) 0.75 (0.71–0.79)
Cornell voltage Overweight 0.26 (0.18–0.35) 0.93 (0.90–0.96) 0.54 (0.40–0.67) 0.88 (0.77–0.84) 0.78 (0.74–0.82)
Cornell voltage Obesity 0.24 (0.18–0.31) 0.95 (0.93–0.96) 0.53 (0.42–0.67) 0.84 (0.81–0.86) 0.81 (0.78–0.83)
Cornell product Overall 0.37 (0.33–0.42) 0.93 (0.92–0.94) 0.64 (0.58–0.69) 0.82 (0.80–0.84) 0.80 (0.78–0.81)
Cornell product Underweight 0.47 (0.32–0.63) 0.94 (0.80–0.98) 0.89 (0.67–0.970) 0.63 (0.49–0.75) 0.70 (0.58–0.80)
Cornell product Normal 0.40 (0.32–0.48) 0.94 (0.91–0.96) 0.75 (0.64–0.83) 0.77 (0.728–0.812) 0.77 (0.73–0.81)
Cornell product Overweight 0.34 (0.26–0.44) 0.934 (0.90–0.96) 0.61 (0.48–0.72) 0.83 (0.78–0.86) 0.80 (0.76–0.83)
Cornell product Obesity 0.35 (0.28–0.42) 0.93 (0.91–0.95) 0.54 (0.45–0.63) 0.85 (0.83–0.88) 0.82 (0.79–0.838)
Romhilt Overall 0.37 (0.33–0.42) 0.89 (0.87–0.90) 0.52 (0.46–0.57) 0.81 (0.79–0.83) 0.76 (0.74–0.78)
Romhilt Underweight 0.50 (0.34–0.66) 0.94 (0.80–0.98) 0.90 (0.69–0.97) 0.65 (0.50–0.77) 0.72 (0.60–0.81)
Romhilt Normal 0.40 (0.33–0.48) 0.88 (0.84–0.91) 0.60 (0.50–0.70) 0.76 (0.72–0.80) 0.73 (0.69–0.77)
Romhilt Overweight 0.36 (0.27–0.46) 0.87 (0.82–0.90) 0.44 (0.34–0.55) 0.82 (0.78–0.86) 0.75 (0.71–0.79)
Romhilt Obesity 0.33 (0.26–0.40) 0.90 (0.88–0.92) 0.44 (0.36–0.53) 0.85 (0.82–0.87) 0.79 (0.76–0.81)
Sokolow Overall 0.25 (0.21–0.29) 0.95 (0.94–0.96) 0.62 (0.55–0.69) 0.80 (0.78–0.82) 0.80 (0.76–0.80)
Sokolow Underweight 0.35 (0.22–0.52) 0.91 (0.76–0.97) 0.80 (0.55–0.93) 0.58 (0.44–0.70) 0.627 (0.51–0.73)
Sokolow Normal 0.24 (0.18–0.32) 0.95 (0.92–0.97) 0.67 (0.54–0.79) 0.73 (0.69–0.77) 0.73 (0.68–0.76)
Sokolow Overweight 0.28 (0.20–0.38) 0.93 (0.90–0.95) 0.55 (0.41–0.68) 0.81 (0.77–0.85) 0.78 (0.74–0.82)
Sokolow Obesity 0.21 (0.16–0.28) 0.96 (0.95–0.98) 0.59 (0.46–0.70) 0.83 (0.81–0.86) 0.82 (0.79–0.84)
Rappaport Overall 0.37 (0.33–0.42) 0.91 (0.89–0.92) 0.560 (0.50–0.61) 0.818 (0.80–0.84) 0.78 (0.76–0.80)
Rappaport Underweight 0.53 (0.37–0.69) 0.79 (0.62–0.89) 0.72 (0.52–0.86) 0.62 (0.47–0.75) 0.66 (0.54–0.76)
Rappaport Normal 0.40 (0.32–0.49) 0.89 (0.85–0.92) 0.62 (0.52–0.72) 0.77 (0.73–0.81) 0.74 (0.70–0.78)
Rappaport Overweight 0.37 (0.28–0.47) 0.89 (0.85–0.92) 0.50 (0.39–0.61) 0.82 (0.78–0.86) 0.77 (0.73–0.81)
Rappaport Obesity 0.32 (0.26–0.39) 0.93 (0.90–0.94) 0.51 (0.42–0.60) 0.85 (0.82–0.87) 0.81 (0.78–0.83)
Open in a new tab

Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy are presented as the calculated value (95% confidence interval).

Rappaport: Sokolow‐Lyon‐Rappaport index; Romhilt: Romhilt‐Estes point score system at score ≥5 for diagnosis of LVH; Sokolow: Sokolow‐Lyon index.

Figure 1.

Figure 1

Open in a new tab

Overall sensitivity (a) and specificity (b) of ECG criteria (*p < 0.05)

The sensitivity and specificity of the five investigated ECG criteria compared among the four different BMI groups are shown in Figure 2. Chi‐square test for trend revealed no significant difference in the specificity of each ECG criterion when compared among the four BMI groups, except for Sokolow‐Lyon‐Rappaport index (p = 0.006). In contrast, chi‐square test for trend showed the sensitivities of Cornell voltage, Romhilt‐Estes point score system, and Sokolow‐Lyon‐Rappaport index to be significantly decreased when BMIs increased. No statistically significant difference was observed for Cornell product or Sokolow‐Lyon index.

Figure 2.

Figure 2

Open in a new tab

Sensitivity (a) and specificity (b) of ECG criteria among BMI groups (*p < 0.05)

Additional analysis was performed that compared the sensitivity and specificity of each ECG criteria between patients with overweight or obese status (BMI ≥23 kg/m2) and patients with normal or underweight status. The results of that analysis are shown in Figure 3. The three most highly sensitive ECG criteria in patients with overweight or obese status were Cornell product, Romhilt‐Estes point score system, and Sokolow‐Lyon‐Rappaport index. The three most highly specific ECG criteria were Cornell voltage, Cornell product, and Sokolow‐Lyon index.

Figure 3.

Figure 3

Open in a new tab

Comparison of sensitivities (a) and specificities (b) among the evaluated ECG criteria for each BMI group (*p < 0.05)

4. DISCUSSION

The present study set forth to investigate the effect of BMI on the diagnostic performance of ECG criteria. To do so, we divided study participants into the following four BMI groups: underweight, normal, overweight, and obese. The results showed a similar specificity, but a decrease in sensitivity in the increasing BMI groups. The three criteria with the highest sensitivity were Cornell product, Romhilt‐Estes point score system, and Sokolow‐Lyon‐Rappaport index. The three criteria with the highest specificity were Cornell voltage, Cornell product, and Sokolow‐Lyon index.

In the present study, the overall specificity was high (0.888–0.951), and the overall sensitivity was low (0.246–0.373) for all ECG criteria, which corresponds with the results reported from previous studies (Buchner et al., 2009; Norman & Levy, 1996; Rodrigues et al., 2016). Norman and Levy (1996) used echocardiography as the gold standard, and they found Cornell product to have a higher sensitivity than Cornell voltage. They also showed that the sensitivity of Cornell product decreases in obesity after fixing the specificity at 95%. Similarly, we found that obesity decreased the sensitivity not only for Cornell voltage, but also for Romhilt‐Estes point score system and Sokolow‐Lyon‐Rappaport index. The sensitivity of Cornell product was slightly higher than the sensitivity of Cornell voltage in our study (p > 0.05).

Rodrigues et al. (2016) compared different ECG criteria for LVH in obese and non‐obese patients and found that obesity decreased both sensitivity and specificity. Our study showed a similar result for sensitivity, but obesity had no effect on specificity. The Rodrigues et al. study recruited patients from hypertensive population who had undergone CMR as part of the investigation for secondary causes of hypertension, and to quantify hypertensive end‐organ damage. In contrast, we recruited from a population of patients who had undergone CMR for any clinical purpose. This difference would naturally result in differences between studies relative to patients’ characteristics and underlying heart diseases.

The lower sensitivity in the high BMI groups was attributed to an increased amount of subcutaneous adipose tissue that increases the distance between the ECG recording electrodes and the heart (Horton, Sherber, & Lakatta, 1977; Nasir et al., 2012). It is, therefore, possible that undetected LVH could exist in overweight and obese patients, and this could adversely affect risk stratification, self‐awareness, and treatment strategy. Further study to adjust ECG criteria for diagnosis of LVH in higher BMI groups, and to improve the diagnostic performance of those criteria should be considered.

In overweight and obese population, our results revealed Cornell product, Romhilt‐Estes point score system, and Sokolow‐Lyon‐Rappaport index to be the most sensitive ECG criteria, and Cornell voltage, Cornell product, and Sokolow‐Lyon index to be the most specific ECG criteria for diagnosing LVH. Cornell product should, therefore, be regarded as the ECG criteria of choice in overweight and obese population. This recommendation is consistent with that proposed by Okin, Jern, Devereux, Kjeldsen, and Dahlof (2000). The main strength of this study compared to the aforementioned studies is the large size of our study population, and the fact that we used CMR as the gold standard for assessment of LVH.

4.1. Limitations

First, the patients included in the present study were patients who underwent CMR for clinical purposes at Siriraj Hospital, which means that a higher prevalence of LVH and other cardiac abnormalities would be observed in our study population than the prevalence of these conditions that would be found in normal population. Therefore, further study in general population needs to be conducted to confirm the generalizability of these findings. Second, increasing BMI may not always indicate increased subcutaneous adipose tissue (Etchison et al., 2011). For example, athletes and body builders have higher BMIs due to a higher percentage of lean muscle mass. Further investigation using skinfold thickness (Durnin & Womersley, 1974) or dual‐energy X‐ray absorptiometry (DEXA; Kelly, Berger, & Richardson, 1998) should be conducted to determine the effect of subcutaneous tissue and its association with ECG diagnostic performance.

5. CONCLUSION

All of the ECG criteria evaluated in this study demonstrated relatively high specificity and relatively low sensitivity. Although the specificity across groups remained similar, higher BMI was found to be significantly associated with decreased sensitivity. The results of this study suggest Cornell product as the ECG criteria of choice in overweight or obese population due to its relatively higher sensitivity and specificity compared to the other evaluated ECG criteria.

CONFLICT OF INTEREST

Both authors declare no personal or professional conflicts of interest relating to any aspect of this study.

Supporting information

 

Click here for additional data file. (34.1KB, docx)

ACKNOWLEDGEMENTS

The authors gratefully acknowledge Mr. Suthipol Udompunthurak and Ms. Khemajira Karaketklang for their assistance with statistical consultation and analysis.

Nomsawadi V, Krittayaphong R. Diagnostic performance of electrocardiographic criteria for left ventricular hypertrophy among various body mass index groups compared to diagnosis by cardiac magnetic resonance imaging. Ann Noninvasive Electrocardiol. 2019;24:e12635 10.1111/anec.12635

REFERENCES

  1. Aguilar, D. , Goldhaber, S. Z. , Gans, D. J. , Levey, A. S. , Porush, J. G. , Lewis, J. B. , … Pfeffer, M. A. (2004). Clinically unrecognized Q‐wave myocardial infarction in patients with diabetes mellitus, systemic hypertension, and nephropathy. American Journal of Cardiology, 94(3), 337–339. 10.1016/j.amjcard.2004.04.028 [DOI] [PubMed] [Google Scholar]
  2. Bikkina, M. , Levy, D. , Evans, J. C. , Larson, M. G. , Benjamin, E. J. , Wolf, P. A. , & Castelli, W. P. (1994). Left ventricular mass and risk of stroke in an elderly cohort. The Framingham Heart Study. JAMA, 272(1), 33–36. 10.1001/jama.1994.03520010045030 [DOI] [PubMed] [Google Scholar]
  3. Buchner, S. , Debl, K. , Haimerl, J. , Djavidani, B. , Poschenrieder, F. , Feuerbach, S. , … Luchner, A. (2009). Electrocardiographic diagnosis of left ventricular hypertrophy in aortic valve disease: Evaluation of ECG criteria by cardiovascular magnetic resonance. Journal of Cardiovascular Magnetic Resonance, 11(1), 18 10.1186/1532-429x-11-18 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Casale, P. N. , Devereux, R. B. , Alonso, D. R. , Campo, E. , & Kligfield, P. (1987). Improved sex‐specific criteria of left ventricular hypertrophy for clinical and computer interpretation of electrocardiograms: Validation with autopsy findings. Circulation, 75(3), 565–572. 10.1161/01.CIR.75.3.565 [DOI] [PubMed] [Google Scholar]
  5. Casale, P. N. , Devereux, R. B. , Kligfield, P. , Eisenberg, R. R. , Miller, D. H. , Chaudhary, B. S. , & Phillips, M. C. (1985). Electrocardiographic detection of left ventricular hypertrophy: Development and prospective validation of improved criteria. Journal of the American College of Cardiology, 6(3), 572–580. 10.1016/S0735-1097(85)80115-7 [DOI] [PubMed] [Google Scholar]
  6. Dai, S. , Ayres, N. A. , Harrist, R. B. , Bricker, J. T. , & Labarthe, D. R. (1999). Validity of echocardiographic measurement in an epidemiological study. Project HeartBeat!. Hypertension, 34(2), 236–241. 10.1161/01.HYP.34.2.236 [DOI] [PubMed] [Google Scholar]
  7. Durnin, J. V. , & Womersley, J. (1974). Body fat assessed from total body density and its estimation from skinfold thickness: Measurements on 481 men and women aged from 16 to 72 years. British Journal of Nutrition, 32(1), 77–97. 10.1079/BJN19740060 [DOI] [PubMed] [Google Scholar]
  8. Dzau, V. , & Braunwald, E. (1991). Resolved and unresolved issues in the prevention and treatment of coronary artery disease: A workshop consensus statement. American Heart Journal, 121(4 Pt 1), 1244–1263. 10.1016/0002-8703(91)90694-D [DOI] [PubMed] [Google Scholar]
  9. Estes, E. H. Jr , & Jackson, K. P. (2009). The electrocardiogram in left ventricular hypertrophy: Past and future. Journal of Electrocardiology, 42(6), 589–592. 10.1016/j.jelectrocard.2009.06.016 [DOI] [PubMed] [Google Scholar]
  10. Etchison, W. C. , Bloodgood, E. A. , Minton, C. P. , Thompson, N. J. , Collins, M. A. , Hunter, S. C. , & Dai, H. (2011). Body mass index and percentage of body fat as indicators for obesity in an adolescent athletic population. Sports Health, 3(3), 249–252. 10.1177/1941738111404655 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Horton, J. D. , Sherber, H. S. , & Lakatta, E. G. (1977). Distance correction for precordial electrocardiographic voltage in estimating left ventricular mass: An echocardiographic study. Circulation, 55(3), 509–512. 10.1161/01.CIR.55.3.509 [DOI] [PubMed] [Google Scholar]
  12. Ishikawa, J. , Ishikawa, S. , Kabutoya, T. , Gotoh, T. , Kayaba, K. , Schwartz, J. E. , … Kario, K. (2009). Cornell product left ventricular hypertrophy in electrocardiogram and the risk of stroke in a general population. Hypertension, 53(1), 28–34. 10.1161/hypertensionaha.108.118026 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Jissho, S. , Shimada, K. , Taguchi, H. , Yoshida, K. , Fukuda, S. , Tanaka, H. , … Goto, Y. (2010). Impact of electrocardiographic left ventricular hypertrophy on the occurrence of cardiovascular events in elderly hypertensive patients. The Japanese trial to assess optimal systolic blood pressure in elderly hypertensive patients (JATOS). Circulation Journal, 74(5), 938–945. 10.1253/circj.CJ-09-0722 [DOI] [PubMed] [Google Scholar]
  14. Kelly, T. L. , Berger, N. , & Richardson, T. L. (1998). DXA body composition: Theory and practice. Applied Radiation and Isotopes, 49(5–6), 511–513. 10.1016/S0969-8043(97)00226-1 [DOI] [PubMed] [Google Scholar]
  15. Koren, M. J. , Ulin, R. J. , Koren, A. T. , Laragh, J. H. , & Devereux, R. B. (2002). Left ventricular mass change during treatment and outcome in patients with essential hypertension. American Journal of Hypertension, 15(12), 1021–1028. 10.1016/S0895-7061(02)03061-3 [DOI] [PubMed] [Google Scholar]
  16. Krittayaphong, R. , Boonyasirinant, T. , Saiviroonporn, P. , Thanapiboonpol, P. , Nakyen, S. , Ruksakul, K. , & Udompunturak, S. (2009). Prognostic significance of left ventricular mass by magnetic resonance imaging study in patients with known or suspected coronary artery disease. Journal of Hypertension, 27(11), 2249–2256. 10.1097/HJH.0b013e3283309ac4 [DOI] [PubMed] [Google Scholar]
  17. Krittayaphong, R. , Nomsawadi, V. , Muenkaew, M. , Miniphan, M. , Yindeengam, A. , & Udompunturak, S. (2013). Accuracy of ECG criteria for the diagnosis of left ventricular hypertrophy: A comparison with magnetic resonance imaging. Journal of the Medical Association of Thailand, 96(Suppl 2), S124–132. [PubMed] [Google Scholar]
  18. Krittayaphong, R. , Saiviroonporn, P. , Boonyasirinant, T. , Nakyen, S. , & Kangkagate, C. (2004). Gender differences on the left and right ventricular volume, systolic function and mass assessed by cardiac magnetic resonance imaging. Thai Heart Journal, 17, 171–178. [Google Scholar]
  19. Levy, D. , Labib, S. B. , Anderson, K. M. , Christiansen, J. C. , Kannel, W. B. , & Castelli, W. P. (1990). Determinants of sensitivity and specificity of electrocardiographic criteria for left ventricular hypertrophy. Circulation, 81(3), 815–820. 10.1161/01.CIR.81.3.815 [DOI] [PubMed] [Google Scholar]
  20. Molloy, T. J. , Okin, P. M. , Devereux, R. B. , & Kligfield, P. (1992). Electrocardiographic detection of left ventricular hypertrophy by the simple QRS voltage‐duration product. Journal of the American College of Cardiology, 20(5), 1180–1186. 10.1016/0735-1097(92)90376-X [DOI] [PubMed] [Google Scholar]
  21. Mor‐Avi, V. , Sugeng, L. , Weinert, L. , MacEneaney, P. , Caiani, E. G. , Koch, R. , … Lang, R. M. (2004). Fast measurement of left ventricular mass with real‐time three‐dimensional echocardiography. Comparison with magnetic resonance imaging. Circulation, 110(13), 1814–1818. 10.1161/01.cir.0000142670.65971.5f [DOI] [PubMed] [Google Scholar]
  22. Myerson, S. G. , Bellenger, N. G. , & Pennell, D. J. (2002). Assessment of left ventricular mass by cardiovascular magnetic resonance. Hypertension, 39(3), 750–755. 10.1161/hy0302.104674 [DOI] [PubMed] [Google Scholar]
  23. Nasir, J. M. , Rubal, B. J. , Jones, S. O. , & Shah, A. D. (2012). The effects of body mass index on surface electrocardiograms in young adults. Journal of Electrocardiology, 45(6), 646–651. 10.1016/j.jelectrocard.2012.07.022 [DOI] [PubMed] [Google Scholar]
  24. Norman, J. E. Jr , & Levy, D. (1996). Adjustment of ECG left ventricular hypertrophy criteria for body mass index and age improves classification accuracy: The effects of hypertension and obesity. Journal of Electrocardiology, 29(Suppl 1), 241–247. 10.1016/S0022-0736(96)80070-7 [DOI] [PubMed] [Google Scholar]
  25. Okin, P. M. , Jern, S. , Devereux, R. B. , Kjeldsen, S. E. , & Dahlof, B. (2000). Effect of obesity on electrocardiographic left ventricular hypertrophy in hypertensive patients: The losartan intervention for endpoint (LIFE) reduction in hypertension study. Hypertension, 35(1 Pt 1), 13–18. 10.1161/01.HYP.35.1.13 [DOI] [PubMed] [Google Scholar]
  26. Rodrigues, J. C. L. , McIntyre, B. , Dastidar, A. G. , Lyen, S. M. , Ratcliffe, L. E. , Burchell, A. E. , … Manghat, N. E. (2016). The effect of obesity on electrocardiographic detection of hypertensive left ventricular hypertrophy: Recalibration against cardiac magnetic resonance. Journal of Human Hypertension, 30(3), 197–203. 10.1038/jhh.2015.58 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Romhilt, D. W. , Bove, K. E. , Norris, R. J. , Conyers, E. , Conradi, S. , Rowlands, D. T. , & Scott, R. C. (1969). A critical appraisal of the electrocardiographic criteria for the diagnosis of left ventricular hypertrophy. Circulation, 40(2), 185–195. 10.1161/01.CIR.40.2.185 [DOI] [PubMed] [Google Scholar]
  28. Romhilt, D. W. , & Estes, E. H. Jr (1968). A point‐score system for the ECG diagnosis of left ventricular hypertrophy. American Heart Journal, 75(6), 752–758. 10.1016/0002-8703(68)90035-5 [DOI] [PubMed] [Google Scholar]
  29. Schlegel, T. T. , Kulecz, W. B. , Feiveson, A. H. , Greco, E. C. , DePalma, J. L. , Starc, V. , … Pahlm, O. (2010). Accuracy of advanced versus strictly conventional 12‐lead ECG for detection and screening of coronary artery disease, left ventricular hypertrophy and left ventricular systolic dysfunction. BMC Cardiovascular Disorders, 10, 28 10.1186/1471-2261-10-28 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Sokolow, M. , & Lyon, T. P. (1949). The ventricular complex in left ventricular hypertrophy as obtained by unipolar precordial and limb leads. American Heart Journal, 37(2), 161–186. 10.1016/0002-8703(49)90562-1 [DOI] [PubMed] [Google Scholar]
  31. Solberg, H. E. (1987). International Federation of Clinical Chemistry (IFCC), Scientific committee, clinical section, expert panel on theory of reference values, and International Committee for Standardization in Haematology (ICSH), standing committee on reference values. Approved recommendation (1986) on the theory of reference values. Part 1. The concept of reference values. Journal of Clinical Chemistry and Clinical Biochemistry, 25(5), 337–342. [PubMed] [Google Scholar]
  32. Vakili, B. A. , Okin, P. M. , & Devereux, R. B. (2001). Prognostic implications of left ventricular hypertrophy. American Heart Journal, 141(3), 334–341. 10.1067/mhj.2001.113218 [DOI] [PubMed] [Google Scholar]
  33. Verdecchia, P. , Angeli, F. , Borgioni, C. , Gattobigio, R. , de Simone, G. , Devereux, R. B. , & Porcellati, C. (2003). Changes in cardiovascular risk by reduction of left ventricular mass in hypertension: A meta‐analysis. American Journal of Hypertension, 16(11 Pt 1), 895–899. 10.1016/S0895-7061(03)01018-5 [DOI] [PubMed] [Google Scholar]
  34. Verdecchia, P. , Schillaci, G. , Borgioni, C. , Ciucci, A. , Gattobigio, R. , Zampi, I. , … Porcellati, C. (1998). Prognostic significance of serial changes in left ventricular mass in essential hypertension. Circulation, 97(1), 48–54. 10.1161/01.CIR.97.1.48 [DOI] [PubMed] [Google Scholar]
  35. Xie, L. D. , & Wang, Z. L. (2010). Correlation between echocardiographic left ventricular mass index and electrocardiographic variables used in left ventricular hypertrophy criteria in Chinese hypertensive patients. Hellenic Journal of Cardiology, 51(5), 391–401. [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

 

Click here for additional data file. (34.1KB, docx)

Articles from Annals of Noninvasive Electrocardiology : The Official Journal of the International Society for Holter and Noninvasive Electrocardiology, Inc are provided here courtesy of International Society for Holter and Noninvasive Electrocardiology, Inc. and Wiley Periodicals, Inc.

RESOURCES