Graphical abstract
Overview of the study population, methods, and key findings. LVH: left ventricular hypertrophy; fTTE: focused transthoracic echocardiography; LVMI: left ventricular mass index; LVM: left ventricular mass; BSA: body surface area; ECG: electrocardiography.
Figure 1
Study flow chart and selection of participants. Legend—fTTE: focused transthoracic echocardiography; ECG: electrocardiography.
Table 1
Characteristics of electrocardiographic indexes used in this study.
| ECG INDEX | DEFINITION | CUT-OFF VALUE |
|---|---|---|
| Cornell voltage criteria (11) | SV3 + RaVL | >2.0 mV for women and >2.8 mV for men |
| Cornell voltage product (19) | (SV3 + RaVL)*QRS | >244.0 mVms (19) or >243.6 mVms (56) |
| Cornell voltage product, adjusted for women 0.6 mV | Men: (SV3 + RaVL)*QRS Women: (SV3 + RaVL + 0.6 mV)*QRS | >244.0 mVms (8, 48) or >243.6 mVms |
| Cornell voltage product, adjusted for women 0.8 mV | Men: (SV3 + RaVL)*QRS Women: (SV3 + RaVL + 0.8 mV)*QRS | >244.0 mVms (48) or >243.6 mVms (57) |
| Novacode 6.1 LVH (58) | RaVL + SV3 | ≥2,600 μV in men ≥2,200 μV in women |
| RaVL (59, 60) | ≥1.1 mV | |
| General MESA ECG-LVH (61) | SV1 + SV2 + RV5 | ≥4.2 mV |
| Ethnicity- and sex-specific MESA ECG-LVH criterion (11) | SV1 + SV2 + RV5 | ≥4.8 in men ≥4.3 in women |
| Sokolow–Lyon voltage (8) | SV1 + either RV5/RV6 (whichever is higher) | >3.5 mV |
| Sokolow–Lyon voltage product (11) | SV1 + (RV5/RV6)*QRS duration | ≥371,000 μVms (371 mVms ) |
| Framingham-adjusted Cornell voltage (11, 62) | Men: [RaVL + SV3 + 0.0174*(age – 49) + 0.191*(BMI – 26.5)]; Women: [RaVL + SV3 + 0.0387*(age – 50) + 0.212*(BMI – 24.9)] | Men ≥2.8 mV Women ≥2.0 mV |
| Minnesota Code 3.1 (63) | RV5/V6 or RI/II/III/aVF or RaVL | >26 mm >20 mm >12 mm |
| Minnesota Code 3.3 (63) | RI or R V5/V6 plus SV1; no MC 3.1 | >15 mm but ≤20 mm, >35 mm |
| Strain pattern (52, 64) | Coexistence in any of the leads I, II, avL, or V3 to V6: ST-segment horizontal or downward sloping depression ≥ 0.05 mV, plus negative T-wave; down-sloping convex ST-segment with an inverted asymmetrical T-wave opposite to the QRS axis | Yes/No |
[i] RaVL: R wave in aVL; BMI: body mass index; MC: Minnesota Code; ECG: electrocardiography; LVH: left ventricular hypertrophy; MESA: the Multi-Ethnic Study of Atherosclerosis.
Table 2
Baseline demographic and clinical characteristics of the study sample, overall and regarding subgroups with echocardiographic signs for LVH.
| VARIABLES | OVERALL (n = 1125) | fTTE-LVH PRESENT (n = 56) | fTTE-LVH ABSENT (n = 1069) |
|---|---|---|---|
| Tanzania, n (%) | 553 (49.2) | 47 (83.9) | 506 (47.3) |
| Lesotho, n (%) | 572 (50.8) | 9 (16.1) | 563 (52.7) |
| Age, years, median (IQR) | 53.4 (44.6–64.4) | 59.0 (49.1–68.1) | 53.2 (44.3–64.3) |
| Male, n (%) | 314 (27.9) | 8 (14.3) | 306 (28.6) |
| HIV positive, n (%) | 423 (37.6) | 10 (17.9) | 413 (38.6) |
| Previously diagnosed HT, n (%) | 469 (41.7) | 34 (60.7) | 435 (40.7) |
| Diabetes, n (%) | 22 (2.0) | 0 | 22 (2.1) |
| Diabetes never checked, n (%) | 102 (9.1) | 3 (5.4) | 99 (9.3) |
| BMI (kg/m2), median (IQR) | 26.4 (22.8–30.7) | 25.1 (21.5–28.1) | 26.5 (22.8–30.7) |
| BSA (m2), median (IQR) | 1.71 (1.55–1.85) | 1.57 (1.47–1.76) | 1.70 (1.56–1.85) |
| Systolic BP (mmHg), median (IQR) | 149 (141–163) | 165 (152–182) | 148 (140–161) |
| Diastolic BP (mmHg), median (IQR) | 99 (93–106) | 104 (98–116) | 98 (93–105) |
[i] Values are presented as n (%), if applicable, with percentages calculated within column groups. fTTE: focused transthoracic echocardiography; HT: hypertension; LVH: left ventricular hypertrophy; BMI: body mass index; BP: blood pressure; BSA: body surface area.
Table 3
Prevalence of positive ECG markers in the entire cohort, as well as separated for the group with echocardiographic signs of LVH and those without.
| INDEX | OVERALL (n = 1125) | ECHOCARDIOGRAPHIC SIGNS FOR LVH (n = 56) | NO ECHOCARDIOGRAPHIC SIGNS FOR LVH (n = 1070) | p-value |
|---|---|---|---|---|
| Cornell voltage criteria, >2.0 mV (women), >2.8 mV (men), n (%) | 256 (22.8) | 28 (50.0) | 228 (21.3) | <0.001 |
| Cornell voltage product ≥ 243.6 mVms, n (%) | 103 (9.2) | 21 (37.5) | 82 (7.7) | <0.001 |
| Cornell voltage product ≥ 244 mVms, n (%) | 100 (8.9) | 21 (37.5) | 79 (7.4) | <0.001 |
| Cornell voltage product ≥ 243.6 mVms, n (%), adjusted 0.6 mV (women) | 203 (18.0) | 30 (53.6) | 173 (16.2) | <0.001 |
| Cornell voltage product ≥ 244 mVms, n (%), adjusted 0.6 mV (women) | 197 (17.5) | 30 (53.6) | 167 (15.6) | <0.001 |
| Cornell voltage product ≥ 243.6 mVms, n (%), adjusted 0.8 mV (women) | 255 (22.7) | 36 (64.3) | 219 (20.5) | <0.001 |
| Cornell voltage product ≥ 244 mVms, n (%), adjusted 0.8 mV (women) | 249 (22.1) | 34 (60.7) | 215 (20.1) | <0.001 |
| Novacode 6.1 LVH, ≥ 2.2 mV (women), ≥ 2.6 mv (men), n (%) | 240 (21.3) | 30 (53.6) | 210 (19.6) | <0.001 |
| RaVL ≥ 1.1 mV, n (%) | 190 (16.9) | 23 (41.1) | 167 (15.6) | <0.001 |
| general MESA ECG-LVH, ≥ 4.2 mV, n (%) | 633 (56.3) | 33 (58.9) | 600 (56.1) | 0.784 |
| Ethnicity- and sex-specific MESA ECG-LVH criterion (African American), ≥ 4.3 mV (women), ≥ 4.8 mV (men), n (%) | 566 (50.3) | 30 (53.6) | 536 (50.1) | 0.716 |
| Sokolow–Lyon voltage > 3.5 mV, n (%) | 390 (34.7) | 29 (51.8) | 361 (33.8) | 0.009 |
| Sokolow–Lyon voltage product, ≥ 371 mVms, n (%) | 172 (15.3) | 24 (42.9) | 148 (13.8) | <0.001 |
| Framingham-adjusted Cornell voltage, ≥ 2.0 mV (women), ≥ 2.8 mV (men), n (%) | 564 (50.1) | 34 (60.7) | 530 (49.6) | 0.179 |
| Minnesota Code 3.1, n (%) | 270 (24.0) | 30 (53.6) | 240 (22.5) | <0.001 |
| Minnesota Code 3.3, n (%) | 205 (18.2) | 4 (7.1) | 201 (18.8) | 0.419 |
| Strain pattern, n (%) | 44 (3.9) | 12 (21.4) | 32 (3.0) | <0.001 |
[i] ECG: electrocardiography; LVH: left ventricular hypertrophy; MESA: the Multi-Ethnic Study of Atherosclerosis.
Table 4
Performance of standard ECG-LVH criteria in overall study population.
| PARAMETER | SENSITIVITY (95% CI) | SPECIFICITY (95% CI) | PPV (95% CI) | NPV (95% CI) | ACCURACY | LR+ | LR– | TP | FP | TN | FN |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Cornell voltage criteria (>2.0 mV women, >2.8 mV men) | 0.500 (0.373–0.627) | 0.787 (0.761–0.810) | 0.109 (0.077–0.154) | 0.968 (0.954–0.978) | 0.772 | 2.344 | 0.636 | 28 | 228 | 841 | 28 |
| Cornell voltage product (unadjusted) ≥ 243.6 mVms | 0.375 (0.260–0.506) | 0.923 (0.906–0.938) | 0.204 (0.137–0.292) | 0.966 (0.953–0.975) | 0.896 | 4.889 | 0.677 | 21 | 82 | 987 | 35 |
| Cornell voltage product (unadjusted) ≥ 244 mVms | 0.375 (0.260–0.506) | 0.926 (0.909–0.940) | 0.210 (0.142–0.300) | 0.966 (0.953–0.975) | 0.899 | 5.074 | 0.675 | 21 | 79 | 990 | 35 |
| Cornell voltage product (women +0.6 mV) ≥ 243.6 mVms | 0.536 (0.407–0.660) | 0.838 (0.815–0.859) | 0.148 (0.106–0.203) | 0.972 (0.959–0.981) | 0.823 | 3.31 | 0.554 | 30 | 173 | 896 | 26 |
| Cornell voltage product (women +0.6 mV) ≥ 244 mVms | 0.536 (0.407–0.660) | 0.844 (0.821–0.864) | 0.152 (0.109–0.209) | 0.972 (0.959–0.981) | 0.828 | 3.429 | 0.550 | 30 | 167 | 902 | 26 |
| Cornell voltage product (women +0.8 mV) ≥ 243.6 mVms | 0.643 (0.512–0.755) | 0.795 (0.770–0.818) | 0.141 (0.104–0.189) | 0.977 (0.965–0.985) | 0.788 | 3.138 | 0.449 | 36 | 219 | 850 | 20 |
| Cornell voltage product (women +0.8 mV) ≥ 244 mVms | 0.607 (0.476–0.724) | 0.799 (0.774–0.822) | 0.137 (0.099–0.185) | 0.975 (0.962–0.983) | 0.789 | 3.019 | 0.492 | 34 | 215 | 854 | 22 |
| Novacode 6.1 LVH (≥ 2.2 mV women, ≥ 2.6 mV men) | 0.536 (0.407–0.660) | 0.804 (0.779–0.826) | 0.125 (0.089–0.173) | 0.971 (0.957–0.980) | 0.790 | 2.727 | 0.578 | 30 | 210 | 859 | 26 |
| R amplitude aVL ≥ 1.1 mV | 0.411 (0.292–0.541) | 0.844 (0.821–0.864) | 0.121 (0.082–0.175) | 0.965 (0.951–0.975) | 0.822 | 2.629 | 0.698 | 23 | 167 | 902 | 33 |
| General MESA ECG-LVH ≥ 4.2 mV | 0.589 (0.459–0.708) | 0.439 (0.409–0.469) | 0.052 (0.037–0.072) | 0.953 (0.931–0.969) | 0.446 | 1.050 | 0.936 | 33 | 600 | 469 | 23 |
| MESA ECG-LVH African American (≥ 4.3 women, ≥ 4.8 men) | 0.536 (0.407–0.660) | 0.499 (0.469–0.529) | 0.053 (0.037–0.075) | 0.953 (0.933–0.968) | 0.500 | 1.068 | 0.931 | 30 | 536 | 533 | 26 |
| Sokolow–Lyon voltage > 3.5 mV | 0.518 (0.390–0.643) | 0.662 (0.633–0.690) | 0.074 (0.052–0.105) | 0.963 (0.947–0.975) | 0.655 | 1.533 | 0.728 | 29 | 361 | 708 | 27 |
| Sokolow–Lyon voltage product ≥ 371 mVms | 0.429 (0.308–0.559) | 0.862 (0.840–0.881) | 0.140 (0.096–0.199) | 0.966 (0.953–0.976) | 0.840 | 3.096 | 0.663 | 24 | 148 | 921 | 32 |
| Framingham-adjusted Cornell (≥ 2.0 women, ≥ 2.8 men) | 0.607 (0.476–0.724) | 0.504 (0.474–0.534) | 0.060 (0.043–0.083) | 0.961 (0.941–0.974) | 0.509 | 1.225 | 0.779 | 34 | 530 | 539 | 22 |
| Minnesota Code 3.1 | 0.536 (0.407–0.660) | 0.775 (0.750–0.799) | 0.111 (0.079–0.154) | 0.970 (0.956–0.979) | 0.764 | 2.386 | 0.599 | 30 | 240 | 829 | 26 |
| Minnesota Code 3.3 | 0.154 (0.062–0.335) | 0.758 (0.727–0.785) | 0.020 (0.008–0.049) | 0.966 (0.949–0.978) | 0.739 | 0.635 | 1.117 | 4 | 201 | 628 | 22 |
| Strain pattern | 0.214 (0.127–0.338) | 0.970 (0.958–0.979) | 0.273 (0.163–0.418) | 0.959 (0.946–0.970) | 0.932 | 7.158 | 0.810 | 12 | 32 | 1037 | 44 |
[i] NPV: Negative predictive value; PPV: positive predictive value; LR+: positive likelihood ratio (rule-in power); LR–: negative likelihood ratio (rule-out power); TP: true positive; FP: false positive; TN: true negative; FN: false negative.
Figure 2
Receiver-operating curves (ROC) with area under the ROC (AUROC) and precision-recall curves (PRC) with area under the PRC (AUPRC) for continuous ECG parameters. Legend—ECG: electrocardiography; LVH: left ventricular hypertrophy; MESA: Multi-Ethnic Study of Atherosclerosis.
Figure 3
Decision curve analysis of ECG criteria for the detection of LVH. Net benefit is plotted against threshold probability, representing the risk of LVH at which echocardiography would be recommended. Colored curves represent predicted probabilities derived from single-predictor logistic regression models for each ECG criterion. Dashed gray curves indicate reference strategies for treating all patients or treating none. Higher net benefit indicates greater clinical utility.