Electrocardiographic criteria for cardiac remodeling in hypertensive patients

Left ventricular hypertrophy (LVH) and left atrial enlargement (LAE) represent important predictors of cardiovascular morbidity and mortality, as well as all-cause mortality, in hypertensive and global population.1, 2 The gold standard for the assessment of LVH and LAE is an imaging tool that is widely available in everyday clinical practice such as echocardiography. Cardiac magnetic resonance is naturally a more accurate method, but its low availability and high costs significantly limit its use for LVH and LAE. However, the first method that was successfully used for the detection of both entities was electrocardiogram (ECG) that represents a rapid, inexpensive and highly available technique. The main problem with this approach is lower sensitivity and specificity in comparison with echocardiography, particularly with cardiac magnetic resonance. Nevertheless, different studies showed a significant predictive value of electrographically diagnosed LVH in the global population or in the patients with arterial hypertension.3-5 The data regarding predictive importance of LAE detected in ECG are less available, but Okin et al6 reported that abnormal P wave, a marker of LA abnormality, was strongly associated with incident stroke in hypertensive patients.

In this issue of the Journal, Jiang et al used six ECG criteria for LVH detection (Sokolow-Lyon voltage, Cornell voltage, Cornell product, SD+SV4, Manning, and R+S), as well as their combination, and showed that the combination of these criteria significantly increased sensitivity and specificity for diagnosing LVH, having echocardiography as a reference method.7 The combination of SD+SV4 or Cornell product criteria had almost the same sensitivity and higher specificity than the combination of all six ECG criteria for LVH detection (38% vs 43% for sensitivity; and 91% vs 88.5% for specificity).7 The best single ECG criterion for LVH detection was SD+SV4, which represents the deepest S-wave amplitude added to the S-wave amplitude of lead V4, with AUC 0.71 (P = 0.002), sensitivity of 29%, and specificity of 92%.7

At the moment, there are 37 different ECG criteria for LVH endorsed by the American Heart Association.8 This could be confusing for clinicians because they cannot decide which single ECG criterion or their combination is the best for determination of LVH. Previous studies considered that the Cornell voltage criteria had the best accuracy with the specificity of 90% and the sensitivity of only 20%-40%.9 However, recently, several studies showed that SD+SV4, a new ECG criterion for LVH diagnosis, had a significantly higher sensitivity than the Cornell voltage criteria.10, 11 Peguero et al first introduced this ECG criterion for LVH and reported that its sensitivity of 62% was significantly higher than the sensitivity for the Cornell criteria (only 35%) in a small population (n = 94) which consisted of hypertensive and non-hypertensive patients.10 The specificity in both criteria was higher than 90% and did not significantly differ.10 Chao et al involved more hypertensive patients (n = 235) and confirmed superiority of the new ECG criterion for LVH over the traditional Cornell criteria.11 The newly proposed criteria had the higher sensitivity in diagnosing LVH (men: 65.5%; women: 81%) than the Cornell criteria (men: 55.2%; women: 56.9%).11 The specificities of both criteria were higher than 70%, with no significant differences between them.

Jiang et al reported a significantly lower sensitivity of SD+SV4 criteria than the previous studies.10, 11 We believe that the main reason for this low sensitivity is low prevalence of LVH and obesity in the study population. Namely, Jiang et al7 reported that LVH diagnosed by echocardiography (LV mass index for men >115 g/m² and for women >95 g/m²) was present in only 14% of the participants, which was in absolute numbers only 21 patients. The authors proposed the combination of SD+SV4 and Cornell criteria for LVH diagnosis, as this combination showed the highest sensitivity and specificity.

All ECG criteria for LVH diagnosis have a huge problem with low sensitivity. In the real world, this means that they unfortunately have a high rate of false-negative cases. In other words, many patients with LVH remain undetected when ECG criteria are used. The new ECG criterion, which involves measuring the amplitude of the deepest S wave in any single lead and adding it to the S-wave amplitude of lead V4, seems to have significantly better accuracy than previous ECG criteria. Peguero et al who introduced this parameter suggested that the reason for better sensitivity lies in electrophysiology and conducting of impulse in patients with LVH.10 Namely, the majority of ECG criteria for LVH are focused on R wave in different leads, whereas their study showed that S wave was more strongly related with LV mass. Four vectors of depolarization have been described in the human heart. The first two vectors show depolarization of the septum, conduction system, and endomyocardial fibers of the LV, which occurs in the first 30 ms of the ventricular depolarization. The third and fourth vectors demonstrate depolarization of the myocardial and epicardial free wall of the LV, which occurs after 50 ms12 Therefore, it is reasonable to hypothesize that voltage changes in patients with mild to moderate LVH are better represented by S wave, which is the potential reason for better sensitivity of these new criteria for LVH detection.

The possible reason for lower sensitivity of ECG-derived LVH could lie in the fact that diffuse myocardial fibrosis, characteristic for hypertensive LVH, has independent and opposing effects on ECG voltage measurements of LVH.13 LV mass index correlated positively, whereas diffuse myocardial fibrosis correlated negatively with QRS voltage.13 Interestingly, the authors showed that myocardial fibrosis did not correlate with LV mass. Consequently, diffuse myocardial fibrosis can mask the ECG changes typical for LVH.

Further contribution of the study by Jiang et al7 concerns the diagnostic accuracy of different ECG criteria for LAE. The authors claimed that P-wave terminal force in lead V1 >40 ms/mm (PTFV1) was important ECG criteria for LAE diagnosis with sensitivity of 26% and specificity of 91% (AUC 0.68, P = 0.008).7 This parameter showed better accuracy than other ECG criteria for LAE such as P-wave dispersion, P-wave duration, and P-wave duration in lead II divided by PR interval.7 PTFV1 correlated better with LA volume index than other mentioned ECG indexes for LAE.

In the literature, there is no consensus regarding accuracy of ECG-diagnosed LAE. Batra et al14 recently showed that sensitivity and specificity of PTFV1 criteria to detect LAE in ECG were 54% and 57%, respectively. However, the authors used echocardiographically assessed LA diameter and not LA volume, as the gold standard technique,13 which is not recommended any more. Additionally, the authors did not compare various ECG criteria for LAE in order to determine which index had the best accuracy. Other investigators used different imaging methods for evaluation of LA volume (echocardiography and cardiac magnetic resonance) and found high specificity, but low sensitivity of ECG-derived criteria for detection of LAE.15-17

Lee et al15 reported that PTFV1 criteria had worse sensitivity (46% vs 69%), but better specificity (64% vs 49%), than P-wave duration. The investigators used LA volume index derived by echocardiography as the gold standard method. Rodrigues et al16 found that individual ECG criteria of LAE in hypertension were specific, but not sensitive, in the diagnosis of LAE. As the reference imaging method for determination of LA volume, they used cardiac magnetic resonance and concluded that ECG should not be used to exclude LAE in the patients with arterial hypertension.16 The authors compared five different ECG criteria for LAE, including PTFV1, and did not show significant difference in accuracy between them. Tsao et al17 also used cardiac magnetic resonance as the reference method and demonstrated low sensitivity and high specificity of ECG criteria for LAE, including PTFV1.

The previous hypothesis that LAE might induce the prolongation of atrial conduction time was not accurate because it did not consider the possibility that the P-wave prolongation might be a consequence of intra-atrial block, commonly seen in the elderly, and not necessary the result of LAE.18

New ECG criteria for LVH (SD+SV4) might be of great importance for clinical practice, particularly in those regions where imaging techniques are not widely available. For practitioners working in undeveloped countries or rural regions, the new ECG criterion or its combination with already approved ECG criteria (Cornell criteria, Sokolow-Lyon voltage criteria, etc) could serve at least as a kind of filter to determine which hypertensive patients should be referred to further and more sophisticated diagnostics. However, this is a new criterion and further studies with a larger number of patients are necessary to confirm its accuracy in global and hypertensive population, as well as its predictive value on cardiovascular morbidity and mortality. On the other hand, ECG-derived criteria for LAE remain inaccurate because of low specificity and should not be used in evaluation of LAE.

CONFLICT OF INTEREST

None.