The report by Lopez et al revisits Goldberger’s electrocardiographic triad (SV1 or SV2 + RV ₅ or RV ₆ ≥3.5 mV, total QRS amplitude in each of the limb leads ≤0.8 mV, and R/S ratio <1 in lead V ₄ ), proposed 30 years ago as allegedly predictive of congestive heart failure with sensitivity of 70% and specificity of >90%.

The investigators evaluated a consecutive series of 51 patients with dilated mostly nonischemic cardiomyopathy with histories of hypertension and found lower sensitivity of the triad of 2%. They attributed the difference in their findings and those of Goldberger’s to differences in the origin of cardiomyopathy (“burned-out” hypertension in their cases and “idiopathic” in Goldberger’s cohort) and to their use of a single electrocardiogram (ECG), in contrast to the use by Goldberger of all the available ECGs for each patient in his cohort, seeking the triad.

Although Goldberger emphasized that the triad should be applied only in toto, Lopez et al provided also an analysis of the presence of each of the 3 components of the triad in their 51 patients. Curiously, in reference to the electrocardiographic finding of total QRS amplitude ≤0.8 mV, the investigators refer to leads I, II, and III, while Goldberger’s triad requires consideration of all 6 limb leads. Thus, their observation that “total QRS amplitude was ≤0.8 mV in each of leads I, II, and III in 10” indicates that ≤10 patients showed ≤0.8 mV in all 6 limb leads.

The use of all available ECGs by Goldberger in his 27 patients was inappropriate, considering that patients had different numbers of ECGs in their files (≥1), raising issues pertaining to the number of ECGs per patient and to the reproducibility of electrocardiography, a test under study. In contrast, what Lopez et al did (using the single ECG in proximity with the echocardiographic examination) was appropriate.

There are also possible issues of selection biases in Goldberger’s study, because the 27 patients are described as being “located from the records of the Cardiology section of San Diego Veterans Administration Medical Center,” without any details as to how the selection was carried out.

In reference to the investigators comment that “the difference is likely due to [Goldberger’s] patients’ having idiopathic dilated cardiomyopathy, whereas those in this study had hypertensive cardiomyopathy with or without ischemia,” I have great doubts; scrutiny of the retrospective and prospective components of Goldberger’s study shows that he evaluated 27 patients with “idiopathic congestive cardiomyopathy” of nonischemic origin, and probably with histories of hypertension, not different from the patients of the investigators’ cohort (i.e., the seemingly “cryptic” “idiopathic” designation may not mean anything unexplainable, or excluding that there was history of hypertension in these patients), “100 normal subjects selected at random,” “30 preoperative patients with high precordial voltage, aortic stenosis and/or regurgitation, and normal left ventricular ejection fraction,” and “a consecutive series of 2000 patients.”

I believe that the differences between Lopez et al’s and Goldberger’s findings are due to (1) selection biases, (2) the common problem of the nonreproducibility of lead V ₄ with an R/S ratio >1 on serial electrocardiography, and (3) the betterment of care for patients with congestive heart failure in the intervening 30 years, ensuring less lung and peripheral edema. The management of the latter results in augmentation of the voltages in the 6 limb leads, which have led this reader to recommend the use of only the limb leads (to avoid, the nonreproducible precordial V ₁ -V ₆ leads, in comparisons of serial ECGs) in the diagnosis and monitoring of patients with congestive heart failure.