Comparison of Electrocardiographic Criteria for Identifying Left Ventricular Hypertrophy in Athletes from Different...

In athletes, isolated electrocardiogram high voltage criteria are widely used to evaluate left ventricular hypertrophy, but positive findings are thought to represent normal electrocardiogram alterations. However, which electrocardiogram criterion can best detect left ventricular hypertrophy in athletes of various sport modalities remains unknown.

Five electrocardiogram criteria used to detect left ventricular hypertrophy were tested in 180 male athletes grouped according to their sport modality: 67% low-static and high-dynamic components and 33% high-static and high-dynamic components of exercise. The following echocardiogram parameters are the gold standard for diagnosing left ventricular hypertrophy: left ventricular mass index ≥134 g.m-2, relative wall thickness ≥0.42 mm, left ventricular diastolic diameter index ≥32 mm.m-2, septum wall thickness ≥13 mm, and posterior wall thickness ≥13 mm. Results for the various criteria were compared using the kappa coefficient. Significance was established at p<0.05.

Fifty athletes (28%) presented with left ventricular hypertrophy according to electrocardiogram findings, with the following sensitivities and specificities, respectively: 38-53% and 79-83% (Perugia), 22-40% and 89-91% (Cornell), 24-29% and 90% (Romhilt-Estes), 68-87% and 20-23% (Sokolow-Lyon), and 0% and 99% (Gubner). The Perugia and Cornell criteria had higher negative predictive values for the low-static and high-dynamic subgroup. Kappa coefficients were higher for Romhilt-Estes, Cornell and Perugia criteria than for Sokolow-Lyon and Gubner criteria.

All five evaluated criteria are inadequate for detecting left ventricular hypertrophy, but the Perugia, Cornell and Romhilt-Estes criteria are useful for excluding its presence. The Perugia and Cornell criteria were more effective at excluding left ventricular hypertrophy in athletes involved in a sport modality with low-static and high-dynamic component predominance.