# Evaluating ECG Criteria for Diagnosing Left Ventricular Hypertrophy in Anabolic-Androgenic Steroid Users

## TL;DR

This study finds that ECG tests are not reliable for detecting heart thickening in long-term steroid users, suggesting the need for more accurate methods like echocardiography.

## Contribution

The study is the first to evaluate ECG criteria for diagnosing LVH specifically in anabolic-androgenic steroid users.

## Key findings

- ECG criteria failed to detect LVH in most steroid users, with over 60% undetected by the best-performing methods.
- AAS users with LVH showed more severe heart dysfunction compared to those without LVH.
- Peguero-Lo Presti and Romhilt-Estes criteria had the best balance of sensitivity and specificity but still missed most cases.

## Abstract

Anabolic-androgenic steroid (AAS) users represent a high-risk cardiovascular population. Left ventricular (LV) hypertrophy (LVH) is a common complication, yet the utility of electrocardiographic (ECG) criteria for detecting LVH remains unexplored.

The objective of the study was to evaluate the diagnostic performance of six commonly used ECG-LVH criteria in identifying echocardiographic LVH among long-term AAS users.

We included 100 male AAS users, 32 with echocardiographic LVH and 68 without. Six ECG-LVH criteria (Sokolow-Lyon, Modified Sokolow-Lyon, Cornell voltage, Cornell product, Peguero-Lo Presti, and Romhilt-Estes) were assessed for sensitivity, specificity, and predictive values, using echocardiography as the reference standard. LV systolic function was evaluated by LV ejection fraction and global longitudinal strain.

AAS users with echocardiographic LVH exhibited more pronounced LV systolic dysfunction than those without, as shown by a higher number of individuals with LV ejection fraction ≤40% (7 [23%] vs 4 [6%], P = 0.02) and more impaired LV global longitudinal strain (−14.7% ± 2.7% vs −16.5% ± 2.2%, P = 0.002). Among voltage-based indices, Peguero-Lo Presti provided the best balance between sensitivity (31%) and specificity (88%). The point-based Romhilt-Estes using a cutoff of 5 offered comparable performance, with a sensitivity of 34% and specificity of 85%. However, most AAS users with echocardiographic LVH were not identified by ECG, with 69% remaining undetected by Peguero-Lo-Presti and 66% by Romhilt-Estes.

ECG was insufficient for identifying echocardiographic LVH in AAS users who represent a high-risk cardiovascular population. Routine specialist evaluations, including echocardiography, should be considered for long-term AAS users to facilitate early detection of LVH and associated dysfunctions.