Background: Detection of left ventricular hypertrophy (LVH) is clinically important because it can be predictive of adverse clinical outcome. However, the best method for detecting LVH in clinical practice is unclear. The aim of this study was to evaluate electrocardiography (ECG) compared with echocardiography (ECHO) as a screening test to detect LVH in a high risk population. Method: In a prospective, double-blinded, single centre analysis of a population of children with chronic kidney disease, LVH detected using standard 12-lead ECG (ECG-LVH) was compared with that detected with 2D-guided M-mode ECHO (ECHO-LVH). Two electrocardiographic methods (A and B) were used to diagnose ECG-LVH and compared with three different indexation methods to define ECHO-LVH. Results: 70 consecutively enrolled participants had 107 ECG and ECHO studies performed on the same day. The prevalence of ECHO-LVH ranged from 17% to 55% using different indexation methods. Increased R wave amplitude using recent age and gender specific voltage criteria (ECG method B) demonstrated the highest sensitivity (68-76%) and specificity (43-77%) for detecting ECHO-LVH. The negative predictive value (NPV) for R waves using ECG method B was 52.4% (range 40-68.6%). For all other ECG criteria, sensitivity did not exceed 13% irrespective of the ECG or ECHO method used. Conclusions: In children, the standard 12-lead electrocardiogram has low sensitivity and low NPV for detecting LVH. These findings are relevant for physiological LVH and should not be extrapolated to detection of hypertrophic cardiomyopathy. In clinical practice, ECHO alone should be used to exclude LVH.