Healthy children (n = 400; 52% boys, 0-18 years of age) from 2 different European hospitals were studied to derive a simplified formula. Left ventricular mass (LVM) was calculated according to the Devereux formula. The derived approach to index LVM was tested on a validation cohort of 130 healthy children from a different hospital center.
There was a strong nonlinear correlation between height and LVM. LVM was best related to height to a power of 2.16 with a correction factor of 0.09. Analysis of residuals for LVM/[(height2.16) + 0.09] showed an homoscedastic distribution in both sexes throughout the entire height range. A partition value of 45 g/m2.16 was defined as the upper normal limit for LVM index. As opposed to formula suggested by current guidelines (ie, LVM/height2.7) when applying the proposed approach in the validation cohort of 130 healthy participants, no false positives for LVH were found (0% vs 8%; P < .01).
Our data support the possibility to have a single partition (ie, 45 g/m2.16) value across the whole pediatric age range to identify LVH, without the time-consuming need of computing specific percentiles for height and sex.