Letter by Bocelli et al Regarding Article, “A Novel Method of Expressing Left Ventricular Mass Relative to Body Size in Children”
To the Editor:
In the recent interesting article by Foster et al,1 the standard method of indexing left ventricular (LV) mass to height2.7 was found to be inadequate to adjust for body size in a reference population that included healthy individuals from birth to 21 years of age. In particular, LV mass index (LVMI) was relatively independent of height in taller subjects but showed a strong residual correlation with height in subjects <100 cm tall, ie, in newborns, infants, and children in their first years of life. In contrast, a more complex method based on percentile curves provided adequate normalization to body size, with no residual relation between LV mass-for-height z scores and height.
We agree with the intriguing clinical utility of the centile method for adjusting LV mass to body size in children. However, a correct comparison with the simpler LVMI method requires the choice of the most appropriate allometric exponent. Although normalization to height2.7 is widely used in both research and clinical practice, caution is needed in applying this method in studies that include newborns or infants. The value of 2.7 was derived by regression analysis in a previous study in which a large proportion of the study population was represented by adults, adolescents, and children in middle childhood or older.2 Accordingly, the recommendation to use this method to normalize LV mass in the pediatric age group (given by the National High Blood Pressure Education Program Working Group in 2004) refers to children and adolescents but not to newborns or infants.3 As a result, indexation to height2.7 has been used commonly in pediatric patients, but studies including newborns and infants often preferred conventional indexation to body surface area. In the study by Foster et al,1 27% of the population was <2 years old, and nearly 15% was ≤6 months old. The appropriateness of indexing to height2.7 in this context should reasonably be considered with caution, and this may explain the inadequate normalization to body size found in the lower range of height.
Interestingly, the evidence that LVMI increased with decreasing height suggests that the value of 2.7 used as an allometric exponent of height may be too high to ensure adequate adjustment of LV mass to body size in newborns and infants; that is, the assumption of an approximately cubic relation overestimates the real effect of height on LV mass in these subjects, which leads to an excessive increase in LVMI for decreasing values of height. In accordance with this hypothesis, previous analyses of large pediatric populations that included newborns and infants found lower values for the exponent, ranging from 1.7 to 2.3.4,5 Intriguingly, if we recalculate LVMI from median values in the groups with different height distributions studied by Foster et al,1 it can be estimated that the large differences observed across groups after indexation to height2.7 (36.2, 36.1, and 50.8 g/m2.7; entire population 41.9 g/m2.7) can be abolished by use of an allometric exponent included in that range. For example, with an exponent of 2, the corresponding resulting values would be 48.7, 48.2, 48.6, and 49.0 g/m2.