We read with interest the article by Levy et al . titled “Reference Ranges of Left Ventricular Strain Measures by Two-Dimensional Speckle-Tracking Echocardiography in Children: A Systematic Review and Meta-Analysis,” recently published in the Journal of the American Society of Echocardiography .

There is increasing interest in the application of deformation analysis in the clinical management of children and young adults with congenital heart disease. Deformation indices, like any other echocardiographic index in the pediatric age range, need to be interpreted according to normative data for a given age and a given body size. This has led various investigators to calculate pediatric ranges of normality for ventricular strain.

Levy et al . performed a very detailed and accurate literature review of published studies, providing a timely report on pediatric two-dimensional speckle-tracking-derived left ventricular (LV) strain nomograms. In accordance with a recent publication from our group, the investigators outlined heterogeneity and inconsistency among current pediatric LV strain nomograms, which are affected by multiple methodologic and numeric limitations. Methodologic issues stem from differences in study protocols, image acquisitions, techniques used, data normalization (i.e., the division of the population into different growth strata), and data expression (i.e., mean values rather than Z scores). Numeric issues are limited sample size (all studies consisted of <250 subjects) and the lack of data on neonates and infants. Discrepancies of current nomograms hamper the possibility of comparisons among them and may pose serious difficulty for a meta-analysis with acceptable statistical power. Regardless of potential limitations and bias, however, Levy et al . performed an elegant meta-analysis of published studies. Because most of the studies reporting pediatric reference intervals for LV strain have normalized by age, we would have expected Levy et al . to present their results by age group. Instead, as in a previous report on normative data for right ventricular strain, they have proposed a single range of normality (mean ± SD) for the whole pediatric population. For example, the mean value for global longitudinal strain is −20.2% (CI, −19.5% to −20.8%), mean global circumferential strain is −22.3% (CI, −19.9% to −24.6%), and mean global radial strain is 45.2% (CI, 38.3% to 51.7%). Thus, single intervals of normality have been used to fit all pediatric ages.

We believe that this message may be confusing and misleading. First, it implies that contrary to all other dimensional and functional indices, deformation indices do not vary with growth. The authors’ conclusion that LV strain does not vary significantly with age is in contrast with several other previous publications, indicating how maturational changes in LV systolic deformation indices occur during growth. Although some reproducible patterns have been noted in older children, with values not far from those in adults, deformation indices vary greatly during the first years of life because of physiologic variations occurring during growth, especially in neonates and infants. A significant influence of heart rate, particularly the higher rates encountered at younger ages, on systolic deformation indices has also been demonstrated. Therefore, normative data for neonates and infants cannot be the same as those adopted for a 17-year-old adolescent. The results of Levy et al . are also in contrast with those of another recent meta-analysis on normal ranges of LV strain in children.

Second, the idea to adopt a single interval of normality for all pediatric ages represents a concept distant from recent literature on pediatric nomograms. In the past few years, there have been debates about the limitations of current pediatric echocardiographic nomograms and the need for more reliable tools to assess disease severity. Studies are ongoing both in the United States and in Europe to produce more reliable Z scores, on the basis of a wide sample of healthy subjects and calculated using a rigorous statistical approach. The goal of this work is to carefully assess potential differences and confounders to produce specific ranges of normality for different ages, body sizes, genders, and races. Thus, for a given measurement and a child of a given gender, race, age, and body size, a specific Z score should be generated.

Current normative data for LV strain are certainly not perfect tools, but they may give us an idea of what values to expect for a child of a given age. Additional criteria may direct the choice of one nomogram over another. For example, for speckle-tracking LV strain analysis, we advise using the normative data of Marcus et al ., which were calculated from a large data set of healthy children. They presented ranges of normality for five different age groups: 0, 1 to 4, 5 to 9, 10 to 14, and 14 to 19 years. Similarly, Zhang et al . proposed pediatric normal values for three-dimensional LV strain measures from 228 healthy children with different ranges of normality for five age groups (<1, 1–5, 5–9, 9–13, and 9–18 years).

In our opinion, the availability of robust nomograms is essential for a correct interpretation of strain analysis in the pediatric age. Thus, pediatric nomograms for ventricular strain involving a large population of normal children obtained using standardized methodology and evaluating all potential confounders are warranted. In this sense, meta-analysis, although attractive and helpful in the absence of prospective studies involving a large number of children stratified for age, may produce results of questionable value. In particular, the adoption of strain analysis of single intervals of normality for all pediatric ages groups (i.e., from neonates to adolescents) seems to be of limited applicability. Strain analysis in children should be performed by comparing data with available age-specific ranges of normality, cognizant of the strengths and limitations of currently available normative data and of potential differences among different techniques and vendors.