We thank Dr. Cantinotti and his colleagues for their thoughtful comments and appreciate their concerns regarding the use of reference ranges of deformation imaging that reflect “normalcy” for children across a wide range of ages. We wish to clear up any confusion from our recent meta-analysis.

We agree with Cantinotti et al . that what may be “normal” in a neonate may be different from what is normal in a child, and an adolescent, even if some efforts are made to index for age, body size, height, or some other parameter. On the basis of the studies we analyzed, our meta-analysis does include several figures (Figures 2, 3, and 4), tables (Table 3), and appendices (Appendices 4 and 5) in which we summarize normal left ventricular (LV) global and segmental strain measures in children divided by age group (and vendor) and provide data (where available) to address the potential effect of the major sources of bias (i.e., demographics and clinical variables, age, and vendor hardware and software). To specifically assess the contribution of age to the variation in the reported reference values, we stratified the data by age distribution (years) in children: infancy (0–1 years), prepuberty (2–9 years), puberty (10–13 years), and late adolescence (14–21 years). Despite the lack of explanation of these variable (age, vendor, etc) in causing heterogeneity between studies, it should not be misconstrued to mean that these biases do not influence strain. We did not set out to propose a single range of normality (mean ± SD) for the whole pediatric population, but notwithstanding these biases, the strain values from these studies are relatively constant across all pediatric ages.

We recognized that the effect of age on LV strain during growth was unclear from individual studies. Marcus et al . observed a statistically significant “second-order polynomial relation” between global peak systolic strain parameters and age in that deformation patterns were lowest in the youngest and oldest age groups. However, other studies found little to no effect of age on deformation patterns. Zhang et al . and Kaku et al . demonstrated that there were small maturational changes in some strain parameters that were “statistically significant but probably clinically irrelevant.” Our group has previously demonstrated that global longitudinal strain did not change significantly with maturation and declining heart rate from birth to 18 years of age. Klitsie et al . showed no linear relation between age and most global peak strain parameters derived by two-dimensional (2D) speckle-tracking echocardiography (STE). Labombarda et al . demonstrated that in healthy control subjects, strain was preserved throughout maturation irrespective of age. Finally, Jashari et al . concluded from their recent meta-analysis that only longitudinal systolic strain rate was significantly determined by age, but this conclusion was limited by the methodology and small number of data sets included in that study.

We appreciate the observation by Cantinotti et al . that strain values should be interpreted according to normative data for a given age and body size. However, not all pediatric echocardiographic measures are interpreted according to normative data for a given age or indexed (i.e., fractional shortening and ejection fraction have reference ranges that are applicable for all pediatric age groups). Deformation imaging measures intrinsic myocardial contractile properties that are established during the second half of pregnancy and remain constant throughout gestation and after birth. The individual unit of the myocardium, the myocyte, does not significantly vary in length from 1.67 μm in diastole to 1.0 μm in systole with age or change in body growth under the same resting physiologic conditions. Deformation values are therefore normalized against their own values from diastole to systole, and strain is a unitless dimension that is indexed against itself.

With regard to the concerns of Cantinotti et al . over the potential methodologic limitations of meta-analysis-derived LV strain nomograms, we used a strict quality analysis detailed in Appendix 3 to address these recognized limitations. The eligible data sets met >75% of the quality checklist items. Furthermore, we contacted all the authors of the eligible studies by e-mail to understand the potential differences in methodologies and attempt to decrease heterogeneity among studies. Cantinotti et al . performed a nice “systematic search” in PubMed only to review the published nomograms of LV strain derived by Doppler tissue imaging (DTI), 2D STE, and three-dimensional echocardiography in children, and many of the methodologic limitations outlined are due to different techniques to acquire strain imaging. Our study provided a comprehensive range of 2D STE–derived strain values on the basis of vendor and age by pooling a large population of healthy children, but we did not address strain derived by DTI or magnetic resonance imaging.

In response to Cantinotti et al .’s concern about the inclusion of studies with small numbers of patients, the reality is that pediatric studies have smaller numbers of patients than studies of adults, as was evident by the inclusion of 2,325 children from 43 data sets in our meta-analysis compared with the 2,597 subjects from 24 studies from Yingchoncharoen et al .’s original meta-analysis of the normal ranges of LV strain in adults. We agree with Cantinotti et al . that a large multicenter prospective study of a large number of children across a wide age range, stratified by age and vendor, and adjusting for all of the potential confounding factors that may contribute to variance in reference measures, is needed. Although meta-analysis research is dependent on the quality of the published studies available and is not a substitute for a carefully designed prospective study, a carefully done meta-analysis can have value when it provides observations that help refine the questions that need to be addressed in future studies.

In this poststandardization era of deformation imaging, reference ranges of strain values established with meta-analyses coupled with forthcoming work from the European Association of Cardiovascular Imaging and American Society Echocardiography industry task force to standardize strain imaging (and potentially build reliable and clinically relevant Z scores) will permit strain imaging to be used more routinely to assess clinical changes in myocardial function and provide a valid basis that allows comparison among studies. Although all of the individual studies included in our meta-analysis were well done, we would not endorse relying on only one study over another from the prestandardization strain era because of the inherent limitations of vendor and software reproducibility among individual studies.

Finally, we agree that there is an influence of heart rate on strain imaging, but we would argue that frame rate and the ratio of frame rate to heart rate are just as important determinants with 2D STE–derived myocardial strain. Cantinotti et al . identified that DTI-derived strain and in particular strain rate are influenced by elevated heart rate (as evident by the DTI strain work of Pena et al . and James et al . in neonates), but our meta-analysis focused only on 2D STE, and frame rate and its ratio to heart rate were appropriately tested to determine if they influenced the variability in reporting of normal strain and strain rate measures in children.