Researchers have identified a series of reference intervals and z-scores for 10 left ventricle (LV) echocardiographic parameters, which they say could aid in clinical decision making for infants with cardiopulmonary conditions. The study, carried out on a large cohort of 13,454 healthy newborns by a Denmark and U.S.-based team, created a set of body surface area (BSA)-specific z-scores for each parameter of LV size and function. Although there were no significant sex-specific differences observed, the team noted significant age-specific differences, especially for the measures of LV systolic function and mitral inflow velocities. These differences were significant enough that separate reference intervals were provided for infants younger than 7 days and those aged 7 to 30 days, concluded the study, published Monday online and in the June 6 issue of the Journal of the American College of Cardiology. “Our reference intervals, which are based on a homogeneous cohort of healthy newborns at least an order of magnitude larger than any previous study population, including 2,343 neonates examined within the first 6 days of life, should provide greater stability and precision in z-score estimation in newborns,” the study says. Copenhagen Baby Heart Study (CBHS) data The researchers used findings from the Copenhagen Baby Heart Study (CBHS), a prospective population-based cohort study, that gathered cardiac structure and function data from infancy onward. The study population included healthy CBHS participants (51% male) born at term (37-42 completed weeks’ gestation) and examined within 30 days of birth after uncomplicated pregnancies. Newborn characteristics of the study population included a median age of 12 days for both boys (interquartile range [IQR]: 8- 15 days) and girls (IQR: 8-12 days) when a transthoracic echocardiography (TTE) was performed. The 10 LV echocardiographic parameters assessed in the newborn cohort were interventricular septal thickness in end diastole (IVSd); left ventricle posterior wall thickness in end diastole (LVPWd) and left ventricle internal diameter in end diastole (LVIDd). Other parameters include left ventricle internal diameter in end-systole (LVIDs), end-diastolic volume (EDV), end-systolic volume (ESV), stroke volume (SV), transmitral early peak inflow velocity (MvE), transmitral atrial peak inflow velocity (MvA) and mitral valve deceleration time (MV Dect). The team then estimated BSA-specific z-scores for each LV parameter using an established model with four essential components: the M-curve, S-curve, G-curve and D-curve, which are related to the mean; the SD; the inverse of skewness; and kurtosis of the distribution of an LV outcome of interest. For each LV parameter, the research team also estimated BSA-specific z-scores within age strata. To achieve this, the three intercepts/constants in the M-, S-, and G-curves were made stratum-specific (age-specific). No stratification needed on infant sex Results by the team, led by R. Ottilia B. Vøgg, MD, PhD, from Copenhagen University Hospital Herlev and Statens Serum Institut, both in Copenhagen, Denmark, found that all normative reference intervals performed well in both sexes without stratification on infant sex. In contrast, creation of separate reference models for infants examined at <7 days of age and those examined at 7 to 30 days of age was necessary to optimize the performance of the reference intervals. “Our finding that a single model for any given LV parameter rarely fit well across the entire neonatal age span (0 to 30 days) suggesting that the inclusion of additional explanatory variables could improve model accuracy,” the researchers said. “Indeed, including age-specific intercepts for the mean, SD, and skewness improved model fit considerably. To optimize the clinical usefulness of our z-scores, we therefore recommend using separate reference intervals for the first week of life and weeks.” ‘The quest for perfection’ Commenting on the study’s implications, Ritu Sachdeva, MBBS, from the Emory University School of Medicine and Children’s Healthcare of Atlanta, described the references values as a “a step forward in the quest for perfection.” She highlighted the strengths of the study, which include the prospective nature, using a standardized protocol for image acquisition by trained sonographers, and the large sample size, with adequate number of neonates within each age range. The study also used robust statistical methods to address the limitations noted in previous studies, she added. Authoring the accompanying editorial comment, Sachdeva also drew attention to the use of the parasternal long-axis view (PLAX) view for LV measurements, arguing that acquiring PLAX views was much easier than parasternal short-axis (PSAX) view. This approach opposes the American Society of Echocardiography (ASE) guidelines and previous studies, which have used the PSAX view. “There are no studies comparing LV measurements taken in these two views and validating if they are interchangeable,” Sachdeva pointed out. Looking ahead, Sachdeva emphasized the use of artificial intelligence (AI) as potentially improving the accuracy and efficiency of measuring echocardiograms in clinical practice. “The large dataset of normal neonatal echocardiograms in the CBHS could be used to report reference values for other cardiac structures as well as to train and validate AI algorithms,” she concluded. Sources: Vøgg ROB, Sillesen A-S, Wohlfahrt J, et al. Normative Echocardiographic Left Ventricular Parameters and Reference Intervals in Infants. J Am Coll Cardiol. 2023;81:2175–2185. Sachdeva R. Complexities of Normative Database for Neonatal Echocardiography: The Quest for Perfection Continues. J Am Coll Cardiol. 2023;81:2186–2188. Image Credit: Syda Productions – stock.adobe.com