Echocardiography nomograms in Black South African neonates

Abstract

The subspecialty of Paediatric cardiology has adapted and refined techniques adopted from adult cardiology to manage children, (Mahle, Blaine and Silverman 2009:33). Although a quantitative evaluation of cardiac chambers, valve annulus and great vessel dimensions in paediatric echocardiography is important, nomograms for these structures are limited. Several studies have already provided normal values in the paediatric population that represent most populations of the world but there is lack of data that has been published in sub-Saharan Africans, (Lemmer et al. 2014:1 and Roge et al. 1978:285). AIMS AND OBJECTIVES Aim: To establish reliable echocardiography nomograms for cardiac chambers, valve annuli, great vessels and thymus dimensions in the Black South African neonatal population. Objectives: To determine normal values of cardiac chambers, valve annuli, great vessel and thymus dimensions. To determine inter-observer variability. To determine the effect of confounding factors such as gender and type of delivery on the measurements obtained. METHODS This is a descriptive, cross-sectional study evaluating cardiac chambers, valve annuli, thymus and great vessel dimensions in 386 African neonates with normal hearts using echocardiography. The study data consists of two arms, a retrospective arm utilizing echocardiographic data acquired during a previous study entitled “To assess the accuracy of pulse oximetry screening as a tool to detect critical congenital heart disease in asymptomatic newborns” and a prospective arm. INCLUSION CRITERIA: Healthy newborns at an age of 12 hours or more before discharge. Patients without any heart disease, with a “normal heart” by echocardiography excluding hemodynamically non-significant patent ductus arteriosus and patent foramen ovale. Black South African patients. Full term babies delivered by caesarian section and normal vertex. EXCLUSION CRITERIA: Patients less than 12 hours from birth. Patients with known structural heart lesion. Patients who are non- black South Africans. Pre-term neonates. DATA COLLECTION AND ANALYSIS: Collected data was entered on excel spreadsheet and analyzed using excel, XLSTAT 2019 and STATISTICA version 13.5.0 statistical packages for analysis by the principal researcher (Nondumiso M. Hadebe). A p-value of less than 0.05 and significance level of 95% was considered statistically significant. A professional biostatistician at University of the Witwatersrand was consulted for assistance in data analysis. Homoscedasticity and heteroscedasticity were tested using Shapiro-Wilk, Kolmogorov-Smirnov, Breusch-Pagan and White tests. The inter-observer variability was tested with intraclass correlation coefficient using Pearson’s correlation coefficient to detect bias. Weight was used to express measurements to body size and to predict mean values of each echocardiographic measurement that were expressed as Z-scores. RESULTS A total of 386 patients from both arms were enrolled with almost equal gender distribution with a slightly higher percentage of females, 191 (49%) were males and 195 (51%) were females. The study involved neonates born through normal vaginal delivery (NVD) and through caesarian section (C/S) which showed equal distribution. The assumption of normality was tested which showed most measurements being homoscedastic. Heteroscedasticity was tested which showed most measurements to be homoscedastic. Effects of confounding factors were tested which showed that body weight has a significant effect on all cardiovascular dimension measurements. Mode of Delivery (MOD) had a significant effect on the size of atrioventricular valve, pulmonary artery and pulmonary artery branch measurements. Gender and BSA had no significant effect on most measurements but with some significant effect on a few measurements. There was no significant effects seen for body length (BL) and gestational age (GA). The echocardiography measurements of 168 patients were used to test for inter-observer variability which showed a strong correlation on most measurements. Birth weight was used to express cardiac measurements to body size. All echocardiographic measurements were grouped into 3 categories of weight and are presented as mean and +/-3 SD. Z-scores and its boundaries for all measurements is presented graphically. This study showed slightly higher dimensions to previous studies from other centres. CONCLUSION This study presents nomograms that are reliable because they were acquired from healthy neonates using current recommendations by the American Society of Echocardiography, (Lopez et al. 2010: 465-495). Weight showed significant potential as a confounding factor and as an independent variable for data normalization. This study showed slightly higher dimensions to previous studies emphasizing that it is important to develop and use regional nomograms because of the effects of environmental, economic and social factors of the region. This study covered the gap of knowledge on cardiac chamber, valve annuli, and arterial and thymus dimensions in the neonatal age group. Further studies are required to reinforce these findings that will also involve the right cardiac dimensions. More studies are needed based on African paediatric populations including the all paediatric age groups (0-18 years).