Do We Finally Have the A to Z of Z Scores?

• Editorials
• adult
• confidence intervals
• diastole
• noise
• sample size

In this issue of Circulation: Cardiovascular Imaging, the North American Pediatric Heart Network reports Z scores of 2-dimensional echocardiographic measurements derived from over 3000 subjects.¹ This represents another stride forward in the assessment of pediatric measurements, where Z scores play a crucial role in decision-making. For those in adult practice, the question of normality of an individual measurement is frequently addressed simply by comparison to a normal reference range. However, in pediatric practice, there is the added dimension of somatic growth: a single reference range cannot be applied across patients of vastly different sizes and ages. Z scores express how many SD above (positive values) or below (negative values) a given measurement lies with respect to the size-specific mean, but SDs are not constant across body sizes (heteroscedasticity), further adding to the complexity of analysis. The reliability of such data is crucial because important clinical decisions may be based on the interpretation of these measurements, for example, biventricular versus single ventricle repair or timing of aortic or mitral valve repair or replacement. This type of approach has been widely applied in both pediatric and fetal echocardiography to assist analysis of sequential measurements from the same patient or to compare different groups of patients.

See Article by Lopez et al

Many studies over the past 25 years have sought to establish sets of algorithms that may be used to calculate cardiac Z scores. When estimating a population mean and SD from a sample, the sample size is a critical determinant of the accuracy of the estimation, and the 3215 subjects assessed in the study by Lopez et al¹ (minimum 495 in each age range) are the largest to date. Only Cantinotti et al² come close in terms of sampling density, reviewing 445 subjects in just the …