Reference ranges for fetal ventricular width: a non-normal approach

Abstract

Objectives: Normal ranges for fetal cerebral ventricular width are usually based on parametric methods, which define cut-off values based on means and standard deviations. We aimed to construct reference charts based on a large sample and using a method that better fits with skewed measurements.

Methods: In November 2001-December 2005, a single operator, who was unaware of the subsequent analysis, performed prospectively and routinely fetal cerebral ventricular width measurements. We used the LMS method (which summarizes the changing distribution of a variable of interest according to a covariate (gestational age) using three curves representing the median (M), the coefficient of variation (S) and skewness (L)) to describe the distribution of measurements with gestational age. The LMS method provides a general method for fitting smooth centile curves to reference data. It utilizes the power transformation family of Box and Cox to allow the skewness of the measurement distribution, as well as the median and variability, to vary with age. The theoretical proportion of fetal ventricular widths > or = 10 mm based on the LMS method and on methods based on the assumption of normality were compared with the actual proportion using binomial tests. The influence on ventricular size of gestational age and fetal gender and presentation was assessed using robust regression methods.

Results: Included in the study were 4769 ventricular measurements, 42 (0.88%) of which were > or = 10 mm. The LMS method allowed us to fit the skewness of the measurements and identified slight changes throughout gestation. The actual proportion of measurements > or = 10 mm was not different from that expected according to the LMS method (0.88% vs. 1%, P = 0.25), whereas it was significantly different from that based on the classical normal approximation (0.88% vs. 0.4%, P < 0.0001). There was a significant difference with gender (males: + 0.29 +/- 0.042 mm as compared to females, P < 0.0001), but there was no difference with fetal presentation (P = 0.4).

Conclusion: We present new reference values for fetal ventricular width based on the LMS method. A fixed cut-off value of 10 mm identifies around 1% of the population throughout gestation, which is more than that calculated in previous studies.