Editor,—Midparental height is an important measure in estimating a child's target height—the genetic potential in stature. Height reference values that allow for parental height are more appropriate for growth evaluation in paediatric clinics. We read with interest the recent paper by Wright and Cheetham on the strengths and limitations of parental heights as a predictor of attained height.1 The authors concluded that midparental height was a useful indicator of the expected height for children when their parents were of average stature but misleading when used to assess short children. We have recently reported the same findings based in 2402 Swedish children.2 We observed that the regression coefficient between midparental height and a child's final height was approximately 0.6 in standard deviation scores (it was 0.5 for children 8 years of age in the paper by Wright and Cheetham).

We believe that the linear function of midparental height could be used to estimate a child's target height, rather than midparental or corrected midparental height, which Wright and Cheetham implicitly used to represent a child's genetic target height. The meaning of midparental height is different for children with short, average, and tall parents. The parents' heights not only reflect the parents' genotype in stature, but also mirrors the extrinsic influences the parents experienced during their own growth span. This provides a biologically meaningful explanation of the so called “regression to the mean phenomenon”. For instance, the intrinsic genetic potential in stature of short parents is usually much greater than their measured heights; consequently, the following generation is usually taller due to a better manifestation of the intrinsic growth potential.2

We agree that short children attending paediatric clinics are usually shorter than their target height, whatever method is used for estimation. The height of parents is important for clinical evaluation of short children. A short child with tall parents is certainly more likely to have a pathological cause than a short child of short parents. It is not appropriate to consider midparental height itself as a simple measure of target height. Clearly, midparental height is not misleading for any child if its linear function is used for estimating a child's target height—the genetic potential in stature.