Adult height corrected for shrinking and secular trend

Abstract

Background: If in growth studies on adults of different ages and generations height standard deviation scores (SDS) is to be calculated, individual heights should be compared with the average height for that generation (corrected for secular trend) at that age (corrected for shrinking). Aim: To generate mathematical formulas for calculating adult height SDS corrected for shrinking and secular trend. Subjects and methods: Decline of height by age was modelled based on data derived from the Baltimore Longitudinal Study of Aging. Based on mean height of Dutch conscripts in 1917 and data from four consecutive nationwide growth studies (1955, 1965, 1980 and 1997), and assuming a constant male-female difference of 13 cm, the secular trend was modelled over 80 years. The average coefficient of variation was calculated from the last three Dutch nationwide growth studies. Results: Height at 21 years of age can be estimated as current height + 0.042*(age − 21) + 0.0015*(age − 21)² for males, and current height + 0.039*(age − 21) + 0.0019*(age − 21)² for females. Mean height of 21-year-olds between 1917 and 1997 is characterized by linear regression. The coefficient of variation was close to 3.8%. Adult height SDS is calculated by the equation (A − B)/(0.038B), where A is the estimated individual height at 21 years and B is the average height of 21-year-olds in that generation. Conclusion: For studies on adult height at various ages and in various generations, these equations can be used to reduce experimental noise.