Agreement between measured height, and height predicted from published equations, in adult South African patients

Objectives: Estimation equations based on different body segments are commonly used to predict height in patients whose height cannot be directly measured. This study aimed to assess the agreement between measured (reference) height and height predicted from published equations derived from measurement of body segments, in a South African public hospital setting. Design: A descriptive cross-sectional study was undertaken. Setting: Medical, surgical, pulmonary, orthopaedic, cardiovascular and general wards at three public hospitals in Bloemfontein. Subjects: Admitted patients, 20–50 years old; able to stand upright without assistance and without medical conditions or treatments affecting height. Outcome measures: Stadiometer height, recumbent height, arm span, demi-span, ulna length, knee height, tibia length, fibula length and foot length were measured with standardised techniques. Height, predicted by 12 published equations, was compared with stadiometer height by 95% confidence intervals (CI) and Bland–Altman analysis. Results: The median stadiometer height of the sample (n = 141; 38.3% female; median age 38.8 years, IQR 33.3–44.4 years) was 165.5 cm (males 169.3 cm; females 158.4 cm). Only a set of equations based on knee height and standardised on a large population of adults < 65 years in the United States estimated height without statistically significant deviance from the stadiometer height. Conclusions: Most standardised equations applied to hospitalised adults in a South African public health setting resulted in height estimations that differed significantly from height. Thus, equations standardised on other populations may not be suitable for the South African population, possibly due to differences in genetic and environmental factors.