Validity of equations using knee height to predict overall height among older people in Benin.

Abstract : Rationale: Chumlea ’s formulas are a validated means of predicting overall height from knee height (KH) among people over 60 years old, but no formula is validated for use in African countries, including Benin. The purpose of this study was to compare height provided by predictive formulas using KH to measured height in an elderly population in Benin. Methods: People over 60 years of age in Benin underwent nutritional assessment with determination of weight, body mass index (BMI), height and KH. A Bland and Altman analysis was carried out by gender and age. The percentage of predictions accurate to ± 5cm compared to the measured height was calculated. The tested formulas were Chumlea ’ s formulas for non-Hispanic Black people (CBP) and two formulas for use among Caucasians (CC1, CC2). Results: Data from 396 subjects were analysed. The mean age was 66.6 ± 5.2 years and mean height of 165.4 ± 8.0 cm. The sex-ratio was of 4.3. The three formulas achieved 98.0% accuracy, but with 4.6% risk of error (±2 SD: − 6 to +9 cm), which appeared to make them unfitted for the whole population. The predictions of the three formulas in the total population are presented in Table 1. Nevertheless, if a level of prediction ± 5 cm is considered acceptable in clinical practice, the CBP formula achieved 83.1% accuracy. Moreover, there was no significant difference in BMI calculated with the measured and the predicted height, and the nutritional status based on BMI did not differ. Table 1: Predictions of the three formulas studied. n = 396 Acurate prediction (%) Overestimation (+5 cm) (%) Underestimation (−5 cm) (%) CBP 83.1 12.4 4.5 CC1 80.3 15.9 3.8 CC2 78.5 18.2 3.3 Conclusion: CBP formulas seem applicable in 83% of cases (±5 cm) to assess the height with KH of older people in Benin and do not overestimate the prevalence of malnutrition.