Desenvolvimento de equação para estimar peso corporal de idosos

Abstract

Body weight is an important anthropometric measurement for clinical and nutritional practice. However, there are situations that limit or prevent its collection, as in the case of the bedridden elderly. Equations for weight estimation are an alternative means for obtaining this anthropometric measurement. This thesis presents two articles. The first article, entitled EQUATION FOR ESTIMATING BODY WEIGHT IN THE ELDERLY: A SYSTEMATIC REVIEW, presented a systematic review of the scientific literature related to body weight estimation equations focused on the elderly. To this end, the project design followed the PRISMA recommendations and it was registered on the PROSPERO database (CRD42014013639). The PubMed and LILACS databases were searched for articles using the keywords in English and their Portuguese equivalents, weight/peso, equation(s)/equação(ões), formula(s)/fórmula(s), older and elderly/idoso, with the help of the truncated terms predict* and estimate* and boolean operators “OR”, “AND” and “NOT”. A total of 4719 articles was identified, of which, 11 articles in English and one in Spanish were selected. The earliest article was published in 1988 and the most recent in 2013. Twenty-eight equations were presented, including 20 in an arithmetic form and 8 in arithmetic + nomograms. Eleven equations were developed for men and 12 for women. Anthropometric measurements were used in them all. The number of variables was between two and six. The variable age was present in four equations and gender in only one. It was concluded that the choice of equation for a body weight estimation in the elderly should be based, among other things, on the characteristics of the target population, availability of resources (instruments), and practicality and/or possibility of obtaining the variables. The second article, entitled DEVELOPMENT OF AN EQUATION TO ESTIMATE SENIORS’ BODY WEIGHT, aimed to develop an equation for estimating body weight in the elderly. It resulted from a cross-sectional study carried out with a sample of elderly people admitted to a university hospital. The variables investigated were age, gender, body weight and anthropometric measurements (knee height, arm and crown-rump length, abdominal, arm, thigh, calf and neck circumference). The research project was approved by the Research Ethics Committee of PUCRS (number 643.175 and 790.495). The study included 131 elderly individuals, 75 (57.3%) being male. Mean age was 70.5±7.6 years (60-96 years). Eight multiple linear regression equations with stepwise automatic extraction were developed (R2>0.71 to 0.95). Equation 3 (R2=0.91) was chosen, composed of three anthropometric measurements: -113.98 + (0.683 x abdominal circumference) + (0.969 x crown-rump length) + (0.875 x calf circumference). Excellent homogeneity was observed in the correlation between the weight of the developed equation and the weight of the equation by Chumlea et al. (Intraclass correlation coefficient = 0.925). The Bland-Altman plot showed there was a random distribution of the measurements within the 95% interval. It was concluded that the equation proved to have practicality (being composed of only three anthropometric measurements, easily collected by means of a tape measure and infant anthropometer) and accuracy in estimating the body weight of the elderly.