Correspondence of two procedures to measure abdominal circumference in a convenience sample of urban, middle-class schoolchildren in Guatemala City.

BACKGROUND Given an emerging concern for juvenile overweight and obesity, even in populations of developing countries, techniques to assess central obesity, such as abdominal circumference (AC), are of increasing interest. OBJECTIVE To compare two procedures for measuring AC in schoolchildren from a convenience sample of middle-class, urban children aged 72 to 131 months in Guatemala City. METHODS We conducted a cross-sectional study (August 2004- April 2005) in 363 schoolchildren; 190 boys and 173 girls. The AC was recorded using both the conventional natural waist circumference (NWC) and the umbilical-level abdominal circumference (UAC) methods; each of which was measured over a thin T-shirt. Height and weight were also measured. The NWC and UAC methods were compared between genders and age groups. In addition, correlation coefficients and Bland-Altman regression were performed. A reproducibility study was performed on a subsample of 50 children. RESULTS Mean NWC was 63.2+/-10.0 cm for boys and 60.2+/-7.7 cm for girls (p=0.001), and mean UAC was 65.9+/-11.0 cm for boys and 63.2+/-8.8 cm for girls (p=0.008). Both measures of AC generally increased with age. The UAC - NWC difference increased with abdominal girth. The Pearson correlation coefficient of BMI vs. NWC and BMI vs. UAC was r=0.92, CV=0.46+/-0.27%. A high test-retest reproducibility (r > 0.98, CV= 0.50+/-0.24%) of the NWC and UAC measurements was found. CONCLUSION Reliable NWC and UAC measurements may be obtained by applying a correction term to account for light clothing. Both measurements are applicable methodologies for the collection of data in populations with cultural limitations.