Variability of abdominal circumference of premature infants.

Serial measurements of abdominal circumference were made in 27 premature infants 1 hour before feeding, immediately before and after feeding, and 1 and 2 hours after feeding. In some infants these measurements were repeated for 2 (n = 18) or 3 (n = 3) consecutive days. All 258 measurements were performed by the same examiner; 124 were also repeated blindly by another examiner to test interobserver reliability. The coefficient of variation of repeated measurements of any one infant by the same examiner on a single day was 2.7%. Ninety percent of repeated measurements agreed within 1.5 cm. The interobserver reliability was 0.99. Abdominal circumference was positively correlated with birth weight (P = .0001) and time from last defecation (P = .0001) and negatively correlated with time from last feeding (P = .04). Increases in abdominal circumference of less than 1.5 cm occur normally and, in the absence of other clinical signs, should not be considered indicative of disease.