To evaluate the accuracy of pre- and postfeeding weights to estimate enteral feeding volumes in preterm infants.

Single-center prospective cohort study of infants 28-36 weeks' corrected age receiving gavage feedings. For each test weight, 3 pre- and 3 postgavage feeding weights were obtained by study personnel, blinded to feeding volume, via a specific protocol. The correlation between test weight difference and actual volume ingested was assessed by the use of summary statistics, Spearman rho, and graphical analyses. The relationship between categorical predictive variables and a predefined acceptable difference (±5 mL) was assessed with the χ² or Fisher exact test.

A total of 101 test weights were performed in 68 infants. Estimated and actual feeding volumes were highly correlated (r = 0.94, P < .001), with a mean absolute difference of 2.95 mL (SD: 2.70; range: 0, 12.3 mL; 5th, 95th percentile: 0, 9.3); 85% of test weights were within ±5 mL of actual feeding volume and did not vary significantly by corrected age, feeding tube or respiratory support type, feeding duration or volume, formula vs breast milk, or caloric density. With adherence to study protocol, 89% of test weights (66/74) were within ±5 mL of actual volume, compared with 71% (19/27, P = .04) when concerns about protocol adherence were noted (eg, difficulty securing oxygen tubing).

Via the use of a standard protocol, feeding volumes can be estimated accurately by pre- and postfeeding weights. Test weighing could be a valuable tool to support direct breastfeeding in the neonatal intensive care unit.