Bedside tibial ultrasound was reproducible, but not always measurable

In 102 extremely preterm or extremely low-birth-weight infants, complete tibial ultrasound measurements had inter-rater ICCs of 0.84 at 1 month and 0.94 at 36 weeks' PMA. Yet 12 of 98 infants available at 1 month had no reading, often because of instability or device limits. Repeatability is promising; diagnostic accuracy remains untested.

Hot take

Intra-rater reliability at 1 mo post-birth and 36 wk PMA was excellent across all operators.

Bedside tibial quantitative ultrasound cleared one engineering hurdle: operators usually obtained highly similar speed-of-sound readings. But reliability was calculated only from complete measurements, and the measurement pathway was least complete among the smallest and sickest infants. This study supports repeatability, not diagnostic accuracy.

What the team tested

This was a prospective reliability substudy nested in an ongoing vitamin D trial at one level IV NICU. The vitamin D assignment was blinded and irrelevant to this analysis because every participant was scheduled for the same ultrasound protocol. The substudy enrolled 102 infants 24 to 96 hours after birth and did not have its own sample-size calculation.