Background and aim The waveform amplitude produced by pulse oximeters can be expressed as an index of pulsatile vs. non-pulsatile signal. This perfusion index (PI) has been shown to correlate with cardiac output, stroke volume, and superior vena cava flow. The aim was to gather PI reference data in preterm infants and to explore if the PI is associated with common clinical parameters.

Patients/methods The PI was recorded in 312 neonates <32 weeks GA during the first 72 h of life. Mixed-effects modelling was applied with PI as the dependent variable and the individual patient as a random factor. Subsequently the association with clinical parameters (i.e. GA, birth weight, IVH, PDA, inotropes) was explored.

Results Mean GA was 28.5 weeks (SD ± 2.1). A quadratic model (0–24 h) combined with a linear model (24–72 h) provided the best fit. The lowest PI was reached 12–18 h after birth, thereafter gradually increasing until 72 h postnatal age. For the first 24 h PI was associated with gender (coefficient 0.05, p = 0.04), inotrope administration (-0.123, p < 0.0001), pulse pressure (0.014, p < 0.0001), SaO₂ (-0.015, p < 0.0001), MABP (-0.013, p < 0.0001), and GA (0.014, p = 0.0168). After the first day, only associations with, inotrope administration (-0.17, p < 0.0001), pulse pressure (0.007, p < 0.0001), MABP (-0.014, p < 0.0001), and SaO₂ (-0.01, p < 0.0001) remained. No association was found with, IVH, PDA, fluid boluses, or birth weight.

Conclusions The evolution of PI values over time probably reflects transitional physiology. The associations with pulspressure, MABP, and inotrope administration suggest that the PI might have an application in blood pressure management.