Background: The perfusion index (PI) derived from pulse oximetry readings represents the ratio of pulsatile (arterial blood) and nonpulsatile contributors to infrared light absorption. PI has been shown to correlate with cardiac performance. In theory, PI is readily available on every pulse oximeter; therefore, no additional sensors or infant handling are required. Currently, reference values are lacking in (preterm) neonates and the association with common clinical conditions is unclear. Objectives: To establish reference values for the PI in premature infants and at the same time determine the influence of common clinical conditions. Methods: PI was prospectively monitored on the lower limb for 72 h in 311 neonates born with a gestational age <32 weeks between January 2011 and December 2013. Longitudinal mixed-effects modeling was used. Linear, quadratic, and cubic models were explored. Main effects and interactions were investigated. Results: A squared model (0-24 h) followed by a linear model (24-72 h) provided the best fit of the data. PI was lowest around 12-18 h after birth and showed a steady increase thereafter. PI was positively related with female gender, gestational age, and pulse pressure. Negative associations were found with SIMV/HFOV ventilation, dopamine administration, mean arterial blood pressure, and arterial oxygen saturation. Although more complex, the general association with a patent ductus arteriosus was positive. Conclusion: PI varied according to several clinical conditions. The association with common clinical factors suggests that PI might be used for monitoring neonatal hemodynamics and possibly as an additional guidance for interventions.