Justus Hayes / Shoes on Wires / shoesonwires.comWhile we are well-equipped with nomograms and evidence-based studies and guidelines to determine when to put a jaundiced infant under phototherapy lights, we have less evidence about when to safely stop phototherapy without significant rebound levels. If only someone could give us a prediction rule for estimating the probability of rebound hyperbilirubinemia occurring when phototherapy is discontinued. Fortunately Chang et al. (10.1542/peds.2016-2896) have devised such a rule based on a study that we are publishing this month involving a retrospective cohort of infants >= 35 weeks gestation enrolled over two years at 17 Kaiser Permanente northern California hospitals who were treated with phototherapy before 2 weeks of life. The authors used stepwise logistic regression to determine possible predictors of rebound (bilirubin rise to phototherapy threshold) hyperbilirubinemia. The authors found three variables did a great job of predicting a score that could indicate when the rebound bilirubinemia might be worrisome enough to not discontinue phototherapy. These three variables consisted of an infant’s gestational age, the age of starting phototherapy and the total serum bilirubin relative to the treatment threshold at phototherapy termination. Just how the scoring system works and how good a predictor it is awaits your linking to this study to learn more.
Suffice it to say, a validation dataset used to confirm the authors derived prediction rule based on these three variables worked extremely well enabling the authors to predict whether a baby would have a high or low probability of rebound before they turned off the lights. To shine some further light (perhaps ultraviolet light) on how you can use this study to help determine your own patients’ risk for a rebound, Dr. Jeffrey Maisels, one of the leading national and international experts on all aspects of hyperbilirubinemia (10.1542/peds.2016-3832) provides some input and further interpretation of the findings in this study through an accompanying commentary. Read this study and commentary and then consider if you want to apply the prediction rule introduced in the Chang et al. article on your next jaundiced infant. Let us know by responding to this blog, sharing a comment attached to the study, or posting on our Facebook or Twitter sites if you find it is as valid and reliable as the authors did in your own practice.