The relationship between skin color and illness severity in the newborn remains untested. We have evaluated the predictive value of skin color readings for illness severity in a population of high-risk newborn infants. A prospective study was conducted on 107 white newborns in the intensive care unit, which were categorized as either high or low severity of illness, defined by the presence of severe neonatal morbidity. Illness severity was also determined using a Score for Neonatal Acute Physiology (SNAP). Color readings were obtained at 10 different body sites using a portable tristimulus colorimeter during the first 24 h, and color was expressed using the standard CIE L*a*b* system. Skin CIE b* values were significantly lower in the high severity group (p < 0.0001), and a significant inverse correlation with SNAP was observed (r(s) range, -0.37 to -0.71, p < 0.0001). In particular a low b* value for the abdomen was found to be a significant predictor of illness severity (92.6% sensitivity; 96.6% specificity; 96% positive predictive value; 93.7% negative predictive value; adjusted odds ratio, 14.7; 95% confidence interval, 6.4 to 33.8). Our findings indicate that skin color reflects clinical severity in the newborn and that skin colorimetry can accurately identify higher risk infants.