Estimating the risk of in-hospital mortality in the newborn intensive care unit can provide important information for health-care providers, and illness severity scores have been devised to provide mortality risk estimates. Calculation of illness severity scores is time-consuming, and the information used to predict mortality is collected only for the first 12 to 24 h of life. A noninvasive continuous measure that uses information collected throughout the hospitalization and that requires no data entry could be less costly and more informative. We have previously shown that the abnormal heart rate characteristics (HRC) of reduced variability and transient decelerations accompany neonatal illness such as late-onset sepsis. We hypothesized that more frequent and severe abnormal HRC are associated with an increased risk of death. We tested this hypothesis in two ways. Using data on infants older than 7 d of age, we first determined the association of the HRC index with death in the next week. Second, we devised a cumulative HRC score and determined its association with in-hospital death. There were 37 deaths in the 685 patients. The major findings were 1) the HRC index showed highly significant association with death in the succeeding 7 d (receiver-operating characteristic area > 0.7, p < 0.001), and 2) the cumulative HRC was highly significantly associated with neonatal in-hospital mortality (receiver-operating characteristic area > 0.80, p < 0.001). In both analyses, HRC added information to birth weight, gestational age, and postnatal age (p < 0.01). The HRC index provides independent information about the risk of neonatal death in the upcoming 7 d, and the cumulative HRC is an estimate of the risk of in-hospital neonatal mortality.