Hypothermia, or low body temperature, is concerning in infants, and there are a number of potential causes. It can be difficult to determine if the low temperature is due to inadequate temperature regulation, environmental exposure, or an underlying bacterial infection.1 The strength of low temperature’s association with true bacterial illness has remained uncertain in previous literature.2,3 Understanding whether low temperature meaningfully predicts bacterial infection, especially in relation to key developmental age groups, is important for informing evaluation pathways and avoiding unnecessary testing.
We studied 128,979 U.S. infants younger than 8 weeks old who had an emergency department (ED) visit between January 1, 2017, and September 30, 2025, and who had a blood, cerebrospinal fluid, or urine culture performed; a documented temperature during the visit; and a birth record in Cosmos. Infants were classified as having a low temperature (<36°C), normal temperature (36–38°C), or elevated temperature (>38°C) based on the lowest or highest temperature recorded during the visit. We then looked for an abnormal blood, cerebrospinal fluid, or urine culture as confirmation of a bacterial infection. Infants were grouped into those who were born prematurely (<37 weeks) or who had a NICU stay near the time of birth and those who were full term without a NICU stay. We accounted for social vulnerability based on residence, rurality based on residence, demographics, ED acuity level, census region of residence, and age group (birth to 3 weeks, 4 weeks, 5 to 8 weeks). Clinical guidelines for managing newborns with an abnormal temperature vary based on the infant’s age.4
Across all age groups, elevated temperature was a strong predictor of bacterial infection, whereas low temperature demonstrated an age-dependent pattern, as shown in Figure 1. Among 5- to 8-week-old infants, those with a low temperature were 42% more likely to have a bacterial infection if they were full term without a NICU stay and 53% more likely if they were preterm or had a NICU stay, compared with infants who had a normal temperature. Among 4-week-old infants, low temperature was not associated with a significant change in infection likelihood in either group. Among the youngest infants (birth to 3 weeks old), there was no clinically meaningful increase in infection likelihood with low temperature, which was associated with only a 9% increase in likelihood for full term infants without a NICU stay and a statistically insignificant increase for infants who were premature or had a NICU stay.
These findings suggest that low body temperature might be a meaningful predictor of bacterial infections for infants starting at 5 weeks old.
