Association of stress hyperglycemia with severe outcomes in ill-appearing children

We aimed to determine the frequency of stress hyperglycemia (SH) among ill-appearing children, identify clinical correlates of hyperglycemia at presentation, and evaluate the association between glucose levels and life-saving interventions (LSIs), severe outcomes and length of stay (LOS).

This retrospective study included ill-appearing children identified by the Pediatric Assessment Triangle who presented to a tertiary PED between January 1 and December 31, 2025. Initial capillary glucose was analyzed as a continuous variable. Multivariable models were adjusted for age, sex, comorbidities, high-risk vital signs, and GCS score. Outcomes included LSIs (airway, hemodynamic, medication, electrical and procedural), severe outcomes (PICU admission or death), and LOS.

Among 1116 children, 10.6% had hyperglycemia (≥150 mg/dL). Patients with diabetic ketoacidosis (n = 15) were excluded. Overall, 10.8% were admitted to the PICU and 0.4% died. Glucose showed modest discrimination for severe outcomes (AUC 0.654; optimal cutoff 125 mg/dL; sensitivity 43.3%, specificity 80.4%). After adjustment, higher glucose levels were independently associated with severe outcomes (aOR 1.13 per 10 mg/dL increase; p < 0.001) and airway intervention (aOR 1.10; p = 0.001), but not with receipt of any LSI overall (p = 0.42). Although glucose was not associated with mean LOS, its effect increased across longer stays, corresponding to increases of 0.9, 2.5, and 5.3 h per 10 mg/dL increase at the 25th, 50th, and 75th percentiles, respectively (all p < 0.001).

Initial glucose may serve as a useful marker of clinical risk in ill-appearing children. Higher glucose levels were independently associated with airway intervention, severe clinical outcomes and longer hospitalization. These findings suggest that glucose measurement may contribute to early risk assessment in the PED and warrant further prospective validation.