Objectives: To evaluate the diagnostic accuracy and predictive ability of various clinical signs of shock for identifying children requiring immediate intervention.
Methods: The authors conducted this prospective study (2019-2023) in children aged 2 mo to ≤ 17 y presenting to the pediatric emergency with an acute illness. Children with shock, defined by the presence of one/more clinical signs and, requirement for fluid boluses/vasoactive agents within the first hour were compared with children without shock. The primary outcome was to evaluate the diagnostic accuracy of the clinical signs; further stratified by malnutrition and severe anemia.
Results: Two thousand five hundred seventy-nine (2,579) children were enrolled, 288 in the shock group and 2291 in the non-shock group. Feeble pulses demonstrated the best sensitivity and specificity (> 85%) with a diagnostic odds ratio of 38.8 (95% CI: 27.2, 55). Tachycardia and capillary refill time (CRT) > 2 s demonstrated sensitivity and specificity of > 70% each. In children with malnutrition and severe anemia, hypotension and coma were the strongest predictors of shock. Using Classification and Regression Tree (CART) analysis, the authors developed a model and validated it in internal and external datasets. Three primary scenarios emerged in the CART analysis, comprising of a combination of feeble pulses, tachypnea, hypotension and/or cool peripheries.
Conclusions: In children presenting with acute illness, presence of feeble pulses was the best predictor of shock followed by tachycardia and prolonged CRT. The CART model, which showed three possible scenarios, needs further validation and may be used as a bedside approach to identify patients with shock requiring interventions.
Keywords: Children; Clinical signs; Fluid bolus; Impaired circulation; Shock.
© 2026. The Author(s), under exclusive licence to Dr. K C Chaudhuri Foundation.