Pupillometry for pain assessment in noncommunicating children in the pediatric intensive care unit: a prospective accuracy study

Pain assessment in noncommunicating children in pediatric intensive care unit (PICU) remains a significant clinical challenge, particularly in deeply sedated patients. Pupillometry, a noninvasive objective technique, showed promise in anesthesia settings, but its relevance in PICU is unclear. We assessed feasibility and diagnostic performance of video pupillometry for pain assessment in noncommunicating intubated children in PICU. We conducted a prospective, single-center diagnostic accuracy study in the French PICU of Lyon (France). Children aged 0-18 years under mechanical ventilation and sedation were eligible. Pupillary diameter variation was measured using the AlgiScan® video pupillometer. Pain was defined as a COMFORT-B score > 17, assessed by nurses blinded to pupillometry results. Diagnostic performance was evaluated using the area under the ROC curve (AUC). Specificity was estimated at a sensitivity threshold of 90%. Sixty-six patients were included and 231 individual measurements analyzed; 52 patients (85%) had at least one complete series. The AUC was 73.6% [95% CI 61.1-82.6]. A weak correlation between COMFORT-B scores and pupillary diameter variation (Spearman's ρ = 0.4, p < 0.05) was found. At 90% sensitivity, the threshold was 0.145, with a specificity of 52.2%. Feasibility was limited in neonates under 5 kg and agitated children. CONCLUSION: To ensure a sensitivity at 90%, the corresponding specificity was quite low and correlation with COMFORT-B was also weak. Feasibility in young infants remains limited due to technical constraints. Further multicenter studies are warranted to clarify its role in analgesia assessment in PICU, as an adjunct to other techniques. TRIAL REGISTRATION: ClinicalTrials.gov identifier: NCT02847195. WHAT IS KNOWN: • Pain assessment in noncommunicating children in the pediatric intensive care unit (PICU) remains a major clinical challenge, particularly in sedated or critically ill patients. Additional tools are needed to better assess pain in this population. WHAT IS NEW: • A weak correlation was found between pupil variation measured by video pupillometry and the COMFORT-B score in PICU patients. Pupillometry may complement existing tools as part of a multimodal approach to pain assessment in this population.