Purpose: To evaluate urine fluorescence as a diagnostic tool.
Procedures: Using a Wood lamp, 60 physicians, assigned to group 1 or 2, independently rated 150 urine specimens from nonpoisoned children as fluorescent or nonfluorescent. Interobserver and intraobserver agreements were assessed. Physician ratings were compared with fluorometry results. The prevalence of urine fluorescence was determined by fluorometry.
Main findings: Group 1 reported fluorescence in 80.7% (95% CI 73.4%-86.6%) of urine specimens; group 2 reported fluorescence in 69.3% (95% CI 61.3%-76.5%). Interrater agreement was poor (72.5%, kappa = 0.25, 95% CI 0.13-0.37); intrarater agreement was good (physician group 1: 97.9%, kappa = 0.93, 95% CI 0.77-1.00; physician group 2: 93.3%, kappa = 0.85, 95% CI 0.69-1.00). The prevalence of urine fluorescence was 100% (95% CI 98.1%-100%).
Conclusion: Our data suggest that determination of urine fluorescence using a Wood lamp is a poor screening tool for suspected antifreeze ingestion in children.