Reliability of the Beit CURE classification for the management of chronic hematogenous osteomyelitis of the long bones in children.

Abstract

Chronic hematogenous osteomyelitis (CHOM) of long bones in children remains a major clinical challenge in low- and middle-income countries. The Beit CURE classification has been used to guide treatment, but its reliability across different centers and raters has not been established. Forty-four children with CHOM were evaluated using digital radiographs by 16 fellowship-trained pediatric orthopedic surgeons from five continents. All the raters were trained in Beit CURE classification and assessed each case twice, 4 weeks apart. Interobserver and intraobserver reliability were calculated using intraclass correlation coefficients (ICCs). The raters also predicted anticipated surgical management and the number of procedures needed. Interobserver reliability was fair overall (ICC: 0.467, 95% confidence interval: 0.36-0.59). Agreement was highest for types A and B4 (>80% consensus) and lowest for types B2, B3, and C (<25%). The reliability for growth plate involvement was moderate (ICC: 0.628) and was poor for site involvement (ICC: 0.269). Predictions regarding surgical planning were inconsistent, especially for B2, B3, and C lesions. The Beit CURE classification demonstrated only limited reproducibility among experienced pediatric orthopedic surgeons. While types A and B4 were reliably identified, other subtypes showed poor agreement, undermining clinical decision-making. Refinements - such as merging B2 and B3, providing clearer definitions of sclerosis and involucrum quality, and integrating clinical parameters - are needed before the Beit CURE classification system can serve as a reliable guide for treatment planning.

Level of evidence: Level 2.