Systematic assessment of constipation on plain abdominal radiographs in children

Background: Constipation in childhood is common and its clinical assessment is often difficult. Plain abdominal radiography is simple and used to quantify constipation. Three scoring systems, those of Barr et al., Leech et al. and Blethyn et al., have been developed to quantify fecal loading on the abdominal radiograph. Objective: In order to determine which method is the most useful in clinical practice, we assessed the reproducibility of the three scoring systems. Materials and methods: Plain abdominal radiographs from 40 clinically constipated children were retrospectively reviewed by two paediatric radiologists on two separate occasions. The radiographs were scored according to three different systems developed by Barr et al., Leech et al., Blethyn et al. Intraobserver variability and interobserver reproducibility were determined for each system. Kappa coefficients were calculated as indicators of inter- and intraobserver agreement for categorical outcome variables. Results: The Leech score showed the highest reproducibility: The intraobserver agreement was high for both observers (κ values of 0.88 and 1.00, respectively, P<0.05). Furthermore, the interobserver agreement was also high: κ 0.91 in the first round and 0.84 in the second. Conclusions: The Leech score proved to be a highly reproducible tool for assessment of childhood constipation and is of value in clinical practice for systematic assessment of constipation on plain abdominal radiographs in children.