Blinded comparison of faecal loading on plain radiography versus radio-opaque marker transit studies in the assessment of constipation

AIM: To compare faecal loading on plain radiography versus radio-opaque marker transit studies in the assessment of constipation. METHODS: The study group was a convenience sample of patients attending the Durham Constipation Clinic. All patients underwent transit studies according to an established protocol, and severity of constipation was assessed contemporaneously using a validated questionnaire (PAC-SYM). Transit studies were performed using radio-opaque markers that were ingested over 3 consecutive days, with a radiograph taken on the fourth day. Digital images of the radiograph were digitally altered to remove all traces of the transit markers without affecting the underlying pattern of faecal loading. Four observers assessed faecal loading independently; two clinicians (C1 and C2) and two radiologists (R1 and R2). C1 and R1 used a previously described format scoring method of assessing faecal loading, whereas C2 and R2 assessed the images as if they were in a clinic or reporting session, grading the faecal loading as mild, moderate, or severe. RESULTS: One hundred patients were recruited out of 186 presenting in a 2-year period. All patients completed assessments. The correlation between observers was only fair to moderate (r ranging from 0.34-0.51). There were some surprisingly marked disagreements in 10-18% of assessments. The correlation between faecal loading and transit was weak for all observers (r ranging from 0.261-0.311). Symptom severity did not correlate with faecal loading. CONCLUSION: These results suggest that there is considerable inter-observer variation in the radiological assessment of faecal loading, irrespective of the training or method used by the observer, and that there is very poor correlation with colonic transit. The diagnosis of constipation, and the assessment of severity, is best performed clinically. (C) 2008 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.