A NOVEL SCORE ASSESSING DUODENAL ENDOSCOPIC MARKERS FOR PREDICTING VILLOUS ATROPHY

Aims:

Several endoscopic markers (EM) have been described in patients with villous atrophy (VA), both in the bulb and distal duodenum, each of them of different significance in predicting VA. Our aim was to develop a scoring system for EM in predicting VA.

Methods:

We consecutively recruited patients in whom EM suggestive of VA were described in their endoscopy reports. We noted atrophy, mosaic, nodularity and fissures in the bulb and distal duodenum, and in the latter we also counted scalloping and loss/reduction of Kerckring folds. According to the histopathology report we divided patients into two groups – with and without VA, and assessed the diagnostic accuracy of each EM. We calculated odds ratio (OR) for each EM and we added the corresponding values for each patient, reaching a total EM score/patient; we then analyzed the predictive value of the score for VA.

Results:

Altogether 235 patients were recruited, of which 50 were excluded from the final analysis because duodenal biopsies were either not taken or results were not available. Among the remaining 185 patients, mean age 45 ± 14 years, 61% female, 61% had VA on biopsy while the other 39% were non-atrophic. Of the 113 patients with VA, 51% were finally diagnosed as CD (of which 2% seronegative). The highest OR values were 19.14, 10.86, 10.32 and 8.46 for duodenal fissures, mosaic, nodularity and scalloping, respectively. Average number of markers was 2.3 for patients with VA and 1.3 for non-atrophic. Average EM score was 6 in non-atrophic and 41 in VA patients. The score had an AUROC of 0.703 (95% CI 0.630 – 0.776) in predicting VA.

Conclusions:

This novel endoscopic scoring system has good accuracy in predicting VA and could be used to better select patients with probability of VA, in whom duodenal biopsies should be performed.