PTU-174 Can gastroparesis cardinal symptom index (GCSI) sub-scores and total score predict delayed gastric emptying?

Abstract

Introduction The Gastroparesis Cardinal Symptom Index (GCSI) is a validated tool for monitoring treatment induced changes in the symptoms of gastroparesis. For diagnosis, a gastric emptying scintigraphy (GES) study must be done to demonstrate the delayed gastric emptying (DGE) that separates gastroparesis from its functional differentials. Previous studies noted associations between GES results and GCSI values but these were not sufficiently predictive to allow reliable diagnosis. This study aimed to corroborate these findings. If GCSI could be proved to be predictive of DGE, it may allow for less GES studies, especially as part of follow up, resulting in reduced costs and less radiation exposure for the patient. Method Forty one patients with symptoms suggestive of gastroparesis were identified at clinic. Patients filled out the GCSI questionnaire and underwent a GES study as part of standard investigation of their symptoms. Linear regressions were used to identify if the GCSI sub-scores or the total GCSI score were predictive of% gastric retention across all imaging times. Mann-Whitney U tests were used to identify any significant differences between the distributions of the GCSI components in those with DGE and those without. Both tests were done for the whole population, then for the three main aetiologies (idiopathic, diabetes mellitus, post-surgical). Results For the whole population, regression analysis identified three components as positive predictors at 2 hrs (total GCSI score, post-prandial fullness and bloating sub-scores). The R2values were all ≤0.16 and the only significant Mann-Whitney U test was for the total GCSI score. For idiopathic cases there was only one significant positive predictor, again at 2 hrs (total GCSI score, R2= 0.244). There were significant differences in the distributions of the average score and the nausea and vomiting sub-score. Post-operative cases had five significant negative predictors: nausea and vomiting sub-score (R2= 0.389) at ½ hr; nausea and vomiting sub-score (R2= 0.547), total GCSI score (R2= 0.428), bloating sub-score (R2= 0.364) at 1 hr; nausea and vomiting sub-score at 4 hrs (R2= 0.468). All of these components had associated significant Mann-Whitney U tests. There were no significant Mann-Whitney U tests in the diabetic cohort, but there were three significant positive predictors: bloating sub-score at 1 hr (R2= 0.577) and 2 hrs (R2= 0.590), post-prandial fullness sub-score at 2 hrs (R2= 0.612). Conclusion All GCSI components studied were associated with significant linear regression findings but their individual predictive power tends to be poor. Thus, it is unlikely the GCSI components in isolation would be clinically useful as diagnostic predictors of gastroparesis. This concurs with the conclusions of previous studies in this field. Disclosure of interest None Declared.