Predictors of Response in Gastroparesis

Abstract

Purpose: Natural history and outcomes of gastroparesis are not well known and the response to treatment is variable. We aim to determine which clinical features could help identify patients with response to therapy. Methods: 69 patients were included. Symptoms were scored using the Gastroparesis Cardinal Symptom Index, a validated scale of severity that utilizes three clusters (nausea/vomiting, post-prandial fullness/satiety, and bloating). Gastric emptying was measured using a standard 4-hour low fat solid meal. Treatment included conservative measures like antiemetics, prokinetics, antidepressants, analgesics. Interventions included enteral feeding and pacing in selected patients. Response to treatment was defined as overall change in GCSI score of 33% or more as compared with baseline. Results: 29 out of 69 patients had diabetes. 40 had idiopathic gastroparesis. 49 were responders (71%). 20 were non responders (29%). Amongst responders 17 (34.69%) patients required enteral feeding and 10 (20.41%) patients underwent enterra placement as opposed to 8(42.11%) and 4(21.05%) respectively in non responders. 52 patients had delayed gastric emptying (75%). Proportion of responders was not different amongst patients with delayed or normal emptying. Demographic characteristics or subtype of gastroparesis (diabetic/idiopathic) did not significantly differ among responders and non-responders. Baseline scores for nausea, vomiting and retching were similar. However, baseline scores for early satiety (P= 0.02), postprandial fullness (P= 0.05), stomach distension (P= 0.0001) and bloating (P= 0.01) were significantly greater in non-responders. Conclusion: Majority of patients with gastroparesis improve with therapy. Specific symptoms seem to be of a predictive value, with severity of early satiety, postprandial fullness, stomach distension and bloating being significantly higher in non responders than responders. On the other hand, neither the type of gastroparesis nor gastric emptying pattern appeared to predict response. Future studies are needed to investigate whether the clinical phenotype of non-responders is associated with distinctive changes in gastric pathophysiology.