Gastric electrical stimulation for refractory gastroparesis: predictors of response and redefining a successful outcome.

Predictors of a favorable response and measures of success with gastric electrical stimulation (GES) for gastroparesis remain elusive. Published results remain inconsistent with respect to patient perceived benefit, despite statistical improvements in objective measures of symptom severity. We performed a retrospective analysis of 56 patients with gastroparesis who underwent insertion of a gastric electrical stimulator during the study period. Data included demographics, symptoms, total symptom severity score (TSS, range 0-24, initial and most recent), and gastric emptying times. TSS were grouped into four severity categories (0-10, 11-14, 15-18, 19-24). TSS improvement was defined as movement to a lower severity category. Perception of improvement was compared with that of TSS score improvement using χ(2) test. Etiology as a predictor of improvement was measured using logistic regression. Initial mean TSS was 21, and post-treatment TSS was 13.5. Improvement was significant for individual symptoms and in reduction of TSS for both diabetic/idiopathic etiologies (P ≤ 0.001). No correlation was noted between likelihood of success/failure and gastric emptying times (P = 0.32). Thirty-eight improved (moved to lower TSS category), whereas 18 failed (remained in same category) (P ≤ 0.001), which correlated with perception of improvement. Of 18 failures, 14 (77.7%) were idiopathic. On logistic regression, diabetics were more likely than idiopathic patients to move to a lower TSS category (odds ratio 14, P = 0.003) and even more likely to improve based on patient perception (odds ratio 45, P = 0.005). GES produces far more consistent improvement in diabetics. Further study of GES in idiopathic gastroparesis is needed. Application of the proposed TSS severity categories allowed differentiation of small, statistically significant (but clinically insignificant) reductions in TSS from larger, clinically significant reductions, thereby permitting more reliable application of TSS to the evaluation of GES efficacy.