Effectiveness of gastric electrical stimulation in gastroparesis: Results from a large prospectively collected database of national gastroparesis registries

Abstract

Background: Gastric electrical stimulation (GES) for treating gastroparesis symptoms is controversial.

Methods: We studied 319 idiopathic or diabetic gastroparesis symptom patients from the Gastroparesis Clinical Research Consortium (GpCRC) observational studies: 238 without GES and 81 with GES. We assessed the effects of GES using change in GCSI total score and nausea/vomiting subscales between baseline and 48 weeks. We used propensity score methods to control for imbalances in patient characteristics between comparison groups.

Key results: GES patients were clinically worse (40% severe vs. 18% for non-GES; P < .001); worse PAGI-QOL (2.2. vs. 2.6; P = .003); and worse GCSI total scores (3.5 vs. 2.8; P < .001). We observed improvements in 48-week GCSI total scores for GES vs. non-GES: improvement by ≥ 1-point (RR = 1.63; 95% CI = (1.14, 2.33); P = .01) and change from enrollment (difference = -0.5 (-0.8, -0.3); P < .001). When adjusting for patient characteristics, symptom scores were smaller and not statistically significant: improvement by ≥ 1-point (RR = 1.29 (0.88, 1.90); P = .20) and change from the enrollment (difference = -0.3 (-0.6, 0.0); P = .07). Of the individual items, the nausea improved by ≥ 1 point (RR = 1.31 (1.03, 1.67); P = .04). Patients with GCSI score ≥ 3.0 tended to improve more than those with score < 3.0. (Adjusted P = 0.02).

Conclusions and inferences: This multicenter study of gastroparesis patients found significant improvements in gastroparesis symptoms among GES patients. Accounting for imbalances in patient characteristics, only nausea remained significant. Patients with greater symptoms at baseline improved more after GES. A much larger sample of patients is needed to fully evaluate symptomatic responses and to identify patients likely to respond to GES.

Keywords: abdominal Pain; gastric electrical stimulation; gastroparesis; nausea; vomiting.

© 2019 John Wiley & Sons Ltd.

Figures

Figure 1

Selection of study population from Consort Diagram * Six (6) patients participated in both study, and their records of the GpR study were used if they had never had the GES. If the GES was implanted during the follow-up, records of the first study when the GES was implanted were used.

Figure 2

Time from enrollment to GES implantation (N = 319; 81 GES patients, 238 non-GES patients) Notes: Time of GES implantation was interpolated as the midpoint between two visits. The follow-up time in GES patients with the GES system was 63% of the maximum possible follow-up time if the GES system had been implanted at enrollment. Among GES patients, 58%, 62% and 84% had the GES system implanted by 16, 24, 36 weeks, respectively; median and mean weeks to the GES implantation were 12 weeks and 17.7 weeks, respectively.

Figure 3

Change of PAGI-SYM scores from study enrollment to 48 weeks, GES vs. non-GES patients (N = 319; 81 GES patients, 238 non-GES patients) Notes: The vertical bars are 95% confidence intervals. The 4-df test for the difference in change in GCSI Total Score between GES and non-GES patients was P = 0.005, and the test for nausea/vomiting subscale was P = 0.01. The significance was computed from generalized estimating equations (GEE) linear regression with robust variance estimation, modeling change in GCSI total score as a function of GES implantation, visit indicator (16-, 24-, 32-, or 48-week), and GES implantation by visit indicator interaction.