Optimal measurement of gastric emptying of solids in gastroparesis or functional dyspepsia: evidence to establish standard test

Abstract

Objective: Symptoms in gastroparesis (Gp) and functional dyspepsia (FD) overlap; using egg protein substitute to measure gastric emptying of solids (GES), ~40% of patients are reclassified from Gp to FD, and vice versa. Our aim was to assess inter-individual and intra-individual coefficients of variation (COV) in GES in symptomatic patients with Gp or FD with documented slow or normal GES, respectively.

Design: Scintigraphic GES (T_1/2 and GE% at 2 and 4 hours) using a 320 kcal real egg meal (30% fat) was tested in the following: single measurements in 20 patients with diabetes mellitus (10 each type 1 and type 2); repeat GES to estimate COV_intra measured: 3 days apart in 9 Gp, 4 weeks apart in 21 Gp and 18 with FD with normal GE assigned to placebo and in 70 patients at 94.3 weeks (median) apart.

Results: COV_inter for GE% at 4 hours and GE T_1/2 were respectively 14.2% and 23.5% in FD and 27.5% and 33% in Gp; COV_intra for GE% at 4 hours and GE T_1/2 up to 4 weeks apart were 23.4% and 37.9% in FD and 20.1% and 33% in Gp. GE% at 2 hours showed less consistent results. However, >85% retained original diagnosis as normal or delayed. From clinical GES to baseline research for Gp group, repeat GES (after treatment) showed the COV_intra for GE% at 4 hours was 37.3% at median 94.3 weeks, with 26/70 changed diagnoses.

Conclusion: The 320 kcal (30% fat) GES scintigraphic test provides consistent diagnosis in >85% and should be the standard test for suspected gastric emptying disorders.

Keywords: T1/2; emptying; percent; prokinetic.

Figure 1.

Parameters of gastric emptying (GE) of solids in patients with functional dyspepsia and documented normal gastric emptying showing interquartile range, 5% and 95% confidence interval (CI) and outliers at baseline and after treatment with placebo for 4 weeks (upper panel). Lower panel shows correlations with 95% CI of regression and upper limit of normal range shown by interrupted lines (174 minutes for GE T_1/2, 25% for GE at 2 hours, and 75% for GE at 4 hours). Note 3 patients had delayed emptying on repeat test.

Figure 2.

Bland-Altman plot showing consistency of gastric emptying (GE) data for the majority of patients with functional dyspepsia. The vast majority of the replicate data are within 40 minutes difference for GE T_1/2, 20% difference for GE at 2 hours, and 10% difference for GE at 4 hours.

Figure 3.

Parameters of gastric emptying (GE) of solids in patients with gastroparesis and documented slow GE showing interquartile range, 5% and 95% confidence interval (CI) and outliers at baseline and after treatment with placebo for 4 weeks (upper panel). Lower panel shows correlations with 95% CI of regression and upper limit of normal range shown by interrupted lines (174 minutes for GE T_1/2, 25% for GE at 2 hours, and 75% for GE at 4 hours). Note: Very few patients’ GE results are misclassified based on these parameters, in particular GE T_1/2 and GE at 4 hours.

Figure 4.

Bland-Altman plot showing consistency of gastric emptying (GE) data for the majority of patients with gastroparesis. The vast majority of the replicate data are within 60 minutes difference for GE T_1/2, 20% difference for GE at 2 hours, and 20% difference for GE at 4 hours. Importantly, despite the variation, only 2 patients were inconsistently classified on repeat test (see regression analysis in figure 3).

Figure 5.

Comparison of earlier clinical diagnostic gastric emptying (GE) test conducted a median 94 weeks prior to the baseline test performed at entry to a research trial. Note significant difference in GE% at 4 hours (upper panel) with greater GE% emptied at the second test and the faster GE in 24 of the 70 patients (lower panel, right) during the second test.

Figure 6.

Plot showing difference in gastric emptying (GE) % at 4 hours based on interval in weeks between the two tests (clinical diagnostic vs. research). Note: The consistency in the difference is best demonstrated in the time interval 0–200 weeks.