Association between changes in symptoms and gastric emptying in gastroparetic patients treated with gastric electrical stimulation


  • Portions of the study were presented in abstract format at 2006 Southern Society for Clinical Investigation Regional Meeting, 3–5 March, Atlanta, GA.

Richard W. McCallum MD, Department of Medicine, Mail Stop: 1058, University of Kansas Medical Center, 3901 Rainbow Blvd, Kansas City, KS 66160, USA.
Tel: 913 588 3842; fax: 913 588 3856;


Abstract  To investigate whether there is an association between gastric emptying rate and symptom improvement in gastroparetic patients treated with gastric electrical stimulation (GES), we retrospectively reviewed 63 gastroparetic patients who received GES therapy for at least 1 year. Patient characteristics, seven upper gastrointestinal (GI) symptoms and 4-h standardized gastric emptying test (GET) were evaluated at baseline and at 1 year of GES. All symptoms were significantly reduced (< 0.001). Mean gastric retention was reduced by 7% (= 0.102) for measurement at 4 h. Of the 63 patients, 14 had their GET normalized and 49 remained delayed after 1 year. Normalized GET patients had a similar symptom improvement as those whose GET remained delayed. Of all upper GI symptoms, the improvements in vomiting (= 0.04), nausea (= 0.002) and epigastric pain (= 0.001) were significantly correlated with reduction in 4-h gastric retention between baseline and 12 months of GES therapy for patients with normalized gastric emptying but there were no correlations with any symptoms and change in gastric emptying for those patients who remained delayed. In summary, overall gastric emptying is not significantly accelerated at 4 h after successful symptomatic improvement with GES but nausea, vomiting and epigastric pain can be correlated with normalization of GET in a subset of patients.




gastric electrical stimulation


gastric emptying test


total symptom score


Recently, gastric electrical stimulation (GES) has been investigated as a new approach for treatment of medically refractory gastroparesis.1,2 Gastric electrical stimulation is achieved by delivering electric current via electrodes to gastric smooth muscle. Two methods of GES for treatment of gastroparesis have been reported in the literature. The first utilizes electrical stimulation with low frequency (i.e. frequency similar to or slightly higher than that of the native slow wave, approximately 3 cpm) and high-energy pulses (width of approximately 300 ms). Studies in both animals and humans have indicated that this method is able to entrain the gastric slow wave and normalize gastric dysrhythmias,3–5 and significantly improve gastric emptying and dyspeptic symptoms in patients with refractory gastroparesis.3 Although implantable devices are being developed that are capable of generating long pulses the external devices reported in human studies only allowed for 3- to 6-month use and were somewhat cumbersome.3 In contrast, an implantable neurostimulator with a high frequency (12 cpm), low-energy (pulse width of 330 μs) output, manufactured by Medtronic (Minneapolis, MN, USA) and approved by the FDA in 2000 under Humanitarian Device Exemption as Enterra Therapy™, can be permanently implanted into the abdomen via laparotomy or laparoscopy.6 Published reports7–12 showed a significant improvement in symptoms and health-related quality of life, and a variable improvement in gastric emptying in patients with refractory gastroparesis after 6 and 12 months of GES. However, the association between symptom improvement and change in gastric emptying with Enterra Therapy™ is still not clarified. Therefore, the aim of this study was to investigate whether there is an association between gastric emptying rate and symptom improvement in gastroparetic patients treated with GES.



Sixty-three gastroparetic patients who had documented delayed gastric emptying of a solid meal and received Enterra Therapy™ (Medtronic, Inc.) for at least 1 year were included in this study. The key inclusion criteria were (i) documented diagnosis of gastroparesis for more than 1 year and refractoriness to antiemetics and prokinetics; (ii) over seven emetic episodes per week and (iii) delayed gastric emptying (gastric retention exceeding 10% at 4 h) based on a 4-h standardized radionuclide solid meal.13 Patients were excluded if they had organic or pseudo-obstruction, primary eating or swallowing disorders, chemical dependency, positive pregnancy test or psychogenic vomiting. The study protocol was approved by the Human Subjects Committee at University of Kansas Medical Center, and written consent forms were obtained from all subjects before the study.

Study protocol

This study consisted of (i) a baseline evaluation of medical history and upper gastrointestinal (GI) symptoms, gastric emptying test, pregnancy testing and blood chemistries to determine the qualification for enrolment; (ii) surgical placement of the GES system by laparotomy, removal of any parenteral nutrition, gastric decompression or feeding tubes and gastric-jejunal feeding devices and in some cases placement of a feeding jejunostomy tube if not already in place and (iii) follow-up at 12 months after implantation to repeat baseline measurements. In addition, adverse events, including hospitalizations, were monitored throughout the follow-up period. The detailed descriptions of the GES system, surgical technique and stimulation parameters have been published previously.8,10,11

Assessment of symptoms

Each patient completed a Symptoms Interview Form at baseline and at the 12-month follow-up visit to assess the vomiting, nausea, early satiety, bloating, postprandial fullness and epigastric pain and burning during the last 2 weeks before the interview. The severity of each symptom was graded as 0 = absent, 1 = mild (not influencing the usual activities), 2 = moderate (diverting from, but not urging modifications, of usual activities), 3 = severe (influencing usual activities, severely enough to urge modifications) and 4 = extremely severe (requiring bed-rest). The sum of the seven symptom subscores was used as an overall total symptom score (TSS) of severity.

Measurement of gastric emptying

Gastric emptying (GE) scintigraphy was performed in the morning after an overnight fast as previously described13 with prokinetics stopped for at least 3 days. This standardized method for gastric emptying consists of a scrambled egg substitute (120 g of Free Cholesterol & Fat Fee Egg, Sunny Fresh Foods, Inc., Monticello, MN, USA; 60 kcal) labelled with 99mTc sulphur-colloid (1 mCi), two slices of whole wheat bread (120 kcal), 30 g jelly (75 kcal), 120 cm3 of water.13 The meal has a total caloric value of 255 kcal (nutritional composition: 72% carbohydrate, 24% protein, 2% fat and 2% fibre). Anterior and posterior images of the stomach were taken immediately after eating, and then hourly for 4 h. Gastric retention of gamma counts was calculated by the Department of Nuclear Medicine using geometric and decay correction. Blood glucose was measured before the GE and the test cancelled if glucose levels were <75 or >275 mg dL−1.

Delayed gastric emptying was defined as the percentage of gastric retention ≥60% at 2 h and ≥10% at 4 h or both.13

Statistical analysis

To minimize the error inflation effect by Intra Cluster Correlation, paired t-test between baseline and 12 months was used to compare changes in symptom scores and changes in gastric emptying at 2 and 4 h. Association of changes in symptom score and gastric emptying were examined with Spearman’s correlation test for all 63 patients, patients with normalized gastric emptying (= 14), and patients with delayed gastric emptying after 1 year (= 49) as well. To cross-examine with correlation analysis and to control for possible covariate effects, repeated measure analysis14,15 was used for testing differences between baseline and 1-year measurement for per cent retention and for testing the linear effects by changes in symptom score, patient age, gender and body mass index (BMI). Solutions to the general linear model with repeated measure14,15 was used to quantify the relationship between per cent retention measured at 4 h and changes in symptom scores by adjusting the effects by time of measurement and gastric emptying status (normalized vs remained delayed). All analysis was conducted using sas for windows (Version 9). Statistical significance was assigned for < 0.05.


Changes in upper GI symptom scores and gastric empting over time

Of the 63 patients (51 female and 12 male; mean age: 41 years, range: 21–66) evaluated in this report, 38 were diabetic with mean duration of 18.9 years (range: 1–39), 11 were idiopathic and 14 were postsurgical gastroparesis. The mean duration of gastroparetic symptoms was 5.9 years (range: 1–20). All patients had delayed solid gastric emptying at baseline (see Table 1).

Table 1.   Summary of results of upper GI symptom severity and gastric emptying for all patients
Severity score (0–4)Baseline12 months ChangeICCt
  1. Gastric emptying results (% retention) and symptom scores are reported as mean (SD). t represents P-values by paired t-test.

  2. ICC, Intra Cluster Correlation (14); TSS, total symptom score; GI, gastrointestine.

Vomiting2.8 (1.43) 1.1 (1.24)1.7 (1.64)0.635<0.001
Nausea3.6 (0.62) 1.6 (1.29)2.0 (1.34)0.497<0.001
Early satiety3.0 (1.00) 1.4 (1.27)1.7 (1.43)0.588<0.001
Bloating2.8 (1.16) 1.4 (1.24)1.5 (1.33)0.623<0.001
Postprandial fullness2.9 (1.03) 1.5 (1.23)1.4 (1.33)0.601<0.001
Epigastric pain2.7 (1.33) 1.3 (1.37)1.4 (1.64)0.652<0.001
Eigastric burning2.0 (1.50) 0.9 (1.20)1.1 (1.54)0.653<0.001
TSS of severity (0–28)19.9 (4.84) 9.1 (6.67)10.8 (7.17)0.971<0.001
Gastric retention (%)Baseline12 monthsDifferenceICC> t
At 2 h73 (18)63 (27)10 (32)0.9970.018
At 4 h46 (25)39 (29) 7 (32)0.9980.102

All symptoms were significantly reduced after 1 year of GES as indicated by the decreases in symptom subscores and TSS compared with those at baseline (< 0.001) (Table 1). Overall, gastric emptying still remained delayed. Mean gastric retention was reduced by 10% (= 0.018) for measurement at 2 h, and 7% (= 0.102) for measurement at 4 h. Of the 63 patients studied, 14 patients (22%) had normalized GE and 49 were still delayed. Both groups had a significant reduction in symptoms (< 0.001; see Tables 2 and 3). No significant effects were found by patient age, gender and BMI on the per cent retention (R%) measured at 2 and 4 h and hence these factors were not included in the linear model with repeated measure analysis.

Table 2.   Summary of results of upper GI symptom severity and gastric emptying for patients with normalized gastric emptying at 1 year (= 14)
Severity score (0–4)Baseline12 months Changet
  1. Gastric emptying results (% retention) and symptom scores are reported as mean (SD). t represents P-values by paired t-test.

  2. TSS, total symptom score; GI, gastrointestine.

Vomiting2.85 (1.61)0.78 (0.89)2.07 (1.68)0.0005
Nausea3.85 (0.36)1.42 (1.34)2.43 (1.39)<0.0001
Early satiety3.21 (1.19)1.42 (1.22)1.79 (1.31)0.0002
Bloating3.28 (0.83)1.78 (1.42)1.50 (1.09)0.0002
Postprandial fullness3.07 (1.07)1.50 (1.22)1.57 (1.09)0.0001
Epigastric pain3.50 (0.94)1.71 (1.44)1.79 (1.53)0.0008
Eigastric burning2.21 (1.58)0.78 (0.97)1.43 (1.38)0.0011
TSS of severity (0–28)22.00 (3.94)9.42 (6.04)12.57 (6.02)<0.0001
Gastric retention (%)Baseline12 months Differencet
At 2 h76.21 (12.8)26.97 (21.3)49.24 (21.2)<0.001
At 4 h42.13 (27.7)2.79 (2.66)39.34 (25.7)<0.001
Table 3.   Summary of results of upper GI symptom severity and gastric emptying for patients with delayed gastric emptying at 1 year (= 49)
Severity score (0–4)Baseline12 months Changet
  1. Gastric emptying results (% retention) and symptom scores are reported as mean (SD). t represents P-values by paired t-test.

  2. TSS, total symptom score; GI, gastrointestine.

Vomiting2.82 (1.39)1.18 (1.32)1.63 (1.64)<0.0001
Nausea3.47 (0.65)1.61 (1.29)1.86 (1.31)<0.0001
Early satiety3.02 (0.94)1.41 (1.31)1.61 (1.48)<0.0001
Bloating2.73 (1.22)1.29 (1.17)1.45 (1.40)<0.0001
Postprandial fullness2.86 (1.02)1.47 (1.24)1.39 (1.40)<0.0001
Epigastric pain2.45 (1.34)1.12 (1.33)1.33 (1.68)<0.0001
Eigastric burning2.00 (1.49)0.94 (1.26)1.06 (1.61)<0.0001
TSS of severity (0–28)19.35 (4.95)9.02 (6.90)10.33 (7.45)<0.0001
Gastric retention (%)Baseline12 months Differencet
At 2 h72.92 (2.77)74.49 (2.67)−1.57 (3.50)0.6555
At 4 h47.66 (3.34)50.17 (3.37)−2.51 (4.01)0.5337

Associations between changes in gastric empting and upper GI symptom scores

Of all seven upper GI symptoms, Spearman’s correlation test revealed that only improvement in nausea was significantly correlated with reduction in the mean R% at 4 h for all 63 patients (= 0.02). A closer examination suggests that a correlation exists between changes in symptom score and R% for patients with normalized gastric emptying (= 14, Tables 2 and 4), specifically in vomiting (= 0.04), nausea (= 0.002), epigastric pain (= 0.001) and TSS (= 0.03). However, none of the symptom score improvements was accompanied with a reduction in R% for patients whose gastric emptying remained delayed after 1 year of follow-up (= 49, Tables 3 and 4). Fig. 1 shows the correlation between changes in TSS and in 4-h gastric retention for patients in the normalized group (left panel) and in the delayed group at 12 months (right panel). Repeated measure analysis indicated that only the improvement in vomiting and nausea is significantly correlated with reduction in 4-h gastric retention at 12 months for all patients (< 0.05; Table 5). These statistics are driven by the subgroup whose gastric emptying normalized because the magnitudes of these relations are quite different for patients with normalized GE than those of patients remained delayed GE (Table 6). For example, an improvement in vomiting or nausea score by 1 point is accompanied by 45% or 50% reduction in R% for patients with normalized GE, while the corresponding reduction for patients with delayed GE are only 10%.15

Table 4.   Correlation analysis between changes in symptom score and gastric retention at 4 h
  1. *P-values.

  2. Spearman correlation coefficients.

  3. TSS, total symptom score.

  4. Significant at 5% level expressed in bold.

Overall (= 63)0.1870.2910.04210.0470.0110.0160.0140.136
Normalized (= 14)0.5740.7410.25550.1980.1290.807−0.0320.580
Delayed (= 49)0.0710.1580.0090.051−0.024−0.151−0.0480.016
Figure 1.

 Correlation between the changes in total symptom score and in 4-h gastric retention after 12 months of gastric electrical stimulation therapy for patients in the normalized group (left panel; Pearson correlation coefficient = 0.58, P-value = 0.03) and in the delayed group (right panel; Pearson correlation coefficient = 0.02, P-value = 0.92).

Table 5.   Test for fixed effect on GE at 4 h by change in symptom scores with repeated measure analysis
EffectNumerator d.f.Denominator d.f.F-valueProbability  > F
  1. TSS, total symptom score; GE, gastric emptying.

GE status16024.570<0.001
GE status16026.935<0.001
GE status16026.134<0.001
GE status16022.096<0.001
Early satiety1600.2550.6157
GE status16021.976<0.001
GE status16022.571<0.001
GE status16022.587<0.001
GE status16022.51<.0001
Table 6.   Statistical solution for fixed effect by repeated measure analysis
EffectTime (year)GE statusParameter estimateStandard errord.f.T-valueProbability > T
  1. GE, gastric emptying; d.f., degrees of freedom of denominator.

Intercept  18.4295.888603.1300.0027
Time1 −6.7874.09262−1.6590.1023
GE status Delayed28.0255.400605.190<0.001
Vomiting  −3.5851.37360−2.6110.0114
Intercept  17.2846.744602.5630.0129
Time1 −6.7874.09262−1.6590.1023
GE status Delayed28.4695.569605.112<0.001
Nausea  −3.5291.74460−2.0230.0475

No significant associations were found between changes in any symptoms and improvement in gastric emptying at 2 h after meal in any groups of patients.


In this study, we confirmed that GES by a permanently implantable system (Enterra Therapy™) significantly reduced severity of all upper GI symptoms assessed, including vomiting, nausea, early satiety, bloating, postprandial fullness and epigastric pain and burning. This is in agreement with previous studies on GES in refractory gastroparesis.7–12

The findings on the effects of GES with Enterra Therapy™ on gastric emptying have been controversial. A few studies showed that GES with Enterra Therapy™ significantly improved gastric emptying, although this improvement varied widely,10,16 other reports demonstrated that gastric emptying was only minimally and not significantly faster after 1 year of Enterra Therapy™.2,8 Our data in this study showed that the mean gastric retention measured by a standardized 4-h scintigraph of a solid meal was not significantly reduced after 12 months of GES. Although numerically less than baseline, the 4-h gastric retention at 12 months was still nearly three times greater than the normal value, i.e. markedly delayed. This observation is consistent with our previous reports2,8 but not in agreement with results in other studies where both 2- and 4-h gastric retention were reported as significantly decreased at 12 months of GES.10,16 There were no differences in the procedures used for GES (i.e. Enterra Therapy™) and gastric emptying test in all these studies, but the patient populations were very different. In our study, 60% of patients had diabetic gastroparesis and 17% were idiopathic, while 48% and 71%, respectively, of the patients in the other two studies mentioned above10,16 had idiopathic gastroparesis. It is recognized that a spontaneous resolution or improvement in gastric emptying does occur over time for most patients with idiopathic gastroparsis.17 The different data analyses methods used in these studies might also be another factor for the different results for gastric emptying with Enterra Therapy™.

One of the new findings in the present study is that the improvement in nausea score was significantly correlated with the reduction in 4-h gastric retention as indicated by Spearman’s correlation test while other gastroparetic symptoms did not achieve this correlation. A previous multicentre study on Enterra Therapy™ reported no such association between changes in symptoms and gastric emptying using Pearson’s correlation test.10 One of the reasons for this difference is that the correlation analysis in that study was only applied to changes in vomiting frequency and changes in 2- or 4-h gastric emptying between baseline and 6 or 12 months and no correlations were sought for changes in nausea and other symptoms.10 Another reason might be due to the different patient populations as described above. Further analysis by separating patients into two groups indicated that strong correlations between changes in the symptom score and R% existed not only for nausea score, but also for vomiting, epigastric pain, and, as a consequence, TSS in that subset (22% of patients) who normalized their gastric emptying at 12 months (see Table 4). This suggests that improvement in gastric empting does contribute to the improvement of those symptoms in this subset of patients. These results are further confirmed by application of repeated measure analysis.14 Unlike correlation testing, repeated measure analysis examines the trend of the primary variable over time and quantifies its relationship with other factors in the perspective of a general linear model. In our study, in addition to considering changes in each symptom score and the time effect on R%, a GE status was also included for differentiating patients whose GE normalized after 1 year in the repeated measure analysis. As a result, the quantitative relationship between changes in symptoms and gastric empting can be established (see Table 6).

The exact mechanisms of action of GES remain to be elucidated. Studies using a gastric barostat showed that high-frequency GES decreased the sensitivity to gastric distention and enhanced gastric accommodation to a meal in patients with severe gastroparesis.18,19 Spectral analysis of heart rate variability in response to high-frequency GES at the proximal stomach suggested enhanced vagal activity.20–22 Increased bilateral thalamic activity on PET imaging during GES indicates there are central effects that could mediate inhibition of nausea and vomiting.23

This present study was not placebo controlled. A previous double-blind study10 did show that 1-month of GES was significantly better than 1-month sham stimulation. Recently published long-term data indicate that the symptomatic improvement with GES is sustained in the majority of initial responders with studies extending beyond 3 years.12 It is highly unlikely that a placebo effect in our study could explain 1 year of sustained clinical improvement in a group of gastroparetics who had failed months and years of standard medical therapy.

In conclusion, the improvements in nausea, vomiting, epigastric pain and TSS are significantly correlated with reduction in gastric retention between baseline and 12 months for the minority of patients with normalized gastric emptying after 12 months of Enterra Therapy™. There was no correlation of any symptoms with gastric emptying for the majority of patients in this study who remained with delayed gastric emptying but had the same significant reduction in symptoms as the 22% who normalized their GET. These results further our understanding of how GES with Enterra Therapy™ improves symptoms in gastroparetic patients.


This work was supported in part by Medtronic, Inc. (Minneapolis, MN). The authors would like to acknowledge the following individuals for their contributions: Katherine Roeser, Pernilla Foran, Teri Lavenbarg and Irene King and faculty, fellows and nursing staff in the Center for Gastrointestinal Nerve and Function and Division of Gastroenterology and Janet Kelly, RN and the operating room staff and surgical residents. We also wish to acknowledge Dr Reginald Dusing, the Director of Nuclear Medicine and his staff and assistants, and Dr Joseph Kyner and the staff of the Cray Diabetic Center at the University of Kansas Hospital for their role in the care of the patients with diabetes mellitus and Dr Twillman in the area of psychology and pain management.