Skip to main content
NCBI home page
JAMA Network logoLink to JAMA Network
. 2026 May 27;161(7):719–726. doi: 10.1001/jamasurg.2026.1593

Glycemic Control and Diabetes Outcomes After Surgical Therapy for Diabetic Gastroparesis

Vineeth R Sadda 1,2, Rhea Verma 1,2, Ahmed E Aly 1,2, Ping Zheng 1,2, Shahin Ayazi 1,2,3,✉
PMCID: PMC13217260  PMID: 42201730

This cohort study examines data from a national electronic health record database to evaluate the association between surgical therapy for gastroparesis and glycemic control and diabetes-related complications in adults with diabetic gastroparesis.

Key Points

Question

Is surgical therapy for gastroparesis associated with improved glycemic control and reduced diabetes-related complications in patients with diabetic gastroparesis?

Findings

In this cohort study including 95 328 patients, surgical therapy was associated with sustained reduction in hemoglobin A1c level, lower insulin utilization, and reduced 5-year diabetes-related complications compared with nonoperative management.

Meaning

Surgical therapy for diabetic gastroparesis may confer metabolic benefits beyond symptom control and could influence long-term diabetes-related outcomes.

Abstract

Importance

Surgical therapy for gastroparesis is effective for refractory symptoms; however, its association with glycemic control and diabetes-related complications in patients with diabetic gastroparesis remains uncertain.

Objective

To evaluate the association between surgical therapy for gastroparesis and glycemic control and diabetes-related complications in adults with diabetic gastroparesis.

Design, Setting, and Participants

This retrospective multicenter cohort study using a national electronic health record database included more than 100 million patients from more than 70 US health care organizations. Adults with concurrent diagnoses of diabetes and gastroparesis between 2010 and 2023 were included. Propensity score matching (1:1) was performed. Queries and analyses were performed in January and February 2026.

Exposure

Surgical therapy for gastroparesis, including gastric electrical stimulation, pyloroplasty, and gastric peroral endoscopic myotomy.

Main Outcomes and Measures

The primary outcome was mean glycated hemoglobin (HbA1c) level over follow-up. Secondary outcomes included probability of adequate glycemic control (HbA1c <7%), severe hyperglycemia (HbA1c ≥10%), insulin utilization, diabetes-related complications, health care utilization, and mortality.

Results

A total of 95 328 eligible patients were identified, including 2272 (2.4%) who underwent surgical therapy. The mean (SD) age was 54.9 (15.2) years; 61 245 (64.2%) were female and 34 083 (35.6%) male. After propensity score matching, 2272 patients remained in each cohort. The mean (SD) follow-up time was 3.5 (1.7) years. Baseline HbA1c was similar between groups. Over follow-up, mean HbA1c decreased by 0.51 percentage points in the surgical cohort and increased by 0.28 percentage points in the nonoperative cohort. At 5 years, mean (SD) HbA1c was 6.29% (1.72) in the surgical group and 7.21% (2.02) in the nonoperative group (between-group difference, 0.92 percentage points; P < .001). The surgical cohort had a higher probability of achieving adequate glycemic control, lower probability of severe hyperglycemia, and lower initiation of insulin therapy. At 5 years, diabetes-related complications occurred in 1209 surgical patients (53.2%) and 1438 nonoperative patients (63.3%; odds ratio, 1.52; 95% CI, 1.34-1.71; P < .001). Emergency visits (947 [41.7%] vs 1243 [54.7%]) and hospitalizations (1034 [45.5%] vs 1340 [59.0%]) were also less frequent (both P < .001). Five-year mortality did not differ (232 [10.4%] vs 232 [10.4%]; P > .99).

Conclusions and Relevance

In this national multicenter cohort, surgical therapy was associated with improved glycemic control, lower insulin utilization, and fewer diabetes-related complications. Metabolic outcomes may warrant consideration alongside symptom relief when evaluating patients for surgical therapy for gastroparesis.

Introduction

Diabetes affects more than 800 million individuals worldwide and is a major contributor to morbidity and health care utilization through its chronic complications.1 Diabetic gastroparesis is one such complication, arising from autonomic neuropathy due to chronic hyperglycemia and resulting in delayed gastric emptying with early satiety, nausea, vomiting, and impaired oral intake.2 Altered gastric emptying disrupts the timing of nutrient absorption, producing insulin-nutrient mismatch, glycemic variability, and inconsistent medication administration. These effects establish a bidirectional relationship between gastric motility and diabetes control.

Initial management of diabetic gastroparesis includes dietary modification and pharmacologic therapy. For patients with persistent symptoms, surgical options, including gastric electrical stimulation (GES), and pylorus-directed procedures, such as pyloroplasty and gastric peroral endoscopic myotomy (G-POEM), improve symptoms and gastric emptying and are reserved for medically refractory disease.2,3

Despite its efficacy for refractory gastrointestinal symptoms, the effect of surgical intervention on diabetes outcomes remains uncertain. Prior studies have primarily examined gastrointestinal end points, with limited evaluation of glycemic control or diabetes-related complications.4,5,6,7 We therefore evaluated the association between surgical treatment of diabetic gastroparesis, short- and long-term glycemic control, and diabetes-related outcomes in a national multicenter cohort.

Methods

Study Design and Data Source

This was a multicenter, retrospective cohort study using data from the TriNetX database to assess diabetes-related outcomes in patients with diabetic gastroparesis who underwent surgical therapy for gastroparesis compared with those who were managed nonoperatively. TriNetX is a global database containing data from electronic health record systems of participating organizations. These data include demographics, diagnosis codes from the International Statistical Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM), Current Procedural Terminology procedure codes, laboratory values, and medication codes. Data are continuously updated and are deidentified by TriNetX. We queried the US Collaborative Network, which contains data from 70 health care organizations across the US with more than 100 million patients. Queries and analyses were performed in January and February 2026.

In accordance with institutional policy and federal guidance, studies using exclusively deidentified data are considered exempt from human participants research oversight and do not require formal institutional review board review. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies.

Study Population

Adults aged 18 years or older with diabetes and gastroparesis were identified by concurrent diagnostic codes or by use of prokinetic therapy (erythromycin or metoclopramide). To standardize disease duration, the first concurrent documentation of diabetes and gastroparesis was required to occur within 5 years before the start of follow-up. The study period extended from January 1, 2010, through December 31, 2023.

Patients were categorized according to receipt of surgical therapy for gastroparesis, including GES, pyloroplasty, or G-POEM. For the surgical cohort, the index date was defined as the date of surgery. For the nonoperative cohort, the index date was defined as a health care encounter during the study period. Subgroup analyses were performed according to diabetes type (type 1 vs type 2) and procedure type (GES vs pyloric intervention).

Outcomes and Follow-Up

The primary outcome was mean glycated hemoglobin (HbA1c) level at the end of follow-up. Secondary glycemic outcomes included achievement of adequate glycemic control (HbA1c <7.0%) among patients with HbA1c 7.0% or greater at baseline, inadequate glycemic control (HbA1c ≥7.0%), severe hyperglycemia (HbA1c ≥10%), and initiation of fast-acting or long-acting insulin therapy.

Acute metabolic events included diabetic ketoacidosis (DKA) or hyperosmolar hyperglycemic state (HHS) and hypoglycemia (blood glucose <54 mg/dL). Additional outcomes included diabetes-related complications categorized by organ system and health care utilization, defined as hospitalization or emergency department visit. Insulin use was identified by medication records using Anatomical Therapeutic Chemical classification codes. The complete list of diagnostic and medication codes is provided in eTable 1 in Supplement 1.

Follow-up began 1 month after the index date and continued for up to 5 years. Mean HbA1c levels were assessed annually using each patient’s most recent value from the previous year. Glycemic control outcomes were evaluated throughout follow-up, and diabetes-related complications were assessed at a short-term (1 year) and long-term (5 years) follow-up.

Statistical Analysis

Analyses were performed using the TriNetX platform. Propensity score matching (1:1) balanced cohorts on demographic details, laboratory values, comorbidities, and medication use during the year before the index date (eTable 2 in Supplement 1). Covariate balance was assessed using standardized mean differences less than 0.1. Subgroup analyses were performed without matching.

For continuous outcomes, cohort means at each time point were calculated using the most recent value within the preceding year. Longitudinal HbA1c trends were visualized using penalized smoothing splines applied to interval-specific means, with 95% CIs derived from interval-level standard errors; statistical testing used unsmoothed data. Continuous variables are reported as mean (SD) and compared using t tests. Categorical variables are reported as frequencies with percentages and compared using χ2 tests, with odds ratios (ORs) and 95% CIs reported.

Time-to-event analyses for glycemic outcomes and insulin initiation included patients without the outcome at baseline. Kaplan-Meier curves were compared using log-rank tests, with hazard ratios (HRs) and 95% CIs estimated using Cox proportional hazards models.

Results

Study population

A total of 95 328 patients met the inclusion criteria, of whom 2272 (2.4%) underwent surgical therapy for gastroparesis. The mean (SD) age was 54.9 (15.2) years; 61 245 (64.2%) were female and 34 083 (35.6%) male. Prior to propensity matching, the nonoperative cohort was older, had a higher baseline mean HbA1c level, and had a greater burden of diabetes-related complications and insulin use (eTable 2 in the Supplement). After 1:1 propensity score matching, there were 2272 patients in each cohort. Baseline characteristics were well balanced, including age, diabetes type, medication use, complication burden, and mean (SD) HbA1c level (6.93% [2.05] in the nonoperative group vs 6.80% [1.89] in the surgical group; P = .10).

Glycemic Control

In the matched cohorts, mean (SD) follow-up was 3.46 (1.70) years in the surgical group and 3.58 (1.67) years in the nonoperative group (P = .02). The mean HbA1c level was lower in the surgical group at 1 year and remained lower throughout follow-up, with an absolute difference of 0.92 percentage points between groups at study end (P < .001). At 1 year, the mean (SD) HbA1c level was 7.20% (2.16) in the nonoperative group and 6.16% (1.69) in the surgical group; at 5 years, it was 7.21% (2.02) vs 6.29% (1.72), respectively (Figure 1). From baseline to end of follow-up, HbA1c decreased by 0.51 percentage points (7.5%) in the surgical group and increased by 0.28 percentage points (4.0%) in the nonoperative group.

Figure 1. Line Graph of Mean Glycated Hemoglobin (HbA1c) Levels in the Surgical and Nonoperative Cohorts.

Line chart of mean H b A 1 c percent over 0 to 60 months for two cohorts. Vertical axis labeled Mean H b A 1 c, percent, with tick labels from 6 point 0 at the bottom to 7 point 4 at the top, and light horizontal gridlines. Horizontal axis labeled Time from index date, mo, with tick labels at 0, 12, 24, 36, 48, and 60. Two time series appear. A dark teal series in the upper half uses large filled circular markers at each time point and a thick smooth curve running through them; a semi transparent gray band surrounds the curve across the full time range. Teal marker values are approximately 6 point 9 to 7 point 0 at 0 months, about 7 point 2 at 12 months, about 7 point 1 at 24 months, about 7 point 2 to 7 point 3 at 36 months, about 7 point 2 at 48 months, and about 7 point 2 at 60 months. The text label Nonoperative appears near the right side of the plot aligned with this upper teal series. An orange series in the lower half uses large filled circular markers and a thick smooth curve; a semi transparent light orange band surrounds the curve. Orange marker values are approximately 6 point 8 at 0 months, about 6 point 2 at 12 months, about 6 point 3 at 24 months, about 6 point 2 at 36 months, about 6 point 3 at 48 months, and about 6 point 3 at 60 months. The text label Surgical appears near the right side of the plot aligned with this lower orange series. No legend box is visible; cohort names are embedded as text near the corresponding curves.

Open in a new tab

Mean (SD) HbA1c was similar between groups at baseline (6.80% [2.02] for the surgical cohort vs 6.93% [1.89] for the nonoperative cohort; P = .10) and diverged by the end of follow-up (6.29% [1.72] vs 7.21% [2.02], respectively; P < .001). Solid lines depict penalized smoothing splines applied to 1-year interval level means, and shaded areas indicate 95% CIs.

Individuals with inadequate baseline glycemic control in the nonoperative group were less likely to achieve adequate control (HR, 0.69; 95% CI, 0.57-0.83; P < .001) and more likely to develop inadequate glycemic control (HR, 2.26; 95% CI, 1.74-2.93; P < .001). In those with HbA1c less than 10%, severe hyperglycemia occurred more frequently in the nonoperative group than in the surgical group (HR, 1.89; 95% CI, 1.46-2.45; P < .001) (Table 1).

Table 1. Probability of Glycemic Control Outcomes and New Insulin Use in Propensity-Matched Cohorts.

Outcome No. (%) HR (95% CI) P value
Nonoperative Surgical
Glycemic control
Adequate 194 (32.6) 252 (42.4) 0.69 (0.57-0.83) <.001
Inadequate 172 (21.1) 85 (10.4) 2.26 (1.74-2.93) <.001
Severe hyperglycemia 163 (11.7) 90 (6.4) 1.89 (1.46-2.45) <.001
Insulin use
Fast-acting insulin use 200 (27.9) 100 (11.9) 2.54 (2.00-3.23) <.001
Long-acting insulin use 233 (23.2) 109 (9.9) 2.50 (1.99-3.13) <.001
Open in a new tab

Abbreviation: HR, hazard ratio.

Insulin Use

In participants not using insulin at baseline, initiation of fast-acting insulin was more frequent in the nonoperative group than in the surgical group (HR, 2.54; 95% CI, 2.00-3.23; P < .001). Initiation of long-acting insulin was also more frequent in the nonoperative group (HR, 2.50; 95% CI, 1.99-3.13; P < .001) (Figure 2).

Figure 2. Kaplan-Meier Curves for Probability of Glycemic Outcomes During Follow-Up Between Patients Managed Surgically and Nonoperatively.

Four Kaplan-Meier panels comparing surgical vs nonoperative probabilities over 0 to 60 months. Four-panel data figure with panels A and B on the top row and panels C and D on the bottom row. Each panel contains two step curves with shaded bands: a dark teal curve labeled Nonoperative and an orange curve labeled Surgical. All panels share a horizontal axis labeled Time from index date, mo, ranging from 0 to 60 with ticks at 0, 12, 24, 36, 48, and 60. All vertical axes are labeled Probability with panel-specific ranges. Panel A title at upper left: Inadequate glycemic control. Vertical axis ranges from 0 to 0.30. The teal curve rises to about 0.27 to 0.28 by 60 months; the orange curve rises to about 0.15 to 0.16 by 60 months. Numbers at risk under panel A: Nonoperative 816, 591, 465, 388, 344, 319; Surgical 820, 718, 591, 509, 391, 289. Panel B title: Severe hypoglycemia. Vertical axis ranges from 0 to 0.20. The teal curve rises to about 0.16 to 0.17 at 60 months; the orange curve rises to about 0.08 to 0.09 at 60 months. Numbers at risk under panel B: Nonoperative 1399, 1261, 1068, 925, 734, 535; Surgical 1403, 1330, 1161, 1003, 785, 594. Panel C title: Fast-acting insulin use. Vertical axis ranges from 0 to 0.40. The teal curve increases to about 0.33 to 0.35 by 60 months; the orange curve increases to about 0.16 to 0.18 by 60 months. Numbers at risk under panel C: Nonoperative 730, 578, 466, 384, 271, 177; Surgical 783, 711, 600, 527, 401, 302. Panel D title: Long-acting insulin use. Vertical axis ranges from 0 to 0.40. The teal curve increases to about 0.28 to 0.30 by 60 months; the orange curve increases to about 0.13 to 0.15 by 60 months. Numbers at risk under panel D: Nonoperative 1034, 838, 671, 565, 413, 273; Surgical 1084, 981, 824, 735, 563, 429.

Open in a new tab

Estimates of the probability of inadequate glycemic control (glycated hemoglobin [HbA1c] ≥7%) among patients with baseline HbA1c less than 7% (A); severe hypoglycemia (HbA1c ≥10%) among patients with baseline HbA1c less than 10% (B); and fast-acting insulin use (C) and long-acting insulin use (D) among patients who were nonusers at baseline. Log-rank P < .001 between the surgical and nonoperative cohort curves for all outcomes.

Diabetic Complications

Short-term and long-term rates of diabetes-related complications were higher in the nonoperative cohort (Table 2). At 5 years, complications occurred in 1438 patients (63.3%) in the nonoperative group and 1209 patients (53.2%) in the surgical group, corresponding to an absolute risk difference of 10.1 percentage points (OR, 1.52; 95% CI, 1.34-1.71; P < .001).

Table 2. Short- and Long-Term Rates of Diabetes-Related Complications.

Outcome No. (%) OR (95% CI) P value
Nonoperative (n = 2272) Surgical (n = 2272)
Diabetes with any complication
Short term 1027 (45.2) 936 (41.2) 1.18 (1.05-1.32) .007
Long term 1438 (63.3) 1209 (53.2) 1.52 (1.34-1.71) <.001
Kidney complications
Chronic kidney disease
Short term 431 (19.0) 310 (13.7) 1.48 (1.26-1.74) <.001
Long term 616 (27.1) 379 (16.7) 1.85 (1.60-2.14) <.001
End-stage kidney disease
Short term 113 (5.0) 83 (3.7) 1.38 (1.03-1.84) .03
Long term 186 (8.2) 118 (5.2) 1.63 (1.28-2.08) <.001
Kidney transplant
Short term 50 (2.2) 26 (1.1) 1.94 (1.21-3.13) .006
Long term 52 (2.3) 30 (1.3) 1.84 (1.16-2.93) .009
Circulatory complications
Heart failure
Short term 259 (11.4) 179 (7.9) 1.51 (1.23-1.84) <.001
Long term 443 (19.5) 275 (12.1) 1.76 (1.49-2.09) <.001
Myocardial infarction
Short term 84 (3.7) 61 (2.7) 1.39 (1.00-1.95) .05
Long term 177 (7.8) 118 (5.2) 1.53 (1.19-1.97) .001
Stroke
Short term 66 (2.9) 40 (1.8) 1.67 (1.12-2.48) .01
Long term 161 (7.1) 93 (4.1) 1.79 (1.36-2.35) <.001
Ophthalmic complications
Retinopathy
Short term 114 (5.0) 68 (3.0) 1.69 (1.23-2.32) .001
Long term 302 (13.3) 161 (7.1) 2.01 (1.57-2.58) <.001
Glaucoma
Short term 48 (2.1) 30 (1.3) 1.58 (0.98-2.55) <.001
Long term 118 (5.2) 75 (3.3) 1.62 (1.12-2.34) .01
Glycemic complications
DKA/HHS
Short term 119 (5.2) 62 (2.7) 1.97 (1.44-2.69) <.001
Long term 225 (9.9) 107 (4.7) 2.22 (1.74-2.84) <.001
Hypoglycemia
Short term 105 (4.6) 95 (4.2) 1.11 (0.84-1.48) .47
Long term 279 (12.3) 189 (8.3) 1.54 (1.26-1.88) <.001
Open in a new tab

Abbreviations: DKA/HHS, diabetic ketoacidosis or hyperosmolar hyperglycemic state; OR, odds ratio.

Kidney complications were the most frequent diabetes-related complication and occurred more often in the nonoperative group at both short- and long-term follow-up. At 5 years, circulatory and ophthalmic complications were less common overall but occurred more frequently in the nonoperative group. The long-term incidence of DKA/HHS and hypoglycemia was also higher in the nonoperative group (Table 2).

Trends in diabetes-related clinical parameters are shown in Figure 3. Mean glomerular filtration rate declined over time in both groups and remained higher in the surgical group throughout follow-up. Levels of low-density lipoprotein cholesterol were lower in the surgical group across follow-up, whereas left ventricular ejection fraction did not differ between groups. Systolic blood pressure decreased postoperatively in the surgical cohort and remained modestly lower than in the nonoperative group across follow-up.

Figure 3. Line Graphs of Diabetes-Related Parameters During Follow-Up Between Patients Managed Surgically and Nonoperatively.

Four-panel line graphs of G F R, systolic blood pressure, L D L, and L V E F over 60 months. Four panels arranged two by two, labeled A through D in small boxed letters at the upper left of each panel. Each panel contains two line series with circular markers and vertical error bars; one series uses orange markers and the other uses dark teal markers. Panel A title G F R. Vertical axis labeled Mean G F R, mL per min per one point seven three m squared, ranging from 70 to 100. Horizontal axis labeled Time from index date, mo, with ticks at 0, 12, 24, 36, 48, 60. Orange series labeled Surgical near the right side, with approximate values 87 at 0, 89 at 12, 85 at 24, 82 at 36, 82 at 48, 81 at 60. Teal series labeled Nonoperative near the lower right, with approximate values 83 at 0, 78 at 12, 77 at 24, 76 at 36, 75 at 48, 73 at 60. Panel B title Systolic blood pressure. Vertical axis labeled Mean systolic blood pressure, mm Hg, ranging from 120 to 135. Orange Surgical values about 129 at 0, 126 at 12, 127 at 24, 126 point 5 at 36, 127 point 5 at 48, 128 at 60. Teal Nonoperative values about 129 at 0, 129 at 12, 128 point 5 at 24, 128 point 7 at 36, 129 point 5 at 48, 130 at 60. Panel C title Serum L D L. Vertical axis labeled Mean serum L D L, mg per dL, ranging from 70 to 100. Orange Surgical values about 93 at 0, 77 at 12, 78 at 24, 78 at 36, 79 at 48, 78 at 60. Teal Nonoperative values about 93 point 5 at 0, 90 at 12, 88 point 5 at 24, 87 point 5 at 36, 89 point 5 at 48, 87 point 5 at 60. Panel D title Left ventricular ejection fraction. Vertical axis labeled Mean L V E F, percent, ranging from 54 to 64. Orange Surgical values about 60 at 0, 58 point 5 at 12, 59 point 7 at 24, 58 point 3 at 36, 61 point 5 at 48, 59 point 2 at 60. Teal Nonoperative values about 58 point 8 at 0, 57 at 12, 59 point 8 at 24, 56 point 9 at 36, 56 point 3 at 48, 59 point 3 at 60.

Open in a new tab

Trends are shown for mean glomerular filtration rate (GFR) (A); systolic blood pressure (B); serum low-density lipoprotein (LDL) (C); and left ventricular ejection fraction (LVEF) (D) during follow-up. Error bars indicate SE.

Health Care Utilization

At 5 years, 1138 patients (54.7%) in the nonoperative group and 901 (41.7%) in the surgical group had an emergency department visit (OR, 1.69; 95% CI, 1.50-1.91; P < .001). Hospitalization occurred in 1276 patients (59.0%) in the nonoperative group and 984 (45.5%) in the surgical group (OR, 1.72, 95% CI, 1.53-1.94; P < .001). Five-year mortality did not differ between groups (10.4% in both groups; P > .99).

Type 1 Diabetes vs Type 2 Diabetes

Of the 2272 surgical participants included in subgroup analyses, 419 had type 1 diabetes and 1853 had type 2 diabetes. At baseline, participants with type 1 diabetes had higher mean (SD) HbA1c (9.95% [2.14] vs 6.59% [1.70]; P < .001) and more frequent long-acting insulin use (221 [52.7%] vs 350 [18.9%]; P < .001). Mean HbA1c decreased by 0.48 percentage points in the type 1 diabetes group and 0.59 percentage points in the type 2 diabetes group over 5 years; however, values remained higher in participants with type 1 diabetes at study end (7.47% [2.33] vs 5.85% [1.35], respectively; P < .001).

GES vs Pyloric Drainage Procedure

In the surgical group, 321 participants underwent GES and 1848 underwent a pyloric drainage procedure. The 103 patients who underwent both procedures were excluded from this analysis. Baseline mean (SD) HbA1c was higher in the GES group (7.88% [2.11] vs 6.60% [1.76], respectively; P < .001). HbA1c decreased at 5 years in both groups, with a decrease of 0.46 percentage points in the GES group and 0.52 percentage points in the pyloric drainage group. Compared with the pyloric drainage group, the GES group was less likely to achieve adequate glycemic control (HR, 0.25; 95% CI, 0.18-0.37; P = .001).

Discussion

Gastroparesis is a recognized complication of diabetes; however, the metabolic consequences of surgical treatment for diabetic gastroparesis remain poorly defined, and current guidelines emphasize symptom control as the primary indication for surgical intervention.2,3 In this large multicenter cohort, surgical therapy was associated with sustained improvements in glycemic control, reduced insulin use, and lower rates of diabetes-related complications compared with nonoperative management. Notably, patients with type 1 diabetes and those undergoing GES had more severe baseline metabolic disease yet experienced postoperative improvements in glycemic control. These findings suggest that metabolic outcomes may warrant consideration alongside symptom relief in discussions of surgical management for diabetic gastroparesis.

In this multicenter cohort, surgical therapy for diabetic gastroparesis was associated with sustained improvements in glycemic control, reflected by durable reductions in mean HbA1c, higher probability of achieving adequate glycemic control, and reduced initiation of insulin therapy. The physiologic relationship between gastric emptying and glycemia is complex and bidirectional. Acute hyperglycemia delays gastric emptying, and interindividual differences in gastric emptying account for up to 35% of variability in postprandial glycemia.8 Variable gastric emptying in insulin-dependent diabetes may complicate the timing of postprandial insulin administration as well as precipitate both early hypoglycemia and delayed hyperglycemia.9 Modulation of nutrient delivery has been shown to reduce postprandial glycemic variability in diabetic gastroparesis, suggesting that stabilization of gastric emptying may influence downstream metabolic control.10

Surgical literature on gastroparesis has focused predominantly on symptomatic outcomes, with glycemic parameters reported inconsistently and typically as secondary measures.6,7,11,12 Studies of GES and combined procedures have described HbA1c reductions ranging from 0.5% to 1%, although results have been heterogeneous and often limited by short follow-up.4,13,14 Earlier observational series suggested larger HbA1c reductions with longer follow-up durations.15,16 Surgical therapy in this cohort was associated with a sustained 0.5–percentage point reduction in HbA1c, accompanied by a higher probability of achieving adequate glycemic control over extended follow-up. These findings are consistent with the hypothesis that surgical correction of impaired gastric emptying may contribute to stabilization of metabolic control. An upcoming French trial evaluating glycemic outcomes after G-POEM may further clarify the relationship between postoperative gastric emptying and metabolic control (ClinicalTrials.gov, NCT06965543).

Large randomized trials have established HbA1c as a modifiable determinant of long-term diabetes outcomes. In the UK Prospective Diabetes Study, each 1% reduction in HbA1c was associated with a 21% reduction in diabetes-related end points and a 37% reduction in microvascular complications, with no evident glycemic threshold below which benefit ceased.17 The Diabetes Control and Complications Trial similarly demonstrated that intensive glycemic control significantly reduced retinopathy, nephropathy, and neuropathy in type 1 diabetes, underscoring the importance of cumulative glycemic exposure.18 Against this backdrop, the sustained 0.5% reduction in HbA1c observed in the present cohort, accompanied by lower insulin initiation and reduced rates of diabetes-related complications, is likely to be clinically meaningful. Although the magnitude of HbA1c reduction is modest compared with that achieved in intensive medical therapy trials, its durability over several years as well as concordant insulin use and complication signals support the biologic plausibility that surgical correction of impaired gastric emptying may favorably influence long-term metabolic trajectories.

Importantly, the surgical cohort demonstrated a substantially lower likelihood of insulin initiation. Progression to insulin dependence represents a meaningful milestone in diabetes, often reflecting declining beta-cell reserve and advancing disease burden.19,20 The approximately 2.5-fold higher hazard of insulin initiation in the nonoperative cohort suggests that surgery may delay metabolic progression. Reduced insulin exposure also carries implications beyond glycemic control, including lower risk of hypoglycemia, weight gain, and health care utilization.21 Collectively, these findings indicate that the observed metabolic effect reflects a clinically consequential shift in glycemic trajectory rather than a modest laboratory change.

Surgical therapy was associated with lower rates of diabetes-related complications, including kidney, cardiovascular, ophthalmic, and metabolic outcomes, compared with nonoperative management. Prior studies evaluating glycemic outcomes after gastroparesis surgery have been limited by short follow-up durations, precluding assessment of long-term complication risk.4,13 Microvascular and macrovascular complications of diabetes result from cumulative exposure to chronic hyperglycemia.22 The extended follow-up in this analysis enabled evaluation of potential downstream effects of sustained metabolic improvement. The lower incidence of DKA and HHS in surgically treated patients further supports a reduced hyperglycemic burden. Although pylorus-directed interventions carry a theoretical risk of postprandial hypoglycemia, no increase in hypoglycemia was observed after surgery. Higher long-term hypoglycemia rates in the nonoperative group were observed in the setting of greater insulin utilization, suggesting that differences in treatment intensity may account for this finding.

Emergency department visits and all-cause hospitalizations were less frequent among surgically treated patients, with absolute reductions of 13.0% and 14.3%, respectively, compared with nonoperative management. Prior retrospective analyses have similarly reported reductions in hospitalization burden after GES.16,23 The indications for health care encounters could not be ascertained within this database; therefore, whether reduced utilization reflects improvement in gastroparesis-related symptoms, enhanced metabolic stability, or attenuation of comorbidity burden cannot be determined. Nevertheless, with the substantial cost of diabetes-related care in the US, reduced health care utilization in a highly comorbid population may carry ramifications beyond patient-level outcomes.24

Given the established association between glycemic control and mortality in diabetes, differences in long-term survival would be biologically plausible in the setting of lower complication rates.25,26 However, no difference in 5-year all-cause mortality was observed between groups. Perioperative mortality in the surgical cohort was low, suggesting that early surgical risk does not explain this finding. The prolonged natural history of diabetic microvascular and macrovascular disease may require longer follow-up to detect potential survival differences.

Patients with type 1 diabetes and those treated with GES had more severe baseline glycemic profiles and higher postoperative HbA1c values compared with their respective counterparts. The greater baseline dysglycemia in the type 1 subgroup is consistent with the longer disease duration and cumulative complication burden typically associated with type 1 diabetes.27 The higher baseline HbA1c observed in the GES group likely reflects differences in case selection, including a greater proportion of patients with type 1 diabetes. Despite these baseline differences, both subgroups experienced postoperative reductions in HbA1c. These findings suggest that although baseline metabolic severity influences absolute postoperative glycemic levels, it does not preclude measurable metabolic improvement after gastroparesis surgery. Prospective studies are needed to clarify whether diabetes phenotype or procedural approach modifies long-term glycemic outcomes.

Limitations

This analysis has several limitations inherent to large administrative databases. Reliance on coded diagnoses and procedural data introduces the possibility of misclassification and incomplete clinical characterization. Because no ICD-10-CM code specifically identifies diabetic gastroparesis, the cohort was defined by concurrent diagnoses of diabetes and gastroparesis, which may have resulted in inclusion of patients with alternative causes of delayed gastric emptying. The database also lacks gastroparesis-specific clinical variables, including symptom severity scores and objective gastric emptying measurements, precluding assessment of baseline physiologic impairment and procedural efficacy. Although propensity score matching balanced measured demographic and metabolic covariates, residual confounding from unmeasured factors, including diabetes duration, socioeconomic status, and referral patterns, remains possible. Because this is an observational study, causal inference cannot be established. Nevertheless, the consistency of associations across multiple metabolic and complication-related outcomes over extended follow-up supports the overall coherence of the findings.

Conclusions

This cohort study found that surgical therapy for diabetic gastroparesis was associated with improved glycemic control, reduced insulin utilization, and lower rates of diabetes-related complications compared with nonoperative management. Patients with type 1 diabetes and those treated with GES exhibited more severe baseline metabolic profiles but still experienced postoperative glycemic improvement. These findings support further prospective investigation to define the relationship between gastric emptying and glycemic regulation, ideally incorporating objective gastric emptying assessments and continuous glucose monitoring. If confirmed, improved metabolic control may represent an additional consideration when evaluating surgical therapy for diabetic gastroparesis.

Supplement 1.

eTable 1. International Classification of Disease, Tenth Revision, Clinical Modification (ICD-10-CM), Current Procedural Terminology, Fourth Edition (CPT-4), Logical Observation Identifiers and Names (LOINC), and Anatomic Therapeutic Chemical classification (ATC) medication codes used for cohort definitions, outcomes, and covariates

eTable 2. Comparison of baseline characteristics between the surgical cohort and nonoperative cohort, before and after propensity score matching

jamasurg-e261593-s001.pdf (243KB, pdf)
Supplement 2.

Data sharing statement

jamasurg-e261593-s002.pdf (15.1KB, pdf)

References

  • 1.Zhou B, Rayner AW, Gregg EW, et al. ; NCD Risk Factor Collaboration (NCD-RisC) . Worldwide trends in diabetes prevalence and treatment from 1990 to 2022: a pooled analysis of 1108 population-representative studies with 141 million participants. Lancet. 2024;404(10467):2077-2093. doi: 10.1016/S0140-6736(24)02317-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Camilleri M, Kuo B, Nguyen L, et al. ACG clinical guideline: gastroparesis. Am J Gastroenterol. 2022;117(8):1197-1220. doi: 10.14309/ajg.0000000000001874 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Staller K, Parkman HP, Greer KB, et al. ; AGA Clinical Guidelines Committee . AGA clinical practice guideline on management of gastroparesis. Gastroenterology. 2025;169(5):828-861. doi: 10.1053/j.gastro.2025.08.004 [DOI] [PubMed] [Google Scholar]
  • 4.Ducrotte P, Coffin B, Bonaz B, et al. ; ENTERRA Research Group . Gastric electrical stimulation reduces refractory vomiting in a randomized crossover trial. Gastroenterology. 2020;158(3):506-514.e2. doi: 10.1053/j.gastro.2019.10.018 [DOI] [PubMed] [Google Scholar]
  • 5.Martinek J, Hustak R, Mares J, et al. Endoscopic pyloromyotomy for the treatment of severe and refractory gastroparesis: a pilot, randomised, sham-controlled trial. Gut. 2022;71(11):2170-2178. doi: 10.1136/gutjnl-2022-326904 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Shada AL, Dunst CM, Pescarus R, et al. Laparoscopic pyloroplasty is a safe and effective first-line surgical therapy for refractory gastroparesis. Surg Endosc. 2016;30(4):1326-1332. doi: 10.1007/s00464-015-4385-5 [DOI] [PubMed] [Google Scholar]
  • 7.Eriksson SE, Chrysostome ML, Sarici IS, Nguyen J, Zheng P, Ayazi S. Pyloric drainage interventions for gastroparesis: a comparison of laparoscopic pyloroplasty and gastric peroral endoscopic myotomy (G-POEM) outcomes. Surg Endosc. 2025;39(6):3514-3524. doi: 10.1007/s00464-025-11731-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Phillips LK, Deane AM, Jones KL, Rayner CK, Horowitz M. Gastric emptying and glycaemia in health and diabetes mellitus. Nat Rev Endocrinol. 2015;11(2):112-128. doi: 10.1038/nrendo.2014.202 [DOI] [PubMed] [Google Scholar]
  • 9.Goyal RK. Gastric emptying abnormalities in diabetes mellitus. N Engl J Med. 2021;384(18):1742-1751. doi: 10.1056/NEJMra2020927 [DOI] [PubMed] [Google Scholar]
  • 10.Betônico CC, Cobello AV, Santos-Bezerra DP, et al. Diet consistency modification improves postprandial glycemic and gastroparesis symptoms. J Diabetes Metab Disord. 2022;21(2):1661-1667. doi: 10.1007/s40200-022-01117-w [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Khashab MA, Wang AY, Cai Q. AGA clinical practice update on gastric peroral endoscopic myotomy for gastroparesis: commentary. Gastroenterology. 2023;164(7):1329-1335.e1. doi: 10.1053/j.gastro.2023.02.027 [DOI] [PubMed] [Google Scholar]
  • 12.Marowski S, Xu Y, Greenberg JA, Funk LM, Lidor AO, Shada AL. Both gastric electrical stimulation and pyloric surgery offer long-term symptom improvement in patients with gastroparesis. Surg Endosc. 2021;35(8):4794-4804. doi: 10.1007/s00464-020-07960-3 [DOI] [PubMed] [Google Scholar]
  • 13.Sarosiek I, Bashashati M, Davis BR, et al. Combined Gastric electrical stimulation and pyloroplasty in gastroparesis: a randomized clinical trial. JAMA Netw Open. 2025;8(12):e2546332-e2546332. doi: 10.1001/jamanetworkopen.2025.46332 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Sarosiek I, Bashashati M, Gonzalez Z, et al. The long-term outcomes of combining pyloroplasty with gastric electrical stimulation in drug-refractory gastroparesis: a prospective single-arm trial. J Investig Med. Published online August 8, 2025. doi: 10.1177/10815589251366904 [DOI] [PubMed] [Google Scholar]
  • 15.van der Voort IR, Becker JC, Dietl KH, Konturek JW, Domschke W, Pohle T. Gastric electrical stimulation results in improved metabolic control in diabetic patients suffering from gastroparesis. Exp Clin Endocrinol Diabetes. 2005;113(1):38-42. doi: 10.1055/s-2004-830525 [DOI] [PubMed] [Google Scholar]
  • 16.Lin Z, Forster J, Sarosiek I, McCallum RW. Treatment of diabetic gastroparesis by high-frequency gastric electrical stimulation. Diabetes Care. 2004;27(5):1071-1076. doi: 10.2337/diacare.27.5.1071 [DOI] [PubMed] [Google Scholar]
  • 17.Stratton IM, Adler AI, Neil HA, et al. Association of glycaemia with macrovascular and microvascular complications of type 2 diabetes (UKPDS 35): prospective observational study. BMJ. 2000;321(7258):405-412. doi: 10.1136/bmj.321.7258.405 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Nathan DM, Cleary PA, Backlund JY, et al. ; Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC) Study Research Group . Intensive diabetes treatment and cardiovascular disease in patients with type 1 diabetes. N Engl J Med. 2005;353(25):2643-2653. doi: 10.1056/NEJMoa052187 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Porte D Jr, Kahn SE. Beta-cell dysfunction and failure in type 2 diabetes: potential mechanisms. Diabetes. 2001;50(suppl 1):S160-S163. doi: 10.2337/diabetes.50.2007.S160 [DOI] [PubMed] [Google Scholar]
  • 20.Lin YJ, Liu PP, Huang HK, et al. Risk of major adverse cardiovascular events and all-cause mortality in type 2 diabetic patients receiving insulin versus non-insulin treatment intensification. Sci Rep. 2025;15(1):29694. doi: 10.1038/s41598-025-13254-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.McCoy RG, Lipska KJ, Van Houten HK, Shah ND. Association of cumulative multimorbidity, glycemic control, and medication use with hypoglycemia-related emergency department visits and hospitalizations among adults with diabetes. JAMA Netw Open. 2020;3(1):e1919099. doi: 10.1001/jamanetworkopen.2019.19099 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Orasanu G, Plutzky J. The pathologic continuum of diabetic vascular disease. J Am Coll Cardiol. 2009;53(5)(suppl):S35-S42. doi: 10.1016/j.jacc.2008.09.055 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Lin Z, McElhinney C, Sarosiek I, Forster J, McCallum R. Chronic gastric electrical stimulation for gastroparesis reduces the use of prokinetic and/or antiemetic medications and the need for hospitalizations. Dig Dis Sci. 2005;50(7):1328-1334. doi: 10.1007/s10620-005-2782-7 [DOI] [PubMed] [Google Scholar]
  • 24.Parker ED, Lin J, Mahoney T, et al. Economic costs of diabetes in the U.S. in 2022. Diabetes Care. 2024;47(1):26-43. doi: 10.2337/dci23-0085 [DOI] [PubMed] [Google Scholar]
  • 25.Gregg EW, Cheng YJ, Srinivasan M, et al. Trends in cause-specific mortality among adults with and without diagnosed diabetes in the USA: an epidemiological analysis of linked national survey and vital statistics data. Lancet. 2018;391(10138):2430-2440. doi: 10.1016/S0140-6736(18)30314-3 [DOI] [PubMed] [Google Scholar]
  • 26.Rawshani A, Rawshani A, Franzén S, et al. Risk factors, mortality, and cardiovascular outcomes in patients with type 2 diabetes. N Engl J Med. 2018;379(7):633-644. doi: 10.1056/NEJMoa1800256 [DOI] [PubMed] [Google Scholar]
  • 27.Patsoukaki V, Lind L, Lampa E, et al. Risk differences and underlying factors of cardiovascular events and mortality in patients with type 2 diabetes versus type 1 diabetes: a longitudinal cohort study of Swedish nationwide register data. Lancet Diabetes Endocrinol. 2025;13(10):848-862. doi: 10.1016/S2213-8587(25)00165-2 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement 1.

eTable 1. International Classification of Disease, Tenth Revision, Clinical Modification (ICD-10-CM), Current Procedural Terminology, Fourth Edition (CPT-4), Logical Observation Identifiers and Names (LOINC), and Anatomic Therapeutic Chemical classification (ATC) medication codes used for cohort definitions, outcomes, and covariates

eTable 2. Comparison of baseline characteristics between the surgical cohort and nonoperative cohort, before and after propensity score matching

jamasurg-e261593-s001.pdf (243KB, pdf)
Supplement 2.

Data sharing statement

jamasurg-e261593-s002.pdf (15.1KB, pdf)

Articles from JAMA Surgery are provided here courtesy of American Medical Association

RESOURCES