Diabetic gastroparesis is a complication that often occurs in patients with long-standing diabetes, and it is characterized by chronic delayed gastric emptying without mechanical obstruction and upper GI symptoms. Diabetic gastroparesis occurs in about 25-55% of patients with diabetes [
Prokinetic agents have been used in managing the symptoms of diabetic gastroparesis [
The meta-analysis was conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [
Citations and abstracts of all retrieved studies were downloaded to Endnote X8.1 citation management software (Thomson Reuters, Philadelphia, PA, USA). After removing duplicated titles and abstracts, the retrieved articles were independently reviewed by two authors (Y.J. Kim and W.C. Chung). The full text of the relevant articles was checked against inclusion criteria and discrepancies, and any issues were resolved by consensus.
In the meta-analysis, the inclusion criteria were as follows: (1) RCTs, (2) studies on adult diabetic and prediabetic state patients, (3) glycemic control measured by HbA1c or FBS, (4) the control group received placebo for the same period as the treatment group, and (5) the treatment group received prokinetic agents for at least 1 week without any other GI medications such as gastric acid inhibitors or mucoprotective drugs. Studies were excluded if they were available only as an abstract, a review study, a case report, a study without raw data available for retrieval, a duplicate publication, a non-English publication, a crossover study design, or studies without a control group or data from a single experiment. Data without a mean value for the outcome between the two groups were also excluded.
The aim of this study was to analyze the effects of prokinetics compared to placebo on glycemic control defined by SMD of HbA1c and FBS. A fasting insulin level was also compared according to prokinetics usage. The effect of prokinetics on glycemic control according to the ingredients was additionally evaluated. Each drug component was categorized as a serotonergic agonist, dopamine antagonist, motilin agonist, and cholinergic agonist.
The pooled SMD with 95% CIs was quantitatively evaluated. Heterogeneity among the studies was measured using Higgins’
The process of literature search and study selection is shown in Figure
Flow diagram of the study.
The characteristics of the eligible studies are summarized in Table
Main characteristics.
Studies | Treatment | Mechanism of action | Treatment period |
---|---|---|---|
Melga, P. et al., 1997 | Levosulpiride | Dopamine antagonist | 24 weeks |
Nam, J. S. et al., 2010 | Mosapride | 5-HT4 agonist | 2 weeks |
Ueno, N. et al., 2000 | Erythromycin | Motilin agonist | 4 weeks |
Ueno, N. et al., 2001 | Erythromycin | Motilin agonist | 4 weeks |
Ueno, N. et al., 2002 | Mosapride | 5-HT4 agonist | 8 weeks |
5-HT4: 5-hydroxytryptamine receptor 4.
Baseline characteristics.
Studies | Age (years) | Sex (male/female) | BMI (kg/m2) | Duration of diabetes (years) | ||||
---|---|---|---|---|---|---|---|---|
T | C | T | C | T | C | T | C | |
Melga, P. et al., 1997 |
8/12 | 9/11 | ||||||
Nam, J. S. et al., 2010† | 14/6 | 8/2 | — | — | ||||
Ueno, N. et al., 2000 |
14/5 | 8/7 | ||||||
Ueno, N. et al., 2001† | 12/18 | 10/12 | ||||||
Ueno, N. et al., 2002 |
7/10 | 8/9 |
The effect of prokinetics on HbA1c was examined, and four of five studies reported a statistically significant improvement in HbA1c between the prokinetics and control groups (Table
The results of meta-analysis for HbA1c.
Studies | Treatment | Control | Standardized mean diff. | 95% lower CI | 95% upper CI | Weights | |||||
---|---|---|---|---|---|---|---|---|---|---|---|
Mean | SD | Mean | SD | ||||||||
Melga, P. et al., 1997 | 20 | 5.7 | 0.95 | 20 | 6.8 | 0.89 | -1.171 | -1.848 | -0.495 | 20.5% | |
Nam, J. S. et al., 2010 | 20 | 5.4 | 0.5 | 10 | 5.55 | 1.2 | -0.184 | -0.944 | 0.577 | 19.5% | |
Ueno, N. et al., 2000 | 19 | 7.6 | 0.87 | 15 | 8.6 | 1.16 | -0.969 | -1.689 | -0.249 | 20.0% | |
Ueno, N. et al., 2001 | 30 | 7.8 | 0.2 | 22 | 8.3 | 0.2 | -2.462 | -3.200 | -1.724 | 19.8% | |
Ueno, N. et al., 2002 | 17 | 7.67 | 0.78 | 17 | 8.5 | 0.99 | -0.909 | -1.620 | -0.199 | 20.1% | |
Total (random effect model) | -1.141 | -1.843 | -0.438 | <0.01 | |||||||
Subgroup 1 (Nam, J. S. et al., 2010; Ueno, N. et al., 2002) (random effect model) | -0.560 | -1.090 | -0.052 | 0.12 | |||||||
Subgroup 2 (Ueno, N. et al. 2000; Ueno, N. et al., 2001) (random effect model) | -1.714 | -3.177 | -0.251 | 0.02 |
The results of the meta-analysis for HbA1C.
Four studies reported results on FBS by comparing the prokinetics group with placebo (Table
The results of meta-analysis for fasting blood glucose.
Studies | Treatment | Control | Standardized mean diff. | 95% lower CI | 95% upper CI | Weights | |||||
---|---|---|---|---|---|---|---|---|---|---|---|
Mean | SD | Mean | SD | ||||||||
Nam, J. S. et al., 2010 | 20 | 6 | 0.9 | 10 | 6.3 | 1.1 | -0.301 | -1.065 | 0.462 | 25.0% | |
Ueno, N. et al., 2000 | 19 | 9.7 | 2.06 | 15 | 11.1 | 3.1 | -0.532 | -1.223 | 0.158 | 25.4% | |
Ueno, N. et al., 2001 | 30 | 8.2 | 0.9 | 22 | 11 | 0.5 | -3.368 | -4.549 | -2.276 | 24.2% | |
Ueno, N. et al., 2002 | 17 | 8.51 | 1.4 | 17 | 9.6 | 1.61 | -0.705 | -1.401 | -0.010 | 25.4% | |
Total (random effect model) | -1.270 | -2.613 | 0.074 | 0.06 | |||||||
Subgroup 1 (Nam, J. S. et al., 2010; Ueno, N. et al., 2002) (random effect model) | -0.522 | -1.036 | -0.010 | 0.04 | |||||||
Subgroup 2 (Ueno, N. et al., 2000; Ueno, N. et al., 2001) (random effect model) | -2.070 | -5.113 | 0.973 | 0.18 |
The results of the meta-analysis for FBS.
The effect of prokinetic agents on FINS was examined in three studies. Two studies (Nam et al. [
The results of meta-analysis for insulin.
Studies | Treatment | Control | Standardized mean diff. | 95% lower CI | 95% upper CI | Weights | |||||
---|---|---|---|---|---|---|---|---|---|---|---|
Mean | SD | Mean | SD | ||||||||
Nam, J. S. et al., 2010 | 20 | 34.51 | 20.09 | 10 | 38.03 | 25.11 | -0.157 | -0.917 | 0.603 | 33.0% | |
Ueno, N. et al., 2001 | 30 | 42 | 5.8 | 22 | 31.9 | 4.6 | 1.867 | 1.202 | 2.532 | 33.6% | |
Ueno, N. et al., 2002 | 17 | 40.8 | 22.68 | 17 | 54.7 | 19.05 | -0.648 | -1.340 | 0.044 | 33.4% | |
Total (random effect model) | 0.359 | -1.205 | 1.923 | 0.65 | |||||||
Subgroup 1 (Nam, J. S. et al., 2010; Ueno, N. et al., 2002) (random effect model) | -0.426 | -0.937 | 0.086 | 0.10 |
The results of the meta-analysis for fasting insulin.
The gut plays a crucial role in glucose homeostasis by aiding digestion, absorption, and assimilation of ingested nutrients [
The mechanisms by which prokinetics improve glucose metabolism are unclear, although several hypotheses have been proposed, such as improvements in insulin sensitivity or much greater secretion of plasma insulin to increase intraduodenal glucose loads. Some animal studies showed that erythromycin and motilin stimulate GI motility and the cyclic release of insulin and pancreatic polypeptides from the pancreas through the vagal-cholinergic muscarinic pathway [
Gastric emptying of solid food is known to consist of two phases: the lag phase (the meal is transported from the fundus to the antrum) and the postlag phase (the solid food particles are propelled through the pylorus) [
The Diabetes Control and Complications Trial (DCCT)/Epidemiology of Diabetes Interventions and Complications (EDIC) study showed that delayed gastric emptying was associated with early and long-term hyperglycemia [
Gastric emptying involves a complex interplay among the GI smooth muscle, gastric pacemaker cell networks, the so-called interstitial cells of Cajal, and neurohormonal systems, particularly inhibitory feedback arising from the interaction of nutrients with the small intestine [
Despite controversies, several current studies showed that an acceleration of the gastric emptying occurs following obesity surgery and more antral resections can lead to faster gastric emptying time [
This systemic review on the efficacy of prokinetics on glycemic control suggests that prokinetics reduced the glycemic marker, HbA1c, with a significant reduction in HbA1c implying long-term fluctuations in blood glucose concentration. However, prokinetics had no stimulatory effect on insulin secretion. This indicated that prokinetics have significant effects in reducing glucose metabolism in diabetic and prediabetic patients, and this action suggests that a mechanism other than insulin secretion is involved.
Our study showed a decrease in HbA1c with prokinetics use, which is consistent with a decrease in overall glucose and not with fasting glucose. There is a subtle difference in the assessment of the metabolic status of blood glucose as a consequence of FBS and HbA1c used in the diagnosis of diabetes. Fasting blood glucose involves evaluation of the state of glucose that remains stable in the body after a temporary increase in exogenously injected glucose is resolved, whereas HbA1c reflects the mean blood glucose level not just at the time of sampling but immediately before the test because it increases with the concentration and time of glucose in contact with hemoglobin. The FBS test is limited by low sensitivity and relatively large fluctuations [
Although the studies in this meta-analysis excluded one-time studies, there were few studies that used long-term medication. Given the nature of the chronic disease, long-term research will likely need to demonstrate the benefits of prokinetics on glycemic control. Generally, prokinetics have few side effects associated with long-term maintenance. However, careful consideration is needed in choosing prokinetics. Attention should be paid to the CNS effect of levosulpiride and metoclopramide and the association between long-term oral erythromycin and poor tolerance, modest efficacy, and the development of tachyphylaxis. Prokinetics added to diabetic medication are expected to be noninvasive compared to newly designed invasive methods for the treatment of the metabolic syndrome. In particular, the dual effects of chronic dyspepsia and glycemic control are to be expected and it is also remarkable in terms of cost-effectiveness.
There are some limitations to consider when interpreting our findings. First, the studies included in this analysis had small sample sizes. In addition, the number of studies for subgroup analysis according to the ingredients of the prokinetics was also limited. Second, we could not directly assess the change in gastric or intestinal emptying by prokinetics because of the difference in the test methods used in each study. The effects of gastrointestinal motility may be different in each subgroup of prokinetics, and this result may be related to glucose levels, but it was difficult to compare this directly in this meta-analysis. Third, the treatment period in all studies was relatively short, such as 2-8 weeks, except one study with erythromycin. The glycemic marker, HbA1c, may determine the average blood glucose levels within the previous three weeks. Comparisons of HbA1c in the absence of an adequate treatment period may have the potential to underestimate the values.
The meta-analysis shows that prokinetics may be effective in significantly improving both gastroparesis symptoms and glycemic control. A further well-designed large-scale prospective study should be performed to determine the long-term effect of prokinetics on glycemic control.
No potential conflict of interest relevant to this article was reported. The authors declare that they have no proprietary, commercial, or financial interests that could be construed to have inappropriately influenced this study.