Gastric Emptying Delay with Pre-Meal Administration of Metformin and Anagliptin in Type 2 Diabetes: Effects on Postprandial Glucose and Triglyceride Levels in a Two-Case Report

Introduction

Dyslipidemia is common in patients with type 2 diabetes, affecting approximately 50% of this group of patients [1]. Diabetic dyslipidemia with insulin resistance is marked by elevated triglyceride (TG) concentrations, reduced high-density lipoprotein cholesterol levels, and the presence of chylomicron particles. Regarding hypertriglyceridemia, approximately 30% to 40% of patients have TG concentrations >200 mg/dL. It is now considered a potential residual risk factor for cardiovascular disease after statin therapy, and postprandial hypertriglyceridemia is especially associated with a high risk of cardiovascular disease.

Metformin is used to treat type 2 diabetes. In addition to its effects on glucose metabolism, its TG-lowering effects in the fasting and postprandial states have been recognized since the 1990s [2,3]. These effects are reported to be dose-dependent, as is its glucose-lowering effect [4]. Recently, we reported that the administration of a high dose of metformin before fat loading significantly reduces the postprandial plasma TG concentration and slows gastric emptying, compared with administration after fat loading, in an animal model [5]. Subsequently, we conducted a small randomized crossover trial, showing that metformin administration 30 min before a meal significantly reduces the postprandial plasma TG concentration, without markedly exacerbating its gastrointestinal adverse effects, compared with administration after a meal [6], suggesting that changing the timing of metformin administration can enhance its TG-lowering effects in patients with type 2 diabetes and postprandial hypertriglyceridemia. Metformin has been demonstrated to stimulate glucagon-like peptide-1 (GLP-1) secretion [7–9]. Because GLP-1 is known to reduce gastric emptying, we hypothesized GLP-1 may be the factor reducing plasma TG levels through gastric emptying. However, in our preliminary study, we have found no difference in GLP-1 levels between pre- and post-meal administration of metformin [6]. Therefore, we hypothesized that if metformin were administered in combination with a dipeptidyl peptidase-4 (DPP-4) inhibitor, which prevents the degradation of circulating GLP-1, it would enhance the delay in gastric peristaltic emptying. To test this hypothesis, we initiated a clinical trial to compare the effects of pre- and post-meal administration of a tablet containing a combination of metformin and anagliptin, a DPP-4 inhibitor, on glucose and lipid metabolism and gastric emptying measured using 13C-acetate. The study was approved by the Ethics Committee of Shiga University of Medical Science (approval number CRB5180008) and was registered with the Japan Registry of Clinical Trials (registration number jRCTs051200098). However, because the trial had to be terminated because of the COVID-19 pandemic, we report the details of the 2 patients for whom data were obtained.

Case Reports

We planned a randomized, open-label, 2-arm, crossover study to investigate the effects of a combination of metformin and anagliptin on the postprandial hypertriglyceridemia of patients with type 2 diabetes (Figure 1). The primary endpoint was the difference in the change in postprandial plasma TG concentration associated with pre-meal (pre-Met/Ana) and post-meal (post-Met/Ana) metformin and anagliptin administration. The key secondary outcomes were changes in the blood glucose and plasma insulin and active GLP-1 concentrations. The study was conducted in accordance with the principles of the Declaration of Helsinki. The research protocol was approved by the Ethics Committee of Shiga University of Medical Science on November 25, 2020 (approval number CRB5180008). All the participants provided their written informed consent to participate. Owing to the difficulty of continuing the study during the COVID-19 pandemic, the study was terminated on May 24, 2022.

The inclusion criteria for the study were as follows: a diagnosis of type 2 diabetes, hospitalization to facilitate glycemic control at the Shiga University of Medical Science Hospital, age less than 75 years, and a non-fasting TG concentration ranging from 200 to 1000 mg/dL. All evaluations were conducted once glycemic control had been achieved. The principal exclusion criteria included a history of acute pancreatitis caused by hypertriglyceridemia and a history of slow gastric emptying, which could be due to severe diabetic autonomic neuropathy, adhesion ileus, or the administration of drugs that affect gastrointestinal motility. We enrolled 3 patients who were taking a 1000-mg dose of metformin daily, but 1 patient was excluded because of a plasma TG concentration that was below the stipulated range. Therefore, only 2 patients completed the study protocol.

Two sets of meal tolerance tests were conducted 2 days apart after the patients had achieved good glycemic control: fasting plasma glucose concentration <140 mg/dL and 2-h postprandial concentration <200 mg/dL (Figure 1A). To evaluate the effects of fat loading, a meal test using cookies was performed after overnight fasting for 10 h. The cookies contained 75 g carbohydrate (flour starch and maltose), 28.5 g fat (butter), and 8 g protein, and provided a total of 592 kcal per pack (Saraya Corp, Osaka, Japan). At the first meal test, the patients were randomized to replace their 1000-mg metformin with 2 metformin/anagliptin combination tablets containing 500 mg of metformin and 100 mg of anagliptin that were taken either (1) 30 min before the test meal (pre-meal Met/Ana protocol) or (2) 15 min after starting the test meal (post-meal Met/Ana protocol). For the pre-meal Met/Ana protocol, a fasting blood sample (t=0 min) was taken immediately before tablet administration, and the patients were asked to eat the cookies with water within a 15-min period. For the post-meal Met/Ana protocol, a fasting blood sample (t=0 min) was taken, the cookies were eaten with water, and the tablet was administered 15 min after the start of ingestion. For both groups, blood samples were drawn 30, 60, 120, 180, and 240 min after the patients started to eat the cookies. The patients’ plasma TG, blood glucose, plasma active GLP-1, and plasma insulin concentrations were measured at the 0, 30, 60, 120, 180, and 240-min time points (Figure 1B). The first and second meal tests were performed using a 2-day washout period. During this period, the patients continued their assigned treatment. For the second meal test, each patient followed the other protocol. All laboratory testing was performed at SRL Laboratories (SRL Inc, Tokyo, Japan). Plasma insulin concentration was measured using a chemiluminescence immunoassay (Architect insulin assay; Abbott Laboratories, Abbott Park, IL, USA), and plasma active GLP-1 concentration was measured using an enzyme-linked immunosorbent assay kit (EMD Millipore Co, St. Louis, MO, USA).

The gastric peristaltic emptying time was evaluated by measuring the exhaled 13CO2 content of the breath after the patients had drunk liquid containing sodium 13C-acetate. During the pre-meal and post-meal Met/Ana protocols, an exhaled breath sample (t=0 min) was obtained before cookie ingestion commenced. The patients were then asked to drink the sodium 13C-acetate-containing liquid 5 min after the start of cookie ingestion. Exhaled breath samples were then obtained 15, 30, 45, 60, 75, 90, 105, 120, 150, 180, 210, and 240 min after cookie ingestion commenced. The 13CO2 content of exhaled air was measured at each time point, and we evaluated the differences in the values of the following variables between the 2 protocols before and after 13C-acetate loading: the time taken to reach the maximum 13CO2/12CO2 ratio in the breath (timing of peak 13CO2 excretion: Tmax), the time elapsed until half of the total amount of 13C imbibed had been eliminated in the breath (T½), and the gastric emptying coefficient.

A 62-year-old man with type 2 diabetes (patient 1) was recruited (Table 1). He was 182.0 cm tall and weighed 118.9 kg (body mass index 35.9 kg/m2). He had had diabetes for 17 years and was on insulin therapy. His non-fasting plasma TG concentration was 254 mg/dL, and his glycated hemoglobin level was 7.9%. He was taking the antidiabetic medications insulin glargine, glimepiride, miglitol, and ipragliflozin, and the antihyperlipidemic medications rosuvastatin and ethyl icosapentate.

A 46-year-old man with type 2 diabetes (patient 2) was also recruited (Table 1). He was 164.6 cm tall and weighed 86.0 kg (body mass index 31.7 kg/m2). He had received a diagnosis of type 2 diabetes 1 month previously. His non-fasting plasma TG concentration was 209 mg/dL, and his glycated hemoglobin level was 10.3%. He was taking the antidiabetic medication ipragliflozin and the antihyperlipidemic medication rosuvastatin.

The effects of pre- and post-meal Met/Ana administration on postprandial triglyceridemia were evaluated using a meal test. The TG, glucose, insulin, and active GLP-1 concentrations are shown in Figure 2. The TG concentration peaked 180 min postprandially when the drugs were administered after the meal (Figure 2A; patient 1: 155 mg/dL, patient 2: 163 mg/dL) and 240 min postprandially when they were administered before the meal (Figure 2A; patient 1: 154 mg/dL, patient 2: 147 mg/dL). Thus, the peak in TG concentration seemed to have been delayed by pre-meal administration.

The peak blood glucose concentration of patient 1 was higher following post-meal administration (Figure 2B; peak value 168 mg/dL) than pre-meal administration (Figure 2B; peak value 139 mg/dL). The blood glucose concentration of patient 2 peaked 120 min postprandially when the drugs were administered after the meal (Figure 2B; peak value 154 mg/dL) and 240 min postprandially when they were administered before the meal (Figure 2B; peak value 151 mg/dL). Thus, pre-meal administration seemed to inhibit or delay the peak in blood glucose concentration. There were no differences in the insulin concentrations of the patients between the 2 timings of administration (Figure 2C). In both patients, the active GLP-1 concentration remained higher following pre-meal administration (Figure 2D; peak value 40 pmol/L for patient 1 and 14.5 pmol/L for patient 2) than post-meal administration (Figure 2D; peak value 30.5 pmol/L for patient 1 and 7.9 pmol/L for patient 2). No adverse events, including the development of disease, were observed.

To assess the gastric peristaltic emptying time, we measured the exhaled 13CO2 concentration after the patients imbibed sodium 13C-acetate. The 13CO2 concentrations in exhaled breath during meal tests are shown in Figure 3. The time taken to reach the Tmax was longer following pre-meal administration (0.43 h for patient 1 and 1.07 h for patient 2) than following post-meal administration (0.33 h for patient 1 and 1.02 h for patient 2). The T½ was also longer following pre-meal administration (2.31 h for patient 1 and 1.85 h for patient 2) than following post-meal administration (1.80 h for patient 1 and 1.84 h for patient 2). The gastric emptying coefficients were 3.05 for patient 1 and 3.35 for patient 2 following post-meal administration, and 2.39 for patient 1 and 2.34 for patient 2 following pre-meal administration, implying that gastric peristaltic emptying was delayed following the pre-meal administration of the drugs in both patients.

Discussion

We made 2 significant findings in the 2 patients studied. First, pre-meal Met/Ana administration was associated with lower postprandial TG and glucose concentrations and a higher active GLP-1 concentration than was post-meal Met/Ana administration. Second, pre-meal Met/Ana administration was associated with a delay in gastric peristaltic emptying.

Although we studied only 2 patients, we found that pre-meal Met/Ana administration was associated with lower postprandial TG than was post-meal administration. These findings are consistent with those of our previous experiment performed in mice [5] and our previous clinical study [6] using metformin alone. However, unlike in the previous clinical study, we found that pre-meal Met/Ana administration seemed to increase the active GLP-1 concentration (Figure 2D). We speculate that the addition of a DPP-4 inhibitor to the metformin treatment may have contributed to this potential effect. Several previous studies have shown that GLP-1 analogs delay gastric emptying and significantly reduce the postprandial TG concentrations of patients with type 2 diabetes [10–13]. These findings imply that GLP-1 might at least in part reduce postprandial TG concentration by delaying gastric emptying. The identified difference in plasma GLP-1 concentration between pre-meal and post-meal Met/Ana administration may be associated with differential lowering of the postprandial TG and blood glucose concentrations. These results are consistent with those of a previous study conducted in Denmark, which showed the effect of metformin on postprandial GLP-1 secretion and glucose concentration in humans [14]. In a recent study, the glucose-lowering effect of metformin in patients with type 2 diabetes was shown to be larger when it is administered before, rather than with, enteral glucose, and this is associated with a larger GLP-1 response [15], which is in agreement with our finding that pre-meal Met/Ana administration seemed to increase active GLP-1 concentration. However, although a previous meta-analysis showed that metformin reduces the plasma TG concentration [16], another trial showed no effect of pre-meal Met on postprandial TG excursions in patients with metformin-naïve type 2 diabetes [17].

In the present study, we used non-fasting TGs in the inclusion criteria for 2 reasons. The first is that postprandial TG levels have been reported to be a stronger risk factor for cardiovascular disease than have fasting TG levels [18]. Although TG levels are typically obtained in the fasting state, non-fasting TG levels were associated with incident cardiovascular events in this cohort study. Second, in our study, TG levels used during the screening phase were derived from clinical laboratory tests performed either shortly after hospital admission or during routine outpatient visits. In many of these cases, fasting conditions could not be strictly ensured, and as such, non-fasting TG values were used. This reflects the practical constraints of real-world clinical data acquisition, in which standardized fasting protocols are often difficult to implement at the screening stage. However, we would like to emphasize that in the randomized controlled trial phase of the study, all relevant biochemical assessments, including TG levels, were conducted after several days of hospitalization, once glycemic control had been stabilized. These tests were performed under standardized overnight fasting conditions to ensure consistency and comparability of metabolic parameters.

The main limitation of this study is the small sample size. Originally, the study was designed to include 20 patients. However, due to the COVID-19 pandemic, including hospital restrictions and limited patient access, we were forced to terminate the study early after only 2 patients were enrolled. Despite the limited number, we considered the observed cases to be clinically valuable and therefore decided to present the findings as a descriptive case report.

Thus, we have shown an acute effect of metformin and anagliptin on the postprandial glucose and TG concentrations of patients with type 2 diabetes, although the small sample size means that the findings of the study should be interpreted with care.

Conclusions

Pre-meal administration of metformin and anagliptin reduced postprandial glucose and TG concentrations, potentially through a delay in gastric emptying. Thus, changing the timing of medication may improve their therapeutic effects.