Dr O. Cynshi, Chugai Pharmaceutical Co., Ltd, 1-1 Nihonbashi-Muromachi 2 Chome, Chuo-ku, Tokyo 103-8324, Japan. E-mail: firstname.lastname@example.org
Background Mitemcinal is an orally active motilin agonist that could potentially improve gastric emptying.
Aim To investigate the effect of mitemcinal on gastric emptying in patients with idiopathic and diabetic gastroparesis.
Methods In a randomized, double-blind design, 106 patients were randomized into four dosing regimens (22 to placebo and 21 each to mitemcinal 10 mg, 20 mg, 30 mg bid or 20 mg tid) for 28 days. A standardized scintigraphic gastric emptying test was performed at screening and again after completing the 4-week protocol.
Results All doses of mitemcinal showed prokinetic activity. A significant improvement in meal retention at 240 min was noted even in the lowest dose group with the greatest improvement observed with 30 mg bid group (75% vs. 10% in placebo group). Diabetic patients responded better than the idiopathic subgroup. In diabetic patients, blood glucose at 1 h after a meal showed dose-dependent elevation. Although gastroparetic symptoms improved with both mitemcinal and placebo, the prominent placebo effect was not statistically exceeded by mitemcinal. Baseline scintigraphy results exhibited no clear correlation between the severity of gastroparetic symptoms and the status of gastric emptying.
Conclusion Mitemcinal is capable of accelerating gastric emptying in both diabetic and idiopathic patients with gastroparesis.
Motilin is an endogenous peptide hormone of 22 amino acids, stimulating gastric emptying (GE) in response to sham feeding or gastric distension and involved in the induction of the migrating motor complex during the interdigestive period.1, 2 Intravenous infusion of motilin accelerated GE in healthy subjects3 and in patients with diabetic gastroparesis.4 Motilin is a peptide hormone that cannot be administered orally, and so pharmaceutical research has focused on the discovery of motilin receptor agonists that enable oral administration in patients with gastrointestinal (GI) diseases. Erythromycin, one of the macrolide antibiotics, is known as a motilin receptor agonist,5, 6 and studies have shown that erythromycin, administered either intravenously or orally, is capable of accelerating GE in patients with diabetic gastroparesis.7, 8 However, erythromycin has many weak points for clinical use; it is not acid-stable, its antibacterial properties can disturb the intestinal bacterial flora by promoting resistance in bacterial strains, and it has risk of inducing cardiac arrhythmias.9
Gastroparesis is a serious and well-recognized disorder characterized by symptoms related to inadequate or delayed emptying of the stomach in the absence of mechanical obstruction. It can occur acutely in response to surgery or medications, or chronically where the most common aetiologies are in the setting of diabetes mellitus or in an idiopathic form. Treatment for gastroparesis usually involves prokinetic agents that enhance and coordinate GI motility and transit of material in the GI tract (e.g. cisapride, domperidone, metoclopramide and erythromycin).10
To date, macrolide motilin agonist have not yet demonstrated clinical benefit for patients in appropriately controlled clinical trials. For example, the motilin agonist ABT-229 failed to show efficacy in relieving the symptoms of gastroparesis and functional dyspepsia in patients with or without delayed GE.11, 12 It has been hypothesized that a long half-life and an induction of desensitization for motilin receptors leading to dose tolerance could have contributed to the disappointing outcome in the clinical studies.13
Mitemcinal is a macrolide analogue currently in clinical development,14 which accelerates gastric motility in animals.15, 16 Recently, it has been reported to improve the symptoms in a subset of diabetic gastroparesis;17 however, its prokinetic effect in human remains unknown. In a double-blind study, we investigate if mitemcinal can accelerate GE in patients with delayed GE.
For all participating sites, Institutional Review Board (IRB) approval for the protocol and patient informed consent were obtained before the site began screening patients. Written informed consent was obtained from all patients before initiation of the study procedures described in the protocol. The study was conducted in accordance with the ethical principles that have their origins in the Declaration of Helsinki and with the standards of good clinical practice.
This randomized, double-blind, placebo-controlled, parallel-group study comparing four dosing regimens of mitemcinal was conducted in 21 sites in the US and two sites in Canada. Randomization was stratified by diagnosis for idiopathic gastroparesis and diabetic gastroparesis. Study drug or placebo was administered 30 min before breakfast, lunch and dinner for 28 days. A standardized scintigraphic GE protocol was performed at screening on day 28. Blood samples were taken at specified times on days 0, 14, 28 and 35 (follow-up).
Eligible patients were male or female, 18–70 years of age (inclusive) with idiopathic or diabetic gastroparesis; body weight <250 pounds (114 kg) and at least one of the following symptoms attributable to idiopathic or diabetic gastroparesis: anorexia, nausea, vomiting, early satiety, bloating or persistent fullness after eating. Patients without a previous diagnosis of gastroparesis must have had symptoms for ≥8 weeks prior to screening and patients with a previous diagnosis of gastroparesis must have had recurrence of symptoms occurring more than once and persisting for ≥5 days following discontinuation of current therapies. Delayed GE (retention ≥6% of isotope at 4 h) was documented on a standardized, morning scintigraphic examination; and there was no evidence of mechanical gastric obstruction.
Scintigraphic determination of GE was performed at screening and at the end of the dosing period (day 28). To standardize this measurement, a validated protocol was used at all sites.18 This protocol involved ingestion of a low fat (2%) meal (120 g eggbeaters labelled with 1 mCi of [99m-Tc]-sulphur colloid, two slices white bread, 28 g jelly, 120 mL bottled water) with anterior and posterior images obtained at 0, 60, 120 and 240 min. The key measurement was the percentage of radiolabel remaining in the stomach at 240 min after completion of the meal. All scintigraphy data were analysed by a blinded, central reading laboratory. Delayed GE was defined as retention of ≥6% of label at 4 h. If the patient vomited before the images at 4 h were taken (one patient in mitemcinal 20 mg tid group), then gastroparesis was defined as retention of ≥60% of label at 2 h.
A symptom questionnaire, adopted from the validated dyspepsia questionnaire19 at screening, was completed on days 0, 14, 28 and 35. The questionnaire referred to symptoms experienced in the previous week. For patients who had experienced any of the symptoms, a symptom composite score was determined using the following ratings and point system for nausea, vomiting, diminished appetite, bloating, early satiety and persistent fullness after eating if the patient experienced any of these symptoms:
Composite symptom score (CSS) = Severity* + frequency + duration.
* Severity was scored as a 5-point Likert scale.
Plasma samples were taken for pharmacokinetic analysis prior to the morning dose of study drug prior to mid-day dose on days 0, 14 and 28. For diabetics, haemoglobin A1c (HbA1c) was drawn on days 0 and 28 and at the follow-up visit.
Safety and tolerability were monitored throughout the trial. On days 0, 14, 28, and at follow-up, patients were asked about adverse events and received a physical examination including measurement of vital signs and weight, collection of routine clinical laboratories and a 12-lead ECG.
Efficacy analyses were performed on four populations: intent-to-treat (ITT), evaluable population (patients who were randomized, received at least one dose of study drug, had GE time measured at baseline and at day 28, and had taken at least 80% of the study medication), evaluable diabetic gastroparesis population, and evaluable idiopathic gastroparesis population.
The primary efficacy variables were absolute change and percentage change from baseline to day 28 in meal retention using scintigraphy measured at 240 min following meal intake. Percentage change from baseline in meal retention was to be compared among the five treatment groups using two-way analysis of variance (anova) for scintigraphy taken at 240 min. The statistical analyses for multiple pairwise comparisons were used with the Dunnett’s adjustment. Furthermore, a retrospective analysis was performed to determine the percentage of patients whose GE returned to normal levels (i.e. <6% at 240 min after a test meal) using the Cochran-Mantel-Haenszel test.
In the diabetic population, blood glucose and HbA1c were analysed at baseline and at day 28. The statistical significance of dose–response was confirmed retrospectively by the Jonckheere’s method and evaluated by the Williams’ method.
The study was conducted between 6 November 1998 and 10 February 2000 at 21 sites in the US and two sites in Canada. A total of 265 patients were screened for this study: 159 patients were screening failures in whom 103 patients showed non-delayed GE (retention <6% of isotope at 4 h), resulting in 106 patients who were randomized. The majority of patients randomized (98 of 106, 93%) completed on day 28, and the resulting evaluable population was 87% (90 of 104) of the ITT population (Table 1). Two patients (one receiving mitemcinal 20 mg bid and the other 20 mg tid) did not have gastroparesis.
Table 1. Summary of patient disposition
10 mg bid
20 mg bid
30 mg bid
20 mg tid
* One patient in this group did not have gastroparesis and was excluded from the intent-to-treat (ITT) population.
Number of patients randomized
Number (%) completed on day 28
Reason for discontinuation before day 28
Demographics and baseline characteristics
The patient population was adequately balanced across the four treatment groups with regard to demographic characteristics (Table 2). In the evaluable population, gastroparesis was attributed to diabetes in 59% of these patients and was idiopathic in 41%. The most troublesome symptoms were bloating and nausea. Across all treatment groups, 63–100% of patients had nausea, diminished appetite, bloating, early satiety and persistent fullness at baseline; 10–38% had vomiting. Symptoms at baseline reportedly occurred most often after dinner (range across treatment groups: 50–65%) and the most frequent severity grade of symptoms appeared as moderate. There were no statistically significant differences in the baseline symptoms among treatment groups.
Table 2. Summary of demographic and baseline characteristics in evaluable patients
Patient demographics and baseline characteristics
Placebo (n = 19)
10 mg bid (n = 16)
20 mg bid (n = 18)
30 mg bid (n = 20)
20 mg tid (n = 17)
* Composite scores for each symptom were calculated from the total points for severity, frequency and duration derived from the scoring system.
Sex, N (%)
Mean age in years (s.d.)
Mean weight in kg (s.d.)
Ethnic origin, n (%)
Aetiology of gastroparesis, n (%)
Duration of gastroparesis, mean (s.d.; years)
Duration of symptoms, mean (s.d.; years)
Most troublesome symptoms, n (%)
Mean composite symptom score* (s.d.)
Few patients had previous non-drug treatment for gastroparesis and approximately 60% had previously been diagnosed using scintigraphy. The most common prokinetic drugs used by patients, across all treatment groups and patient populations, were cisapride (40–100%), metoclopramide (17–63%) and erythromycin (14–57%).
A number of physiological factors, including the blood glucose level for diabetics and the stage of the menstrual cycle for females, may affect GE. The mean blood glucose prior to the GE test (GET) for self-monitoring patients and the mean number of days between the GET and day 1 of the menstrual cycle were similar across treatment groups and are not likely to have had an impact on the outcome of the trial. Screening meal retention assessed at 240 min was approximately 34% for the 10 mg bid group compared with 21–24% in the other treatment groups in this population. For this reason, primary analysis of meal retention was the percentage change rather than the absolute change from baseline.
Table 3 summarizes the mean percentage change in meal retention at 240 min for all patient populations. In the evaluable population, a potent prokinetic response was apparent. This response was significant even in the lowest dose (10 mg bid) and greatest in the highest dose (30 mg bid), which had percentage changes (mean ± S.E.) in improvement from baseline of 70 ± 11% and 75 ± 8%, respectively, compared with the placebo group, which had a mean improvement from baseline of 10 ± 21%. The diabetic patients showed a potent prokinetic response that was the greatest in the 20 mg bid group which had a mean improvement in percentage change in meal retention of 92 ± 2%, compared with the placebo group (18 ± 21%).
Table 3. Percentage change in meal retention at 240 min from screening to day 28
Mitemcinal treatment schedule
10 mg bid
20 mg bid
30 mg bid
20 mg tid
P-values for pairwise comparisons against placebo are based on Dunnett’s adjustment.
Mean change (s.d.)
Mean change (s.d.)
Mean change (s.d.)
Mean change (s.d.)
A retrospective analysis was performed to determine the percentage of patients whose GE returned to normal levels (i.e. <6% at 240 min after a meal) after treatment with mitemcinal or placebo for 28 days. The placebo group had the lowest percentage of patients (26%) whose GET returned to normal levels, while the mitemcinal treatment groups had substantially higher percentages of patients (69–85%) returning to normal levels, although the dose–response was almost a plateau (Figure 1).
Comparing the idiopathic with the diabetic patients, the magnitude of the prokinetic response was generally similar in the 10 and 30 mg bid treatment groups, while the effect appeared less potent in the idiopathic 20 mg bid and 20 mg tid groups (Table 3). This finding may reflect the less homogeneous nature of patients with idiopathic gastroparesis. Data for the placebo and both 20 mg groups showed high variability; however, the median percentage improvement showed results similar to the mean percentage improvement for diabetic patients. These data suggest that mitemcinal is capable of accelerating GE in both diabetic and idiopathic patients.
In the diabetic patients, glycaemic control was investigated after mitemcinal treatment for 28 days. Blood glucose at 1 h after a low fat (2%) meal showed dose-dependent elevation and was significantly elevated in the 30 mg bid group compared with the placebo group (Figure 2). This finding may indicate that the acceleration of GE led to an elevation of postprandial blood glucose. There was no apparent dose-related effect on HbA1c, which is not surprising given the short duration of the study.
Composite symptom scores (CSS) – the sum of severity, frequency and duration scores for each of the six symptoms – evaluated on baseline are shown in Table 2. The CSS (mean ± S.E.) overall was the lowest (31 ± 3) in the placebo group and the highest (39 ± 3) in the 20 mg tid group. The mean change from baseline to day 28 for the CSS overall and for each of the symptoms in evaluable patients showed a prominent placebo response that was not statistically exceeded by any of the mitemcinal groups (Table 4). The 20 mg bid group appeared different from the others, having the smallest response for all symptoms and the highest CSS at baseline.
Table 4. Change in composite symptom scores from screening to day 28 in evaluable patients
Placebo (n = 19)
10 mg bid (n = 15)
20 mg bid (n = 18)
30 mg bid (n = 20)
20 mg tid (n = 17)
All (n = 70)
Number (n) of patients represents evaluable patients at day 28 and does not reflect the total number of patients randomized to each treatment.
−3.2 (4.3) (n = 18)
−0.5 (2.9) (n = 17)
−2.3 (3.7) (n = 69)
There was no clear relationship between severity of symptoms and degree of gastric dysfunction for any of the screened patients. Regarding the correlation between changes in GE and CSS, some patients in the placebo group showed worsening in GE from baseline while others showed improvement. The range of values in this group probably reflects the natural variability of the disease; what appears to be a prominent placebo response may represent spontaneous improvement in the condition. In the mitemcinal treatment groups, GE worsened for only one patient in each group, except in the 20 mg bid group in which two patients worsened from baseline; all other patients improved in GE. In each of the 10 mg bid and 20 mg tid groups, one patient showed worsening CSS and all others experienced improvement from baseline; however, the magnitude of improvement in CSS did not appear to correlate with the magnitude of improvement in GE. In the 20 mg bid and 30 mg bid groups, although all but one or two patients improved in GE, five or six patients had worse CSS on day 28. Of the five patients treated with mitemcinal whose GE worsened from baseline, three nevertheless had improved CSS.
The estimated pharmacokinetic parameters are shown in Table 5. In comparison with Cmax on the first day of dosing, Cmax on day 14 was higher in the majority of patients with a mean ratio of Cmax (day 14/day 0) of approximately 2. This result was expected because of the predicted plasma kinetics. The mean ratio of the trough concentrations on days 14 and 28 was approximately 1, which would suggest that a mean steady-state had been achieved by day 14.
Table 5. Pharmacokinetic parameters of mitemcinal on days 0 and 14
AUC (ng × h/mL)
Drug plasma concentration was determined at days 1 and 14 of the 28-day mitemcinal treatment from the first 40 patients randomized.
Values are expressed as mean (s.d.).
10 mg bid (n = 9)
20 mg bid (n = 8)
30 mg bid (n = 11)
20 mg tid (n = 10)
Safety and tolerability
Adverse events that were considered severe were experienced in 17% of patients treated with mitemcinal and 27% of patients treated with placebo. Severe events reported for more than one patient treated with mitemcinal were diarrhoea (three patients), dyspepsia (two patients) and hypoglycaemia (two patients). There was no apparent increase in severity of adverse events associated with administration of mitemcinal. The percentages of patients with one or more adverse event were 55%, 76%, 71%, 48%, and 91% in placebo, 10 mg bid, 20 mg bid, 30 mg bid, and 20 mg tid groups, respectively. There was no clear dose–response in the number of events; however, the highest number of events was in the 20 mg tid group. The adverse event with the highest incidence across all active treatment groups was hypoglycaemia (21%) in comparison with the placebo (9%), though there was no clear dose–response in the incidences.
Concerning the laboratory results other than blood glucose, the only apparently dose-related effect of mitemcinal was on hepatic enzymes. ALT, AST, alkaline phosphatase and GGT were slightly increased in all active treatment groups compared with placebo. Although both 20 mg tid and 30 mg bid groups received 60 mg of mitemcinal per day, the tid regimen appeared to have a more pronounced effect on these parameters. There was no evidence of any meaningful effect of mitemcinal on laboratory results, vital signs, physical examination findings or body weight. Additionally, this study with mitemcinal did not show any evidence of ECG abnormality.
The primary efficacy parameter in this study was GE time with scintigraphic determination. Changes in percentage retention following meal intake from baseline to the 28-day evaluation were prominent in all dose groups after the treatment of mitemcinal for 28 days. These results indicated that mitemcinal is a promising orally active prokinetic drug that accelerates GE. Although a well-known motilin receptor agonist, erythromycin, has been reported to accelerate GE in patients with gastroparesis,20 the usage of erythromycin is restricted in patients because of its antibacterial action and its potential cardiac concerns. Another motilin receptor agonist, ABT-229, had been developed as an orally active drug to overcome the shortcomings of erythromycin and reported to accelerate GE with a single dose in humans.21 However, ABT-229 failed to demonstrate acceleration of GE after a 7-day treatment.22 Furthermore, it failed to demonstrate improved gastroparetic symptoms in diabetic patients.11 The reason for these failures was considered to be tachyphylaxis induced by ABT-229.23 In fact, Thielemans et al., from a comparison between erythromycin and ABT-229 on the internalization of motilin receptors, unveiled a difference in the desensitization of motilin receptors leading to tachyphylaxis.24 The pharmacokinetic analysis in this study indicated a steady exposure of mitemcinal up to 28 days and our finding for GE suggested that, unlike ABT-229, mitemcinal was effective from repeated usage in accelerating GE without tachyphylaxis. The results of the dosing regimens of mitemcinal applied in this study showed no suggestion of an advantage in tid over bid dosing, although long-term use of 60 mg of mitemcinal per day might induce a more pronounced effect on hepatic enzymes.
In idiopathic patients, the dose effect is surprising with a lower efficacy of 20 mg vs. 10 or 30 mg as shown in Table 3. The reason was that the incidences of worsened GE after treatment were unbalanced among the groups. In fact, the number of the patients with worsened GE was unbalanced as 2/8, 0/6, 2/7, 0/10 and 1/7 for placebo, mitemcinal 10 mg bid, 20 mg bid, 30 mg bid and 20 mg tid, respectively. Nevertheless, the normalization rate of GE was dose-dependent similarly to Figure 1 as 38%, 67%, 71%, 90% and 57% for placebo, mitemcinal 10 mg bid, 20 mg bid, 30 mg bid and 20 mg tid, respectively, despite a small number of idiopathic patients. A larger number for the future clinical study would be required to clarify the dose–response in the patients with idiopathic gastroparesis.
In diabetic patients, erythromycin was reported to reduce postprandial blood glucose by stimulating insulin secretion.25 On the other hand, the acceleration of GE may reasonably make the absorption of postprandial glucose earlier, thus increasing postprandial blood glucose. This phenomenon was reported with a erythromycin derivative, EM523 L.26 Both studies were investigated with a small number of diabetic patients and the results were controversial; therefore it remained unclear which was the case of mitemcinal. As illustrated in Figure 2, mitemcinal increased postprandial blood glucose in a dose-dependent manner. Together with the observation of the increased hypoglycaemic events, mitemcinal might cause to increase postprandial blood glucose and decrease fasting blood glucose if patient used insulin with the same amount and at the same timing as before nevertheless the GE was accelerated. In this study, there was no evidence or the expectation of the stimulation of insulin secretion from repeated administration. To avoid hypoglycaemic events in diabetic patients administered mitemcinal in future, an adjustment in insulin usage as a measure to counterbalance gastric motility should be the recommended.
Secondary efficacy parameters were symptoms as evaluated by questionnaire. Symptom data were analysed by CSS and individually by symptom. There was no clear relationship between gastroparetic symptoms scored according to CSS and degree of gastric dysfunction for patients at the baseline and also no clear relationship between improvement in CSS and change in GE time. Horowitz et al. have also reported that the relationship between gastric symptoms and GE is inconsistent.27, 28 Although we have already reported the symptomatic improvement in diabetic gastroparesis,17 there was no correlation between gastric symptoms and GE in that study. Our finding in this study provides more evidence showing that there is no clear relationship between gastric symptoms and GE and therefore the acceleration of GE does not guarantee a beneficial effect for the symptoms. The most frequent CSS for all individual symptoms was zero = ‘none’ and the average CSS at background was the lowest in the placebo group; thus the prominent placebo response in the symptoms might be related to the background. For future study to add to our previous one,17 various factors such as heterogeneity of the patient population, small sample size, short duration of the study and the prominent placebo effect should be considered in trying to explain and demonstrate the improvement of gastroparesis symptoms.
In conclusion, our study found that all doses of mitemcinal showed good prokinetic activity and that even the lowest dose (10 mg bid) exhibited significant improvement in GE. In addition, mitemcinal accelerated GE in both diabetic and idiopathic patients. There was no attenuation in the plasma mitemcinal concentration after repeated administration for 28 days and mitemcinal exerted a clear acceleration of GE after repeated exposure.
We thank Ms Ford Frances for her excellent editorial assistance.
Declaration of personal interests: R. W. McCallum has received research grant support from Chugai Pharmaceutical Co., Ltd, Novartis Pharmaceuticals Corp., Salix Pharmaceuticals, Medtronic, Inc., Dynogen Pharmaceuticals, Inc. and The SmartPill Corporation. Dr McCallum has also served on the advisory board for The SmartPill Corporation, as a consultant for Takeda, Chugai, Novartis and Medtronic, and received honoraria as a speaker from Takeda Pharmaceuticals, Novartis and TAP Pharmaceuticals. Declaration of funding interests: Funding for this research was provided by Chugai Pharmaceutical Co., Ltd. O. Cynshi is an employee of Chugai.
The following investigators participated in the study:
S. J. Gordon, Allegheny University Hospitals, PA; M. S. Kipnes, Diabetes and Glandular Disease Clinic, TX; D. Einhorn, Diabetes and Endocrine Associates, CA; C. Clinkingbeard, Pacific Coast Clinical Coordinators, ID; M. M. Schuster, the Johns Hopkins Bayview Medical Center, MD; G. Tougas, McMaster University Medical Centre, Canada; R. S. Patel, Medical University of South Carolina, SC; P. B. Miner Jr, Oklahoma Foundation for Digestive Research, OK; B. M. Miskin, Palm Beach Research Center, FL; J. G. Moore, Salt Lake City VA Medical Center, UT; S. R. Weiss, San Diego Endocrine and Medical Group, CA; S. A. Brietzke, Uniformed Services University of the Health Sciences, MD; J. K. DiBaise, University of Nebraska Medical Center, NE; D. J. Patterson, Virginia Mason Medical Center, WA; D. S. Riff, Associated Gastroenterology Medical Group, CA; D. J. Pambianco, VA; C. F. Barish, Wake Research Associates, NC; G. N. Verne, University of Florida, FL; C. A. Sninsky, Gainesville VA Medical Center, FL; R. Hardi, Metropolitan Gastroenterology Group, MD; M. A. Safdi, Consultants for Clinical Research, OH; R. E. Tepper, Long Island Clinical Research Associates, NY; T. L. Abell, University of Tennessee Medical Center, TN; W. J. Snape Jr, Long Beach Gastroenterology Associates, CA; S. P. Marcuard, Carolina Physicians, NC; T. V. Nowak, St Vincent Hospitals and Health Services, IN; C. M. Schmitt, Southeastern Clinical Research, TN; J. A. Pino, Alabama Research Center, AL; P. Poitras, Centre de Recherche du CHUM, Canada; M. Olyaee & S. Sarosuh, University of Kansas, KS.