Improvement of Gastric Emptying in Diabetic Gastroparesis by Erythromycin — Preliminary Studies

We studied 10 patients with diabetes mellitus (3 men and 7 women whose mean age was 51 years [range, 30 to 75]) and 10 healthy subjects matched for age and sex. All gave informed consent to participate in the study, which had been approved by the University Hospital's ethics committee. The patients had had insulin-dependent diabetes for 5 to 41 years (mean, 24.4). All had normal serum creatinine concentrations, and their mean hemoglobin A_1c value was 8.0 percent (range, 5.3 to 11.6; normal range, 3.6 to 6.4). From previous scintigraphic studies, all the patients were known to have prolonged gastric emptying for liquids and even longer emptying for solids, due to diabetic gastroparesis. Four patients had symptoms of gastric obstruction; eight patients had evidence of peripheral neuropathy, and nine of cardiac autonomie neuropathy.

The effect on gastric emptying of erythromycin (200 mg of erythromycin lactobionate, Abbott) and placebo was tested in random order in a double-blind, crossover study on two separate days, with a one-day interval between the tests, after an overnight fast of at least eight hours. The healthy subjects received neither drug nor placebo. The patients' blood glucose concentrations were maintained between 5.5 and 8.3 mmol per liter by combined infusions of insulin and glucose during the fast and the subsequent study period. The simultaneous gastric emptying of liquids and solids was determined scintigraphically with a double-isotope technique. The technique used a standardized meal consisting of one scrambled egg labeled with 500 μCi of [^99mTc]sulfur colloid, two slices of bread, and 150 ml of water containing 100 μCi of [¹¹¹In]diethylenetriamine penta-acetic acid. The weight of the solids was 110 g, and they contained 0.966 MJ (231 kcal), consisting of 35 percent fat, 47 percent carbohydrate, and 18 percent protein. The meals were eaten in a mean (±SE) period of 8±2 minutes.

Immediately after finishing the meal, the patients and healthy subjects were seated between the two heads of a dual-headed gamma camera fitted with medium-energy collimators. In the patients, 200 mg of erythromycin (or placebo) was infused intravenously over a 15-minute period after the meal. Simultaneous anterior and posterior 1-minute images were obtained in both the technetium-99m and indium-111 energy windows every 10 minutes for one hour and then every 15 minutes for another hour. The regions of interest were marked by hand around the stomach on each anterior and posterior image. The results were corrected with use of the geometric mean for the physical decay of the technetium, the downscatter of the indium into the technetium window, and the changes in depth as the meal moved from the fundus to the antrum. Results were expressed as the percentages of solids and liquids remaining in the stomach over time after the completion of the meal.

All 10 patients were subsequently treated for four weeks with erythromycin (250 mg of erythromycin ethylsuccinate [Abbott] orally three times a day, 30 minutes before meals). We studied the effect on gastric emptying on the morning of the last day of treatment using the same scintigraphic technique, with the double-isotope test meal eaten 30 minutes after the oral administration of 250 mg of erythromycin. The results were analyzed with Student's t-test for paired or unpaired data. P values lower than 0.05 were considered significant.

Figure 1. Figure 1. Speed of Gastric Emptying in Patients Taking Placebo or Erythromycin and Healthy Subjects.

Panel A shows the rate of emptying of the solid part of the test meal, expressed as the mean (±SE) percentage of isotope remaining in the stomach at various times after the ingestion of the meal, in 10 patients with diabetes after the intravenous administration of placebo (◇—◇) or 200 mg of erythromycin (◇—◇) and in 10 healthy subjects (◇----◇). Panel B shows the rate of emptying of the liquid part of the test meal, expressed as the percentage of isotope remaining in 10 patients with diabetes after the administration of placebo (○—○) or erythromycin (•—•) and in 10 healthy subjects (○----○).

Table 1. Table 1. Mean (±SE) Percentage of Simultaneously Ingested Solid and Liquid Food Retained in the Stomach 60 and 120 Minutes after the Completion of the Meal in Patients with Diabetes and Healthy Subjects.

Figure 1A shows the percentage of the solid part of the meal that remained in the stomach over time in the 10 patients with diabetes after the intravenous administration of placebo and erythromycin (200 mg) and in the 10 healthy subjects matched for age and sex. The mean values after 60 and 120 minutes are shown in Table 1. Erythromycin markedly accelerated the extremely slow gastric emptying of solids in the patients with diabetic gastroparesis; after the infusion of erythromycin, the emptying of solids was more rapid in the patients than in the healthy subjects, and the speed approximated that of the emptying of liquids in the healthy subjects.

Figure 1B shows the percentage of the liquid part of the meal retained in the stomach over time in the 10 patients with diabetes after the intravenous administration of placebo and erythromycin and in the 10 healthy subjects. The mean values 60 and 120 minutes after the meal are shown in Table 1. Erythromycin accelerated the severely impaired emptying of liquids in the patients with diabetes to normal speed.

The effect of four weeks of treatment with oral erythromycin on gastric emptying in the patients with diabetes is also shown in Table 1. The emptying of both solids and liquids was accelerated (but less than after a single intravenous dose), and the rates of emptying for solids and liquids remained different.

There were no side effects of erythromycin administration or changes in the results of liver-function or other laboratory tests during the four-week treatment period in the patients with diabetes. Among the four patients who reported symptoms of gastric obstruction, one noted improvement and two others, who had tolerated only tube feedings, were able to resume oral feeding. Three patients reported fewer hypoglycémie attacks. The mean hemoglobin A_lc value in the 10 patients at the end of the four-week period was 7.6 percent (range, 5.1 to 10.0). The dose of insulin did not change appreciably.

We found that a dose of 200 mg of erythromycin administered intravenously during the immediate postprandial period markedly improved the severely impaired gastric emptying of both liquids and solids in the 10 patients with diabetic gastroparesis. The drug was also effective when given orally for four weeks. This observation period, however, was too short to allow valid conclusions about the effect of the drug on long-term control of diabetes. Interestingly, after the intravenous administration of erythromycin, the well-known difference in the emptying rate of liquids and solids was no longer discernible: the rate for solids was as rapid as that for liquids and was similar to the rate for liquids in healthy subjects. The mechanism by which erythromycin induces this strong gastrokinetic action is probably related to its motilin-agonistic properties.

Itoh and collaborators found that 1 to 3 mg of erythromycin per kilogram of body weight per hour, administered intravenously over a 15-minute period during fasting, induced rhythmic phase 3 contractile activity of the migrating motor complex in dogs¹ and humans.² The erythromycin-induced phase 3 activity always started in the stomach and migrated through the small intestine, with the same characteristics as spontaneous phase 3 activity. Zara et al. have confirmed these results.^{10 , 11} Erythromycin thus mimics the effect of exogenous motilin, which also induces phase 3 activity that starts in the stomach.¹² In contrast, somatostatin¹³ and met-enkephalin¹⁴ induce migrating motor complexes at the level of the small intestine. Pancreatic-polypeptide infusions lower plasma motilin levels and abolish only the gastric component of phase 3 activity, which can be reinduced by adding motilin to the pancreatic-polypeptide infusion.¹⁵ Moreover, Bormans et al. found that in humans only phase 3 activity that starts in the stomach is accompanied by a rise in plasma motilin levels.¹⁶ It has therefore been suggested that motilin is involved in the generation of migrating motor complexes at the gastroduodenal level. This concept is in agreement with the finding that motilin has no effect on an autotransplanted jejunal loop¹⁷; with the results of in vitro experiments showing a decreased responsiveness of the lower gut to motilin¹⁸; and with the results of Peeters et al.,⁶ who found that motilin receptors are most dense in the gastric antrum and upper duodenum and that the number of receptors distal to Treitz's ligament is very limited.

Several studies suggest that erythromycin may act as a motilin agonist through a direct action on motilin receptors. Erythromycin stimulates the contraction of rabbit duodenal tissue in vitro, an action not blocked by tetrodotoxin or atropine.¹⁹ Erythromycin and related 14-member macrolide compounds have the same regional and species specificity as motilin; they act on rabbit and human duodenum, but not canine duodenum or rabbit or guinea pig ileum.²⁰ They also inhibit the binding of labeled motilin to rabbit antral-smooth-muscle homogenates,^{3 , 4} and their ability to inhibit motilin binding correlates with their potency in inducing contractions.²⁰

The main action of motilin seems to be confined to the interdigestive period, and whether it has actions during the postprandial period is uncertain. A physiologic role for motilin in the regulation of gastric emptying is unlikely, since plasma motilin levels tend to decrease after meals.^{21 , 22} These data, however, do not exclude the possibility that large doses of motilin have a pharmacologic effect on gastric emptying. Such an effect may be more pronounced in cases of antral hypomotility, as in patients with diabetic gastroparesis, but this has not been tested because of the high cost of such experiments. If erythromycin and related compounds act as motilin agonists, they may constitute a new group of gastrokinetic agents with possible clinical applications. Our study of the effect of erythromycin on gastric emptying in patients with severe diabetic gastroparesis seems to confirm the drug's strong gastrokinetic effect. The dose we used (200 mg) was about 3 to 10 times higher than that used by Itoh et al. to induce phase 3 contractile activity during fasting.² Preliminary radiocinematographic observations in our laboratory show that erythromycin induces very powerful, lumen-obliterating peristaltic contractions in the gastric antrum of these patients (as in healthy subjects), which rapidly empty the stomach of all contrast material. The lumen-obliterating character of these peristaltic contractions may explain the similar rates of emptying for liquids and solids seen in our patients after the administration of erythromycin. Whether the drug increases the diameter of the pylorus during an antral contraction, thus allowing the emptying of larger food particles, remains to be examined.

Whatever the mechanism involved, erythromycin markedly improves impaired gastric emptying in patients with severe diabetic gastroparesis, and the drug is active when given orally for four weeks. This demonstration of a motility-stimulating effect of erythromycin in humans indicates that long-term, placebo-controlled studies of the oral administration of erythromycin should be undertaken to determine its therapeutic efficacy in patients with disordered gastric motility.