Gastroparesis in adults: oral erythromycin

Evidence review: efficacy

One systematic review was identified (Maganti et al. 2003) that included clinical studies evaluating the efficacy of oral erythromycin for the symptomatic relief of gastroparesis. Only studies including assessment of symptoms as an end point were included. Symptom improvement was defined as a 25% or greater decrease in total symptom score. A total of 35 "clinical trials" evaluating erythromycin in gastroparesis were identified, 5 of which were stated as including assessment of symptoms as an end point, although none had symptom improvement as the primary end point. These 5 small studies were reported individually; meta-analysis was not undertaken and it is not reported whether the authors considered this.

The study by Samsom et al. (1997) was a double-blind, crossover study in 12 people with type 1 diabetes and dyspepsia. Participants were assigned to 2 weeks' oral erythromycin stearate (250 mg, 3 times daily) or placebo followed by a 1-week washout then crossover to 2 weeks of the alternative. Symptoms (early satiety, fullness, abdominal pain, bloating, nausea and vomiting) were recorded daily by participants in a symptom diary using a 4-point Likert scale. There was no statistically significant difference in mean symptom scores between the erythromycin and placebo groups (1.53±0.67 compared with 1.81±0.86; p=0.16). Individual symptom scores were not presented.

Ambulatory antroduodenal manometry was performed at the end of each 2-week treatment period. On fasting assessments, erythromycin statistically significantly decreased the duration of gastric motor activity (86.2±25.3 compared with 118.9±46.0 minutes; p=0.03), decreased phase II periods (48.5±19.4 compared with 68.7±23.5 minutes; p<0.05), and increased phase III periods (6.3±1.7 compared with 5.2±1.4 minutes; p<0.05). There was no statistically significant increase in the number of phase III events. Postprandially, erythromycin statistically significantly increased the number of antral contractions (p<0.01) and antral motility index (p<0.03).

No adverse effects were observed in the study.

The study by Erbas et al. (1993) was a single-blind crossover study comparing oral erythromycin (250 mg, 3 times daily) with metoclopramide (10 mg, 3 times daily) in 13 people with diabetes and gastroparesis. Treatment was for 3 weeks followed by a 3-week washout then crossover to the alternative. Symptoms (gastric retention, early satiety, abdominal pain, bloating, nausea and vomiting, constipation, diarrhoea and anorexia) were assessed on a 4-point Likert scale. Median baseline symptom score was 8 (range 2 to 11) out of a possible total of 24 points, with a higher score indicating more severe symptoms. Symptom scores improved in both treatment groups, but erythromycin improved symptoms statistically significantly more from baseline than metoclopramide (median symptom score 2 points [range, 0−5] compared with 3 points [0−11] respectively; p<0.05).

The half-time of gastric emptying (T_1/2) was assessed at baseline and after each treatment using a radiolabeled meal. Both treatments significantly improved gastric emptying. T_1/2 at baseline was a median 110 minutes (77−120) which reduced to 55 minutes (28−115) after 3 weeks of erythromycin (50% decrease), and to 67 minutes (5−115) after 3 weeks of metoclopramide (39% decrease). When compared with baseline values, the authors report that significant differences in gastric emptying parameters were found after both erythromycin and metoclopramide treatment compared with baseline, but there was no difference between the groups. No statistical values were reported in the original research paper.

No adverse effects were reported after erythromycin, although with metoclopramide 2 people experienced weakness, sedation and leg cramps, 1 had palpitations and 1 other had drowsiness.

The study by Richards et al. (1993) was open-label and uncontrolled evaluating 4 weeks' oral erythromycin (500 mg, 4 times daily) in 10 people with idiopathic gastroparesis and 4 people with diabetic gastroparesis. Oral therapy followed an initial dose of intravenous erythromycin. After completion of the 4-week treatment period, participants were able to enter a long-term treatment study.

Only 10 of the 14 participants completed the study; 4 withdrew (2 because of rash, 1 with cramps and vomiting, and 1 because of other medical problems). Analysis was per protocol. Symptoms (early satiety, abdominal pain, bloating, nausea, vomiting, heartburn and anorexia) and global wellbeing were assessed at baseline and at the end of the treatment period (plus every 8 weeks for those who continued oral erythromycin in the long-term study) using a 10-point Likert scale with higher scores indicating more severe symptoms. Erythromycin was associated with a statistically significant improvement in the global assessment score (mean score decrease from 7.2±2.1 to 4.7±2.9; p=0.03) but did not statistically significantly improve symptom score (mean score decrease from 30±16 to 24±13; p=0.2). The systematic review reports that 2 people experienced a 25% or greater reduction in their symptoms, a result that was not identified in the original research paper. Gastric emptying was assessed at baseline, after the initial dose of intravenous erythromycin and after 4 weeks of oral treatment using radioisotope imaging. The percentage of radiolabeled meal retained in the stomach at 2 hours decreased from 85±11% at baseline to 48±21% after 4 weeks (p<0.01). Of the 10 people who completed the study and continued with oral erythromycin, 5 experienced sufficient symptom relief to continue in the longer term (average 8.4 months). No toxicity was experienced by these 5 people who continued erythromycin. The other 5 people stopped using erythromycin because of no or inadequate improvement in symptoms. No attempt was made to apply statistical analysis to the longer-term symptom scores, although the authors report that in general scores did not change much from those recorded at the end of the 4-week trial.

The study by Ramirez et al. (1994) was open-label and uncontrolled evaluating 2 weeks' erythromycin ethylsuccinate oral suspension (125 to 250 mg, 3 times daily, depending on the optimal tolerated dose) in 9 people with gastroparesis after truncal vagotomy and antrectomy for refractory peptic ulcer disease. Oral therapy followed an initial trial with a single dose of intravenous and/or oral erythromycin. The reported results are for the 2-week oral trial. Symptoms (postprandial fullness, abdominal pain, nausea and vomiting) were assessed on a 4-point Likert scale (maximum score of 12) at baseline and after treatment. Erythromycin did not significantly improve symptoms in the group overall (mean symptom score 9.2±0.5 at baseline compared with 7.8±0.6 after treatment; p value not reported). In the 3 people who did show symptom improvement, their score decreased from a mean 11.0±0.3 at baseline to 5.7±0.3 after treatment (p<0.01). Individual symptom scores were not reported.

There were reported to have been no adverse effects during the study.

The study by Fiorucci et al. (1994) was reported to be an open-label uncontrolled study evaluating 4 weeks' oral erythromycin (250 mg, 3 times daily) in 12 people with systemic sclerosis and gastroparesis. The publication referenced by Maganti et al. (2003) evaluated only a single intravenous erythromycin infusion and reported only the effect of gastric and gallbladder activity and gastric motility. None of the results in the referenced publication are consistent with those reported in the systematic review; it is unknown whether a 4-week oral study followed the intravenous trial and where the symptomatic results in the report of this study in the systematic review were published.

In addition to the systematic review, a small retrospective chart review (Dhir and Richter 2004) was identified that included 25 people (mean age 49±18 years, range 19 to 86 years) with impaired gastric emptying and dyspepsia symptoms. All participants received erythromycin oral suspension 50-100 mg before each meal and at bedtime. Symptoms at baseline and at short-term follow-up (6−8 weeks) were reported retrospectively from patient charts. Symptoms on long-term follow-up were evaluated through telephone interview. Symptoms were assessed as worse, unchanged, some improvement (<50% of baseline) and dramatic improvement (>50% of baseline) at both short- and long-term follow-up. Of the 18 people who were followed up in the short term, 15 (83%) experienced symptom improvement with erythromycin, of whom 12 (66%) reported dramatic improvement. Three people (17%) reported no improvement or worsening of symptoms and differences reported were all statistically significant (p=0.005). At the telephone follow-up, 12 of 18 people (67%) had experienced symptom improvement during long-term (11±7 months, range 1 to 29 months) erythromycin use compared with 6 (33%) who experienced no improvement or worsening of symptoms (p=0.16). Only 3 people (17%) were still using erythromycin at the time of the telephone follow-up. These 3 people reported complete symptom resolution. The most common reason for stopping erythromycin was failure to completely relieve symptoms.

During short-term follow-up, 1 person stopped erythromycin because of severe nausea and vomiting. In the longer term, 5 people reported adverse effects of erythromycin including headaches, nausea, abdominal cramps and loose bowel movements.