Erythromycin Enhances Fasting and Postprandial Proximal Gastric Tone in Humans

Transcription

1 GASTRONTROLOGY 1995;109:32-39 rythromycin nhances Fasting and Postprandial Proximal Gastric Tone in Humans STANISLAS BRULY DS VARANNS,* VleRONIQU PARYS,* ALAIN ROPRT,* JAN ALAIN CHAYVIALL, t CLAUD ROZle, g and JAN PAUL GALMICH* *Laboratoire Fonctions Digestives et Nutrition, Centre de Recherche sur Volontaires, Nantes; ~INSRM Unit6 45, HSpital. Herriot, Lyon; and ~INSRM Uni% 410, Facult6 X Bichat, Paris, France Background & Aims: Low doses of erythromycin induce antral contractions and accelerate gastric emptying. However, the effect of erythromycin on the proximal stomach remains unknown. The aim of this study was to assess the effect and mechanism(s) of action of erythromycin on proximal gastric tone in humans. Methods: Gastric tone was measured using an electronic barostat in two groups of 6 subjects both in the fasting state and after a 200-kcal meal. On different occasions, subjects received saline, atropine alone (6 l~g- kg 1. h-1 for 30 minutes), erythromycin alone (1.5 mg/kg in the fasting state and 1.5 and 3.0 mg/kg in the postprandial state), and erythromycin plus atropine. Results: Low-dose (1.5 mg/kg) erythromycin enhanced fasting gastric tone, but only the 3.0-mg/kg dose reduced the duration of meal-induced relaxation (37 _+ 14 vs minutes; P < 0.01). Atropine did not change the fasting or postprandial gastric tone as well as the erythromycin-induced responses. Plasma motilin levels were unaffected by erythromycin infusion. No correlation was observed between gastric tone and plasma motilin or erythromycin levels. Conclusions: rythromycin enhances fasting and postprandial proximal gastric tone in humans by a mechanism that does not seem to involve endogenous motilin release or a cholinergic pathway. rythromycin, a macrolide antibiotic, can induce gastrointestinal side effects such as nausea, vomiting, epigastric pain, and diarrhea. Such effects have been re- 1 lated to several gastric and intestinal motor patterns. However, at low, microbiologically ineffective doses, erythromycin strongly accelerates gastric emptying in patients with diabetic gastroparesis 2 and shortens the lag phase in healthy subjects 3 (for review, see Peeters4). In humans, the effect of low doses of erythromycin on antroduodenal motor activity has been well documented by manometric studies. Depending on the doses, erythromycin administered in fasting subjects induces either true gastric phase IIIs or phase III-like activities that do not migrate toward the small intestine and are not followed by phase I but by a prolonged period of antral contractile activity) '~ After a meal, erythromycin increases the amplitude of antral contractions, improves antroduodenal coordination, and induces an earlier reappearance of fasting motility/ However, the effect of erythromycin on proximal gastric tone remains unknown. The use of an electronic barostat to record proximal gastric tone in both the fasting and postprandial state allowed us to gain new insights on the physiology and pathophysiology of the proximal stomach. 8 We therefore applied such a technique as previously described 9'1 to study the effects of erythromycin on the human proximal stomach. The mechanisms by which erythromycin acts on gut motility remain incompletely understood. In vitro studies showed erythromycin to act as a motilin receptor agonist, 1I'12 whereas some in vivo studies suggest an indirect action via a cholinergic pathway Whether erythromycin acts via endogenous motilin release or not is still a source of controversy. Indeed, some studies point out an erythromycin-induced motilin increase, 14'16 whereas most of them showed no effect. 1'5-7 The purpose of this study was (1) to investigate the effects of low doses of erythromycin on proximal gastric tone both in the fasting and postprandial states in healthy humans and (2) to evaluate the role of cholinergic and/ or endogenous motilin release on erythromycin gastric motor response. Materials Subjects and Methods ighteen healthy volunteers without gastrointestinal symptoms were included in this study: 6 in a preliminary study (see below) and 12 in the main study. All gave written informed consent to the protocol, which had been previously approved by the local Human thical Committee. In the main study, to completely assess the effects on gastric tone of erythromycin alone as well as atropine alone and atropine combined with erythromycin both in the fasting state and post- Abbreviations used in this paper: AUC, area under the curve by the American Gastroenterological Association /95/$3.00

2 July 1995 RYTHROMYCIN AND PROXIMAL GASTRIC TON 33 I :=:,' I I IPostprandiall I studies [ [ Meal... quilibration 0 I I I placebo-placebo I M (1.5 mg/kg)-placebo Group 1 M (1.5 mg/kg)-atropine placebo-placebo placebo-atropine Group 2 I quilibration" "01 I I 310 Time (rain) ~t ~t ~t ~t ~t ~t Blood drawn placebo-placebo I M (3.0 mg/kg)-placebo Group 1 M (3,0 mg/kg)-atropine. placebo-placebo I placebo-atropine M (1,5 mg/kg)-placebo Group 2 M (1.5 mg/kg)-atropine V I I I I Time (rain) Figure 1. Schematic representation ofthe procedure design. Proximal gastric tone was measured using an electronic barostat in 2 groups of 6 healthy subjects in fasting conditions and following a balanced liquid meal. rythromycin lactobionate (M) or its placebo was intravenously infused during a lo-minute period (shaded area), whereas atropine or its placebo was infused during a 30-minute period. prandial conditions, a total of 12 sets of experiments was necessary (Figure 1). Because of the large number of experiments to be performed, our population of 12 healthy volunteers was divided into 2 groups of 6 subjects each. Therefore, every subject in each group was investigated on 6 separate occasions. Group 1 (3 fasting and 3 postprandial assessments) included 5 men and 1 woman ranging in age from 21 to 26 years (mean _+ SD, years). Group 2 (2 fasting and 4 postprandial assessments) included 3 men and 3 women ranging in age from 19 to 26 years (mean + SD, years). Proximal Gastric Tone Recording Gastric tone was measured using an electronic barostat according to a previously published technique. 8,0 Highly compliant polyethylene bags with a maximal volume capacity of ml were used. The bag was connected to the pressure transducer and the injection pump by a 16F-diameter single-lumen polyvinyl tube (i.e., the same lumen was connected both to the barostat strain gauge to detect changes in intragastric pressure and to the air injection system). Bag volume was continuously recorded on a potentiometric recorder running at 10 mm/min (Servotrace; IP Sefram, Paris, France). Meal The 200-mL liquid meal had a caloric content of 200 kcal. We previously showed that this meal induces profound, long-lasting relaxation of proximal stomach.i Meal composition (per 1000 ml) consisted of 650 ml of partially skimmed milk (fat, 15.5%), 60 g of liquid protein solution (Alburone; Sopharga, Puteaux, France), 150 g of dextrin maltose (Nutripharm lgi, Levallois, France), and 250 m of water. Total energy content was 1 kcal/ml, provided as 60% carbohydrate, 20% fat, and 20% protein. Meal osmolarity was 270 mosm/l. Pilot Study With Increasing rythromycin Doses To determine the lowest dose of intravenous erythromycin inducing a consistent increase of fasting proximal gastric tone, a preliminary study was performed in 6 healthy subjects ranging in age from 19 to 39 years (mean _+ SD, years), iv Gastric tone was assessed by an electronic barostat on the same occasion according to four different doses (0.2, 0.37, 0.75, and 1.5 mg/kg) of intravenous erythromycin infused in random order in the fasting state. The first dose was infused 20 minutes after a spontaneously occurring phase III of the migrating motor complex recorded by a manometric probe situated in the antroduodenal area. Only and 1.5-mg/ kg doses induced a significant decrease of barostat bag volume (P < 0.05) in all subjects. These responses were poorly dose related. Because the responses to erythromycin were more homogeneous with 1.5 mg/kg, we finally chose that dose for the main study. For postprandial study, we also studied a 3.0-mg/kg dose because previous studies suggested that higher erythromycin doses than in the fasting state were required to affect gastrointestinal motility postprandially. 1~1. Moreover, the 3.0- mg/kg dose is similar to doses used in many studies on gastric emptying and antral motility. 2'3'v'19'2 Main Study In the first group of volunteers (group 1), two series of three recordings were performed in the fasting state and after ingestion of the 200-mL liquid mixed meal (Figure 1). For each study, proximal gastric tone was measured in response to intravenous infusion of (1) erythromycin lactobionate (rythrocin, 1000 rag/20 ml; Abbott, Rungis, France) and atropine (Atropine; Chaix et Du Marais, Paris, France), (2) erythromycin lactobionate and placebo atropine (0.15 mol/l NaC1), or (3) placebo erythromycin lactobionate (0.15 mol/l NaC1) and placebo atropine. The three regimens were administered in random order and single-blind fashion. At 8 AM, after an overnight fast, the folded barostat bag was introduced into the stomach through the mouth, slowly unfolded by gently injecting 200 ml of air, and then completely deflated. A first catheter was inserted into an antecubital vein on the left arm for intravenous infusions. A second heparinated catheter was inserted into the right arm for blood samples. The barostat bag was then connected to the barostat. Using a pressure selection dial, intrabag pressure was gradually increased by 1- mm Hg stepwise increments every 2 minutes until minimal intragastric distending pressure, defined as the lowest pressure level providing a bag volume of >30 ml, 1 was reached.

3 34 DS VARANNS T AL. GASTRONTROLOGY Vol. 109, No. 1 Thereafter, intrabag pressure was set at 2 mm Hg and 1 mm Hg above minimal distending pressure, respectively, for fasting and postprandial studies. Volume recordings were performed with the subject in a sitting position. In the fasting studies, after a 20-minute stabilization period and a 10-minute control period, erythromycin lactobionate (1.5 mg/kg) or placebo was infused for 10 minutes. Simultaneously, atropine (6 gg" kg -1" h -1) or placebo was infused during a 30-minute period. ach dose was dissolved in 0.15 mol/l NaC1 to a total volume of 10 m for erythromycin and 20 ml for atropine. Proximal gastric volume was recorded until the completion of atropine infusion (30 minutes). In the postprandial studies, at the end of the control period (10 minutes), subjects were given the liquid meal at a rate of 100 ml/min using a straw to reduce movements. At the end of the meal, erythromycin (3.0 mg/kg) or placebo and atropine or placebo were infused as previously indicated. Proximal gastric volume was recorded during a period of 120 minutes after the meal. In both series, blood samples (14 ml) were collected at 10-minute intervals starting before the erythromycin-atropine (or placebo) infusions to 30 minutes and then (in postprandial studies) at 30- minute intervals to 120 minutes for peripheral measurements of plasma motilin (4 m) and erythromycin (10 ml). Samples were immediately centrifuged at 6000 rpm for 10 minutes at 4 C. Supernatants were separated and stored at -20 C. In the second group of volunteers (group 2), two series of experiments were performed (Figure 1). The first series was performed in the fasting (interdigestive) state. Placebo (0.15 mol/l NaC1) or atropine (6 gg" kg -l" h 1) was given intravenously for 30 minutes to healthy subjects in a random order according to a single-blind design, and proximal gastric tone was recorded as described above. The second set of recordings was performed after the ingestion of the same 200-mL liquid mixed meal (i.e., in the postprandial state). Volunteers received in a random order one of the following regimens (Figure 1): (1) placebo erythromycin lactobionate (0.15 mol/l NaC1) + placebo atropine (0.15 mol/l NaC1), (2) placebo erythromycin lactobionate + atropine (6 ~tg'kg -I" h -1 for 30 minutes), (3) erythromycin lactobionate (1.5 mg/kg for 10 minutes) + placebo atropine, and (4) erythromycin lactobionate + atropine. The drugs were administered intravenously just after the end of the meal. As in group 1, blood samples were collected at 10-minute intervals, immediately before drug infusion to 30 minutes, and then (in the postprandial study) at 30-minute intervals to 120 minutes for measurements of plasma motilin. Motilin and rythromycin Assays For all subjects, plasma motilin concentrations were measured at the same time point by a specific radioimmunoassay (intra-assay variation, 8%) using antiserum 2AD obtained in the rabbit after 5 injections of porcine motilin coupled with bovine serum albumin (final dilution, 1/2,000,000; inhibition dose, 50%; 4.5 fmol per tube). Reactivity was N-terminal type (100% for full motilin and motilin 1-11, 72% for motilin 1-20, and 0.3% for motilin 2-22). Plasma erythromycin concentrations were measured by high-performance liquid chromatography using electrochemical detection as previously described for roxithromycin determination. 21 Briefly, a dichloromethane extract of the sample was chromatographed on a C 18 reversed-phase column (bt Bondapak; ID, mm; particle size, 10 gm; Waters Assoc., Milford, MA) with acetonitrile/83 mmol/l ammonium acetate/methanol (63.6, 19.4, 17.0) adjusted to ph 7.5 with acetic acid as the mobile phase. rythromycin and the internal standard, roxithromycin, were detected by a dual-electrode electrochemical detector (SA Model 5100 A; Coulochem nvironmental Sciences, Bedford, MA) with a detection limit of O. 1 }.tg/ml (C~max, Bois Guillaume, France). Area under the curve (AUC) from 0 to 10, 20, and 30 minutes after erythromycin administration was calculated for both erythromycin and motilin concentrations. Data Analysis Barostat bag volume. A volume wave was defined as a change in bag volume >30 ml that reverted in <2 minutes to a volume within +50% of the previous level. 1 Baseline volume was defined as the mean volume excluding volume waves. Baseline volume will subsequently be referred to here as volume. Volume was measured every 0.5 minutes and averaged over 1-minute intervals. Fasting recordings. The following variables were defined and calculated. Control volume was the mean bag volume during the 10-minute control period preceding erythromycin infusion, volume change was the variation below control volume, beginning of response was defined as the time when bag volume differed by 30 ml from control volume, maximal response was the maximal change in volume observed after infusion, end of maximal response was the time when bag volume increased by >30 ml from maximal response, and end of response was the time when bag volume reached a level within 30 ml of control volume. Based on these definitions, we measured latency (i.e., the interval between the start of infusion and the beginning of response), maximal response duration, and AUC from the start of infusion to 30 minutes. Postprandial recordings. The following variables were defined and calculated. Control volume was mean bag volume during the 10-minute control period preceding the meal; volume change was the variation over control volume; positive response to the meal was a change in bag volume meeting most of the criteria set by Azpiroz and Malagelada, 8 i.e., (1) beginning < 5 minutes after the end of the meal, (2) larger than 30 ml, and (3) lasting for at least 10 minutes; beginning of the response was the time when bag volume first differed by 30 m from control volume; maximal response was the maximal and uniform (variations <30 ml) change in volume observed after the meal; and end of maximal response was the time when bag volume decreased by >30 ml from maximal volume. 8 Based on these definitions, we measured latency (i.e., the interval between the beginning of the meal and the beginning of response), maximal response and its duration, duration of response, and AUC from the start of the meal to 120 minutes.

4 July 1995 RYTHROMYCIN AND PROXIMAL GASTRIC TON A O J 4 ID, I -100 > <3 100 b g lb 1~ =b 2~ ab Time (min) B M ~'~ ~ ATROPIN Time (min) Figure 2. ffects of atropine (6 ~Lg" kg 1. h 1 in 30 minutes) and 1.5 mg/kg erythromycin (MI.5) given alone or combined on the fasting tone of the proximal stomach. xperiments were conducted in 2 groups of healthy subjects (A, group 2; B, group 1). Gastric tone was measured using a barostat and expressed as changes (mean volume change over control volume) in the volume of the bag., Placeboplacebo; A, placebo-atropine; e, M1.5-placebo, A; Ml.5-atropine. Finally, a positive response to erythromycin on proximal gastric response to the meal was defined as a decrease of bag volume >60 ml lasting for at least 2 minutes and beginning <45 minutes after the end of the meal (value corresponding to the total response duration under placebo less 3SD). The absence of relaxation within 5 minutes after the beginning of the meal was considered as an erythromycin effect (value corresponding to mean latency under placebo plus 3SD). Statistics Results are expressed as means + SD. Means were compared by three-way analysis of variance (subject, infusion, and time). Because the subject was considered to be random P and infusion and time to be fixed, the effect of infusion and time was tested by reference to mean squares of the interactions, respectively, subject X infusion and subject X time. When analysis of variance showed a significant effect, multiple comparisons were performed using the one-way Dunnett's test. Simple regression was calculated to determine correlations between plasma motilin and erythromycin levels and A volumes. Calculations were performed using the A values (increment above the basal value) and the AUC of plasma motilin plus erythromycin and the AUC of volume. A P value of <0.05 was considered significant. Results ffect of rythromycin on Fasting Gastric Tone In all of the subjects studied, 1.5 mg/kg erythromycin decreased intragastric bag volume compared with placebo (Figure 2), but the effect fell short of statistical significance (P = 0.06). Response characteristics (latency, maximal volume change, maximal response duration, and AUC) are shown in Table 1. The number and characteristics of volume waves were not significantly changed by erythromycin. ffect of rythromycin on Postprandial Gastric Tone During placebo experiments, the 200-mL meal was followed by marked, sustained relaxation (i.e., bag volume increase) in all subjects both in group 1 and in group 2 (a representative tracing is shown in Figure 3A) Bag volume quickly reached its maximal value and then slowly returned to control level. rythromycin (1.5 mg/ kg) did not significantly change the gastric response to the meal (Figure 4A). Neither the maximal volume change nor the response duration were significantly modified by erythromycin (Table 2), although there was a nonsignificant trend to a shortened postcibal response (Figure 4A). On the contrary, 3.0 mg/kg erythromycin given after Table 1. ffects of Atropine and rythromycin Alone and Combined With Atropine on Fasting Gastric Tone in 2 Groups of 6 Healthy Subjects Latency (min) Maximal response duration (min) AVma (ml) AUC (103 ml. min) Regimens (groups) Placebo-placebo (2) No response No response 70 _ _+ 2.3 Atropine-placebo (2) No response No response 55 _ Placebo-placebo (1) No response No response -63 _ M1.5-placebo (1) _ a 2.9 _+ 2.3 ~ Ml.5-atropine (1) _+ 112 ~ 4.0 _+ 2.0 a NOT. Results are expressed as mean + SD. Atropine was given as 6 ~g. kg 1. h 1 during a 30-minute period, and erythromycin was given as 1.5 mg/kg. See Figure 1. AVma, maximal volume variation in the barostat bag. For definitions of latency and maximal response duration, see Materials and Methods. ~P < 0.05 compared with placebo (analysis of variance). M1.5-placebo vs. M1.5-atropine, not significant.

5 36 DS VARANNS T AL. GASTRONTROLOGY Vol. 109, No. I the meal did modify gastric response in all subjects (P < 0.001) (Figures 3B and C show representative tracings, and Figure 4B shows average responses). Bag volume then quickly returned to fasting level, giving a shortened postcibal response ( vs minutes, P < 0.001, for erythromycin and placebo, respectively) (Figure 4B and Table 2). However, maximal volume change was not significantly modified by erythromycin (Table 2). The response was observed immediately in 4 subjects (i.e., during erythromycin infusion) (Figure 3C), whereas it was delayed in two others (11 and 15 minutes, i.e., 1 and 5 minutes after the end oferythromycin infusion) (Figure 3B). Finally, after 120 minutes (end of recording), there was still a significant difference in proximal gastric volume between erythromycin and placebo (respectively, vs ml; P < 0.05) (Figure 4B), suggesting a long-lasting effect. ffect of Atropine on Basal and rythromycin-lnduced Gastric Motor Response Atropine did not significantly affect either fasting (Figure 2) or postprandial intragastric bag volume compared with placebo (data not shown). Similarly, atropine did not significantly affect erythromycin effects (1.5 and 3.0 mg/kg) either in the fasting (Table 1) or postprandial state (Table 2 and Figure 4). Plasma rythromycin Concentration In fasting and postprandial studies performed in group 1, maximal plasma erythromycin concentrations occurred for all subjects 10 minutes after drug administration (Figure 5). Atropine did not significantly modify either the peak level or time course of plasma erythromy- A J = 500 A -I 100 == 300 " 500 m O > 700 A placebo _ ~.,,,~- Meal l ~ ~v~,~ * - ~ '. ; ' ;o' ;o ' 1'20 B M 3.0 mg/kg C " Meal I I I ' ' :io 60 M 3.0 mg/kg O > 700 Meal ' 6 ' :io' 60' Time (rain) Figure 3. Individual recordings of barostat bag volume in 3 different healthy subjects in response to a meal with (A) placebo or (B and C) erythromycin (M; 3.0 mg/kg). The liquid meal (200 ml, 200 kcal) was ingested during a 2-minute period; infusions started at the end of the meal and lasted 10 minutes. (A) After the meal, there was a large rapid and sustained volume increase, i.e., a proximal gastric tone decrease or relaxation, to a maximal value followed by a slow return to fasting level. (B) rythromycin infusion did not affect early proximal gastric relaxation but induced a fast return to preprandial values. (C) rythromycin infusion delayed the initial part of proximal gastric relaxation, which was shortened. i2o i2o -100' 0."100 > ' 40O B ~'100 > ' 40O Postprandial time (min) ~ ~ ~ ~ ~ = ~ M1.5 I " - placebo placeb ~i~ atropine z= M1.5-atropine 120 Postprandial time (min) I i I i I. I ' = _..~~ = M3.0 - placebo M ~ & I:;KAq N ~ ~ ~ M3.0-.'afrnnin= atropine atropine Figure 4. ffects of two doses of erythromycin (A, 1.5 mg/kg MI.5; B, 3.0 mg/kg M3.0) given alone or combined with atropine on the postprandial tone of the proximal stomach. xperiments were conducted and expressed in the same manner as in Figure 2. Atropine alone is not shown for clarity, but results were not different from placebo. MI.5 did not significantly change postprandial proximal relaxations of the proximal stomach, whereas M3.0 resulted in an early return of proximal gastric tone to preprandial level. This effect of erythromycin was not affected by atropine.

6 July 1995 RYTHROMYCIN AND PROXIMAL GASTRIC TON 37 Table 2, ffects of rythromycin Alone and Combined With Atropine on Postprandial Gastric Tone in 2 Groups of 6 Healthy Subjects Latency (min) Total response duration (min) AVm~ (ml) AUC (10 3 ml. min) Regimens (groups) Placebo-placebo (2) 1.5 _ _ _+ 10 Placebo-atropine (2) 3.2 _ _ _+ 11 Ml.5-placebo (2) 2.0 ± _ _ _+ 11 M1.5-atropine (2) 1.5 _ ± ± 11 Placebo-placebo (1) ± _ _+ 7 M3.0-placebo (1) 4.3 ± ± 14 a _+ 5 b M3.0-atropine (1) 2.5 +_ _+ 14 a ± 9 b NOT. Results are expressed as mean _+ SD. rythromycin was given as 1.5 and 3.0 mg/kg, and atropine was given as 6 LLg" kg -l" h ~ during a 30-minute period. See Figure 1. AVmax, maximal volume variation in the barostat bag; AUC, during the 120 minutes after the meal; M, erythromycin. For definitions of latency and total response duration, see Materials and Methods. ap < compared with placebo; bp < compared with placebo (analysis of variance). M3.0-placebo vs. M3.0-atropine, not significant. cin (Figure 5). There was no statistically significant correlation between plasma erythromycin levels or AUCs and variations in bag volumes for the fasting study (r = 0.31; n = 18; P = 0.21). The same holds true for the postprandial study. Plasma Motilin Concentration In all studies, erythromycin (1.5 and 3.0 mg/kg) alone or erythromycin combined with atropine did not significantly change plasma motilin concentrations (Figure 6). The meal-induced decrease of plasma motilin levels was not significantly modified by any of the treatments. 3 ~ 1o, ~8 e "6 6 e- ~4 Q 52 m o. 0 M1.5 M3.0 "'" %' Fasting 12 " Postprandial lo 6 4- i i i Postinfusion i i i i i i time (min) Figure 5. Plasma erythromycin concentrations (mean ± SM) in fasting and postprandial states in 6 healthy subjects. In the fasting state, erythromycin (1.5 mg/kg) was infused from 0 to 10 minutes. Placebo or atropine (6 pg. kg ~. h l) was infused from 0 to 30 minutes. In the postprandial study, the liquid meal (200 ml, 200 kcal) was ingested within 2 minutes before erythromycin infusion, which lasted 10 minutes (3.0 mg/kg). Regardless of the type of infusion, plasma erythromycin peaks occurred at 10 minutes in all subjects. Atropine and meal ingestion did not modify plasma erythromycin levels., M-placebo; t, M-atropine. No significant correlation was observed in any study between proximal gastric tone and plasma erythromycin or plasma motilin concentrations. Discussion This report shows that low, microbiologically ineffective doses oferythromycin markedly enhanced proximal gastric tone in fasting and postprandial states in humans This effect was not eliminated by atropine or preceded by motilin release. The barostat technique used to monitor gastric tone has already been validated by other investigators 22 and has been recently applied by us to the study of postprandial gastric relaxation. * In the present study, we avoided a large liquid meal to minimize any interference of the meal volume with the barostat bag within the stomach. Under these experimental conditions, our previous findings showing evident relaxations of the proximal stomach were consistently reproduced (Figure 4). Our results, although not surprising, show for the first time in humans that erythromycin dramatically affects gastric tone, confirming the findings of our less sophisticated pilot study. 17 In the postprandial state, only the higher dose (3.0 mg/kg) of erythromycin induced a significant response, whereas 1.5 mg/kg did not significantly change gastric tone. This result confirms that higher doses oferythromycin are needed to change gastric tone postprandially than during fasting. 1J8 However, lower (i.e., 1.5 mg/kg) doses of the antibiotic proved to also be capable of shortening, although not significantly, the extent of postprandial relaxation A large interindividual variation could account for the lack of statistical significance. Of greater interest was the fact that erythromycin strongly reduced the duration of postprandial relaxation.

7 I 38 DS VARANNS T AL. GASTRONTROLOGY Vol. 109, No. 1 A ,- 100, m ~ 50 B 250 m " T,-c,.- placebo-placebo,], T..e-.. M1.5-placebo - - M1.5-atropine n - - u - -! Time (min) 50" -O-- placebo-placebo "o'" M3.0-placebo - - M3.0-atropine n - - n - - n - n - - n " " n = Time (min) Figure 6. Lack of effect of erythromycin (A, 1.5 mg/kg M1.5; B, 3.0 mg/kg M3.0) given alone or combined with atropine on postprandial plasma motilin levels. This motor response occurred very rapidly (sometimes immediately) during erythromycin infusion and occasionally was short-lasting. In postprandial recordings, bag volume was influenced by two opposing effects: meal-induced relaxation (latency, minutes) and erythromycin-induced contraction (latency, 3-15 minutes). Although the intragastric bag was shown to be located mainly into the proximal stomach, ~ any contraction of the distal end could, of course, interfere with bag volume measurement. Therefore, an effect of erythromycin on antral contractility might have influenced the recording of its effect on proximal gastric tone. However, by using simultaneous double isotopic labeling (bag and meal), we recently showed that the bag remained well positioned in the proximal stomach without overlapping the meal located in the antrum, m Moreover, we never recorded a 3 W/min frequency in the proximal stomach, which is characteristic of the antral contractility. 23 Several mechanisms may account for the erythromycin-induced increase in proximal gastric tone: (1) the direct action of erythromycin on smooth muscle motilin receptors present both on the proximal and distal stomach as shown by in vitro studies, 11'12'24 (2) motilin re- lease, 1<25 and/or (3) activation of other endocrine or neural pathways. In humans, as in animals, it has been shown that atropine administered in doses ranging from 6 to 12 btg" kg -1" h-* can also inhibit the effects oferythromycin on esophageal and gallbladder motility. 13'.4 However, this potential cholinergic-mediated effect has not been carefully investigated as far as gastric motility is concerned. In our study, atropine (6 ~g" kg -l' h -1) did not significantly affect either the postprandial or fasting response of gastric tone to erythromycin. This quite low dose was chosen not only to reduce the risk of side effects in healthy subjects but also because it has been shown to actually inhibit the erythromycin-induced gallbladder emptying) 4 However, a different atropine sensitivity of the erythromycin-mediated effects on gastric and gallbladder smooth muscles cannot be excluded and may be related either to a different sensitivity to cholinergic blockade or to a different receptors' density. In fact, the dose used in our study could have been insufficient to achieve complete blockade of muscarinic receptors. Consequently, the lack of effect of this dose of atropine on gastric tone does not allow us to definitely exclude the involvement of cholinergic pathways in erythromycininduced increase of gastric tone. As in our pilot study, no motilin release was observed before motor response in fasting or postprandial conditions. It is therefore unlikely that motilin release played a role in the motor effect of erythromycin. This interpretation is reinforced by the fact that no correlation was found between motilin plasma levels and motor parameters. However, a local paracrine motilin release cannot be excluded on the basis of our experiments. It may be concluded that erythromycin, given at low, microbiologically ineffective doses, enhances fasting tone and dramatically reduces the duration of postprandial relaxation. The lack of motilin release preceding erythromycin-induced motor response may suggest a direct effect on motilin receptors present on the circular smooth muscle of the proximal stomach, 24 especially because no effect of atropine was evidenced. However, because of the relatively low dose of atropine used in this study, some involvement of cholinergic pathways in the erythromycin action on the proximal gastric tone cannot be definitely excluded. Further studies with a large range of doses of both the macrolid antibiotic and selective muscarinic antagonists will allow us to gain new insights into the erythromycin mechanism of action. References 1. Sarna SK, Soergel KH, Koch TR, Stone J, Wood CM, Ryan RP, Arndorfer RC, Cavanaugh JH, Nellans HN, Lee MB. Gastrointesti-

8 July 1995 RYTHROMYCIN AND PROXIMAL GASTRIC TON 39 nal motor effects of erythromycin in humans. Gastroenterology 1991; 101: Janssens J, Peeters TL, Vantrappen G, Tack J, Urbain JL, De Roe M, Muls, Bouillon R. Improvement of gastric emptying in diabetic gastroparesis by erythromycin. N ngl J Med 1990;322: delbroek M, Horowitz M, Wishart J, Akkermans L. rythromycin increases alcohol absorption by accelerating gastric emptying but slows small intestinal transit in normal subjects (abstr). Gastroenterology 1992; 102:A Peeters TL. rythromycin and other macrolides as prokinetic agents. Gastroenterology 1993; 105: Tomomasa T, Kuroume T, Arai H, Wakabayashi K, Itoh Z. rythromycin induces migrating motor complex in human gastrointestinal tract. Dig Dis Sci 1986;31: Tack J, Janssens J, Vantrappen G, Peeters T, Annese V, Depoortere I, Muls, Bouillon R. ffect of erythromycin on gastric motility in controls and in diabetic gastroparesis. Gastroenterology 1992; 103: Annese V, Janssens J, Vantrappen G, Tack J, Peeters TL, Willemse P, Van Cutsem. rythromycin accelerates gastric emptying by inducing antral contractions and improved gastroduodenal coordination. Gastroenterology 1992; 102: Azpiroz F, Malagelada JR. Gastric tone measured by an electronic barostat in health and postsurgical gastroparesis. Gastroenterology 1987; 92: Parys V, Bruley des Varannes S, Ropert A, Roz6 C, Galmiche JP. Utilisation d'un barostat 61ectronique pour la mesure de la r~ponse motrice de I'estomac proximal & I'alimentation et ~ diff6rents stimuli nerveux chez I'homme. Gastroenterol Clin Biol 1993;17: Ropert A, Bruley des Varannes S, Bizais Y, Roz6 C, Galmiche JP. Simultaneous assessment of liquid emptying and proximal gastric tone in humans. Gastroenterology 1993; 105: Peeters TL, Matthijs G, Depoortere I, Cachet T, Hoogmartens J, Vantrappen G. rythromycin is a receptor motilin agonist. Am J Physiol 1989;257:G470-G Depoortere I, Peeters TL, Vantrappen G. The erythromycin derivative M-523 is a potent motilin agonist in man and in rabbit. Peptides 1990; 11: Chaussade S, Michopoulos S, Sogni P, Guerre J, Couturier D. Motilin agonist erythromycin increases human lower esophageal sphincter pressure by stimulation of cholinergic nerves. Dig Dis Sci 1994;39: Fiorucci S, Bosso R, Morelli A. rythromycin stimulates gallbladder emptying and motilin release by atropine-sensitive pathways. Dig Dis Sci 1992;37: Holle G, Steinbach, Forth W. ffects of erythromycin in the dog upper gastrointestinal tract. Am J Physiol 1992;263:G52- G Kawamura O, Sekiguchi T, Kusano M, Nishioka T, Itoh Z. ffect of erythromycin on interdigestive gastrointestinal contractile activity and plasma motilin concentration in humans. Dig Dis Sci 1993;38: Parys V, Bruley des Varannes S, Chayvialle JA, Galmiche JP. ffects of erythromycin on gastric tone. Hepatogastroenterology 1992;39: Inatomi N, Satoh H, Maki Y, Hashimoto N, Itoh Z, Omura S. An erythromycin derivative, M-523, induces motilin-like gastrointestinal motility in dogs. J Pharmacol xp Ther 1989; 251: Mantides A, Xynos, Chrysos, Georgopoulos N, Vassilakis JS. The effect of erythromycin in gastric emptying of solids and hypertonic liquids in healthy subjects. Am J Gastroenterol 1993;88: Fraser R, Shearer T, Fuller J, Horowitz M, Dent J. Intravenous erythromycin overcomes small intestinal feedback on antral, pyloric, and duodenal motility. Gastroenterology 1992;103: Demotes-Mainard FM, Vincon GA, Jarry CH, Albin HC. Micromethod for the determination of roxithromycin in human plasma and urine by high performance liquid chromatography using electrochemical detection. J Chromatogr 1989;490: Azpiroz F, Malagelada JR. Physiological variations in canine gastric tone measured by an electronic barostat. Am J Physiol 1985; 248:G229-G Urbain JLC, Van Cutsem, Siegel JA, Mayeur S, Vandecruys A, Janssens J, de Roe M, Vantrappen G. Visualization and characterization of gastric contractions using a radionuclide technique. Am J Physiol 1990;259:G1062-G LOdtke F, MOiler H, Golenhofen K. Direct effects of motilin on isolated smooth muscle from various regions of the human stomach. Pflegers Arch 1989;414: Itoh Z, Nakaya M, Suzuki T, Arai H, Wakabayashi K. rythromycin mimics exogenous motilin in gastrointestinal contractile activity in the dog. Am J Physiol 1984; 247:G688-G694. Received April 14, Accepted March 13, Address requests for reprints to" Jean Paul Galmiche, M.D., Laboratoire Fonctions Digestives et Nutrition, Centre de Recherche sur Vo- Iontaires, CHU G et R La~nnec, Nantes Cedex, France. Fax: (33) Supported by a grant from Lilly France Ltd. in collaboration with the Club Fran ais de Motricit~ Digestive (barostat buildup) and an INSRM grant (93CN28). The authors thank Professor Jean Pierre Auget for his help in statistical analysis and Professor Carmelo Scarpignato (Institute of Pharmacology, University of Parma, Parma, Italy) for his strong criticism of the manuscript and helpful suggestions.