- Email: [email protected]
Metformin in the digestive tract
Diabetes Research and Clinical Practice, 4 (1988) 223-229 Elsevier
223
DRC 00181
Metformin in the digestive tract N. Vidon 1, S. Chaussade 1, M. Noel -~, C. Franchisseur 1, B. Huchet 1 and J. J. Bernier 1 I INSERM U.290. H6pital Sablt-Lazare. 75010 Paris. France and :S.E.R.P.A., 92152 Suresnes Cedex, France
(Received I April 1987, revision received 9 September 1987, accepted 14 September 1987)
Kt:v words." Metformin; Digestive tract
Summary After ingestion of metformin, a drug of the biguanide class, there are gastrointestinal side effects in the form of nausea and vomiting, and about 30% of the drug is recovered in feces. The purpose of this work was to explain these two phenomena. Two sets of experiments were carried o u t . Study 1 evaluated the gastroduodenal (GD) absorption in six healthy volunteers by means of an intubation method, employing a twin-lumen tube introduced into the intestine and another into the stomach. Metformin l g was introduced into the stomach with a homogenized meal containing a non-absorbable marker, ~4CPEG 4000: another marker, PEG 4000, was perfused continuously into the duodenum at the ampulla of Vater. Samples of GD contents were collected every 15 rain during 4 h. Metformin was poorly absorbed from the stomach, about 10% over a 4-h period. It did not modify the gastric emptying of a meal but induced a duodeno-gastric reflux in five out of six subjects. About 20% of the amount of drug emptied from the stomach were absorbed from the duodenum. The delivery process was the rate-limiting factor for metformin absorption from the duodenum. The AUC/24 h increased as the absorption rate from the duodenum increased. Study 2 investigated in six healthy volunteers, using another intestinal perfusion technique, the jejunal and ileal absorption of metformin. Metformin 400 mg in saline solution was perfused, over a 2-h period, below an inflated balloon, directly into either the jejunum or the ileum. The mean amount of drug absorbed along a 25-cm segment was low, and similar from the jejunum and ileum: 10.8% and 8.8% respectively. When the drug was perfused into the jejunum, the AUC values were about 2.5 times higher than the values when the drug was perfused into the ileum. These results suggest that the whole intestine is necessary for a sufficient absorption of the drug.
Introduction Address for correspondence: Miss N. Vidon, INSERM U.290, H6pital Saint-Lazare, 107 bis rue du faubourg Saint-Denis, 75010 Paris, France.
Metformin has been recommended for many years in the treatment of maturity-onset dmbetes. It is now the most commonly prescribed oral hypogly-
0168-8227/88/$03.50 ~C~ 1988 Elsevier Science Publishers B.V. (Biomedical Division)
224
cemic drug of the biguanide class. In spite of its widespread use, little is known about the pharmacokinetics of this drug. Previous investigations [14] have shown that after ingestion of metformin, the cumulative excretion in urine over 48 h is about 50% of the ingested drug amount while the recovery in feces is about 30%. Moreover, it has been suggested that metformin could induce inhibition of gastric emptying in the rat [5] and the dog [6] and that this inhibition could be responsible for the gastrointestinal side effects of biguanides (nausea, vomiting). The purpose of this work was to study in healthy volunteers (1) the effect of metformin on gastric emptying of a meal, and (2) the gastrointestinal absorption of this drug by a direct method. Materials and methods
Subjects Studies were performed on 12 healthy volunteers (all men), aged 25.0 ± 1.2 years (mean -4- SEM). No subject had a history of bowel disease or diabetes. All subjects gave written informed consent, the protocol having been approved by the Ethical Subcommittee of the Saint-Lazare Hospital. Their weight was 71.9 ± 1.3 kg and their height was 178.9 + 1.4 cm. The subjects were instructed to take no drug during the 8 days preceding the study and none other than metformin during the 3 days of intubation.
General procedure The intubation techniques facilitate investigation of the gastrointestinal absorption of drugs in man [7-9]. A complete description of the absorption of metformin from stomach to colon is given in this paper. The effect of metformin on postprandial gastric function is also reported. Two sets of experiments were performed. In the first study (A), metformin was ingested with an homogenized meal; the gastric emptying of metformin and its gastroduodenal absorption were studied by applying the technique described by Bernier et al. [10], as improved by Malagelada et al. [1 l] and Vidon et al. [12]. This involved, after an intubation of the stomach and duodenum, measuring gastric
functions by dilution of two aqueous non-absorbable markers, one present in the meal, the other simultaneously perfused into the duodenum. The second study (B) involved perfusing metformin, dissolved in a saline solution, directly into the upper jejunum or the upper ileum; the drug absorption was measured from the 25-cm segment below each perfusion point.
Solutions Study A. I. Unlabelled polyethylene glycol (PEG 4000, 10 g/l) in normal saline solution was perfused into the duodenum as a duodenal recovery marker at a flow rate of 2 ml/min. II. The homogenized meal consisted of 90 g tenderloin steak, 70 g white bread, 14.5 g olive oil, 90 g pear sherbet and 190 ml water containing 30 ~tCi of ~4C-labelled polyethylene glycol 4000 (~4CPEG). The total caloric value was 490 kcal, made up of approximately 50% carbohydrates, 30% fat and 20% proteins. After homogenizing, the volume, osmolality and pH of the meal were 400 ml, 490 mOsm/kg and 5.6, respectively.
Study B. The perfused solution contained: NaCI 130 mM, KCI 5 mM, mannitol 30 mM, 14C-PEG 15/~Ci/1 and metformin 700 mg/l.
Exper#nental design Study A. Experiments were carried out over a 3day period in six subjects aged 23.7 + 0.2 years. On the first day, subjects were intubated with a double-lumen tube. The meal was given on two consecutive days (days 2 and 3), with or without 1 g of metformin in a random order. After an overnight fast, the tube was positioned under fluoroscopic control with the perfusion site of unlabelled PEG 4000 at the ampulla of Vater, the aspiration site being located 20 cm distally, near the ligament of Treitz. A gastric tube was then positioned with its tip in the antrum. The volunteers adopted a sitting position for the duration of the study. The meal was introduced into the stomach by the gastric tube. Duodenal perfusion of PEG 4000 was started immediately before the ingestion of the
225 meal. The meal was introduced over a period of 164 + 15 s. Gastric and duodenal contents were sampled every 15 min for 4 h, duodenal contents being aspirated at a rate of 1 ml/min. At the end of the experiment, the gastric contents were completely aspirated. The stomach was then rinsed with 250 ml of normal saline solution to recover all the meal marker. Blood samples were taken at 0, 15, 30, 60, 90, 120, 150, 180, 210, 240, 360, 480 min and 24 h when the meal contained drug.
Study B. Experiments were carried out over a 3day period in six subjects aged 25.9 + 2.1 years. On the first day, subjects were intubated with a four-lumen tube which also incorporated an occlusive balloon [13]. The technique has been described elsewhere [8]. The segment length was 25 cm. The solution was perfused at a rate of 5 ml/min over 2 h. After an equilibration period to obtain a hydrodynamic steady state, five successive 15-min samples were collected. The jejunal and ileal absorption was studied on days 2 and 3, respectively. Blood samples were taken at 0, 15, 30, 45, 60, 90, 120, 150, 180, 210, 240, 360 and 480 min. For the jejunal study only, a blood sample was also taken at 24 h.
Analytical method Study A. In each gastric and intestinal sample, the cold PEG concentration was measured by the turbidimetric method of Hyden [14]; IgC-PEG was measured in a scintillation counter; metformin concentrations were measured by a high-performance liquid chromatography (HPLC) method.
Study B. In each intestinal sample, cold PEG and metformin concentrations were measured as in study A. In both studies, metformin plasma concentrations were measured by HPLC. Calculations and statistical analysis Study A. The postprandial volume of gastric contents, its fraction emptied into the duodenum, the
gastric emptying of the meal and drug and the gastric secretion were measured using previously reported procedures [11]. The duodenal absorption of the drug was calculated using measured gastric and duodenal concentrations of markers and metforrain. Gastric absorption of the drug was estimated from the difference between the amount given with the meal and that leaving the stomach (amount passing the pylorus + amount removed with gastric samples). Areas under the plasma concentration-time curves (AUC) were calculated by the trapezoidal rule. The peak plasma concentration of metformin (Cma,) and times to reach the peak level (Tmax) were determined.
Stud), B. The absorption of metformin along the studied segment per unit of time was calculated according to the usual formulae for perfusions [15]. Absorption was defined as the difference between the amount entering and that recovered at the sampling site. Data are expressed as means + SEM. Wilcoxon's t-test for paired or non-sequential and correlation analysis were used for statistical comparisons.
Results
Study A Gastric absorption of metformin.
Total ~4C-PEG recovery was measured by adding 14C-PEG contained in gastric samples, the final aspiration, the gastric lavage and the amount passing the ligament of Treitz. Recovery was nearly complete, being 95.1 4- 2.7 and 95.7 + 1.4% with and without metforrain, respectively. The amount of drug lost from the stomach by gastric emptying and sampling over the study was 81.9% of the amount ingested. The metformin to lgC-PEG ratio in each gastric sample over 4 h postprandially also reflects gastric absorption of the drug; it was 86.7 + 1.0%. These results indicate that about 13% of metformin were absorbed from the stomach over a 4-h period.
Gastric emptying.
Gastric emptying of the meal
226 TABLE I INTRAGASTRIC VOLUME OF MEAL EXPRESSED AS PERCENT OF INGESTED MEAL VOLUME IN RELATION TO TIME lntragastric volume of meal Time after meal ingestion
1h
2h
3h
4h
With metformin Without metformin
66.5 + 1.6 67.2 4- 2.5
36.9 4- 3.4 37.4 4- 3.0
12.9 4- 3.3 13.9 4- 4.8
1.6 4- 0.9 5.5 4- 5.0
varied little between subjects, the intragastric volume of meal in relation to time was not statistically different with or without metformin (Table 1). At the end of the 4 h, almost all of the initial meal had left the stomach.
Gastric secretion.
Total gastric volume is made up of the volume of the meal present in the stomach plus the volume of salivary and gastric secretions. These salivary and gastric secretions varied considerably between subjects but were not significantly modified by metformin. Their values were 1040 • 230 and 977 + 154 ml with and without metformin, respectively.
Duodenogastric reflux.
In five subjects out of six, there was a duodenogastric reflux when the meal contained metformin. Indeed, gastric juice was
Gastric emptying o1 M e t f o r m i n mg 15" 100-
50
stained yellow by bile acids as early as 20 min after ingestion of the meal in four subjects and continued so over more than 2 h. Cold P E G was detected in three gastric samples from five subjects. When metformin was not ingested, gastric juice was never stained by bile acids. In one subject, cold P E G was detected during 3 h and 45 min after ingesting the meal with and without drug respectively; no side effects were associated with duodenogastric reflux.
Gastric emptyhTg and duodenal absorption of metformhT. Gastric emptying of metformin did not vary between subjects. The amount of drug emptied within a period of 4 h was 78.4 + 2.2% of the total amount that could have been emptied, i.e., the amount ingested minus the amounts sampled over 4h. The a m o u n t of drug emptied every 15 min varied considerably during the study (Fig. 1). The duodenal absorption of this drug (Y) is closely related to the gastric emptying rate (X) (Fig. 2). The relation is linear (r = 0.73): Y m g / 1 5 min/20 cm = 0.22Xmg/15 m~n -- 1.1; the slope of 0.22 for the regression line indicates that only 22% of the drug emptied from the stomach was absorbed from the duodenum. Individual absorptive capacities for this segment ranged from 11 to 31% of that delivered at the pylorus over 4 h.
Plasma metform#7 concentrations. 0
'
i
15 30
'
/
I
60
90
'
I
120
•
I
I
I
I
150
180
210
240
Minutes
Fig. 1. Gastric emptying of drug in relation to time over a 4-1a period after ingestion with a meal (mean :t: SEM).
The highest maximum plasma concentrations of metformin (Table 2) were recorded for subjects 5 and 6 who exhibited the highest mean duodenal absorption of
227 Metformin
in order that sufficient metformin can be detected in the plasma to give an A U C value.
duodenal
absorption mg/15
min./20
cm
Study B
100
50
•
I
"-° ~
•
J)
I
•
• eo •
• ~
"e
50
O
• e j l ~ . - - /
•
100
Metformin Gastric delivery mg/15 min.
Fig. 2. Duodenal absorption of metformin. Relationship between gastric delivery and duodenal absorption rate. "y' is the duodenal absorption rate, "x' the delivery rate from the stomach (mean of six subjects).
drug over 4 h, 27 and 31% respectively. L o w Cmax and low A U C values were associated with low rates o f absorption. Interindividual differences in Cmax and bioavailability (Fig. 3) were related to the a m o u n t o f m e t f o r m i n absorbed in the d u o d e n u m . A m i n i m u m o f 50 mg o f metformin must be absorbed in the d u o d e n u m over 4 h, i.e., 3 mg/15 min
TABLE 2
Discussion The first study showed that metformin did not modify the rate o f gastric emptying in healthy
AUC
mg.min.l "1
INDIVIDUAL VALUES OF C. . . . . Tmax AND AUC/24 h AFTER INGESTION OF METFORMIN WITH A MEAL (STUDY A) Subject
The difference between absorption rates o f metformin measured in a 25-cm segment o f the jejunum and ileum was not statistically significant. The absorption rates were 10.8 + 2.6 and 8.8 + 1.1% o f the a m o u n t o f drug perfused into the jejunum and ileum, respectively. The peak plasma concentration o f metformin (Fig. 4) and the A U C / 2 4 h (Table 3) when the drug was perfused into the jejunum were significantly (P < 0.05) higher than the values when the drug was perfused into the ileum. The respective A U C values were 280 + 24 mg • min • 1 -l and 123 + 14 m g • min • 1- t. In subject 3, on day 3, the perfusion point was in the transverse colon. Metformin was not detected in plasma over 8 h except at 150 min; at that time the drug plasma concentration was only 0.15 mg/l which is the m i n i m u m detectable level by this m e t h o d o f analysis.
Cm,~ (rag/l)
T,,,, (h)
AUC/24 h (mg.min.1-1)
1 2 3 4 5 6
1.17 0.96 0.62 1.18 1.70 1.71
3 3 4 4 3.5 6
478 253 422 656 1062 1141
Mean SEM
1.22 0.17
669 147
100-
50-
I
100
I
200
mg
Duodenal absorption
of metformin
over 4h,
Fig. 3. Duodenal absorption of metformin. Relationship between AUC/24 h and duodenal absorption rates over 4 h after ingestion with a meal. r = 0.989.
228 Metformine
plasma
concentration
•
Ing/mt)
JCJ iJn.~l Or l}eal p~rfu~lon
"! 111 0.6 0.5
/
,,.
tTt o.,°"2 ]t/¢/~tt t t
t
t
t Iminutesl
50
IOO
t50
200 250 SO0 350 400 450 500
Fig. 4. Metformin plasma concentrations in relation to time when drug is perfused into jejunum (O) or ileum ( e ) .
human volunteers. This absence of effect is in contrast with an animal study [5,6] which showed that biguanides could induce inhibition of gastric emptying. However, metformin did induce duodenogastric reflux in five subjects. The mechanism of this reflux is not known but could be the consequence of the action of biguanides on pyloric activity. The drug was absorbed from the stomach, duodenum, jejunum and ileum but probably not from the colon; its absorption rate throughout the gastrointestinal tract was relatively weak, specially in
the stomach where it was about only 10% over 4 h. In the second study, the amount of drug perfused into the jejunum was similar to that reaching the jejunum when 1 g of metformin was ingested with the homogenized meal in study A. The amounts of drug absorbed appeared to be directly related to the amounts delivered to the intestine by the stomach. Extrapolation of these results to the total small bowel length (about 180 cm between the ligament of Treitz and the cecum) suggests that about 80% of the 1 g of ingested drug with the meal would be absorbed in the gastrointestinal tract, the remaining 20% would reach the colon and would not be absorbed. The same method of extrapolation when drug in saline solution was perfused into the jejunum or ileum shows that 50% and 30%, respectively, would be absorbed, suggesting that the length of the small intestine would be a limiting factor of metformin absorption; that explains the difference between the observed AUC/24 h during the jejunal and ileal perfusion. These results are in agreement with those of Pentikainen et al. [3] who obtained a bioavailability of 50-60% when the drug was given with 200 ml water after an overnight fast.
TABLE 3 INDIVIDUAL VALUES OF C,,.x, T,.~x AND AUC/24 h AFTER PERFUSION OF METFORMIN (STUDY B) Cm,x and AUC values were calculated for a dose of 400 mg metformin presented to the absorption site (the amount of metformin perfused minus the amounts sampled during 2 h). Subject
Perfusion into Jejunum
Ileum
C,.,. (mg/l)
T.,.~ (min)
AUC/24 h (mg.min.l- i)
Cm.~ (mg/1)
T,,,~. (min)
AUC/24 h (mg.min.l- 1)
1 2 3 4 5 6
0.52 0.57 1.04 0.53 0.59 0.62
180 120 150 180 180 150
236 256 397 263 266 264
0.29 0.30
180 150
133 104
0.33 0.23 0.42
150 210 180
89 122 169
Mean SEM
0.65 0.08
160 10
280 24
0.31 0.03
174 11
123 14
229
These results could also explain the difference in bioavailability observed when subjects ingested an aqueous solution of metformin, two tablets of Glucophage retard or a formulation which provided a slower release of the drug [2,16]. The bioavailability will be higher when the drug is ingested with (at the beginning of) a meal than under fasting conditions. The exact mechanism by which biguanides exert their hypoglycemic activity in diabetic patients has been of interest for many years and several mechanisms of action have been proposed. It has been suggested that biguanides could induce an inhibition of glucose absorption [17,18]. It has also been shown that biguanides improve glucose tolerance after ingestion of glucose whereas they do not after intravenous administration of glucose [19]. The biguanides could inhibit glucose absorption by inhibiting Na+-glucose transport across brush-border enterocytes. The degree of inhibition of glucose transport could be linked to the concentration ofbiguanides in the small intestine. Thus the weak absorption rate of metformin from the duodenum, jejunum and ileum would maintain a relatively high concentration of the drug along the small intestine which could be involved in the inhibition of glucose absorption. If this is indeed the case, their improvements in absorption rate could lead to reduced hypoglycemic activity.
5
6
7
8
9
10
11
12
13
Acknowledgement
14
The authors are indebted to Dr. G. L. Nicholson for reviewing the English text.
15
16
References I Sirtori, C. R., Franceschini, G., Galli-Kienle, M. et al. (1978) Disposition of metformin (NN-dimethylbiguanide) in man. Clin. Pharmacol. Ther. 24, 683-693. 2 Noel, M. (1980) Kinetic study of normal and sustained release dosage forms of metformin in normal subjects. J. Int. Biomed. Data 1, 9-20. 3 Pentikainen, P.J., Neuvonen, P.J. and Penttila, A. (1979) Pharmacokinetics of metformin after intravenous and oral administration to man. Eur. J. Clin. Pharmacol. 16, 195-202. 4 Tucker, G. T., Casey, C., Phillips, P.J., Connor, H., Ward, J.D. and Woods, H.F. (1981) Metformin kinetics in healthy
17
18 19
subjects and in patients with diabetes mellitus. Br. J. Clin. Pharmacol. 12, 235-246. Ghareb, A., Botros, M., Saba, J.A. et al. (1969) Mechanism of action of biguanides on glucose metabolism. Ains Shams Med. J. 20, 313-318. Forster, H., Hager, E. and Mehnert, H. (1965) Der Einfluss yon Butylbiguanid im Tierversuch auf die Resorption yon Glukose and Fruktose. Arzneimittelforschung (Drug Res.) 15, 1340-1344. Jobin, G., Cortot, A., Godbillon, J. et al. (1985) Investigation of drug absorption from the gastrointestinal tract of man. I. Metoprolol in the stomach, duodenum and jejunum. Br. J. Clin. Pharmacol. 19, 97S-105S. Vidon, N., Evard, D., Godbillon, J. et al. (1985) Investigation of drug absorption from the gastrointestinal tract of man. II. Metoprolol in the jejunum and ileum. Br. J. Clin. Pharmacol. 19, 107S-112S. Godbillon, J., Evard, D., Vidon, N. et aI. (1985) Investigation of drug absorption from the gastrointestinal tract of man. III. Metoprolol in the colon. Br. J. Clin. Pharmacol. 19, 113S-118S. Bernier, J.J. and Lebert, A. (1971) Vitesse de l'~vacuation de l'estomac et du duodenum au cours de l'hyperglyc6mie provoqu6e per os. Biol. Gastroent6rol. 4, 351-352. Malagelada, J.R., Longstreth, G.E., Summerskill, W.H.J. and Go, V.L.W. (1976) Measurement of gastric functions during digestion of ordinary solid meals in man. Gastroenterology 70, 203-210. Vidon, N., Muschart, J.M., Cosnes, J., Ruskon6, A. and Bernier, J.J. (1979) Etude critique de l'estimation de la vidange gastrique par la m6thode de perfusion duod~nale d'une substance non absorbable ;i faible d6bit. Gastroent~rol. Clin. Biol. 3, 549 552. Phillips, S. F. and Summerskill, W. H. J. (1966) Occlusion of the jejunum for intestinal perfusion in man. Mayo Clin. Proc. 41,224-231. Hyden, S (1955) A turbidimetric method for the determination of higher polyethylene glycols in biological materials. Ann. R. Agric. Coll. Sweden 22, 139-145. Modigliani, R., Rambaud, J.C. and Bernier, J.J. (1973)The method of intraluminal perfusion of the human small intestine. I. Principle and technique. Digestion 9, 176-192. Pentikainen, P.J. (1986) Bioavailability of metformin. Comparison of solution, rapidly dissolving tablet, and three sustained release products. Int. J. Clin. Pharm. Ther. Tox. 24, 213-220. Czyzyk, A., Tawecki, J., Sadowski, J., Ponikowska, I. and Szczepanik, Z. (1968) Effect of biguanides on intestinal absorption of glucose. Diabetes 17, 492-499. Caspary, W.F. (1977) Biguanides and intestinal absorptive function. Acta Hepato-Gastroenterol. 24, 473-480. Hollobaugh, S.L., Rao, B. and Kruger, F.A. (1970) Studies on the site and mechanism of action of phenphormin. I. Evidence for significant "non-peripheral' effects of phenphormin on glucose metabolism in normal subjects. Diabetes 19, 45-49.
223
DRC 00181
Metformin in the digestive tract N. Vidon 1, S. Chaussade 1, M. Noel -~, C. Franchisseur 1, B. Huchet 1 and J. J. Bernier 1 I INSERM U.290. H6pital Sablt-Lazare. 75010 Paris. France and :S.E.R.P.A., 92152 Suresnes Cedex, France
(Received I April 1987, revision received 9 September 1987, accepted 14 September 1987)
Kt:v words." Metformin; Digestive tract
Summary After ingestion of metformin, a drug of the biguanide class, there are gastrointestinal side effects in the form of nausea and vomiting, and about 30% of the drug is recovered in feces. The purpose of this work was to explain these two phenomena. Two sets of experiments were carried o u t . Study 1 evaluated the gastroduodenal (GD) absorption in six healthy volunteers by means of an intubation method, employing a twin-lumen tube introduced into the intestine and another into the stomach. Metformin l g was introduced into the stomach with a homogenized meal containing a non-absorbable marker, ~4CPEG 4000: another marker, PEG 4000, was perfused continuously into the duodenum at the ampulla of Vater. Samples of GD contents were collected every 15 rain during 4 h. Metformin was poorly absorbed from the stomach, about 10% over a 4-h period. It did not modify the gastric emptying of a meal but induced a duodeno-gastric reflux in five out of six subjects. About 20% of the amount of drug emptied from the stomach were absorbed from the duodenum. The delivery process was the rate-limiting factor for metformin absorption from the duodenum. The AUC/24 h increased as the absorption rate from the duodenum increased. Study 2 investigated in six healthy volunteers, using another intestinal perfusion technique, the jejunal and ileal absorption of metformin. Metformin 400 mg in saline solution was perfused, over a 2-h period, below an inflated balloon, directly into either the jejunum or the ileum. The mean amount of drug absorbed along a 25-cm segment was low, and similar from the jejunum and ileum: 10.8% and 8.8% respectively. When the drug was perfused into the jejunum, the AUC values were about 2.5 times higher than the values when the drug was perfused into the ileum. These results suggest that the whole intestine is necessary for a sufficient absorption of the drug.
Introduction Address for correspondence: Miss N. Vidon, INSERM U.290, H6pital Saint-Lazare, 107 bis rue du faubourg Saint-Denis, 75010 Paris, France.
Metformin has been recommended for many years in the treatment of maturity-onset dmbetes. It is now the most commonly prescribed oral hypogly-
0168-8227/88/$03.50 ~C~ 1988 Elsevier Science Publishers B.V. (Biomedical Division)
224
cemic drug of the biguanide class. In spite of its widespread use, little is known about the pharmacokinetics of this drug. Previous investigations [14] have shown that after ingestion of metformin, the cumulative excretion in urine over 48 h is about 50% of the ingested drug amount while the recovery in feces is about 30%. Moreover, it has been suggested that metformin could induce inhibition of gastric emptying in the rat [5] and the dog [6] and that this inhibition could be responsible for the gastrointestinal side effects of biguanides (nausea, vomiting). The purpose of this work was to study in healthy volunteers (1) the effect of metformin on gastric emptying of a meal, and (2) the gastrointestinal absorption of this drug by a direct method. Materials and methods
Subjects Studies were performed on 12 healthy volunteers (all men), aged 25.0 ± 1.2 years (mean -4- SEM). No subject had a history of bowel disease or diabetes. All subjects gave written informed consent, the protocol having been approved by the Ethical Subcommittee of the Saint-Lazare Hospital. Their weight was 71.9 ± 1.3 kg and their height was 178.9 + 1.4 cm. The subjects were instructed to take no drug during the 8 days preceding the study and none other than metformin during the 3 days of intubation.
General procedure The intubation techniques facilitate investigation of the gastrointestinal absorption of drugs in man [7-9]. A complete description of the absorption of metformin from stomach to colon is given in this paper. The effect of metformin on postprandial gastric function is also reported. Two sets of experiments were performed. In the first study (A), metformin was ingested with an homogenized meal; the gastric emptying of metformin and its gastroduodenal absorption were studied by applying the technique described by Bernier et al. [10], as improved by Malagelada et al. [1 l] and Vidon et al. [12]. This involved, after an intubation of the stomach and duodenum, measuring gastric
functions by dilution of two aqueous non-absorbable markers, one present in the meal, the other simultaneously perfused into the duodenum. The second study (B) involved perfusing metformin, dissolved in a saline solution, directly into the upper jejunum or the upper ileum; the drug absorption was measured from the 25-cm segment below each perfusion point.
Solutions Study A. I. Unlabelled polyethylene glycol (PEG 4000, 10 g/l) in normal saline solution was perfused into the duodenum as a duodenal recovery marker at a flow rate of 2 ml/min. II. The homogenized meal consisted of 90 g tenderloin steak, 70 g white bread, 14.5 g olive oil, 90 g pear sherbet and 190 ml water containing 30 ~tCi of ~4C-labelled polyethylene glycol 4000 (~4CPEG). The total caloric value was 490 kcal, made up of approximately 50% carbohydrates, 30% fat and 20% proteins. After homogenizing, the volume, osmolality and pH of the meal were 400 ml, 490 mOsm/kg and 5.6, respectively.
Study B. The perfused solution contained: NaCI 130 mM, KCI 5 mM, mannitol 30 mM, 14C-PEG 15/~Ci/1 and metformin 700 mg/l.
Exper#nental design Study A. Experiments were carried out over a 3day period in six subjects aged 23.7 + 0.2 years. On the first day, subjects were intubated with a double-lumen tube. The meal was given on two consecutive days (days 2 and 3), with or without 1 g of metformin in a random order. After an overnight fast, the tube was positioned under fluoroscopic control with the perfusion site of unlabelled PEG 4000 at the ampulla of Vater, the aspiration site being located 20 cm distally, near the ligament of Treitz. A gastric tube was then positioned with its tip in the antrum. The volunteers adopted a sitting position for the duration of the study. The meal was introduced into the stomach by the gastric tube. Duodenal perfusion of PEG 4000 was started immediately before the ingestion of the
225 meal. The meal was introduced over a period of 164 + 15 s. Gastric and duodenal contents were sampled every 15 min for 4 h, duodenal contents being aspirated at a rate of 1 ml/min. At the end of the experiment, the gastric contents were completely aspirated. The stomach was then rinsed with 250 ml of normal saline solution to recover all the meal marker. Blood samples were taken at 0, 15, 30, 60, 90, 120, 150, 180, 210, 240, 360, 480 min and 24 h when the meal contained drug.
Study B. Experiments were carried out over a 3day period in six subjects aged 25.9 + 2.1 years. On the first day, subjects were intubated with a four-lumen tube which also incorporated an occlusive balloon [13]. The technique has been described elsewhere [8]. The segment length was 25 cm. The solution was perfused at a rate of 5 ml/min over 2 h. After an equilibration period to obtain a hydrodynamic steady state, five successive 15-min samples were collected. The jejunal and ileal absorption was studied on days 2 and 3, respectively. Blood samples were taken at 0, 15, 30, 45, 60, 90, 120, 150, 180, 210, 240, 360 and 480 min. For the jejunal study only, a blood sample was also taken at 24 h.
Analytical method Study A. In each gastric and intestinal sample, the cold PEG concentration was measured by the turbidimetric method of Hyden [14]; IgC-PEG was measured in a scintillation counter; metformin concentrations were measured by a high-performance liquid chromatography (HPLC) method.
Study B. In each intestinal sample, cold PEG and metformin concentrations were measured as in study A. In both studies, metformin plasma concentrations were measured by HPLC. Calculations and statistical analysis Study A. The postprandial volume of gastric contents, its fraction emptied into the duodenum, the
gastric emptying of the meal and drug and the gastric secretion were measured using previously reported procedures [11]. The duodenal absorption of the drug was calculated using measured gastric and duodenal concentrations of markers and metforrain. Gastric absorption of the drug was estimated from the difference between the amount given with the meal and that leaving the stomach (amount passing the pylorus + amount removed with gastric samples). Areas under the plasma concentration-time curves (AUC) were calculated by the trapezoidal rule. The peak plasma concentration of metformin (Cma,) and times to reach the peak level (Tmax) were determined.
Stud), B. The absorption of metformin along the studied segment per unit of time was calculated according to the usual formulae for perfusions [15]. Absorption was defined as the difference between the amount entering and that recovered at the sampling site. Data are expressed as means + SEM. Wilcoxon's t-test for paired or non-sequential and correlation analysis were used for statistical comparisons.
Results
Study A Gastric absorption of metformin.
Total ~4C-PEG recovery was measured by adding 14C-PEG contained in gastric samples, the final aspiration, the gastric lavage and the amount passing the ligament of Treitz. Recovery was nearly complete, being 95.1 4- 2.7 and 95.7 + 1.4% with and without metforrain, respectively. The amount of drug lost from the stomach by gastric emptying and sampling over the study was 81.9% of the amount ingested. The metformin to lgC-PEG ratio in each gastric sample over 4 h postprandially also reflects gastric absorption of the drug; it was 86.7 + 1.0%. These results indicate that about 13% of metformin were absorbed from the stomach over a 4-h period.
Gastric emptying.
Gastric emptying of the meal
226 TABLE I INTRAGASTRIC VOLUME OF MEAL EXPRESSED AS PERCENT OF INGESTED MEAL VOLUME IN RELATION TO TIME lntragastric volume of meal Time after meal ingestion
1h
2h
3h
4h
With metformin Without metformin
66.5 + 1.6 67.2 4- 2.5
36.9 4- 3.4 37.4 4- 3.0
12.9 4- 3.3 13.9 4- 4.8
1.6 4- 0.9 5.5 4- 5.0
varied little between subjects, the intragastric volume of meal in relation to time was not statistically different with or without metformin (Table 1). At the end of the 4 h, almost all of the initial meal had left the stomach.
Gastric secretion.
Total gastric volume is made up of the volume of the meal present in the stomach plus the volume of salivary and gastric secretions. These salivary and gastric secretions varied considerably between subjects but were not significantly modified by metformin. Their values were 1040 • 230 and 977 + 154 ml with and without metformin, respectively.
Duodenogastric reflux.
In five subjects out of six, there was a duodenogastric reflux when the meal contained metformin. Indeed, gastric juice was
Gastric emptying o1 M e t f o r m i n mg 15" 100-
50
stained yellow by bile acids as early as 20 min after ingestion of the meal in four subjects and continued so over more than 2 h. Cold P E G was detected in three gastric samples from five subjects. When metformin was not ingested, gastric juice was never stained by bile acids. In one subject, cold P E G was detected during 3 h and 45 min after ingesting the meal with and without drug respectively; no side effects were associated with duodenogastric reflux.
Gastric emptyhTg and duodenal absorption of metformhT. Gastric emptying of metformin did not vary between subjects. The amount of drug emptied within a period of 4 h was 78.4 + 2.2% of the total amount that could have been emptied, i.e., the amount ingested minus the amounts sampled over 4h. The a m o u n t of drug emptied every 15 min varied considerably during the study (Fig. 1). The duodenal absorption of this drug (Y) is closely related to the gastric emptying rate (X) (Fig. 2). The relation is linear (r = 0.73): Y m g / 1 5 min/20 cm = 0.22Xmg/15 m~n -- 1.1; the slope of 0.22 for the regression line indicates that only 22% of the drug emptied from the stomach was absorbed from the duodenum. Individual absorptive capacities for this segment ranged from 11 to 31% of that delivered at the pylorus over 4 h.
Plasma metform#7 concentrations. 0
'
i
15 30
'
/
I
60
90
'
I
120
•
I
I
I
I
150
180
210
240
Minutes
Fig. 1. Gastric emptying of drug in relation to time over a 4-1a period after ingestion with a meal (mean :t: SEM).
The highest maximum plasma concentrations of metformin (Table 2) were recorded for subjects 5 and 6 who exhibited the highest mean duodenal absorption of
227 Metformin
in order that sufficient metformin can be detected in the plasma to give an A U C value.
duodenal
absorption mg/15
min./20
cm
Study B
100
50
•
I
"-° ~
•
J)
I
•
• eo •
• ~
"e
50
O
• e j l ~ . - - /
•
100
Metformin Gastric delivery mg/15 min.
Fig. 2. Duodenal absorption of metformin. Relationship between gastric delivery and duodenal absorption rate. "y' is the duodenal absorption rate, "x' the delivery rate from the stomach (mean of six subjects).
drug over 4 h, 27 and 31% respectively. L o w Cmax and low A U C values were associated with low rates o f absorption. Interindividual differences in Cmax and bioavailability (Fig. 3) were related to the a m o u n t o f m e t f o r m i n absorbed in the d u o d e n u m . A m i n i m u m o f 50 mg o f metformin must be absorbed in the d u o d e n u m over 4 h, i.e., 3 mg/15 min
TABLE 2
Discussion The first study showed that metformin did not modify the rate o f gastric emptying in healthy
AUC
mg.min.l "1
INDIVIDUAL VALUES OF C. . . . . Tmax AND AUC/24 h AFTER INGESTION OF METFORMIN WITH A MEAL (STUDY A) Subject
The difference between absorption rates o f metformin measured in a 25-cm segment o f the jejunum and ileum was not statistically significant. The absorption rates were 10.8 + 2.6 and 8.8 + 1.1% o f the a m o u n t o f drug perfused into the jejunum and ileum, respectively. The peak plasma concentration o f metformin (Fig. 4) and the A U C / 2 4 h (Table 3) when the drug was perfused into the jejunum were significantly (P < 0.05) higher than the values when the drug was perfused into the ileum. The respective A U C values were 280 + 24 mg • min • 1 -l and 123 + 14 m g • min • 1- t. In subject 3, on day 3, the perfusion point was in the transverse colon. Metformin was not detected in plasma over 8 h except at 150 min; at that time the drug plasma concentration was only 0.15 mg/l which is the m i n i m u m detectable level by this m e t h o d o f analysis.
Cm,~ (rag/l)
T,,,, (h)
AUC/24 h (mg.min.1-1)
1 2 3 4 5 6
1.17 0.96 0.62 1.18 1.70 1.71
3 3 4 4 3.5 6
478 253 422 656 1062 1141
Mean SEM
1.22 0.17
669 147
100-
50-
I
100
I
200
mg
Duodenal absorption
of metformin
over 4h,
Fig. 3. Duodenal absorption of metformin. Relationship between AUC/24 h and duodenal absorption rates over 4 h after ingestion with a meal. r = 0.989.
228 Metformine
plasma
concentration
•
Ing/mt)
JCJ iJn.~l Or l}eal p~rfu~lon
"! 111 0.6 0.5
/
,,.
tTt o.,°"2 ]t/¢/~tt t t
t
t
t Iminutesl
50
IOO
t50
200 250 SO0 350 400 450 500
Fig. 4. Metformin plasma concentrations in relation to time when drug is perfused into jejunum (O) or ileum ( e ) .
human volunteers. This absence of effect is in contrast with an animal study [5,6] which showed that biguanides could induce inhibition of gastric emptying. However, metformin did induce duodenogastric reflux in five subjects. The mechanism of this reflux is not known but could be the consequence of the action of biguanides on pyloric activity. The drug was absorbed from the stomach, duodenum, jejunum and ileum but probably not from the colon; its absorption rate throughout the gastrointestinal tract was relatively weak, specially in
the stomach where it was about only 10% over 4 h. In the second study, the amount of drug perfused into the jejunum was similar to that reaching the jejunum when 1 g of metformin was ingested with the homogenized meal in study A. The amounts of drug absorbed appeared to be directly related to the amounts delivered to the intestine by the stomach. Extrapolation of these results to the total small bowel length (about 180 cm between the ligament of Treitz and the cecum) suggests that about 80% of the 1 g of ingested drug with the meal would be absorbed in the gastrointestinal tract, the remaining 20% would reach the colon and would not be absorbed. The same method of extrapolation when drug in saline solution was perfused into the jejunum or ileum shows that 50% and 30%, respectively, would be absorbed, suggesting that the length of the small intestine would be a limiting factor of metformin absorption; that explains the difference between the observed AUC/24 h during the jejunal and ileal perfusion. These results are in agreement with those of Pentikainen et al. [3] who obtained a bioavailability of 50-60% when the drug was given with 200 ml water after an overnight fast.
TABLE 3 INDIVIDUAL VALUES OF C,,.x, T,.~x AND AUC/24 h AFTER PERFUSION OF METFORMIN (STUDY B) Cm,x and AUC values were calculated for a dose of 400 mg metformin presented to the absorption site (the amount of metformin perfused minus the amounts sampled during 2 h). Subject
Perfusion into Jejunum
Ileum
C,.,. (mg/l)
T.,.~ (min)
AUC/24 h (mg.min.l- i)
Cm.~ (mg/1)
T,,,~. (min)
AUC/24 h (mg.min.l- 1)
1 2 3 4 5 6
0.52 0.57 1.04 0.53 0.59 0.62
180 120 150 180 180 150
236 256 397 263 266 264
0.29 0.30
180 150
133 104
0.33 0.23 0.42
150 210 180
89 122 169
Mean SEM
0.65 0.08
160 10
280 24
0.31 0.03
174 11
123 14
229
These results could also explain the difference in bioavailability observed when subjects ingested an aqueous solution of metformin, two tablets of Glucophage retard or a formulation which provided a slower release of the drug [2,16]. The bioavailability will be higher when the drug is ingested with (at the beginning of) a meal than under fasting conditions. The exact mechanism by which biguanides exert their hypoglycemic activity in diabetic patients has been of interest for many years and several mechanisms of action have been proposed. It has been suggested that biguanides could induce an inhibition of glucose absorption [17,18]. It has also been shown that biguanides improve glucose tolerance after ingestion of glucose whereas they do not after intravenous administration of glucose [19]. The biguanides could inhibit glucose absorption by inhibiting Na+-glucose transport across brush-border enterocytes. The degree of inhibition of glucose transport could be linked to the concentration ofbiguanides in the small intestine. Thus the weak absorption rate of metformin from the duodenum, jejunum and ileum would maintain a relatively high concentration of the drug along the small intestine which could be involved in the inhibition of glucose absorption. If this is indeed the case, their improvements in absorption rate could lead to reduced hypoglycemic activity.
5
6
7
8
9
10
11
12
13
Acknowledgement
14
The authors are indebted to Dr. G. L. Nicholson for reviewing the English text.
15
16
References I Sirtori, C. R., Franceschini, G., Galli-Kienle, M. et al. (1978) Disposition of metformin (NN-dimethylbiguanide) in man. Clin. Pharmacol. Ther. 24, 683-693. 2 Noel, M. (1980) Kinetic study of normal and sustained release dosage forms of metformin in normal subjects. J. Int. Biomed. Data 1, 9-20. 3 Pentikainen, P.J., Neuvonen, P.J. and Penttila, A. (1979) Pharmacokinetics of metformin after intravenous and oral administration to man. Eur. J. Clin. Pharmacol. 16, 195-202. 4 Tucker, G. T., Casey, C., Phillips, P.J., Connor, H., Ward, J.D. and Woods, H.F. (1981) Metformin kinetics in healthy
17
18 19
subjects and in patients with diabetes mellitus. Br. J. Clin. Pharmacol. 12, 235-246. Ghareb, A., Botros, M., Saba, J.A. et al. (1969) Mechanism of action of biguanides on glucose metabolism. Ains Shams Med. J. 20, 313-318. Forster, H., Hager, E. and Mehnert, H. (1965) Der Einfluss yon Butylbiguanid im Tierversuch auf die Resorption yon Glukose and Fruktose. Arzneimittelforschung (Drug Res.) 15, 1340-1344. Jobin, G., Cortot, A., Godbillon, J. et al. (1985) Investigation of drug absorption from the gastrointestinal tract of man. I. Metoprolol in the stomach, duodenum and jejunum. Br. J. Clin. Pharmacol. 19, 97S-105S. Vidon, N., Evard, D., Godbillon, J. et al. (1985) Investigation of drug absorption from the gastrointestinal tract of man. II. Metoprolol in the jejunum and ileum. Br. J. Clin. Pharmacol. 19, 107S-112S. Godbillon, J., Evard, D., Vidon, N. et aI. (1985) Investigation of drug absorption from the gastrointestinal tract of man. III. Metoprolol in the colon. Br. J. Clin. Pharmacol. 19, 113S-118S. Bernier, J.J. and Lebert, A. (1971) Vitesse de l'~vacuation de l'estomac et du duodenum au cours de l'hyperglyc6mie provoqu6e per os. Biol. Gastroent6rol. 4, 351-352. Malagelada, J.R., Longstreth, G.E., Summerskill, W.H.J. and Go, V.L.W. (1976) Measurement of gastric functions during digestion of ordinary solid meals in man. Gastroenterology 70, 203-210. Vidon, N., Muschart, J.M., Cosnes, J., Ruskon6, A. and Bernier, J.J. (1979) Etude critique de l'estimation de la vidange gastrique par la m6thode de perfusion duod~nale d'une substance non absorbable ;i faible d6bit. Gastroent~rol. Clin. Biol. 3, 549 552. Phillips, S. F. and Summerskill, W. H. J. (1966) Occlusion of the jejunum for intestinal perfusion in man. Mayo Clin. Proc. 41,224-231. Hyden, S (1955) A turbidimetric method for the determination of higher polyethylene glycols in biological materials. Ann. R. Agric. Coll. Sweden 22, 139-145. Modigliani, R., Rambaud, J.C. and Bernier, J.J. (1973)The method of intraluminal perfusion of the human small intestine. I. Principle and technique. Digestion 9, 176-192. Pentikainen, P.J. (1986) Bioavailability of metformin. Comparison of solution, rapidly dissolving tablet, and three sustained release products. Int. J. Clin. Pharm. Ther. Tox. 24, 213-220. Czyzyk, A., Tawecki, J., Sadowski, J., Ponikowska, I. and Szczepanik, Z. (1968) Effect of biguanides on intestinal absorption of glucose. Diabetes 17, 492-499. Caspary, W.F. (1977) Biguanides and intestinal absorptive function. Acta Hepato-Gastroenterol. 24, 473-480. Hollobaugh, S.L., Rao, B. and Kruger, F.A. (1970) Studies on the site and mechanism of action of phenphormin. I. Evidence for significant "non-peripheral' effects of phenphormin on glucose metabolism in normal subjects. Diabetes 19, 45-49.