Enhancement of absorption and effect of glipizide by magnesium hydroxide The effects of magnesium hydroxide on the pharmacokinetics and pharmacodynamics of glipizide were studied in eight healthy volunteers in a randomized crossover trial. After an overnight fast, 5 mg glipizide was given with either 150 ml water or water containing 850 mg magnesium hydroxide. Magnesium hydroxide increased the areas under the plasma glipizide concentrationtime curves (AUC) from 0 to 1/2

hour and from 0 to 1 hour by 180% (p < 0.05) and 69% (p < 0.05), respectively. The peak plasma concentration, time to peak, total AUC, elimination half-life, and mean residence time of glipizide remained unchanged. The incremental plasma insulin area from 0 to 1/2 hour increased by 85% (p < 0.05),

and the time to maximal insulin response was reduced (p < 0.05) during the magnesium hydroxide phase. The corresponding decremental plasma glucose area increased fourfold (p < 0.05), and the maximal glucose decrease was 35% greater (p < 0.05) than during the control phase. We conclude that the concomitant ingestion of magnesium hydroxide and glipizide may result in accelerated absorption of glipizide and increased early insulin and glucose responses. (CLIN PHAR/VIACOL THER 1991;49:39-43.)

Kan i T. Kivistó, MD, and Pertti J. Neuvonen, MD Turku, Finland The blood glucoselowering effect of sulfonylureas may be influenced by their rate of gastrointestinal absorption and the time of their appearance in blood relative to the time of meal intake." Food itself can delay the absorption of glipizide, and taking glipizide on an empty stomach has been shown to result in better glucose disposition in diabetic patients. I Accordingly, it is often recommended to take sulfonylureas (e.g., glipizide) 1/2 hour before meals to ensure their rapid absorption.4 Although sparingly soluble in the acidic gastric contents, glipizide is absorbed completely and relatively rapidly.5 However, there is evidence that absorption of glipizide and other sulfonylureas can be unusually slow in some healthy subjects, as well as in some di-

abetic patients." Antacids are known to affect absorption of many drugs, and clinically important interactions have been described. The consequences of drug-antacid interactions depend greatly on the antacid used, but reduced absorption of the affected drug is most often encountered.9'10 It is known that magnesium hydroxide can

From the Department of Pharmacology, University of Turku. Supported in part by the Technology Development Centre (TEKES), Helsinki, Finland. Received for publication May 31, 1990; accepted Sept. 19, 1990. Reprint requests: Kan i T. Kivistö, MD, Kiinamyllynkatu 10, SF-20520 Turku, Finland. 13/1/25579

considerably increase the rate of absorption of certain weakly acidic drugs (e.g., tolfenamic and mefenamic acids). I The purposes of this study were to examine whether magnesium hydroxide could also accelerate the gastrointestinal absorption of glipizide and to characterize the effects of the proposed interaction on insulin and glucose responses.

MATERIAL AND METHODS General design. A randomized crossover design with two phases, at least week apart, was used. Each volunteer was considered to be healthy on the basis of medical history, physical examination, and routine laboratory tests (renal and hepatic functions, fasting glucose, and hemoglobin). The subjects were thoroughly informed, both verbally and in writing, and informed consent was obtained. The study protocol was accepted by the Ethics Committee of the Faculty of Medicine, University of Turku. Eight male volunteers (aged 22 to 29 years; weight 65 to 92 kg) participated in the study. After an overnight fast, they were given 5 mg glipizide (one 5 mg Melizid tablet; Leiras, Finland) with either 150 ml water or 150 ml water containing 10 ml (850 mg) magnesium hydroxide (Milk of Magnesia; Winthrop Laboratories, England). The drugs were swallowed in a 1

sitting position, and the subjects spent the next 2 hours lying in a supine position. The subjects were under direct medical supervision for 2 hours from the beginning of the study. No food was allowed during

39

CI.IN PHARMACOI. THER

40 Kivistä and Neuvonen

JANUARY 1991

500

Table I. Effect of magnesium hydroxide (850 mg) on the pharmacokinetics of glipizide (5 mg)

400

300

tn,a (hr) Cma (ng/ml) AUC(0-1/2) (ng

100

10

Time (hours)

Fig. 1. Effect of magnesium hydroxide (850 mg, solid cir-

cles) on the absorption of glipizide (5 mg), reflected as plasma glipizide concentrations (mean ± SE) in eight subjects. 5p < 0.05 compared with control (open circles).

that time. Breakfast was given after the 2-hour blood sample, whereupon the subjects could move and eat as desired. Hypoglycemic symptoms were monitored during the study days. Glucose for both oral and intravenous use was available in case of severe hypoglycemia, but it was not needed. Timed venous blood samples were collected from an indwelling catheter in heparinized tubes at 0, 15, 22, 30, and 45 minutes and at 1, 11/2, 2, 3, 4, 5, 7,

hr/ml) AUC(0-1) (ng hr/ml) AUC(0-10) (ng hr/ml) AUC (ng hem') MRT (hr)

1.6 -± 0.32

1.2 ± 0.29 0.15 ± 0.02t 445 ± 32.0

1665 ± 198 2048 ± 383

t,12 (hr) Data are mean values

Mg(OH)2

0.32 ± 0.06 416 ± 26.7 27.0 ± 10.5 155 ± 40.4

t1/2,b, (hr)*

200

Control

5.6 ± 0.85 3.4 ± 0.70

75.2 ± 19.41.

262 ± 36.9t 1718 ± 167 2025 ± 300 4.9 -± 0.57

3.1 ± 0.35

SE in eight subjects.

t,, Time lo reach maximum plasma concentration;

absorption

half-life; C maximum plasma concentration; AUC, area Under the plasma concentrationtime curve; MRT, mean residence time; t, elimination halflife. * values of two subjects could not be calculated because of erratic absorption.

tp < 0.05 compared with control.

termined up to the breakfast time (2 hours). Insulin was determined by radioimmunoassay (Phadeseph Insulin RIA; Pharmacia Diagnostics AB, Uppsala, Sweden), and glucose was determined by the glucose oxidase method (Reflotron; Boehringer Mannheim GmbH, Mannheim, Germany). The interassay coeffi-

cients of variation were 6.0% (mean, 12.8 mU/L; n = 5) and 3.3% (mean, 5.6 mmol/L; n = 8), respectively.

and 10 hours after drug ingestion. The tubes were

Pharmacokinetic analysis. The plasma concentra-

chilled on ice both before and after sample collection.

tiontime data for glipizide were fitted to an open

Plasma was separated within 30 minutes at +4° C, and the samples were stored at 20° C until analyzed.

one- or two-compartment model, whichever gave a better fit, by weighted least-squares analysis with the

Analytic methods. Plasma glipizide concentrations were determined by high-performance reversed-phase liquid chromatography, modifying two previously published methods.12'13 To 1 ml of plasma, 0.2 ml of 0.5 mol/L hydrochloric acid and 400 ng tolbutamide,

SIPHAR pharmacokinetic curve-fitting program (SIMED, Creteil, France). The absorption of glipizide

the internal standard, were added. The sample was

then extracted with 5 ml of a mixture of dichloromethane and hexane (1:1) and shaken for 4 minutes. After centrifugation, the organic phase was evaporated to dryness under a stream of nitrogen at 40° C. The residue was redissolved in 70 jJ of the mobile phase, and an aliquot of 20 i.l was injected into the chromatograph. The mobile phase consisted of a mixture of methanol and 0.01 mol/L phosphate buffer with pH 3.5 (55:45), and the flow rate was 1.2 ml/min. The ultraviolet detector was set at 229 nm, and the sensitivity was kept at 0.02 absorbance units full scale. The interassay coefficient of variation was 5.4% (mean, 329 ng/ml; n = 13). Plasma insulin and glucose concentrations were de-

was characterized by the peak time (tma), half-life of absorption (t I/2abs), peak plasma concentration (Cmax), and area under the plasma drug concentration-time curve from 0 to 1/2 hour (AUC(0-1/2)), 0 to hour (AUC(0-1)), 0 to 10 hours (AUC(0-10)), and 0 to infinity (AUC). In addition, estimates of mean residence time (MRT) and elimination plasma half-life (t1,2) were obtained. 1

Insulin and glucose responses. Insulin response was characterized by determining the incremental AUC from 0 to 1/2 hour, 0 to 1 hour, and 0 to 2 hours by the trapezoidal rule. In addition, the maximal increase in insulin concentration together with the tmax were determined. For glucose data, the decremental AUC from 0 to 1/2 hour, 0 to 1 hour, and 0 to 2 hours and the maximal decrease in concentration together with the tmax were determined. Statistical methods. Statistical analyses were made

VOLUME 49

Magnesium hydroxide -ghpizide interaction

NUMBER 1

Table II. Effect of magnesium hydroxide (850 mg) on insulin and glucose responses to glipizide (5 mg)

70

Magnesium hydroxide

50

Control

41

60

40

Insulin Incremental area 0-1/2 hr (mU min/L)

Incremental area 0-1 hr (mU

262 -± 103

485 ± 145*

1243 -± 385

1475 -± 242

1811 -± 493

1941 ± 336

48.7 -± 12.2

49.4 -± 9.5

0.91 ± 0.17

0.56 ± 0.06*

1.96 ± 1.0

7.73 ± 2.0*

33.6 ± 10.1

63.1 ± 12.1

123 ± 25.5

184 ± 20.2

2.19 ± 0.39

2.95 ± 0.28*

30 20

min/L)

Incremental area 0-2 hr (mU min/L) Maximal increase (mU/L) Time of maximal increase (hr)

Glucose Decremental area 0-1/2 hr (mmol min/L) Decremental area 0-1 hr (mmol min/L) Decremental area 0-2 hr (mmol min/L) Maximal decrease (mmol/L) Time of maximal decrease (hr)

10 0 o

30

60

90

120

Time (minutes)

1.2 ± 0.15

1.0 ± 0.07

Data are mean values ± SE in eight subjects. *p < 0.05 compared with control.

Time (minutes)

with the SYSTAT software package (SYSTAT Inc., Evanston, Ill.). The Wilcoxon paired-sample test was used to compare the peak times between the phases. Otherwise, the Student t test (two tailed) for paired values was used. Glipizide concentration values were analyzed by analysis of variance (ANOVA) for repeated measurements. The Student t test (two tailed) for paired values was then employed to determine possible differences in glipizide concentrations between the two phases. The p values < 0.05 were considered to be statistically significant. Results are expressed as SE. mean values

Fig. 2. Plasma insulin and glucose levels (mean ± SE) in eight subjects after intake of 5 mg glipizide with water only (open circles) or with magnesium hydroxide (850 mgsolid circles). *Significant (p < 0.05) difference in change from baseline between treatments.

nesium hydroxide phase at 22, 30, and 45 minutes.

Insulin and glucose responses. The incremental insulin area from 0 to 1/2 hour increased by 85% (p < 0.05) in the magnesium hydroxide phase, but there was no difference in the incremental insulin area from

0 to 1 hour or 0 to 2 hours (Table II; Fig. 2). The RESULTS Glipizide absorption. Magnesium hydroxide accelerated glipizide absorption, as reflected in significant changes in t1/2abs and fractional early AUC values (Table I; Fig. 1). AUC(0-1/2) and AUC(0-1) were increased by 180% and 69%, respectively. There was in the magnesium a tendency toward earlier Cmax, AUC(0-10), (p = 0.07). hydroxide phase

maximal insulin response remained unchanged but occurred earlier (0.56 versus 0.91 hour; p