Pharmacokinetic interaction between proton pump inhibitorsand roxithromycin in volunteers

F. KEES*, A. HOLSTEGE  , K. P. ITTNER*, M. ZIMMERMANN*, G. LOCK  ,

J . SCHOÈ LMERICH  & H. GROBECKER*

*Department of Pharmacology and Clinical Pharmacology, and  Department of Internal Medicine I,

University of Regensburg, Germany

Accepted for publication 13 December 1999

INTRODUCTION

The clinical ef®cacy of a triple therapy for the

treatment of Helicobacter pylori-induced peptic ulcer

disease combining either a proton pump inhibitor or

an H2-receptor antagonist with two antibiotics, one of

which being a macrolide, has been well documented.1

The interaction between proton pump inhibitors and

antibiotics is still of current interest.2 Synergistic

interactions of macrolides with proton pump inhibitors

have been demonstrated in vitro.3 A speci®c antibac-

terial activity of proton pump inhibitors by their own

has been discussed.4 The combination of macrolides

with proton pump inhibitors may result in altered

bioavailability of either drug.5

The aim of the present study was to investigate the

effect of omeprazole or lansoprazole on the steady-

state plasma pharmacokinetics of roxithromycin in

healthy volunteers, and vice versa, and to investigate

the effect of these proton pump inhibitors on the

concentrations of roxithromycin at the infection site in

the stomach.

SUMMARY

Background: Triple therapy including two antibiotics

and a proton pump inhibitor is a rational approach to

the treatment of Helicobacter pylori induced peptic ulcer

disease. The interaction of antimicrobial therapy and

acid suppression is not yet well elucidated.

Aims: To investigate the effects of proton pump inhib-

itors on roxithromycin levels in plasma and gastric

tissue under steady-state conditions in volunteers.

Methods: In two crossover studies omeprazole 20 mg

b.d., lansoprazole 30 mg b.d., roxithromycin 300 mg

b.d., and the combination of roxithromycin with either

omeprazole or lansoprazole were administered to 12

healthy volunteers over 6 days. Blood plasma levels of

the drugs were measured. In addition, roxithromycin

concentrations were also determined in gastric juice

and gastric tissue obtained during endoscopy.

Results: The proton pump inhibitors and roxithromycin

did not alter the blood plasma pharmacokinetics of each

other. When compared to roxithromycin administered

alone, its combination with a proton pump inhibitor

signi®cantly increased the roxithromycin concentra-

tions in gastric juice (3.0±5.0 lg/mL vs. 0.3±0.4 lg/

mL) and gastric tissue (antrum: 3.8±4.0 vs. 2.8 lg/g,

fundus: 5.9±7.4 vs. 4.2±4.4 lg/g).

Conclusions: Proton pump inhibitors and roxithromycin

do not alter the systemic bioavailability of each other.

However, proton pump inhibitors increase the local

concentration of roxithromycin in the stomach which

may contribute to the clinically proven synergic bene-

®cial action in eradication therapy of H. pylori.

Correspondence to: Dr F. Kees, Department of Pharmacology, University of

Regensburg, UniversitaÈtsstrasse 31, D-93053 Regensburg, Germany.E-mail: frieder.kees@chemie.uni-regensburg.de

Aliment Pharmacol Ther 2000; 14: 407±412.

Ó 2000 Blackwell Science Ltd 407

MATERIALS AND METHODS

Subjects

The study was conducted in two parts separated by a

6- to 8-week drug free interval. The combination of

omeprazole with roxithromycin was evaluated in part I,

the combination of lansoprazole with roxithromycin in

part II. Twelve healthy volunteers were enrolled in each

part of the study, seven volunteers participated in both

parts. The volunteers of parts I and II were 24±31 and

21±35 years old, respectively (medians 25 and

27 years, respectively), had body weights of 64±88

and 65±85 kg, respectively (medians 76 and 77 kg,

respectively), and body heights of 174±191 and 175±

188 cm, respectively (medians 183 cm for both). They

had an uneventful medical history, and a normal

physical examination and laboratory pro®le. All volun-

teers were H. pylori-negative as determined by an ELISA

antibody test for IgG and IgA (Pyloriset, Orion Espoo,

Finland). The study was approved by the appropriate

ethics committee and informed written consent was

obtained from all volunteers.

Study design

Each part of the study was conducted according to an

open, randomized, three-way crossover design. In

period 1 omeprazole (Antra, Astra, Wedel, Germany)

20 mg b.d. (part I) or lansoprazole (Lanzor, HMR, Bad

Soden, Germany) 30 mg b.d. (part II) was administered

for 5 days; in period 2, roxithromycin (Rulid, HMR, Bad

Soden, Germany) 300 mg the b.d.; and in period 3,

omeprazole 20 mg/roxithromycin 300 mg b.d. (part I)

or lansoprazole 30 mg/roxithromycin 300 mg b.d.

(part II). The drugs were administered at least 15 min

before food intake together with 150±200 mL of water

except on days of gastroscopy when the volunteers had

to abstain from food until 1 h after gastroscopy. On day

6 of each period the pharmacokinetic pro®le of the

drugs was studied. No other medications were permitted

during the course of each part of the study.

Days of endoscopy

Each subject underwent upper gastrointestinal endo-

scopy 4 h (tissue sampling time: 3.9±5.0 h) following

the morning dose on day 5 of periods 2 and 3. Samples

of venous blood were obtained from each subject before

the morning dose (steady state control), and immedi-

ately before gastroscopy. In addition, 1±2 mL of gastric

juice were collected. Lignocaine spray (30±60 mg) was

routinely administered, and intravenous midazolam

was offered and administered (2±4 mg) in 27 out of

48 cases. Two to four biopsies of gastric mucosa were

collected from both the gastric fundus and the antrum.

Tissue samples were placed on a piece of paper to

remove adherent gastric juice, and weighed immedi-

ately. They amounted to 5±24 mg (mean 14 mg). All

specimens were shock frozen with liquid nitrogen and

stored at )30 °C.

Days of kinetic pro®le

On day 6 of each period the steady state kinetics of

omeprazole, lansoprazole, and roxithromycin in plasma

were established. Six to eight millilitres of venous blood

were obtained prior to drug administration (time 0), and

after 0.33, 0.67, 1, 1.5, 2, 3, 4, 6, 8, 10, 12 and 24 h,

and also after 34 and 48 h for roxithromycin assay. The

drugs were administered while fasting between 07.30

and 08.00 hours with 200 mL of water. Two hours

later a low-fat continental breakfast was served, lunch

4.5 h later, a snack 7 h later, and dinner 10 h after

drug administration. Smoking and drinking alcoholic

and caffeine beverages were not allowed from 12 h

before and up to 24 h after drug administration.

Analytical methods

All samples were analysed by reversed-phase high-

performance liquid chromatography (HPLC) and

photometric (omeprazole, lansoprazole) or coulometric

(roxithromycin) detection according to published

methods.6, 7 In order to determine the stability of

roxithromycin in acidic environment, solutions of

roxithromycin 50 lg/mL were incubated at pH 1.0,

2.0, 3.0, 5.2 and 6.0 (100, 10, 1 mM HCl or 10 mM

sodium phosphate buffer, respectively) at 37 °C in the

autosampler of the HPLC apparatus. Samples were

withdrawn automatically and analysed for roxithromy-

cin by HPLC.

Pharmacokinetic analysis and statistical evaluation

The steady-state pharmacokinetics of omeprazole, lan-

soprazole and roxithromycin were assessed by standard

non-compartmental methods. The parameters, area

under the plasma concentration±time curve at steady

state (AUCss), peak concentration (Cmax) and terminal

408 F. KEES et al.

Ó 2000 Blackwell Science Ltd, Aliment Pharmacol Ther 14, 407±412

half-life (t1/2) as well as the concentrations of roxithro-

mycin in gastric juice and gastric mucosa were

compared by the two-sided paired t-test, regarding

P < 0.05 as signi®cant.

RESULTS

Side-effects and tolerability

In general, study medications were well tolerated. Only

a few mild adverse events were reported during a total

of 504 days of drug treatment: seven volunteers in part

I and four in part II reported episodes of gastrointestinal

disturbance (four in part I had a stomach upset, one

suffered from abdominal fullness, two volunteers in both

parts I and II suffered from nausea, two in part I and

one in part II suffered from bloating, one in part I and

two in part II suffered from diarrhoea, one in part II

suffered from vomiting) and one volunteer from both

parts I and II had an episode of headache.

Pharmacokinetic parameters in plasma

The pharmacokinetic parameters of roxithromycin were

similar in both parts I and II, and were also similar with

and without co-administered proton pump inhibitors

(Table 1). The bioavailability parameters Cmax and the

AUC of either drug were somewhat higher during the

co-administration regimen, but the differences were not

statistically signi®cant compared to the regimen where

the drugs were administered alone.

Concentrations of roxithromycin in gastric mucosa

Both proton pump inhibitors raised gastric pH

(mean � standard deviation) signi®cantly (P < 0.001)

from 2.5 � 1.1 to 6.3 � 0.9 (omeprazole) and from

3.0 � 1.7 to 6.0 � 1.2 (lansoprazole). Gastric tissue

concentrations of roxithromycin were signi®cantly

higher when given together with a proton pump

inhibitor. In addition, biopsy samples taken from the

fundus had signi®cantly higher roxithromycin concen-

trations than antrum biopsies. Furthermore, the roxi-

thromycin concentrations in gastric juice were higher

during concomitant treatment with omeprazole or

lansoprazole (Figure 1).

Stability and lipophilicity of roxithromycin

in aqueous solution

Roxithromycin disintegrated at 37 °C according to ®rst-

order kinetics with a half-life of 5 min at pH 1, and of

1.3 h at pH 2. Only 25% degradation was observed

within the incubation period of 5.5 h at pH 3, and no

degradation at pH 5.2 and pH 6.0.

As depicted in Figure 2, the lipophilicity of roxithro-

mycin is poor at low pH, but increases 1000-fold from

pH 5±8. Log D of octanol±water partition coef®cient

amounts to )0.75 at pH 5, 1.8 at pH 7.4, and 2.5 at

pH 8, i.e. about 15% of roxithromycin are in the octanol

layer at pH 5, about 90% at pH 7.4, and 97% at pH 8.

DISCUSSION

Bioavailability of omeprazole or lansoprazole

and roxithromycin during co-administration

Concurrent administration of roxithromycin and omep-

razole or lansoprazole did not signi®cantly in¯uence

blood plasma concentration±time pro®les of either drug.

In contrast to these ®ndings, the combined administra-

tion of omeprazole and clarithromycin resulted in

higher and more prolonged concentrations of omepraz-

ole in blood plasma than did the ingestion of omeprazole

Table 1. Plasma pharmacokinetic param-

eters (mean � standard deviation) of

omeprazole, lansoprazole and

roxithromycin in 12 healthy male

volunteers on day 6 of treatment alone or

with co-administered roxithromycin or

proton pump inhibitor, respectively

Treatment Drug Cmax (lg/mL) tmax (h) t1/2 (h)

AUCss

(lg ´ h/mL)

Ome Ome 0.59 � 0.28 1.3 � 0.5 1.1 � 0.5 1.34 � 0.92

Ome/Roxi Ome 0.67 � 0.39 1.0 � 0.2 1.2 � 0.7 1.75 � 1.58

Lanso Lanso 0.68 � 0.34 1.4 � 0.6 1.2 � 0.4 1.62 � 0.99

Lanso/Roxi Lanso 0.76 � 0.32 1.4 � 0.6 1.2 � 0.4 1.76 � 1.03

Roxi (part I) Roxi 10.4 � 2.1 1.8 � 0.7 12.3 � 1.3 88 � 21

Roxi/Ome Roxi 11.4 � 2.4 1.1 � 0.8 12.3 � 1.8 102 � 23

Roxi (part II) Roxi 10.5 � 2.5 1.8 � 0.9 13.4 � 1.8 87 � 25

Roxi/Lanso Roxi 10.6 � 1.7 1.9 � 0.8 13.3 � 3.0 93 � 16

Dosage: omeprazole 20 mg, lansoprazole 30 mg, roxithromycin 300 mg, b.d. each.

Ome, omeprazole; Lanso, lansoprazole; Roxi, roxithromycin.

ROXITHROMYCIN AND PROTON PUMP INHIBITORS 409

Ó 2000 Blackwell Science Ltd, Aliment Pharmacol Ther 14, 407±412

alone.5 Apparently, clarithromycin signi®cantly inhibits

the oxidative metabolism of omeprazole by CYP3 A4.

On the other hand, our study con®rms the poor

potential of roxithromycin to interact with the oxidative

metabolism of other drugs.8

Concentrations of roxithromycin in gastric juice

When administered alone, the mean roxithromycin

concentrations after a dose of 300 mg b.d. were

0.2±0.4 lg/mL in gastric juice, but tenfold higher in

the presence of a proton pump inhibitor. The higher

concentrations with co-medication can be explained

both by the reduced gastric juice volume induced by

proton pump inhibitors, and by the instability of

macrolides in an acidic environment.9 Despite much

better acid stability compared to erythromycin, roxi-

thromycin is disintegrated within a few minutes at pH 1

and 37 °C. As with clarithromycin, roxithromycin

shows good stability only at pH > 3.10

Accordingly, the higher concentrations of roxithromy-

cin in gastric juice with the roxithromycin/proton pump

inhibitor regimen may be explained at least partially by

degradation of roxithromycin at low gastric pH. After

intravenous infusion of clarithromycin, equal concen-

trations of the macrolide were measured in gastric juice

when clarithromycin was administered alone or with

concurrent omeprazole.11 However, in that study the

gastric juice was collected in 15-min intervals. There-

fore, inactivation of the macrolide in the acidic envi-

ronment was small compared to our design, where only

one specimen of pooled gastric juice was collected 4 h

after drug administration.

Concentrations of roxithromycin in gastric mucosa

The mean concentrations of roxithromycin were

2.8 lg/g in antral tissue, 4.2±4.4 lg/g in fundic tissue,

Figure 1. Concentrations (mean, standard deviation) of roxi-

thromycin in gastric juice and gastric tissue in 12 healthy male

volunteers 4 h after the morning dose on day 5 of treatment with

roxithromycin 300 mg b.d., when given alone and with co-

administered omeprazole 20 mg b.d (A) or lansoprazole 30 mg

b.d. (B), respectively. Note that the differences between the

concentrations in the antrum and fundus are signi®cant

(P < 0.001). * P < 0.05; ** P < 0.01; *** P < 0.001.

Figure 2. Effect of pH on log Doctanol±water for roxithromycin.

Solutions of roxithromycin 50 lg/mL in 50 mM sodium phos-

phate of pH 3±10 were equilibrated at room temperature with

equal volumes of octanol. Roxithromycin was determined in the

aqueous layer by HPLC with ultraviolet detection at 210 nm

before and after equilibration. From the results the octanol±water

partition coef®cient of roxithromycin was calculated.

410 F. KEES et al.

Ó 2000 Blackwell Science Ltd, Aliment Pharmacol Ther 14, 407±412

and 30±50% higher with concomitant proton pump

inhibitors. Similar results (although the differences were

not statistically signi®cant) were obtained with clari-

thromycin.5 However, the higher plasma concentra-

tions of clarithromycin with concomitant omeprazole

may have accounted for its higher tissue concentra-

tions.

In contrast, the blood plasma concentrations of

roxithromycin sampled immediately before gastroscopy

were only signi®cantly higher (P < 0.05) during the

combined omeprazole/roxithromycin but not during the

combined lansoprazole/roxithromycin regimen, com-

pared to roxithromycin alone. In addition, no statisti-

cally signi®cant difference was found at all sampling

times on day 6 when the plasma concentration±time

pro®le was established (data not shown). Therefore, the

signi®cant difference may have been caused incidentally

by the greater variation and difference in blood vs.

tissue sampling time. We conclude that the higher

concentrations of roxithromycin in gastric tissue during

combined roxithromycin/proton pump inhibitor treat-

ment cannot be explained by a higher systemic

bioavailability of roxithromycin.

It appears that omeprazole or lansoprazole exhibit a

favourable effect on the local concentrations of roxi-

thromycin in the stomach; several factors may contri-

bute to this result. First of all, we cannot exclude that

the higher concentrations of roxithromycin in gastric

juice during proton pump inhibitor treatment may

partially account for the elevated tissue levels of

roxithromycin. It is nearly impossible to separate the

biopsy tissue specimens clearly from adherent mucus

and/or gastric juice. Secondly, in normal physiological

conditions there is a strong pH gradient between blood

plasma or tissue and gastric lumen. Weak bases such as

the macrolides clarithromycin and roxithromycin

(pKa � 9.2) are totally protonated in the acidic gastric

lumen; they change into hydrophilic character, and

mass transfer is uni-directional from plasma or tissue to

gastric lumen. During treatment with a proton pump

inhibitor the luminal pH rises to 6 and even higher. The

macrolides take on lipophilic character, as has been

proven by the octanol±water partition coef®cient. The

driving force of the pH gradient from blood and tissue to

gastric lumen is reduced, and higher concentrations of

macrolides in gastric tissue could result in the presence

of proton pump inhibitors.

The results are in agreement with an ion-trapping

mechanism, as has been employed to explain the high

intracellular accumulation of macrolides and other

basic drugs, or to explain the accumulation of nicotine

in gastric juice after transdermal administration.12±16

The results are also in agreement with an active or

carrier mediated process as has recently been concluded

from studies with clarithromycin.17

We have also shown that tissue concentrations of

roxithromycin in the antrum were signi®cantly lower

than in the fundus which may be explained by (i)

enrichment of roxithromycin in the acid-containing

parietal cells present in the fundus and corpus but not

in the antrum, and (ii) a wash-out effect of gastric

juice which is collected in the corpus and antrum but

not in the fundus. For both reasons higher tissue

concentrations will result in the fundus compared to

the antrum.

In conclusion, consistent results from this two-part

study in healthy volunteers demonstrate that proton

pump inhibitors and roxithromycin do not alter the

systemic bioavailability of each other, for example by

inhibition of the metabolism of the proton pump

inhibitor or by enhancing the intestinal absorption of

roxithromycin. However, proton pump inhibitors

increase the local concentration of roxithromycin in

the stomach which may contribute to the clinically

proven synergism of combined administration of roxi-

thromycin and proton pump inhibitors in eradication

therapy of H. pylori.18±20

ACKNOWLEDGEMENTS

This study was supported by a grant from Hoechst

Marion Roussel, Bad Soden, Germany.

REFERENCES

1 Malfertheiner P, Megraud F, O'Morain C, et al. Current

European concepts in the management of Helicobacter pylori

infectionÐthe Maastricht Consensus Report. Eur J Gastroen-

terol Hepatol 1997; 9: 1±2.

2 Lind T, Megraud F, Unge P, et al. The Mach2 Study: Role of

omeprazole eradication Helicobacter pylori with 1-week triple

therapies. Gastroenterology 1999; 116: 248±53.

3 Malizia T, Tejada M, Marchetti F, et al. Synergic interactions of

macrolides and proton-pump inhibitors against Helicobacter

pylori: a comparative in-vitro study. J Antimicrob Chemother

1998; 41(Suppl. B): 29±35.

4 KuÈ hler TC, Swanson M, Shcherbuchin V, Larsson H, Mellgard

B, SjoÈstroÈm JE. Structure±activity relationship of 2-[[(2-pyr-

idyl) methyl]thio]-1H-benzimidazoles as Helicobacter pylori

agents in vitro and evaluation of their in vivo ef®cacy. J Med

Chem 1998; 41: 1777±88.

ROXITHROMYCIN AND PROTON PUMP INHIBITORS 411

Ó 2000 Blackwell Science Ltd, Aliment Pharmacol Ther 14, 407±412

5 Gustavson LE, Kaiser JF, Edmonds AL, Locke CS, DeBartolo

ML, Schenck DW. Effect of omeprazole on concentrations of

clarithromycin in plasma and gastric tissue at steady state.

Antimicrob Ag Chemother 1995; 39: 2078±83.

6 Karol MD, Granneman GR, Alexander K. Determination of

lansoprazole and ®ve metabolites in plasma by high-perfor-

mance liquid chromatography. J Chromatogr B 1995; 668:

182±6.

7 Kees F, Spangler S, Wellenhofer M. Determination of macro-

lides in biological matrices by high-performance liquid chro-

matography with electrochemical detection. J Chromatogr A

1998; 812: 287±93.

8 V, Rosenstiel NA, Adam D. Macrolide antibacterials. Drug

interactions of clinical signi®cance. Drug Saf 1995; 13:

105±22.

9 Lambert JR. Pharmacology of the gastric mucosa: a rational

approach to Helicobacter polytherapy. Gastroenterology 1996;

111: 521±3.

10 Erah PO, Goddard AF, Barrett DA, Shaw PN, Spiller RC. The

stability of amoxycillin, clarithromycin and metronidazole in

gastric juice: relevance to the treatment of Helicobacter pylori

infection. J Antimicrob Chemother 1997; 39: 5±12.

11 Goddard AF, Jessa MJ, Barrett DA, et al. Effect of omeprazole

on the distribution of metronidazole, amoxicillin, and clari-

thromycin in human gastric juice. Gastroenterology 1996;

111: 358±67.

12 Shore PA, Brodie BB, Hogben CAM. The gastric secretion of

drugs: a pH partition hypothesis. J Pharmacol Exp Ther 1957;

119: 361±8.

13 Knowles DJ. Uptake of erythromycin by McCoy and Hep2

cells: its dependence on cellular pH gradients. J Antimicrob

Chemother 1988; 21: 765±72.

14 Bouvier d'Yvoire MJ, Dresco IA, Tulkens PM. Computer-aided

prediction of macrolide antibiotic concentrations in human

circulating polymorphonuclear leucocytes. J Antimicrob

Chemother 1998; 41(Suppl. B): 63±8.

15 Daniel WA, Bickel MH, Honegger UE. The contribution of

lysosomal trapping in the uptake of desipramine and chlo-

roquine by different tissues. Pharmacol Toxicol 1995; 77:

402±6.

16 Lindell G, Lunell E, Graffner H. Transdermal administered

nicotine accumulates in gastric juice. Eur J Clin Pharmacol

1996; 51: 315±8.

17 Goddard AF. Review article: factors in¯uencing antibiotic

transfer across the gastric mucosa. Aliment Pharmacol Ther

1998; 12: 1175±84.

18 Okada M, Oki K, Shirotani T, et al. A new quadrupole therapy

for the eradication of Helicobacter pylori. Effect of pretreatment

with omeprazole on the cure rate. J Gastroenterol 1998; 33:

640±5.

19 Pohle T, Stoll R, Kirchner T, et al. Eradication of Helicobacter

pylori with lansoprazole, roxithromycin and metronidazole±

an open pilot study. Aliment Pharmacol Ther 1998; 12:

1273±8.

20 Okada M, Nishimura H, Kawashima M, et al. A new quad-

rupole therapy Helicobacter pylori: In¯uence of resistant strains

on the treatment outcome. Aliment Pharmacol Ther 1999;

13: 769±74.

412 F. KEES et al.

Ó 2000 Blackwell Science Ltd, Aliment Pharmacol Ther 14, 407±412