+

Interaction of Norfloxacin with Divalent and Trivalent Pharmaceutical Cations.

In Vitro Complexation and in Vivo Pharmacokinetic Studies in the Dog
STEVEN C. WALLIS*, BRUCE G. CHARLESX, LAWRENCE R. GAHAN*, LUCIO J. FILIPPICH, MEGAN G. BREDHAUER,
PAUL A. DUCKWORTH

AND

Received February 20, 1996, from the *Department of Chemistry, The University of Queensland, Brisbane, Queensland, Australia,
Department of Pharmacy, The University of Queensland, Brisbane, Queensland, Australia, Department of Companion Animal Medicine
and Surgery, The University of Queensland, Brisbane, Queensland, Australia, and the Department of Chemistry, Queensland University of
Technology, Brisbane, Queensland, Australia.
Accepted for publication May 10, 1996X.
Abstract 0 The formation constants of the fluoroquinolones norfloxacin
and ciprofloxacin with Mg2+ (log 1 ) 2.97(4), log 2 ) 5.6(2)), Zn2+
(log 1 ) 3.77(2), log 2 ) 7.59(3)), and Fe2+ (log 1 ) 3.99(5), log 2
) 7.2(5)) were determined by potentiometric titration. The pH at which
precipitation occurred in the titration solutions was compared for the metal
ions Ca2+, Mg2+, Zn2+, Fe2+, Cu2+, and Al3+. The formation constants
were used to predict a rank order of metals that may be expected to
hinder the gastrointestinal absorption of the fluoroquinolones, in vivo. The
effects of metal ions on the pharmacokinetics of orally-administered
norfloxacin in the dog were investigated. Norfloxacin (12 mg/kg) was
administered alone or with equimolar doses of each of the chloride
salts of Ca2+, Mg2+, Zn2+, Fe2+, and Al3+. Statistically significant reductions
in serum norfloxacin concentrations were observed after analysis by HPLC.
The Cmax was reduced 2985%, while the area under the norfloxacin
serum concentrationtime curve (AUC0-) was reduced by 2979%. The
extent of the reduction in AUC0- was correlated with the magnitude of
the formation constant of the 1:1 norfloxacin:metal chelate complex for
the divalent metal ions. On coadministration of 12 mg/kg norfloxacin
with various doses of Mg2+ (chloride) the AUC0- and Cmax decreased
with increasing Mg2+ dose. The interaction peaked at a Mg2+:norfloxacin
ratio of 1:2, suggesting the formation of a 1:2 Mg:norfloxacin complex.
Formation constant data were used to simulate the percentage of
norfloxacin complexed at pH 6.5. Combinations of metal ion and
norfloxacin which result in only a small extent (<20%) of norfloxacin
complex formation can result in relatively large decreases in oral
bioavailability of this antimicrobial agent.

Introduction
A number of orally-administered drugs suffer markedly
impaired absorption when coadministered with products containing divalent and trivalent ions, particularly the metallic
cations contained in iron supplements, antacids, and vitaminmineral preparations. For instance, significant detrimental
effects on the bioavailability of tetracyclines,1 levodopa,2 and
penicillamine3 have been reported on coadministration with
metal ion containing preparations, while a recent review4
noted a range of candidate drugs for potential interactions
with iron salts.
Increasing attention has been focused on the pharmacokinetic interaction between metal ions and the fluoroquinolone
group of drugs.5 These antimicrobials are presently enjoying
widespread clinical application, particularly against some
serious pathogens which previously were difficult to treat.
They offer advantages over many alternative treatments
because they are well tolerated by most patients, they are well
absorbed when taken by mouth, and they need only be
administered twice a day. Norfloxacin, a widely-prescribed
X

Abstract published in Advance ACS Abstracts, July 1, 1996.

1996, American Chemical Society and

American Pharmaceutical Association

representative of this group, suffers reduced bioavailability

when administered with aluminum contained in antacids6 and
sucralfate,7,8 and when given with either ferrous sulfate,9,10
zinc sulfate,11 or magnesium salts.9 Calcium-containing dairy
products12 and calcium carbonate6 also significantly reduced
the oral availability of norfloxacin. Hitherto, it has been
difficult to assess objectively the comparative effects of cations
on norfloxacin pharmacokinetics because of major differences
in experimental design, formulation, and dose among the
various published studies.
The purpose of the present study was, firstly, to measure
the in vitro formation constants of several common pharmaceutical cations (Ca2+, Mg2+, Zn2+, Fe2+, and Al3+) with
norfloxacin and, secondly, to determine the effects of these
ions on the pharmacokinetics of norfloxacin in the dog. Of
special interest was the existence of any correlation between
the formation constants13,14 and parameters describing the
oral absorption of norfloxacin following normal clinical doses.

Experimental Section
Chemicals and SolutionssNorfloxacin was obtained from Sigma
Chemical Co. (St. Louis, MO) and used as received. Water was glassdistilled and further purified by a Milli-Q water purifier. Solutions
of CaCl2, ZnCl2, and MgCl2 were prepared by dissolution of the AR
grade salts in water. Solutions of FeCl2 and AlCl3 were prepared by
dissolution of elemental iron and aluminum, respectively, in concentrated HCl and dilution to 0.1 M with water. The Fe(ClO4)3 solution
was prepared by dissolution of the low chloride Fe(ClO4)3 in 0.1 M
HClO4. Standardization of the metal chloride and perchlorate solutions was performed by EDTA compleximetric titration.15
Formation ConstantssThe formation constants of norfloxacin
with magnesium(II), zinc(II), and iron(II) were determined by potentiometric titration at 25 C. The pKa for norfloxacin (carboxylic acid
pKa ) 6.30(2), piperazinyl amine pKa ) 8.5(1)) and ciprofloxacin
(carboxylic acid pKa ) 6.18(1), piperazinyl amine pKa ) 8.5(1)) were
determined previously.16 Preliminary titrations were also performed
with calcium(II), iron(III), and aluminum(III). The potentiometric
titrations were carried out under an inert atmosphere of watersaturated nitrogen in a water-jacketed vessel maintained at 25.0 C
or 37.0 C. Data were obtained from 10 mL aliquots of solutions
containing 5.0 10-3 M HCl, 0.15 M NaCl, and 1.0 10-3 M
norfloxacin titrated with a standardized solution of NaOH at approximately 0.1 M. Formation constant data were gathered for
solutions to which a 0.1 M metal chloride (CaCl2, MgCl2, ZnCl2, FeCl2,
AlCl3) solution was added so as to give a metal-to-ligand ratio in the
range 2:1 to 1:3. Measurements were commenced at pH 2 and
finished either at pH 11 or on precipitation. The ionic strength was
held constant in the titrations through the presence of 0.15 M NaCl
supporting electrolyte. At least three titrations, with different metalto-ligand ratios, were performed for each system.
Measurements with iron(III) were conducted with the analogous
perchlorate salt solutions, rather than the chloride salts. No supporting electrolyte was used as 0.15 M NaNO3 and 0.15 M NaClO4
reduced the solubility. Although it was observed that in the determination of the formation constants for norfloxacin with copper(II),
the presence of 0.15 M NaCl supporting electrolyte and the

S0022-3549(96)00087-1 CCC: $12.00

Journal of Pharmaceutical Sciences / 803

Vol. 85, No. 8, August 1996

absence of a supporting electrolyte had no effect on the calculated

value of log , the absence of the supporting electrolyte decreased
the pH at which precipitation occurred.
A Metrohm E665 Dosimat autoburet equipped with a 5 cm3 buret
was used to deliver the titrant, and the potential was measured by
an Orion Ross Sure Flow 81-65BN combination electrode (containing
3 M NaCl as the filling solution) connected to an Orion 290A pH
meter. The autoburet and pH meter were interfaced to an IBM
compatible personal computer which controlled the addition of titrant,
using a locally written program, so that successive additions of titrant
caused a decrease of approximately 4 mV. The electrode was
calibrated by a strong acid strong base titration, and the resulting
data were used to calculate E0 and pKw. The pKa, E0, pKw, and
formation constants were determined using Superquad17 running on
an IBM compatible computer.
Animals and Drug AdministrationsWritten ethical approval
was obtained from the University of Queensland Animal Experimental Ethics Committee. Young, adult, mixed-breed dogs weighing
between 13 and 25 kg were selected for the study. All were vaccinated
(Canvac 3/1, CSL Ltd, Parkville, Victoria, Australia) against distemper, infectious hepatitis, and parvovirus. The dogs were acclimatized
for 10 days in the Experimental Dog Ward before the study began.
All were fasted for 16 h before drug administration and allowed to
eat commercial canned dog food 6-8 h after dosing. On a study day,
acepromazine (0.2 mg/kg; Promex 2, Apex Laboratories Pty Ltd, St
Marys, New South Wales, Australia) was administered subcutaneously not less than 1 h prior to dosing as a tranquilizer.
Norfloxacin was dissolved in 100 mL of 0.01 M HCl alone, or in
combination with the metal chloride salts, and administered orally
via a stomach tube, followed by 10 mL of 0.01 M HCl. All
treatments were administered in an open, randomized, crossover
fashion. Administration of individual treatments was separated by
a 7-day washout period (equivalent to >35 norfloxacin terminal
elimination half-lives). There were two parts to the investigation.
In part 1, norfloxacin (12 mg/kg; 37.6 mol/kg) was given alone or
with an equimolar amount of AlCl3, CaCl2, FeCl2, ZnCl2, or MgCl2
(37.6 mol/kg). In part 2, norfloxacin was administered alone (12 mg/
kg; 37.6 mol/kg) or in combination with MgCl2 in molar ratios of
Mg2+:norfloxacin of 0.25:1, 0.5:1, 1:1, 2:1, and 4:1.
Blood SamplingsBlood samples (1.5 mL) were collected via a 20G
catheter inserted into the cephalic vein, or by isolated venipuncture
from a jugular vein. Samples were drawn just before dosing, then at
0.25, 0.5, 0.75, 1, 1.5, 2, 2.5, 3, 4, 6, 8, 12, 24, and 36 h after dosing.
Sera were isolated by centrifugation and stored at <-70 C. At the
completion of the studies the animals were humanely euthanized with
sodium pentobarbital and necropsied.
Drug AssaysSerum norfloxacin concentrations were measured by
a reversed-phase HPLC method developed in this laboratory.18
Briefly, samples (100 L) were extracted with chloroform (1 mL)
containing an internal standard, N-ethylnorfloxacin (2.6 g), and
chromatographed on an RP-18 Spheri-3 column (3.2 mm i.d. 40
mm, 3 m particles; Applied Biosystems, San Jose, CA) in conjunction
with a mobile phase of 11% (v/v) acetonitrile in triethylamine (0.001
M) and pH 2.5 phosphate buffer (0.01 M) pumped at 1 cm3/min. The
detection wavelength was 279 nm, and the limit of quantification was
0.1 mg/L. Within-day and between-day imprecision (coefficient of
variation, %) was e8.6%, and inaccuracy was e5.3%.
Pharmacokinetic and Statistical AnalysissPeak serum concentrations (Cmax) and time to reach the peak (Tmax) were recorded
from the observed data. The terminal elimination half-life (T1/2), mean
residence time (MRT), and area under the serum norfloxacin concentration-time curve (AUC0-) from time zero to infinity were calculated in a spreadsheet format by standard formulas.19 Analyses of
variance (ANOVA) were performed using the partial sums of squares
general linear modeling (GLM) procedure in the SAS/STAT software
package.20 The level of significance (P) was set at 0.05.

Results
Potentiometric TitrationssThe formation constants of
norfloxacin and ciprofloxacin with magnesium(II), zinc(II), and
iron(II) were determined by potentiometric titration (Table
1). The potentiometric titrations commenced at pH 2 and
finished either at pH 11 or on formation of a precipitate. The
804 / Journal of Pharmaceutical Sciences
Vol. 85, No. 8, August 1996

Table 1sFormation Constants (log )a of Major Species for Norfloxacin

with Various Metal Ions and Percentage of Norfloxacin Complexed at pH
6.5 Calculated from Formation Constantsb
Metal
Ca2+
Mg2+
Zn2+
Fe2+
Al3+
Al3+

ML

ML2

ML3

2.2
2.97(4) 5.6(2)
3.77(2) 7.59(3)
3.99(5) 7.2(5)
7.03
12.47
17.92
6.11

M(LH) M(LH)(OH)3 % Compc Refd

0.34

23.34

8.5
35
81
78
99
56

13
TSe
TSe
TSe
13
14

a is the overall formation constant. For the formation of ML , ) [ML ]/

2
2
[M][L]2. b LH and L are the deprotonated and monoprotonated forms of
c
fluoroquinolone, respectively. % comp is the calculated percentage of norfloxacin
complexed at pH 6.5. d Reference for formation constants. e This study.

complexation was modeled on the formation of ML and ML2

species, following the scheme adopted for norfloxacin complexation with copper(II).16 Here, L represents the zwitterionic form of the fluoroquinolone, where the carboxylic acid
is ionized and the piperazinyl secondary amine is protonated.
On deprotonation of the piperazinyl amine of L, the L-H
species forms. With increasing pH it was likely that the
piperazinyl amine of the complexed ligand would be deprotonated, forming M(L-H) from ML, and ML(L-H) and M(LH)2 from ML2 (Figure 1). The deprotonation of either ML or
ML2 with increasing pH was required for a better-fitting
model, although the deprotonation products of ML and ML2
were difficult to distinguish; whichever deprotonation product
was modeled had no effect on the formation constants of the
ML and ML2 species. For some titrations (in particular, those
involving the ferrous ion) precipitation occurred prior to
formation of appreciative amounts of the deprotonation species, hence excluding the species from the model and obviating
the determination of formation constants for these species.
For zinc titrations, the formation of zinc hydrolysis products
was included in the model.21 ML3 species were omitted from
the model as there was no gain from their inclusion, nor was
there evidence for their existence.
Preliminary titrations were also performed with Ca2+, Fe3+,
and Al3+. Titrations with solutions containing Ca2+ produced
negligible deviations from the ligand-alone titration curve,
indicating that the extent of complexation was low. The
titration curves with Fe3+ and Al3+ produced greater deviations from the ligand-alone titration curve than any of the
other metal ions studied, suggesting a greater amount of
ligand complexed. Iron(III) solutions were prepared free of
chloride due to its great affinity for that anion.22
In Vivo StudiessAll treatments were well tolerated by the
study animals, and no abnormalities were noted at necropsy.
The dose of norfloxacin administered was similar to that used
in previous norfloxacin pharmacokinetic studies in dogs23 and
in humans.24 In all studies Tmax was 1.5 h. The mean
terminal elimination half-life was 4.6 h. In the first study,
serum norfloxacin concentrations measured alone and after
coadministration of identical doses of five individual cations
in a 6 6 crossover design were plotted against time (Figure
2). Summary pharmacokinetic results are presented in Table
2. Three-way ANOVAs indicated a statistically significant
reduction among treatments with respect to both the Cmax and
the AUC0- of norfloxacin. The most pronounced differences
were observed after norfloxacin was administered alone and
in conjunction with AlCl3. There was a 7-fold reduction (P <
0.05) in the mean (( standard error) value for Cmax (2.4 ( 0.4
to 0.35 ( 0.06 mg/L), and there was an almost 5-fold decrease
(P < 0.05) in AUC0- (17.0 ( 3.0 to 3.6 ( 0.9 mg/Lh). The
formation constant (log 1) for norfloxacin with each metal
was plotted against the reduction in AUC0- obtained upon
coadministration with that metal (Figure 3); previously

Figure 1sScheme of complexation and deprotonation adopted for determination of formation constants for norfloxacin and ciprofloxacin with Mg2+, Zn2+, and Fe2+ (only
one section of the metalnorfloxacin complex is shown for clarity).

Figure 2sSerum norfloxacin concentration versus time profiles for 12 mg/kg

norfloxacin administered alone and with equimolar amounts of various metal
ions: Ca2+, Mg2+, Zn2+, Fe2+, and Al3+. Values are mean standard error.
Table 2sPharmacokinetics of Norfloxacin in Dogs following
Administration of a Single Oral Dose of Norfloxacin (12 mg/kg ) 37.6
mol/kg) Alone and in Combination with Five Different Cations (37.6
mol/kg)a
Cation

Cmax (mg/L)

Tmax (h)

T1/2 (h)

AUC0- (mg/Lh)

MRT (h)

Noneb
Ca2+
Mg2+
Zn2+
Fe2+
Al3+

2.4 (0.4)
1.7 (0.3)
0.92 (0.05)
0.8 (0.1)
0.58 (0.09)
0.35 (0.06)

1.5 (0.2)
1.6 (0.5)
1.6 (0.3)
2.2 (0.6)
2.5 (0.4)
2.9 (0.4)

4.6 (0.3)
4.7 (0.4)
5.1 (0.6)
6.0 (0.7)
5.0 (0.2)
5.7 (0.8)

17 (3)
12 (2)
6.8 (0.5)
7 (2)
5.0 (0.8)
3.6 (0.9)

7.2 (0.4)
7 (2)
7.9 (0.8)
10 (1)
8.5 (0.3)
9 (1)

Figure 3sPlot of formation constant (log 1) for ML complex of norfloxacin with

metal ions versus reduction in norfloxacin AUC0- after coadministration of
equimolar doses of those metal ions. (Al3+(A), Djurdjevic et al.,14 Al3+(B),
Okabayashi et al.13).

Values are mean standard error. b Norfloxacin in the absence of cations.

published data13,14 for Al3+ was included in the plot. The

respective mean Tmax, MRT, and T1/2 values did not differ
significantly among the six treatments, although a trend to
slightly greater Tmax values and a decrease in Cmax were
apparent. There were no statistically significant differences
among the study animals, nor was there a statistically
significant effect with respect to the order in which the
treatments were administered for any of the pharmacokinetic
parameters reported.
The second part of the study examined the effect of five
different doses of a selected metal ion, Mg2+ (as MgCl2), on
the oral absorption of norfloxacin in the dog. In an incomplete
block, unbalanced design, six dogs received norfloxacin alone
and in combination with a 1:1 molar dose of MgCl2:norfloxacin
on different occasions, while additional data was obtained in
four dogs who received 0.25:1, 0.5:1, 2:1, and 4:1 molar dose
combinations on separate occasions. Both subgroups of
animals received the treatments as a crossover in random
order. Mean serum norfloxacin-time plots are illustrated in
Figure 4, while the pharmacokinetic parameters are shown
in Table 3. The oral absorption of norfloxacin decreased in
rank order with increasing magnesium doses. For example,
a dose of 0.25:1 (Mg2+:norfloxacin) reduced the Cmax obtained

Figure 4sSerum norfloxacin concentration versus time plots after norfloxacin

(12 mg/kg) was administered alone and with magnesium chloride in Mg2+:
norfloxacin ratios of 0:1, 0.25:1, 0.5:1, 1:1, 2:1, and 4:1. Values are mean
standard error.

for norfloxacin alone by one-third, while a 4:1 ratio reduced

the Cmax by an average value of 73%. Likewise, the 0.25:1
dose ratio reduced the average AUC0- for norfloxacin alone
by 41%, while the 4:1 ratio reduced the AUC0- by 76%.
Reductions in norfloxacin AUC0- obtained when different
amounts of Mg2+ were administered with a fixed dose of
norfloxacin are illustrated in Figure 5. The ANOVAs showed
that there were insignificant differences among dogs, and for
the order of treatment administration for any pharmacokinetic
parameter. Furthermore, no significant differences were
observed with respect to treatment for Tmax, T1/2, or MRT.

Journal of Pharmaceutical Sciences / 805

Vol. 85, No. 8, August 1996

Table 3sPharmacokinetics of Norfloxacin in Dogs following

Administration of a Single Dose of Norfloxacin (12 mg/kg ) 37.6
mol/kg) Alone and in Combination with Five Different Molar Ratios of
Magnesium Chloridea
Mg2+:Norfloxacin

Cmax
(mg/L)

Tmax
(h)

T1/2
(h)

AUC0-
(mg/Lh)

MRT
(h)

0.0:1
0.25:1
0.5:1
1.0:1
2.0:1
4.0:1

2.4 (0.4)
1.6 (0.3)
1.2 (0.2)
0.92 (0.05)
0.72 (0.05)
0.55 (0.08)

1.5 (0.2)
0.9 (0.2)
0.8 (0.1)
1.6 (0.3)
1.5 (0.4)
1.8 (0.4)

4.6 (0.3)
4.5 (0.4)
4.4 (0.3)
5.1 (0.6)
5.1 (0.5)
4.8 (0.6)

17 (3)
10 (2)
7 (1)
6.8 (0.5)
5.9 (0.8)
4.0 (0.8)

7.2 (0.4)
6.8 (0.6)
6.7 (0.3)
7.9 (0.8)
8.0 (0.8)
7.6 (0.6)

Values are mean standard error.

Figure 7sPlot of percentage of norfloxacin complexed at pH 6.5 with Mg2+

(calculated from formation constants) versus reduction in AUC0-.

Figure 5sPlot of Mg2+:norfloxacin dose ratio versus reduction in norfloxacin

AUC0-.

Figure 8sPlot of percentage norfloxacin complexed with various metal ions versus
pH.

versus pH are shown for equimolar ratios of six cations with

norfloxacin (Figure 8) and for different ratios of Mg2+ to
norfloxacin (Figure 9).

Discussion
Figure 6sPlot of percentage of norfloxacin complexed at pH 6.5 with various
metal ions (calculated from formation constants) versus reduction in AUC0-
(Al3+(A), Djurdjevic et al.,14 Al3+(B), Okabayashi, et al.13).

The theoretical percentages of norfloxacin complexed in

solution at pH 6.5 were calculated using the formation
constants reported here and elsewhere.13,14 Thus, Figure 6
shows the observed reduction in norfloxacin AUC0- versus
the theoretical percentage of norfloxacin in the complexed
form when norfloxacin was given alone or in equimolar doses
with various cations. The results of a similar analysis for
varying ratios of Mg2+:norfloxacin are shown in Figure 7. Plots
of the theoretical percentages of norfloxacin in complexed form
806 / Journal of Pharmaceutical Sciences
Vol. 85, No. 8, August 1996

The formation constants of norfloxacin with magnesium(II) (log 1 ) 2.97(4), log 2 ) 5.6(2)) were smaller than those
determined for zinc(II) and iron(II), which were very similar
(log 1 ) 3.77(2), log 2 ) 7.59(3); log 1 ) 3.99(5), log 2 )
7.2(5), respectively). As was found in analogous measurements with copper(II),16 the formation constants for ciprofloxacin were of the same magnitude as those for norfloxacin.
Furthermore, there were negligible changes in the formation
constants when the experiments were conducted at 37 C.
However, unlike the results of copper(II) measurements,16
there was no consistent increase in the pH at which precipitation occurred with increasing temperature. Precipitation
occurred in norfloxacin titrations at a higher pH than for
ciprofloxacin titrations with Mg2+ and Zn2+, but not with Fe2+.

Figure 9sPlot of percentage norfloxacin complexed versus pH for various Mg:

norfloxacin ratios.

Norfloxacin formation constants have been reported previously for Ca2+, Mg2+, Cu2+, and Al3+.13,14,16 The formation
constants for norfloxacin with magnesium determined in this
work agree with those published by Okabayashi.13 Our
preliminary studies with Ca2+, Fe3+, and Al3+, in conjunction
with our studies on divalent Mg, Zn, Fe, and Cu, indicated
that the extent of coordination of norfloxacin (or ciprofloxacin)
with these metals follows the order Ca2+ < Mg2+ < Zn2+
Fe2+ < Cu2+ < Al3+ Fe3+. These results agree with the rank
order observed for published formation constants.13,14,16 The
same order was found in the values of log 1 for lomefloxacin13
with divalent Ca, Mg, Fe, Zn, and Cu and Al3+ and for
nalidixic acid with divalent Ca, Mg, Zn, and Cu.25
Formation constants for only the ML complex have been
determined for fluoroquinolones coordinating to calcium. Log
1 values for the complexes with norfloxacin13 (2.22),
lomefloxacin13 (2.08), ofloxacin13 (2.12), nalidixic acid25 (2.2),
and oxolinic acid25 (2.4) have been reported previously.
Extrapolations of in vitro complexation data to the in vivo
situation have had only limited success.26 Rationalization of
the effects of complexation in vitro with effects in vivo attracts
similar reservations as encountered for the pH-partition
hypothesis of drug absorption.27 Variability in pH of the
gastrointestinal tract among individuals, the existence of
microclimates of pH at membrane surfaces, the flow of water
in and out of the gastrointestinal tract, the possibilities of ion
pairing, the common-ion effect, and salting out may influence
the amount of drug complexed. Furthermore, variability in
gastric emptying, motility and volume, transit times through
the gastrointestinal tract, and relative area of the absorptive
surfaces may all affect drug absorption. Differences in the
solubility, lipophilicity, and size of the drug on complexation
with a metal must also be recognized, along with the competitive presence of other ions and the kinetics of complexation
and re-equilibrium. Most significantly, it is the dynamic
nature of absorption that confounds the correlation of in vitro
equilibrium data with in vivo response. Nonetheless, attempts have been made to identify pertinent effects of
complexation on drug absorption.
Ross and Riley28 have shown that the addition of Ca2+,
Mg2+, and Al3+ increased the solubility of lomefloxacin. In
an associated study, it was shown that the 1-octanol-water
partition coefficients of lomefloxacin and fleroxacin were
decreased by the presence of Ca2+, Mg2+, and Al3+.29 Both
these effects would be expected to hinder the absorption of
the fluoroquinolones. It is likely that complexes will precipitate with increasing pH possibly because complexation at

lower pH values produces charged complex species that are

more water-soluble and less lipophilic; these complexes may
be neutralized as the pH increases, thereby reducing the water
solubility.
The effects of complexation on drug absorption may be more
varied and remain somewhat contentious. Complexed species
are typically larger than and shaped differently from the
uncomplexed drug. Tanaka et al. noted a decrease in the
partition coefficient of chelated fluoroquinolones and argued
that adsorption of drug onto aluminum hydroxide reprecipitated in the small intestine may also be a factor in reduced
bioavailability of fluoroquinolones on coadministration with
aluminum-containing antacids.26 It is possible that norfloxacin is absorbed by the paracellular route, i.e., through the tight
junctions between adjacent cells in the intestinal epithelium,
or by facilitated diffusion through porin channels via interaction with the so-called channel proteins. The larger size and
different geometry of the complexes would tend to hinder
diffusion through these relatively narrow pathways, compared
with the uncomplexed norfloxacin. If active transport by an
energy-dependent carrier system is responsible for fluoroquinolone absorption, then it is possible that the complexed drug
is unable to occupy the enzymatic receptors involved in this
process. Whatever the mechanism by which complexation
hinders absorption, it may be reasonable to expect that an
increase in the percentage of fluoroquinolone complexed would
be associated with a decrease in drug bioavailability. It is
possible to simulate the amounts of fluoroquinolone complexed
under specific conditions using the appropriate formation
constants. Although there are many factors that can influence
the actual situation in vivo, a simplified model was used here
to explore the relative influence of equimolar doses of Al3+,
Fe2+, Zn2+, Mg2+, and Ca2+ and different doses of Mg2+ on the
chelation of norfloxacin; the formation constants determined
here and in other studies13,14 have been used (Table 1). For
comparison, the percentages of norfloxacin complexed at pH
6.5 by each cation-fluoroquinolone combination are also
presented, although the possibility of precipitation of the
complexes was not incorporated into the model. The extent
of complexation was greater and complexation commenced at
lower pH values for metal ions with higher formation constants (Figure 6). As the dose of magnesium increased from
0 to 4 times the molar dose of norfloxacin, the extent of
complexation increased, the onset of coordination occurring
at lower pH values, as seen in Figure 7.
Hitherto, the relationship between proportion of norfloxacin
complexed and the effect on oral norfloxacin bioavailability
when administered with metal ions has been very difficult to
elucidate from literature data. Pharmacokinetic studies on
norfloxacin-metal ion coadministration have differed greatly
in terms of dose, type of formulation, subjects, and experimental procedures used. Therefore, we conducted a series of
pharmacokinetic studies designed to complement the in vitro
complexation experiments. Dogs were selected as the experimental animal since they have pharmacokinetics similar to
those of humans for orally-administered norfloxacin.23 All
treatments were administered in 0.01 M HCl solution for
several reasons. First, disintegration and dissolution-related
factors which may exist with norfloxacin tablets and many
vitamin, mineral, and antacid preparations were avoided.
Second, different antacid products contain different salts (e.g.,
carbonates, citrates, silicates, hydroxides), some of which may
have limited solubility over the physiological pH range,
thereby providing substrates for the adsorption of norfloxacin.30 Third, the absorption of norfloxacin may be altered
unpredictably by raised gastrointestinal pH9,31 as would occur
after dosing with antacids. Thus, the administration of all
cations as their chloride salts was deemed to be more

Journal of Pharmaceutical Sciences / 807

Vol. 85, No. 8, August 1996

appropriate, particularly in view of the presence of high

concentrations of chloride ions naturally present in the gut.
A major finding was the marked rank order correlation
between the reduction in oral bioavailability of norfloxacin in
the presence of cations and the magnitude of the norfloxacincation formation constants determined in vitro. In fact, there
was a very good correlation between the formation constant
of the ML complex with divalent metal ions and the reduction
in bioavailability as illustrated in Figure 3. These results
strongly suggested that as the potential for coordination
between norfloxacin and the cation increased, the smaller
became the fraction of the norfloxacin dose absorbed into the
systemic circulation. It should be pointed out, however, that
although Zn2+ and Fe2+ had similar in vitro formation
constants, the greater reduction in norfloxacin absorption was
observed following coadministration with Fe2+. One possible
explanation involves the oxidative processes for iron within
the gastrointestinal tract. Kara et al.32 have shown that the
in vitro oxidation of Fe2+ and Fe3+ is rapid at pH 6. Trivalent
cations such as Fe3+ and Al3+ have an increased propensity
for chelation to norfloxacin11 and would be expected to leave
less of the norfloxacin dose available for absorption. However,
their affinities for chloride and hydroxide, as well as slow
kinetics and a tendency to form poorly soluble polymeric
moieties, are complicating factors in any comparative assessment.
The present results indicated that it was the extent rather
than the rate of norfloxacin absorption which was affected by
the presence of coadministered cations. While the first-order
absorption rate constants could not be calculated because of
insufficient data points in the absorptive phase, there was a
slight tendency toward longer Tmax values when either Fe2+
or Al3+ was administered with norfloxacin, although these
differences were not statistically significant. Similar conclusions have been drawn for the interaction of a number of
fluoroquinolones with several metal cations.5 The serum
concentrations of cations achieved following different treatments were not measured, but the clearance of norfloxacin
apparently was unaffected by their presence as evidenced by
the similar elimination half-lives among treatments. This
suggested that the amounts of these ions absorbed did not
affect liver and renal function, and that the pharmacokinetics
of any norfloxacin that was absorbed and then subsequently
complexed in the systemic circulation were similar to those
of the free drug.
In another part of the study norfloxacin was administered
to dogs alone and in combination with magnesium chloride
in molar ratios ranging from 0.25:1 to 4:1 (Mg2+:norfloxacin).
A rank order decrease in both the Cmax and the AUC0- was
observed with increasing amounts of Mg2+. Interestingly, the
most pronounced effect on relative norfloxacin bioavailability
occurred in the region between 0:1 and 0.5:1. The shape of
the plot may be explained by considering the simplistic model
of the in vivo formation of a metal ion-fluoroquinolone
complex with the M:L ratio of 1:2 (equivalent to 0.5:1). Within
the M:L range of 0 to 1:2 (0.5:1) all the metal ion present would
be able to complex twice its (molar) amount of fluoroquinolone.
In this initial phase, small amounts of metal ion would exert
a larger influence on the amount of fluoroquinolone available
for absorption, leading to a sharp reduction in bioavailability.
It is noteworthy that an M:L molar ratio of 0.5:1 can be readily
achieved when a clinical norfloxacin dose of 400 mg (equivalent to 1.25 mmol) is combined with much less than the
normal doses recommended for many commercial antacid and
mineral preparations. For example, a 10 mL dose of Mucaine
contains 712 mg of aluminum hydroxide (equivalent to 9 mmol
of Al3+) and 195 mg of magnesium hydroxide (equivalent to
3.4 mmol of Mg2+) giving M:L ratios of 7.2:1 and 2.7:1 for Al3+
and Mg2+, respectively. For cation:norfloxacin ratios greater
808 / Journal of Pharmaceutical Sciences
Vol. 85, No. 8, August 1996

than 1:2 (or 0.5:1) additional amounts of cation had little

impact stoichiometrically; thus the extra metal ion would not
be expected to have a commensurate effect on oral bioavailability. However, the above model cannot fully explain Figure
5. It implies that the bioavailability of norfloxacin should be
negligible for an M:L ratio of 1:2 (or 0.5:1) since all the
fluoroquinolone molecules should be complexed when, in fact,
only a 60% reduction in bioavailability was measured. This
discrepancy may be due to the fact that not all of the
norfloxacin present would be complexed because the reaction
does not go to completion, or because of unidentified factors
not accounted for by the simple extrapolation of in vitro
equilibrium calculations to the dynamic in vivo situation.
The latter is particularly relevant to a clinical situation.
Pharmacokinetic variability may be attributed, in part, to the
natural fluctuations in pH (4 to 7) of the upper duodenal
region from which the fluoroquinolones are absorbed.33 In
addition, the wide range of commercially available antacid
preparations may produce quite different physiological effects
in the gastrointestinal tract. Acid-neutralizing capacities may
range from 4 mequiv per 5 mL to 25 mequiv per 5 mL for
Gaviscon and Mylanta-II, respectively.34 At pH 6.5 fluoroquinolones form six-membered, cyclic coordination complexes
with metal cations chiefly via interaction with the ionized
carboxylate (pKa 6) and the adjacent keto group, or at the
secondary piperazinyl amine (pKa 8.5) at higher pH.16,35
Increasing pH at the absorptive site would result in an
increased extent of metal-norfloxacin complex formation
(Figure 8) and, therefore, an antacid with a high acidneutralization capacity would produce more extensive complexation and corresponding reduction in oral norfloxacin
absorption, compared with a less potent preparation. Apart
from pH effects on complexation, differences in the nature and
amount of the ingredients in various antacids may produce
variable effects on gastrointestinal motility and, therefore,
fluoroquinolone absorption. For example, magnesium (MgCO3, Mg(OH)2) has a laxative effect while Al(OH)3 and CaCO3
can cause constipation.34
In summarizing this study we report that a number of
common pharmaceutical cations had a marked detrimental
effect on the extent of canine norfloxacin absorption, even
when the molar dose of this antimicrobial was considerably
higher than that of the cation. The magnitudes of the in vitro
norfloxacin-metal ion formation constants had an inverse
rank order correlation with norfloxacin bioavailability when
norfloxacin was administered with these ions. A relatively
small extent of complexation resulted in disproportionately
large reductions in bioavailability which appeared to reach a
limit in the range of 60-80%, compared with norfloxacin
alone. While there have been no previous attempts to
characterize such interactions in vivo, the present data
indicated the likely formation of a 1:2 metal ion:fluoroquinolone complex at the pH of the upper gastrointestinal tract
from which many drugs, including fluoroquinolones, are
absorbed.

References and Notes

1. Neuvonen, P. J. Drugs 1976, 11, 45-54.
2. Campbell, N. R. C.; Hasinoff, B. Clin. Pharmacol. Ther. 1989,
45, 220-225.
3. Osman, M. A.; Patel, R. B.; Schuna, A.; Sundstrom, W. R.;
Welling, P. G. Clin, Pharmacol. Ther. 1983, 33, 465-470.
4. Campbell, N. R. C.; Hasinoff, B. B. Br. J. Clin. Pharmacol. 1991,
31, 251-255.
5. Lomaestro, B. M.; Bailie, G. R. DICP, Ann. Pharmacother. 1991,
25, 1249-1258.
6. Nix, D. E.; Wilton, J. H.; Ronald, B.; Distlerath, L.; Williams,
V. C.; Norman, A. Antimicrob. Agents Chemother. 1990, 34, 432435.

7. Parpia, S. H.; Nix, D. E.; Hejmanowski, L. G.; Goldstein, H. R.;

Wilton, J. H.; Schentag, J. J. Antimicrob. Agents Chemother.
1989, 33, 99-102.
8. Lehto, P.; Kivisto, K. T. Antimicrob. Agents Chemother. 1994,
38, 248-251.
9. Okhamafe, A. O.; Akerele, J. O.; Chukuka, C. S. Int. J. Pharm.
1991, 68, 11-18.
10. Lehto, P.; Kivisto, K. T.; Neuvonen, P. J. Br. J. Clin. Pharmacol.
1994, 37, 82-85.
11. Campbell, N. R. C.; Kara, M.; Hasinoff, B. B.; Haddara, W. M.;
McKay, D. W. Br. J. Clin. Pharmacol. 1992, 33, 115-116.
12. Kivisto, K. T.; Ojala-Karlsson, P.; Neuvonen, P. J. Antimicrob.
Agents Chemother. 1992, 36, 489-491.
13. Okabayashi, Y.; Hayashi, F.; Terui, Y.; Kitagawa, T. Chem.
Pharm. Bull. 1992, 40, 692-696.
14. Djurdjevic, P. T.; Jelikic-Stankov, M.; Stankov, D. Anal. Chim.
Acta 1995, 300, 253-259.
15. Vogel, A. I. Vogels Textbook of Quantitative Chemical Analysis,
5th ed.; Longman: Essex, 1989; p 415.
16. Wallis, S. C.; Gahan, L. R.; Charles, B. G.; Hambley, T. W.;
Duckworth, P. A. J. Inorg. Biochem. 1996, 62, 1-15.
17. Gans, P.; Sabatini, A.; Vacca, A. J. Chem. Soc., Dalton Trans.
1985, 1195-1200.
18. Wallis, S. C.; Charles, B. G.; Gahan, L. R. J. Chromatogr. B
1995, 674, 306-309.
19. Gibaldi, M. Biopharmaceutics and Clinical Pharmacokinetics,
4th ed.; Lea & Febiger: Philadelphia, 1991; pp 14-23, 377378.
20. SAS Institute Inc. SAS/STAT Users Guide, Release 6.03 ed.;
SAS Institute Inc.: Cary, NC, 1988.
21. Baes, C. F.; Mesmer, R. E. The Hydrolysis of Cations; Robert E.
Krieger: Malabar, FL, 1986; pp 287-294.
22. Cotton, F. A.; Wilkinson, G. Advanced Inorganic Chemistry, 2nd
ed.; Interscience: New York, 1966; p 859.

23. Walker, R. D.; Stein, G. E.; Budsberg, S. C.; Rosser, E. J.;

MacDonald, K. H. Am. J. Vet. Res. 1989, 50, 154-157.
24. Swanson, B. N.; Boppana, V. K.; Vlasses, P. H.; Rotmensch, H.
H.; Ferguson, R. K. Antimicrob. Agents Chemother. 1983, 23,
284-288.
25. Timmers, K.; Sternglanz, R. Bioinorg. Chem. 1978, 9, 145-155.
26. Tanaka, M.; Kurata, T.; Fujisawa, C.; Ohshima, Y.; Aoki, H.;
Okazaki, O.; Hakusui, H. Antimicrob. Agents Chemother. 1993,
37, 2173-2178.
27. Florence, A. T.; Attwood, D. Physicochemical Principles of
Pharmacy; Macmillan: Basingstoke, 1988; pp 342-348.
28. Ross, D. L.; Riley, C. M. Int. J. Pharm. 1992, 87, 203-213.
29. Ross, D. L.; Elkinton, S. K.; Knaub, S. R.; Riley, C. M. Int. J.
Pharm. 1993, 93, 131-138.
30. Deppermann, K.-M.; Lode, H. Drugs 1993, 45, 65-72.
31. Grasela, T. H.; Schentag, J. J.; Sedman, A. J.; Wilton, J. H.;
Thomas, D. J.; Schultz, R. W.; Lebsack, M. E.; Kinkel, A. W.
Antimicrob. Agents Chemother. 1989, 33, 615-617.
32. Kara, M.; Hasinoff, B. B.; McKay, D. W.; Campbell, N. R. G.
Br. J. Clin. Pharmacol. 1991, 31, 257-261.
33. Harder, S.; Fuhr, U.; Beerman, D.; Staib, A. H. Br. J. Clin.
Pharmacol. 1990, 30, 35-39.
34. Brunton, L. L. In The Pharmacological Basis of Therapeutics;
Gilman, A. G., Rall, T. W., Nies, A. S., Taylor, P., Eds.; Pergamon
Press: New York, 1990; pp 897-913.
35. Wallis, S. C.; Gahan, L. R.; Charles, B. G.; Hambley, T. W.
Polyhedron 1995, 14, 2835-2840.

Acknowledgments
Financial support from the Australian Research Council is gratefully acknowledged.

JS960087F

Journal of Pharmaceutical Sciences / 809

Vol. 85, No. 8, August 1996