Analytica Chimica Acta 531 (2005) 105–110

High speed liquid chromatography for in-process control of rifabutin

Domingo Blanco Gomisa,∗ , Noé Sánchez Núñeza , Elena Andrés Garcı́aa ,
Pilar Arias Abrodoa , Miguel Bayod Jasanadab , Ma . Dolores Gutiérrez Álvareza
a Department of Physical and Analytical Chemistry, University of Oviedo, 33006 Oviedo, Spain
b Asturpharma S.A., Silvota, 33192 Llanera, Asturias, Spain

Received 7 July 2004; received in revised form 29 September 2004; accepted 29 September 2004
Available online 23 December 2004

Abstract

A fast HPLC method has been developed for simultaneous determination of rifabutin and its synthesis precursors. The analytes are separated
in 1.8 min by means of a Kromasil 100 C18 column (50 mm × 2.1 mm i.d., 3.5 ␮m) at 30 ◦ C. The mobile phase (A: 5 mM KH2 PO4 adjusted to
pH 6.5 with KOH; B: acetonitrile) was pumped at a flow rate of 0.4 ml min−1 according to the fast gradient mode: 0 min, 58% B; 0–0.4 min,
95% B. Detection was by ultraviolet absorbance at 275 nm. The method was validated in accordance with the International Conference on
Harmonisation (ICH) guidelines and good accuracy, intermediate precision (≤4.6%) and linearity in the range 5–50 mg l−1 were observed
for all compounds. This method is sensitive (limits of detection ranged between 0.1 and 0.3 mg l−1 ) and selective to quantify rifabutin and its
synthesis precursors and could be used for in-process control.
© 2004 Elsevier B.V. All rights reserved.

Keywords: High speed HPLC; In-process control; Rifabutin

1. Introduction protease inhibitors, drugs with narrow therapeutic ranges, are

Rifabutin is a potent spiropiperidyl-rifamycin deriva-
tive with a broad spectrum of antibacterial activity against
Mycobacterium tuberculosis, including rifampicin-resistant
strains, and atypical mycobacteria. Prophylactic treatment
with rifabutin was shown to decrease the incidence of My-
cobacterium avium complex (MAC) by approximately 50%
in patients infected with human immunodeficiency virus [1].
A three drug combination therapy containing rifabutin has
been shown to be more effective in the treatment of MAC bac-
teremia than a four-drug combination therapy containing ri-
fampin [2]. The clinically significant interactions of the newer
antiretroviral agents, belonging to the family of protease in-
hibitors, with rifampin [3] may foster the usage of rifabutin in
some clinical settings, because rifabutin reduces serum con-
centrations of antiretroviral agents, but less so than rifampin.
Consequently, the drug interactions between rifabutin and
∗ Corresponding author. Tel.: +34 985103490; fax: +34 985103125.
E-mail address: dbg@fq.uniovi.es (D. Blanco Gomis).
easier to manage than those with rifampin [4] and, therefore,
rifabutin currently remains a major drug for the management
of MAC in AIDS patients.
Rifabutin analysis has been carried out by different meth-
ods, including HPLC, in biological fluids [5–18] and pharma-
ceutical formulations [19,20]. Nowadays, the United States
Pharmacopoeia (USP) recommends liquid chromatographic
methods for the determination of rifabutin in raw materials
and in its dosage forms. However, there are no published
reports to date concerning the simultaneous analysis of ri-
fabutin, its synthesis intermediates and potential impurities
(synthetic route is outlined in Fig. 1) in a short analysis time.
Controlling the levels of precursors, products and by-
products in pharmaceuticals is one of the most important tasks
in pharmaceutical quality, safety and efficacy. In this sense,
the number of analysis to be carried out in a pharmaceutical
process control could be very high and, therefore, laborato-
ries are demanding fast methods to enhance their productiv-
ity. On the other hand, the use of chromatographic methods in
the pharmaceutical industry is particularly mandatory since

0003-2670/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.aca.2004.09.083
106 D. Blanco Gomis et al. / Analytica Chimica Acta 531 (2005) 105–110
Fig. 1. Synthetic route to rifabutin. Py: pyridine, HMT: hexamethylenetetramine.
they have to adopt the norms of the pharmacopoeia (resolu-
tion, peak asymmetry, repeatability, intermediate precision,
accuracy, etc.). As the classic liquid chromatography is un-
able to solve a complex mixture of substances in seconds or
as maximum in some few minutes, high speed liquid chro-
matography emerges like a solution of agreement between
resolution and analysis time.
In its conventional format, this technique makes use of
columns of standard diameter (4.6 mm) with length up to
10 cm, generally with stationary phases of 3 ␮m particle size
and working at a flow rate of mobile phase higher than
2 ml min−1 . In our opinion, the diameter of these columns
should be lower than 3.2 mm, because such columns allow to
reduce flow rate up to five-fold, approximately. Related mi-
crobore columns increase effectiveness, sensibility and pre-
cision of the chromatographic methods.
In order to improve resolution, modern stationary phases
of fine granulometry (2 or 1.5 ␮m, porous or not porous)
could be used although their current price limits their appli-
cations. Another alternative may be the use of the monolithic
stationary phases, characterized by high porosity and perme-
ability. The analysis time reduction is achieved by increasing
flow rate up to nine times the usual in a conventional column
although the high consumption solvent and the price of the
columns are important drawbacks. Furthermore, these mono-
lithic columns may generate on basic compounds more peak
asymmetry than good classical full-endcapped columns [21].
When the required chromatographic effectiveness is not
very high, a microbore short column with conventional 3 ␮m
stationary phases is the choice, due to the fact that column
and operation price are cheap and the instrumental set-up is
also relatively simple and economical. Basically, modifica-
tions involve the injection system (maximum volume ≤3 ␮l),
the detector cell dead volume (≤3 ␮l), the connecting capil-
lary tubes (length ≤80 cm, i.d. ≤170 ␮m) and when injection
valve turns (manual or automatically) slowly, it is convenient
to install a by-pass that avoids pressure pulses that could de-
teriorate short columns quickly.
The aim of this work is to develop a rapid and reliable
high speed HPLC method for in-process control of rifabutin
in order to enhance laboratory productivity. The validation
parameters stated by ICH guidelines [22,23] have been con-
sidered.
2. Experimental
2.1. Reagents and standards
Rifamycin S (RS), 3-bromorifamycin S (3Br-RS), 3-
aminorifamycin S (3A-RS), 3-amino-4-iminorifamycin S
(3A-4I-RS) and rifabutin (RBT) were kindly supplied by As-
turpharma (Silvota-Llanera, Spain). Potassium dihydrogen
phosphate was obtained from Sigma Chemical Co. (St. Louis,
MO, USA), potassium hydroxide was obtained from Merck
(Darmstadt, Germany) and HPLC gradient quality acetoni-
trile was purchased from Romil (Loughborough, UK). Milli-
Q water (Millipore, Milford, MA, USA) was used throughout.
All other chemical and solvents were of analytical reagent or
HPLC grade.
2.2. Standard solutions
Stock solutions of rifabutin and related compounds were
prepared at a concentration of 100 mg l−1 by dissolving the
appropriate amount of the pure drug in acetonitrile. The
 D. Blanco Gomis et al. / Analytica Chimica Acta 531 (2005) 105–110
solutions were stored at 4 ◦ C and protected from light to min-
imize photodegradation. Under these conditions all solutions
proved to be stable for more than 2 months. Dilutions of
the 100 mg l−1 standards were used to make the appropriate
working solutions of the drugs in the sample solvent (5 mM
potassium dihydrogen phosphate adjusted to pH 6.5 with
potassium hydroxide and 50% acetonitrile). Working solu-
tions have to be prepared in a weaker solvent than the mobile
phase to obtain sharp chromatographic peaks. The solutions
were sonicated in an ultrasonic bath for 1 min and filtered
through a 0.22 ␮m PVDF syringe filter (Lida, Kenosha, WI,
USA). The resulting filtered solution was placed in a HPLC
vial.
2.3. Apparatus and conditions
HPLC analyses were performed on a Shimadzu HPLC
system (Duisburg, Germany) equipped with two LC-10AD
pumps, a UV–vis SPD-M10AD photodiode array detector
with a 2.5 ␮l flow cell and 80 ms of sampling time, a Sil-
10AD automatic injector and DGU-14A degas on-line and
60 cm connecting tubes with an internal diameter of 127 ␮m.
The column used was a Kromasil 100 C18 (50 mm × 2.1 mm
i.d., 3.5 ␮m) (Teknokroma, Barcelona, Spain). An ODS guard
column was used to protect the analytical column. Before use,
the mobile phase was vacuum filtered through a 0.22 ␮m ny-
lon membrane filter (Lida, Kenosha, WI, USA). The chro-
matographic experiments were carried out at 30 ◦ C. The
binary gradient used at a flow rate of 0.4 ml min−1 was
as follows: 0 min, 58% B; 0–0.4 min, 95% B, where sol-
vent A was water with 5 mM potassium dihydrogen phos-
phate adjusted to pH 6.5 with potassium hydroxide and
solvent B was acetonitrile. The sample injection volume
was 1 ␮l.
Identification of the compounds was performed by means
of their retention time and UV spectra. Spectra were taken
at the leading edge, the apex and the tailing edge to monitor
for peak purity. Quantification was carried out at 275 nm by
the external standard method. All measurements were made
using a Shidmadzu CLASS-VP Version 5.032 software. The
asymmetry factor (As) – that it is the width from the frontside
and the backside of the peak to the apex – was calculated at
10% of the peak height. All results were the mean of at least
duplicate injections.
3. Results and discussion
3.1. Optimization of the chromatographic system
At present, the HPLC methods for the analysis of rifabutin,
both in biological fluids and in pharmaceuticals, are based on
the conventional HPLC modality. Thus, the recommended
USP method [19], for the analysis of rifabutin and related
compounds in pharmaceuticals, needs an analysis time higher
than 20 min. In our opinion, this time consuming method is
107
inappropriate for an efficient process control. For this reason,
another alternative should be considered in order to reduce the
analysis time, increase laboratory productivity with a reduced
number of HPLC instruments, and to reduce the operation
costs.
Because the separation of rifabutin, precursors and by-
products seems to be a relative easy problem, we have cho-
sen the most economical alternative to utilize the fast liquid
chromatography, microbore short columns (50 mm × 2 mm
i.d.) packed with 3–3.5 ␮m octyl- or octadecylsilane sta-
tionary phases. In comparison with other HPLC methods
which use conventional columns with lengths ranged between
250 mm for rifabutin analysis in biologic fluids [10,11,16]
and 110–125 mm for rifabutin analysis in pharmaceutical for-
mulations [19,20], i.d. of 4.6 mm and a particle size of 5 ␮m,
the use of short microbore columns increase mass sensitiv-
ity at less in four times and reduce flow rate in five times
approximately, increasing also the effectiveness per unit of
time.
Because some analytes have polar nature, we have as-
sayed several full endcapped stationary phases in order to
attaint the best chromatographic shape evaluating the asym-
metry and resolution. Good results were obtained with Kro-
masil 100 C18 and C8 columns. However, the peak profile
(As < 1.4), resolution (Rs > 1.6) and analysis time (1.8 min)
was better for C18 column and was selected for subsequent
studies. A special advantage of high speed columns with low
void volumes is the reduced time for re-equilibration after
changing the mobile phase. Thus, several column volumes
can be passed through a column in 1–2 min and the column
is then ready for the subsequent analysis, ideally to gradient
mode. At the same time, the minimum of the van Deempter
curve is displaced to superior flows.
The assembly of these high speed columns in a conven-
tional HPLC instrument requires an injection system capable
of injecting volumes of below 3 ␮l. In addition, the use of
smaller particle columns with their inherently higher efficien-
cies and lower void volumes imposes additional constraints
on the design and construction of injection valves as extra col-
umn band broadening must be minimized if the full potential
of these columns is to be realized. Therefore, the injector by-
pass must be constructed so as to eliminate pressure pulses
yet not substantially contribute additional extra-column band
broadening. When properly designed and constructed the in-
jector by-pass extends the lifetime of a high speed column
from an intolerably short period to a reasonably long one
[24]. Injector by-pass was constructed using two union-tees,
the appropriate capillary tubes and fittings to obtain a di-
vision rate of 332. The flow cell and sampling time of the
detector should be lower than 3 ␮l and 100 ms, respectively.
A 2.5 ␮l flow cell and 80 ms sampling time were selected.
The different components of equipment were connected by
using capillary tubes of 60 cm overall length and 127 ␮m in-
ternal diameter to minimize instrumental bandwidth (30 ␮l)
improving the effectiveness of the separation per unit
of time.
108 D. Blanco Gomis et al. / Analytica Chimica Acta 531 (2005) 105–110
Fig. 2. Effect of the mobile phase pH on the capacity factor of the rifabutin
and related compounds. Column: 50 mm × 2.1 mm i.d., 3.5 ␮m, Kromasil
100 C18 ; mobile phase, 25 mM KH2 PO4 , adjusted to pH 6.5 with KOH and
acetonitrile (55%); flow rate: 0.2 ml min−1 ; temperature: 30 ◦ C; detection at
275 nm; (
) rifamycin S; () 3-bromorifamycin S; () 3-aminorifamycin
S; () 3-amino-4-iminorifamycin S; (䊉) rifabutin.
3.2. Optimization of chromatographic conditions
For the fast RP-HPLC method development, the main
chromatographic conditions to be optimized were pH, ionic
strength, percentage organic modifier, temperature, and flow
rate.
The compounds show in the studied range of pH (4.0–7.0)
basic, acid or neutral properties. Consequently, the pH must
be adjusted carefully to ensure an adequate separation. As can
be seen in Fig. 2, as a consequence of the different nature of
the solutes various changes in the elution order are observed.
Consequently, it is not advisable to select values of pH next
to the crossing points. Resolutions higher than 1.5 were only
attained about pH 6.5. The effect of the ionic strength on
the capacity factor, tested between 2.5 and 25 mM, seems to
be negligible. An ionic strength of 5 mM was selected for
subsequent experiments. Likewise, the effect of temperature
on resolution and analysis time is irrelevant to 35 ◦ C. Over
this temperature the resolution falls below 1.5. In order to
minimize the pressure drop across the column and increase
solute mass transfer, 30 ◦ C was selected.
The Kromasil 100 C18 column shows similar efficiency
working with flow rates of between 0.2 and 0.4 ml min−1 and
with the purpose of shorten the analysis time, a flow rate of
0.4 ml min−1 was selected. The pressure drop does not exceed
150 bar. The small flow rate used is lower than the one used
in the conventional HPLC methods [10,11,16,19,20], what
notably reduces the solvent consumption.
The separation time of precursors and products of rifabutin
synthesis in isocratic mode with 55% organic modifier is time
consuming (10.7 min). Therefore, a gradient mode is requi-
site in order to separate the compounds in a few minutes.
The optimized gradient program begins with a high level of
acetonitrile (58%) and increases quickly to 95% in 0.4 min,
as shown in Section 2.3. Under these conditions, analysis
can be carried out in 1.8 min. As can be seen in Fig. 3,
which shows the chromatogram obtained according to these
Fig. 3. Chromatogram of rifabutin, its synthesis intermediates and degra-
dation products obtained in the gradient mode. Column: 50 mm × 2.1 mm
i.d., 3.5 ␮m, Kromasil 100 C18 ; mobile phase, solvent A 5 mM KH2 PO4 , ad-
justed to pH 6.5 with KOH, solvent B acetonitrile; gradient conditions: 0 min,
58% B; 0–0.4 min, 95% B; flow rate: 0.4 ml min−1 ; temperature: 30 ◦ C; de-
tection at 275 nm; injected volume: 1 ␮l; concentration of each compound,
25 mg l−1 ; peak identification: (1) 3-bromorifamycin S; (2) rifamycin S; (3)
3-aminorifamycin S; (4) 3-amino-4-iminorifamycin S; (5) rifabutin.
optimized conditions, adequate resolution (>1.5), including
the by-products, and asymmetry (<1.5) were obtained in a
short analysis time. Comparing this new method with the one
used in the pharmaceutical industry at the moment [19,20],
the chromatography separation time has been reduced about
90%, reducing the solvent consumption per analysis from
20 ml to less than 1 ml, approximately.
3.3. Validation
Calibration curves were constructed by the analysis of trip-
licates of five points in the range from 5 to 50 mg l−1 of ri-
fabutin and all its synthesis precursors. The calibration graphs
for all the compounds showed a good fitting to a linear model
between the peak areas and analyte concentrations, with the
regression coefficients >0.999 in all cases. The linearity of the
calibration graphs was also checked with two different statis-
tical tests: linearity and proportionality tests. For the linearity
test, the linearity of the method was confirmed by showing
that the response factor R.S.D. and slope R.S.D. values were
lower than 5 and 2%, respectively; the values obtained from
the Fisher test (analysis of variance (ANOVA)) were always
 D. Blanco Gomis et al. / Analytica Chimica Acta 531 (2005) 105–110 109

Table 1
Calibration parameters for linearity
Specification Range 5–50 (mg l−1 )
3Br-RS RS 3A-RS 3A-4I-RS RBT
Correlation coefficient ≥0.997 0.9996 0.9996 0.9996 0.9997 0.9997
Standard error 1.5 2.0 1.6 2.2 2.2
Linearity test
Response factor R.S.D. ≤5% 4.8 4.3 4.8 3.4 2.8
Slope 2.87 3.96 3.02 5.55 4.81
Slope S.D. 0.02 0.03 0.03 0.04 0.04
Slope R.S.D. ≤2% 0.8 0.8 0.9 0.7 0.7
Confidence interval 0 not included 2.82–2.92 3.89–4.03 2.96–3.08 5.47–5.63 4.73–4.89
Experimental t texp > ttab 119.958 118.394 116.319 153.753 135.362
ANOVA Fexp < Ftab 0.600 0.454 1.989 0.396 0.351
Proportionality test
Intercept 0.5 −0.2 −0.9 −0.4 1.3
Intercept S.D. 0.7 1.0 0.7 1.0 1.0
Confidence interval 0 included −1.0 to 2.0 −2.3 to 1.9 −2.5 to 0.7 −2.6 to 1.8 −0.9 to 3.5
Experimental t texp < ttab 0.683 0.233 1.266 0.401 1.267

Table 2
Analytical characteristics of the chromatographic method
Specification (%) 3Br-RS RS 3A-RS 3A-4I-RS RBT
Instrumental repeatability ≤2 1.6 1.3 1.5 1.9 1.5
Intermediate precision ≤5 4.1 3.9 3.0 2.4 4.6
Limit of detection (mg l−1 ) 0.2 0.3 0.2 0.2 0.1
Limit of quantification (mg l−1 ) 0.5 1.0 0.7 0.5 0.4

lower than the tabulated ones (α = 0.05). Finally, the slopes of
the linear calibration curves were statistically different from 0
(texp > ttab , α = 0.05). In the proportionality test, it was demon-
strated that the intercept was not statistically different from
0, as the confiance limits include zero and the Student’s t-test
values calculated were always lower than the tabulated values
for the same signification level. This indicates the absence of
systematic error, and the linearity thus being demonstrated
[25]. The calibration data are summarized in Table 1. We
have also tested a linear range from 5 to 100 mg l−1 , check-
ing that it completes all the specifications except the response
factor R.S.D.
The precision of the method was assessed by express-
ing the relative standard deviation of several repeated mea-
surements (Table 2). Instrumental repeatability was estimated
from six replicates at three concentrations – low, medium and
high level inside the lineal range: 5, 25 and 50 mg l−1 , respec-
tively – obtaining values (1.3–1.9%) of below acceptance cri-
terion (≤2%). The estimation of repeatability was performed
during 3 h. The compounds were stable and showed no signif-
icant difference in the peak area after this time. Intermediate
precision was determined by comparing the results obtained
from the analysis of freshly prepared samples on two separate
days. The results, ranging between 2.4 and 4.6%, were also
lower the acceptance criterion (≤5%). Therefore, acceptable
precision was obtained for all preparations.
The detection and quantification limits are shown in the
aforementioned Table 2. These were determined by 10 re-
peated measures of the blank, followed by the preparation of
calibration plots (peak height versus concentration) from 5
to 50 mg l−1 . The detection and quantification limits were in
the range of 0.1–0.3 and 0.4–1.0 mg l−1 , respectively.
Recovery experiments were performed in order to study
the accuracy of the method. A mixture of known concentra-

Table 3
Validation parameters for accuracy
Compounds Recovery (%) texp Gexp

5 mg l−1 25 mg l−1 50 mg l−1

3-Bromorifamycin S 102.0 99.2 100.2 0.4027 0.6289
Rifamycin S 102.7 99.1 101.7 0.8103 0.4819
3-Aminorifamycin S 102.0 98.9 101.5 0.6332 0.5170
3-Amino-4-iminorifamycin S 98.0 99.2 100.1 1.8921 0.6369
Rifabutin 100.7 99.7 100.1 0.3486 0.7778
Specifications: 97–103%, texp < ttab and Gexp < Gtab .
110 D. Blanco Gomis et al. / Analytica Chimica Acta 531 (2005) 105–110
tions of these substances was prepared and analyzed at low,
medium and high calibration ranges by this method on the
same day. All analyses were carried out in triplicate. The av-
erage recoveries obtained ranged between 98.0 and 102.7%.
It was also proven, by means realization of a Student’s t-test
(texp < ttab , α = 0.05), that a significant difference does not
exist between the average recovery and 100%. Furthermore,
carrying out Cochran’s G-test (Gexp < Gtab , α = 0.05), it was
demonstrated that the factor concentration does not influence
the variability of the results. All these results thus testify to
the accuracy of the proposed method (Table 3).
4. Conclusions
The new high speed chromatographic method developed
has been found to be effective for the rapid analysis of phar-
maceuticals. Compared to a conventional column, separa-
tion was performed in approximately 10 times less analysis
time, dramatically reducing solvent consumption and cost per
analysis. These columns can be easily assembled to a conven-
tional HPLC instrument with few modifications. They offer
great potential to replace larger conventional columns for in-
process and quality control in the pharmaceutical industry.
Acknowledgments
The authors wish to acknowledge the financial support of
the Spanish Ministry of Science and Technology, Project No.
PPQ2000-0660.
References
[1] S.D. Nightingale, D.W. Cameron, F.M. Gordin, P.M. Sullam, D.L.
Cohn, R.E. Chaisson, L.J. Eron, P.D. Sparti, B. Bihari, D.L. Kauf-
man, J.J. Stern, D.D. Pearce, W.G. Weinberg, A. LaMarca, F.P. Sie-
gal, N. Engl. J. Med. 329 (1993) 828.
[2] S.D. Shafran, J. Singer, D.P. Zarowny, P. Phillips, I. Salit, S.L. Walm-
sley, I.W. Fong, M.J. Gill, A.R. Rachlis, R.G. Lalonde, M.M. Fan-
ning, C.M. Tsoukas, N. Engl. J. Med. 335 (1996) 377.
[3] J. McCrea, D. Wyss, J. Stone, A. Carides, S. Kusma, C. Kleinbloe-
sem, Y. AlHamdan, K. Yeh, P. Deutsch, Clin. Pharmacol. Ther. 61
(1997) 152.
[4] C.K. Finch, C.R. Chrisman, A.M. Baciewicz, T.F. Self, Arch. Int.
Med. 162 (2002) 985.
[5] B.L. Heifets, M.D. Iseman, J.P. Lindholm, Am. Rev. Respir. Dis.
137 (1988) 711.
[6] B.L. Heifets, J.P. Lindholm, M.D. Iseman, Am. Rev. Respir. Dis.
137 (1988) 719.
[7] C.A. Benson, P.L. Williams, J.S. Currier, F. Holland, J. McKinsey,
Clin. Infect. Dis. 37 (2003) 1234.
[8] F.M. Hamzeh, C. Benson, J. Gerber, C. Flexner, Clin. Pharmacol.
Ther. 73 (2003) 159.
[9] W.K. Kraft, J.B. McCrea, G.A. Winchell, A. Carides, S.A. Waldman,
J. Clin. Pharmacol. 44 (2004) 305.
[10] R.C. Lewis, N.Z. Hatfield, P.K. Narang, Pharm. Res. 8 (1991)
1434.
[11] Y.Y. Lau, G.D. Hanson, B.J. Carel, J. Chromatogr. B 676 (1996)
125.
[12] H. Bartels, R. Bartels, J. Chromatogr. B 686 (1996) 235.
[13] C.B. Trapnell, C. Jamis-Dow, R.W. Klecker, J.M. Collins, Antimi-
crob. Agents Chemother. 41 (1997) 924.
[14] I. Utkin, T. Koudriakova, T. Thompson, C. Cottrell, E. Iatsimirskaia,
J. Barry, P. Vouros, N. Gerber, Drug Metab. Dispos. 25 (1997) 963.
[15] A. Cato, J. Cavanaugh, H. Shi, A. Hsu, J. Leonard, R. Granneman,
Clin. Pharmacol. Ther. 63 (1998) 414.
[16] G. Gatti, C.R. De Pascalis, F. Miletich, R. Casazza, D. Bassetti, J.
Chromatogr. B 728 (1999) 233.
[17] K. Gallicano, Y. Khaliq, G. Carignan, A. Tseng, S. Walmsley, W.
Cameron, Clin. Pharmacol. Ther. 70 (2001) 149.
[18] R.E. Polk, D.F. Brophy, D.S. Israel, R. Patron, B.M. Sadler, G.E.
Chittick, W.T. Symonds, Y. Lou, D. Kristoff, D.S. Stein, Antimicrob.
Agents Chemother. 45 (2001) 502.
[19] United States Pharmacopeia, XXVII Rev., NF XXII, United States
Pharmacopeial Convection, Inc., Rockville, 2004, p. 1647.
[20] Real Farmacopea Española, 2nd ed., Suplemento 2.1, Ministerio de
Sanidad y Consumo, Madrid, 2002, p. 3186.
[21] D.V. McCalley, J. Chromatogr. A 965 (2002) 51.
[22] CPMP/ICH/381/95, Note for Guidance on Validation of Analytical
Methods: Definitions and Terminology, ICH Topic Q2A, Step 5,
CPMP Adopted November 1994.
[23] CPMP/ICH/281/95, Note for Guidance on Validation of Analytical
Procedures: Methodology, ICH Topic Q2B, Step 4, CPMP Adopted
December 1996.
[24] J.L. DiCesare, M.W. Dong, J.R. Grant, Chromatographia 15 (1982)
595.
[25] M.C. Cels, S.G. Fora, M.P. Forn, J.S. Roca, L.V. Pla, Validación
de métodos analı́ticos, Monografı́a de la Asociación Española de
Farmacéuticos de la Industria, Sección Catalana, 1989.