Journal of Chromatographic Science, Vol.

41, January 2003

Quantitative Thin-Layer Chromatographic Method of

Analysis of Azithromycin in Pure and Capsule Forms

Alaa Khedr and Mahmoud Sheha

Downloaded from https://academic.oup.com/chromsci/article/41/1/10/374909 by guest on 07 October 2020

Faculty of Pharmacy, Assiut University, Assiut, Egypt

Abstract Introduction
A validated stability-indicating thin-layer chromatographic (TLC)
method of the analysis of azithromycin (AZT) in bulk and capsule Azithromycin (AZT) (Figure 1) is a subclass of macrolide
forms is developed. Both AZT potential impurity and degradation antibiotics. Physically, AZT as a dihydrate is a white crystalline
products can be selectively and accurately estimated in both raw powder. AZT in an aqueous or chloroform solution has insignif-
material and product onto one precoated silica-gel TLC plate icant UV absorption properties; also, it is very difficult to find
60F254. The development system used is n-hexane–ethyl a suitable active reagent to enhance its UV absorptivity with no
acetate–diethylamine (75:25:10, v/v/v). The separated bands are byproducts or degradation of the intact compound. Attention
detected as brown to brownish red spots after spraying with should also be paid to the fact that AZT dihydrate–acidic solu-
modified Dragendorff’s solution. The Rf values of AZT, tion is rapidly decomposed via intramolecular dehydration to
azaerythromycin A, and the three degradation products are 0.54,
form erythromycin-6,9-hemiketal and then anhydroery-
0.35, 0.40, 0.20, and 0.12, respectively. The optical densities of the
separated spots are found to be linear in proportion to the amount
thromycin (1). Therefore, United States Pharmacopoeia (USP)
used. The stress testing of AZT shows that azaerythromycin A is 24 and most plasma work have used high-performance liquid
the major impurity and degradation product, accompanied by chromatographic (HPLC) methods and tracing with electro-
three other unknown degradation products. The stability of AZT is chemical detectors (2,3). This official method depends pri-
studied under accelerated conditions in order to provide a rapid marily on the liability of certain reaction centers of AZT to
indication of differences that might result from a change in the oxidation without sample pretreatment. Alternatively, only one
manufacturing process or source of the sample. The forced
degradation conditions include the effect of heat, moisture, light,
acid–base hydrolysis, sonication, and oxidation. The compatibility
of AZT with the excipients used is also studied in the presence and
absence of moisture. The amounts of AZT and azaerythromycin A
are calculated from the corresponding linear calibration curve;
however, the amounts of any other generated or detected
unknown impurities are calculated as if it were AZT. This method
shows enough selectivity, sensitivity, accuracy, precision,
linearity–range, and robustness to satisfy Federal Drug
Administration/International Conference of Harmonization
regulatory requirements. The method developed can also be used
for the purity testing of AZT raw material and capsules, content
uniformity testing, dissolution testing, and stability testing of AZT
capsules. The potential impurity profiles of both active AZT
material and capsule forms are found comparable. The linear Figure 1. Azithromycin dihydrate (C38H72N2O12 • 2H2O, fw = 785.0):
range of AZT is between 5 and 30 mcg/spot with a limit of (2R,3S,4R,5R,8R,10R,11R,12S,13S,14R)-13-[(2,6-dideoxy-3-C-methyl-3-
quantitation of 2 mcg/spot. The intraassay relative standard O-methyl-α- L -ribo-hexopyranosyl)oxy]-2-ethyl-3,4,10-trihydroxy-
deviation percentage is not more than 0.54%, and the day-to-day 3,5,6,8,10,12,14-heptamethyl-11-[[3,4,6-trideoxy-3-(dimethylamino)-α-D-
variation is not more than 0.86%, calculated on the amounts of xylo-hexopyranosyl]oxy]-1-oxa-6-azacyclopentadecan-15-one.
AZT RS recovered using different TLC plates.

10 Reproduction (photocopying) of editorial content of this journal is prohibited without publisher’s permission.
Journal of Chromatographic Science, Vol. 41, January 2003

publication has described a precolumn derivatization proce- Chromatography

dure of AZT with 9-fluorenylmethyloxycarbonyl chloride to Preparation of developing system
yield a stoichiometric amount of fluorescence derivative for A mixture of n-hexane (75 mL), ethyl acetate (25 mL), and
separation by an HPLC method (4). This method depends on 99% diethylamine (10 mL) was prepared and 50 mL of this
the stoichiometric reactivity of the five hydroxyl groups with mixture was poured into a TLC tank. The tank was then cov-
the reagent. ered with a lined lid and presaturated with the solvent vapor
However, the spectrophotometric method described by system for at least 30 min at room temperature before use.
Bebawy et al. (5) using 7,7',8,8'-tetracyanoquinodimethane is
based on the measurement of the charge–transfer complex. All Loading
of these methods are not described as stability indicating. How- The samples were applied to the marked start edge of the
ever, the USP 24 assay method described azaerythromycin A as TLC plate (1.5-cm height) using the specified TLC–Hamilton
the only potential impurity expected in the bulk material. In glass syringe (25-µL capacity). The sample volumes for the
order to estimate the shelf life, the suitability of the packaging assay experiments were 10 µL, and the volumes spotted for the

Downloaded from https://academic.oup.com/chromsci/article/41/1/10/374909 by guest on 07 October 2020

material, the compatibility with the combined excipients, or purity or stability testing were 20 µL. The plate was then
the validation of batch processing, a well-validated stability allowed to air-dry for 10 min and then inserted into the TLC
indicating assay method should be used. tank for development. The TLC plate was developed for no less
In order to meet the regulatory guidance of the Federal than a 15-cm migration distance of the solvent from the start
Drug Administration/International Conference of Harmoniza- line.
tion (ICH), the material should be forcibly degraded until
10–30% of the major compound degrades (6). The method Preparation of detection reagent
should be able to detect and quantitatively estimate the degra- Forty grams of potassium iodide was dissolved in 100 mL of
dation products generated. In this work, a stability-indicating water (solution 1), and approximately 850 mg of bismuth sub-
assay procedure is described to estimate the potential impurity nitrate was dissolved in 50 mL glacial acetic acid (20% in
A (azaerythromycin A) and other possible degradation prod- water) (solution 2). Five grams of sodium nitrite was dissolved
ucts. Also, the compatibility, impurities profile, and applica- in 100 mL water (solution 3). Immediately before use, a mix-
bility are investigated. From a regulatory point of view, the ture of 10 mL of solution 1, 10 mL of solution 2, and 20 mL of
method developed can also be used to fulfill the essential sim- glacial acetic acid was prepared and diluted to 100 mL with
ilarity and drug-development criteria as part of the summary water in a 100-mL volumetric flask (Dragendorff’s reagent)
of product characteristics required by notes to an applicant (7). (solution 4).
After drying the developed TLC plate (at room temperature
for approximately 20 min), the detection reagent (solution 4)
was sprayed (~ 8 mL was consumed), the plate was air-dried for
Experimental 10 min, and then the plate was sprayed with solution 3 (5%
sodium nitrite). The entire plate was turned a brown color.
Equipment The plate was allowed to air-dry for no less than 30 min at
The UVP scanner and software (GelWorks 1D Advanced Ver- room temperature. The plate was then scanned and integrated
sion 3.01) were from Ultra Violet Products (Cambridge, U.K.). to allow lanes and bands identification and quantitation. The
A test tube atomizer (12 mL) from Desaga GmbH (Wiesloch, scanning and data integration were performed by using a Gel-
Germany) was connected with the positive-pressure outlet works 1D UVP instrument. The concentrations of the eluted
valve of a membrane pump. A thin-layer chromatographic spots were correlated to the light intensity (pixels).
(TLC) spotting syringe (25 µL) was obtained from Hamilton
(Reno, NV). Nylon-sample filtration discs (0.45 µm) were used. Procedure
A TLC Tank (standard type) and minivials lined with a silylated Standard preparation of AZT RS solution
tetrafluoroethylene cap (1 mL) were from Alltech (Deerfield, Approximately 52.4 mg of AZT dihydrate USP RS was accu-
IL). rately weighed into a 25-mL volumetric flask. It was then dis-
solved and diluted to the mark with dichloromethane (2 µg/µL
Materials as AZT).
TLC aluminum sheets precoated with silica gel 60F254 (20
× 20 cm, 0.25-mm layer thickness) were obtained from E. Standard solution of azaerythromycin A
Merck (Darmstadt, Germany). HPLC-grade solvents and other Approximately 20 mg of azaerythromycin A was accurately
chemicals were of analytical grade. AZT dihydrate RS and weighed into a 25-mL volumetric flask. It was then dissolved
azaerythromycin A (impurity A) RS were from USP RS and diluted to volume with dichloromethane (0.8 µg/µL).
(Rockville, MD). AZT dihydrate active in bulk was a free AZT sample solutions for assay were prepared as described
sample from Orchid Chemicals and Pharmaceutical (Chennai, under standard preparation (0.8 µg/µL).
India). Zithromax (250-mg capsules, batch no. 9114) (Pfizer,
Cairo, Egypt) and Azalid (250-mg capsules, pilot batch) (T3A Sample solution for purity testing
Industrial, Assiut, Egypt) were obtained directly from the Approximately 262 mg of AZT dihydrate (equivalent to 250
manufacturer. mg AZT) was accurately weighed and dissolved in 5 mL

11
 Journal of Chromatographic Science, Vol. 41, January 2003

dichloromethane into a 5-mL volumetric flask (1 mg/20 µL). prepared in water, and 5 mL of this solution was sonicated in
an ultrasonic bath for 30 min (47 kHz). A volume of 20 µL was
Zithromax (250-mg capsule solution) applied to the TLC plate.
The content of one capsule was transferred into a 5-mL vol-
umetric flask. An aliquot of approximately 4 mL dichloro- Heating of 250-mg capsules of Zithromax
methane was then added and the flask shaken for Two glass vials were used, each containing a content of one
approximately 2 min. It was then diluted to the mark with the 250-mg capsule of Zithromax. One was mixed with 10 µL of
same solvent. The contents of the flask were filtered with water. Both vials were capped tightly and allowed to stand in a
syringe filtration disks (0.45 µm) (solution 5). heating incubator at 60°C for 14 days. The vial contents were
An accurate volume of 20 µL was then applied to the TLC extracted by dichloromethane, filtered with syringe filtration
plate for purity testing (1 mg/spot). disks, and diluted to obtain a claimed concentration of 1 mg/20
T wo milliliters of solution 5 was then further diluted to µL AZT. A volume of 20 µL was applied to the TLC plate. The
yield a claimed concentration of 2 mg/mL of AZT. An accurate same procedure was repeated but using 250 mg Azalid.

Downloaded from https://academic.oup.com/chromsci/article/41/1/10/374909 by guest on 07 October 2020

volume of 10 µL of this later solution was applied to the TLC In all of the stress cases, the same plate included the other
for assay of 250-mg capsules of Zithromax. reference spots at three levels within the calibration curve (5,
10, and 25 µg/spot) in addition to the spot of azaerythromycin
Standard calibration of AZT solution A (8 µg/spot). The plate was then allowed to air-dry for 10 min
Five concentration levels of AZT USP RS were prepared in before the development. After development, the plate was then
dichloromethane. A 10-µL volume of each level was applied to allowed to dry, sprayed with both developed reagents (solution
the TLC plate (5, 10, 15, 20, and 30µg/spot). 4 and solution 3 as described previously), and the detected
spots evaluated by using the UVP instrument. The gray scan
Forced degradation and one-dimensional mode was applied. The spot lanes and
Boiling bands were defined, detected, and integrated quantitatively.
A sample solution of AZT for purity testing (1 mg/20 µL) was
prepared in water, and 1 mL of this solution was heated to
100°C in capped minivials inserted in a temperature-controlled
block heater for 30 min. The vials were cooled before opening, Results and Discussion
and a volume of 20 µL was applied to the TLC plate (equivalent
to 1 mg AZT/spot) in order to estimate and detect the degra- AZT has insignificant UV absorbance even in high concen-
dation products generated. Another portion of the boiled tration. Therefore, this work was designed to separate AZT
sample solution was diluted to yield a claimed concentration of from its possible degradation products or its potential impuri-
20 µg/spot of AZT. This was to estimate the remaining amounts ties. We went through three phases to achieve our goals, which
of the intact drug. were the development and optimization, validation, and applic-
ability of the method. Several mobile systems were tried; how-
Acid and alkali hydrolysis of AZT ever, a suitable separation of AZT on a TLC plate could best be
A sample solution of AZT (150 mg/mL) was prepared in achieved by using a development system as described in the
dichloromethane. In a minivial, an equivalent volume (0.2 mL “Experimental” section. In order to detect the eluted spots,
each) of this latter solution was mixed with 0.05N hydrochloric freshly prepared Dragendorff’s reagent was sprayed on the air-
acid. In another vial the same concentration of AZT using an dried plate. Upon using this reagent the spots corresponded to
equivalent volume of 1N sodium hydroxide was prepared. In both AZT only, were colored orange, and rapidly faded to colorless
cases, the prepared solution was heated at 60°C for 30 min with bands. After spraying with Dragendorff’s reagent (solution 4)
intermittent shaking. The samples were cooled to room temper- and re-drying the plate, a solution of 5% sodium nitrite was
ature, neutralized with an amount of acid or base equivalent to sprayed. The plate then became fully brown as a result of the
that previously added, and 20 µL was applied to the TLC plate regeneration of iodine. After standing at room temperature
(the claimed AZT concentration is 1 mg/spot). Furthermore, a for approximately 30 min, a brown band corresponding to AZT
portion of both solutions was diluted to get a claimed concen- and any degradation products persisted on the plate with a
tration of 20 µg/10 µL, and a 10-µL volume of the sample was light yellow background. The separated spots were almost cir-
loaded to estimate the remaining amounts of the intact drug. cular with no tailing, which allowed for the accurate mea-
suring of the Rf value. The validity of this procedure was tested
Oxidation for the following parameters: limit of detection (LOD), limit of
An aliquot of 0.2 mL of an AZT sample solution (150 mg/mL quantitation (LOQ), linearity, range, precision, accuracy, selec-
prepared in CH2Cl2) was mixed with 0.4 mL of a hydrogen tivity, robustness, ruggedness, and sample solution stability.
peroxide solution (10%) in a minivial capped and heated at Before starting the evaluation of the separated bands, the fol-
60°C for 30 min with intermittent shaking. A volume of 20 µL lowing steps have to be followed in order to get an accurate
was applied to the TLC plate. representative chromatogram using Gelworks UVP 1D soft-
ware: (a) open the Gelworks 1D software and download the
Sonication scan of TLC as a new file, (b) start to assign a lane for each band
A sample solution of AZT for purity testing (1 mg/20 µL) was pathway, (c) start band detection, (d) remove the background

12
Journal of Chromatographic Science, Vol. 41, January 2003

in order to get an accurate start and end of each band, (e) the However, the lower LOQ was estimated after the spotting of
obtained chromatogram can then be evaluated quantitatively, AZT RS (1.0 µg/spot) nine times and calculation of the relative
(f) assign standard spot band and define both the strength and standard deviation (RSD) value of the response (raw volume).
measuring unit and type of quantitation (i.e., linear, nonlinear, The RSD value of the band raw volume was no more than 3%.
or matching), and (g) use the calibration curve generated in The LOD value of azaerythromycin A was 1.4 µg/spot, and the
order to calculate all other unknown or unmatched bands that LOQ was estimated on the basis of the expected maximum
were detected. limit (which was 0.8%) to correspond with 8 µg/spot (Table I).

Linearity and range Precision

The results obtained revealed a good linear calibration fit System suitability testing was performed on TLC plates using
between the band volume (concentration in micrograms) and a standard solution of AZT prepared in dichloromethane (20
light intensity as pixels (raw volume) in the range of 5 to 30 µg/spot). The RSD of the raw volumes (pixel intensity) of the
µg/spot. The calibration curve was generated automatically bands separated was no more than 1.22% (Figure 3). Also, the

Downloaded from https://academic.oup.com/chromsci/article/41/1/10/374909 by guest on 07 October 2020

once the method of calculation was defined (Figure 2). The cal- %RSD of the calculated Rf value was close to zero.
ibration equation calculated was:
Selectivity
y = 7.79E–005x + 0.05944 Eq. 1 The proposed method was able to discriminate completely
between the major compound and its potential impurity A and
where y is the strength and x corresponds to the band raw other possible degradation products. The following lanes were
volume (response) with a square regression coefficient value developed and matched together: sample AZT RS, premelted
equal to approximately 0.9957. AZT sample solution, spiked AZT with impurity A, and impu-
The calibration curve was valid only for the calculation of an
unknown AZT concentration separated on the same plate
because the brown color of the bands faded proportionally and
slowly.

LOD and LOQ

The LOD was estimated for nine bands that could be detected
from the TLC background. The LOD found was 0.3 µg/spot.

Figure 3. The repeatability of peak pixel intensity related to the corre-

sponding spot.

Figure 2. Generated calibration curve of onplate derivatized AZT dihydrate.

Table I. LOD and LOQ of Both AZT RS and Impurity A

LOD LOQ
Compound (µg/spot, %RSD*) (µg/spot, %RSD*) Figure 4. Thin-layer chromatogram of pure and degraded AZT: (A) effect of
heat on Zithromax capsule, (B) effect of heat on Zithromax capsule, (C)
AZT RS 0.30, 6.2 1.00, 2.9 effect of moisture on Zithromax capsule, (D) effect of sonication, (E) effect
Impurity A 1.40, 7.3 8.00, 2.3 of light, (F) AZT raw material (all stressed and unstressed AZT concentra-
tions used were 1 mg/spot), (G) 5 µg/spot AZT RS, (H) 10 µg/spot AZT RS,
* %RSD of the response of nine repetitive experiments. (I) 25 µg/spot AZT RS, and (J) 8 µg/spot impurity A RS.

13
 Journal of Chromatographic Science, Vol. 41, January 2003

rity A RS. The strength of all samples

Table II. Accuracy of the Developed Method Using AZT RS
was 1 mg/spot, and impurity A RS
loading was 8 µg/spot. The same chro- Spiked concentration Measured concentration*
matographic profile of the degradation (µg/spot) (µg/spot, mean ± SD) %RSD %Deviation†
products of the stressed AZT was
obtained in order to confirm method 1.0 0.95 ± 0.05 5.26 5.00
selectivity. 5.0 4.88 ± 0.21 4.30 2.40
10.0 9.85 ± 0.30 3.04 1.50
Accuracy 20.0 19.80 ± 0.27 1.36 1.00
Five concentration levels of AZT RS 30.0 30.31 ± 0.27 0.89 1.03
(including LOQ) were prepared and
analyzed by the proposed method and * Mean of 3 experiments.
† %Deviation = [(spiked concentration – mean measured concentration) × 100] / spiked concentration.
calculated from the calibration curve

Downloaded from https://academic.oup.com/chromsci/article/41/1/10/374909 by guest on 07 October 2020

derived on the same TLC plate.
Because the recovered amounts of AZT
RS were within the acceptable range of
100% ± 5% of the claimed amount, the
method was deemed to be accurate
(Table II).

Robustness
In order to measure the extent of the
method robustness, the most critical
parameters were interchanged while
keeping the other parameters
unchanged, and in parallel the chro-
matographic profile was observed and
recorded. The chromatographic para-
meters (including spot area, raw
volume, migration distance, and reso-
lution of the intact drug from its degra-
dation products) were observed. The
parameters interchanged (in the range
of 10%) were included as well as the
composition of the mobile system alter- Figure 6. Linear calibration curve of AZT (A), and one-level calibration fit of standard impurity A (B). The
natively, spray amounts of Dragendorff’s data of both curves were retrieved from the same TLC plate.
reagent, the amount and strength of
sodium nitrite solution, drying time,
and drying temperature. The resolution
and detection of the analyzed spot mate-
rial was relatively acceptable in all con-
ditions, except that the drying time
should be adequate before development,
derivatization, and scanning. Also, the
plate should be left to dry in the open air
at room temperature without any
heating. If the plate were to be heated in
an oven at approximately 40°C, the spot
color and background will fade, but
nonproportionally.

Ruggedness
The ruggedness of the method was
evaluated by applying the analysis of Figure 5. Representative one-lane chromatogram of forced degraded AZT (effect of sonication): (A)
Zithromax capsule solutions using two AZT, (B) unknown impurity, (C) degradation product matched with impurity A, and (D) unknown
different TLC plates from two different impurity.
manufacturers [Merck and Fluka

14
Journal of Chromatographic Science, Vol. 41, January 2003

Chemie GmbH (Reisenhofen, Germany)] of different layer degradation products and impurity profiles in all cases were
thicknesses (0.25 and 0.20 mm). Both plates yielded the same similar with the complete separation of degradation products
resolution efficiency and Rf values; however, the thinner-layer from the major drug spot. In order to estimate the unknown
plate required a longer time for the solvent to reach the marked degradation products, a three-point calibration curve was
front. A compact band zone without tailing was observed in used on the same plate. Also, in order to estimate the known
both cases. impurity (azaerythromycin A), one level concentration was
The percentage recoveries of the capsule content using two used to estimate the matched detected impurity A released
different TLC plates were found comparable (~ 100% ± 1.5%). from the raw material or product. Figure 5 is an example of
Awareness should be paid to the amount of samples delivered how to detect and evaluate the one-lane degradation profile.
from the spotting syringe because this step expresses the mate- Also, as in Figure 6, the calibration curves were calculated
rial concentration. A calibrated autospotting device could be on the same plate. From this investigation, it was clear that
useful to avoid sample volume variation. However, the loaded AZT was very sensitive toward acid, and if the hydrolysis
volume area variation did not make any difference in the time or acid strength increased, the entire compound would

Downloaded from https://academic.oup.com/chromsci/article/41/1/10/374909 by guest on 07 October 2020

results. be completely degraded with no spot corresponding to AZT.
In contrast, the compound was relatively stable in an alkali
Sample solution stability medium. Table III shows the recovered amounts of impurity
As per the experimental procedure described, the develop- A and total impurities in all stress conditions. As per ICH
ment of approximately 20 samples on one plate will consume guidelines and drug dose, the limit of impurity A should not
approximately 30 min, which is not enough time to allow for be more than 1% (0.8% was used as an acceptable limit),
much material degradation. The proposed procedure did not unknown impurities no more than 0.5%, and total impuri-
show any degradation products resulting from experimental ties no more than 2.0%.
error or run time. However, freshly prepared samples were Two strengths of AZT were prepared from the same Zithro-
used within 2 h, and the sample solutions should not be max and Azalid capsules or raw material stock solutions. The
allowed to stand until analysis. higher strength (1 mg/spot) was spotted to evaluate the purity
level of the sample. However, the diluted solution (20 µg/spot)
Forced degradation of AZT was spotted to assay the content of material that remained.
A working standard of AZT was used for this study. This The stress testing results revealed that AZT was compatible
was to confirm that the proposed method was able to detect with the combined excipients and degraded to a limit of no
and analyze the major compound in the presence of any more than 3.55% for 250-mg capsules of both Zithromax and
possible degradation product or in-process impurities Azalid.
resulting from the manufacturing procedure. Both raw
material and capsule form were included in this study. A
relatively high concentration of AZT was prepared in
dichloromethane or water, following the stress conditions Conclusion
applied. As described in the “Experimental” section, different
stress cases were applied in order to include the effect of The method developed was valid for purity testing, a sta-
boiling, heat, acid, base hydrolysis, sonication, oxidation, bility-indicating assay, and content-uniformity testing. Also,
and excipient compatibility with the drug. A chromatogram this method could be used for the quantitation of pure mate-
of stressed samples illustrated in Figure 4 shows that the rial and capsule form. The developed method was able to esti-
mate the dissolution profile of
Table III. Calculated Amount Percentages of Impurity A and Total Impurities* Zithromax capsules applying the USP
method, but using 500 mL dissolution
Calculated amounts percentage media instead of 900 mL. This method
Stress condition Impurity A Total impurities could be useful for the study of pure
material and product shelf life. All pro-
Boiling 2.32 14.28 posed analytical procedures should be
Acid hydrolysis 26.66 38.36 completed at room temperature. This is
to avoid inconsistency of the developed
Alkali hydrolysis 1.47 11.84
brown color.
Oxidation 2.64 14.86
Sonication 0.26 0.81
Light 0.53 0.72
Heat, dry capsule content 1.22 2.11 Acknowledgments
Heat, moisture, capsule content 2.30 3.55
Unstressed sample (1 mg/spot) 0.14 0.26 The authors acknowledge Dr. Tarek
El-Hady, chairman of T3A Industrial, for
* The amount of total impurities was calculated from the calibration curve of AZT RS, and the amount of impurity
A was calculated from the calibration fit of impurity A RS. his encouragement and funding of the
project.

15
 Journal of Chromatographic Science, Vol. 41, January 2003

References formance liquid chromatography using pre-column derivatization

with 9-fluorenylmethyloxycarbonyl chloride and fluorescence
detection. J. Chromatogr. B Biomed. Sci. Appl. 720: 89–97 (1998).
1. E.F. Fiese and S.H. Steffen. Comparison of the acid stability of 5. L.I. Bebawy, K. el Kelani, L. Abdel Fattah, and A.K. Ahmad. Study
azithromycin and erythromycin A. J. Antimicrob. Chemother. of 7,7’,8,8’-tetracyanoquinodimethane charge transfer complexes
25(suppl. A): 39–47 (1990). with some lone-pair-donating drugs. J. Pharm. Sci. 86: 1030–33
2. United States Pharmacopoeia 24, USP-NF, Rockville, MD, 2000, (1997).
p. 185. 6. “Stability Testing of New Active Substances and Medicinal Prod-
3. F. Kees, S. Spangler, and M. Wellenhofer. Determination of ucts, ICH Topic Q1A”. ICH Harmonized Tripartite Guidelines.
macrolides in biological matrices by high-performance liquid ICH, Surrey, U.K., 1996. CPMP/ICH/280/95.
chromatography with electrochemical detection. J. Chromatogr. 7. “Eudralex, Notice to Applicants, Medicinal Products for Human
A 812: 287–93 (1998). Use”. Presentation and Content of Dossier, October 25, 2001.
4. J.S. Torano and H.J. Guchelaar. Quantitative determination of European Commission, Brussels, Belgium, Vol. 2B.
the macrolide antibiotics erythromycin, roxithromycin,
azithromycin and clarithromycin in human serum by high-per- Manuscript accepted July 30, 2002.

Downloaded from https://academic.oup.com/chromsci/article/41/1/10/374909 by guest on 07 October 2020

16