`Solutions Containing Tyloxapol by Solid-Phase Extraction and
`Reversed-Phase High-Performance Liquid Chromatography
`
`TONY Y. FAN’ AND G. MICHAEL WALL
`Received September 3, 1992, from Analytical Chemistry, Alcon Laboratories, Inc., 6201 South freeway, fort Worth, ?X 76134.
`publication February 25, 1993.
`
`Accepted for
`
`Abstract 0 A procedure using solid-phase extraction (Supelcoclean
`CN) followed by HPLC [Beckman Ultrasphere CN, acetonitri1e:phos-
`phate solution (60:40, vlv)] was developed and validated to quantitate the
`quaternary ammonium presetvative benzalkonium chloride in an exper-
`imental ophthalmic formulation containing the polymeric material tylox-
`apol. This procedure makes routine determinations of benzalkonium
`chloride at concentrations of 0.0035 to 0.01 %simpler than the traditional
`ion-pairing colorimetric methods. This method is quick, specific, and
`especially useful for drug product stability studies. In addition, because
`the method distinguishes each homologue, it can be extended to
`routinely determine the homologue ratio for quality control purposes.
`
`Benzalkoniuml chloride (BAC; 1) is widely used as an
`antimicrobial preservative in aqueous pharmaceutical prep-
`arations, especially in ophthalmic solutions. BAC is actually
`a mixture of n-alkylbenzyldimethyl ammonium chlorides
`with n-alkyl chain lengths varying from C, to Cl,.l Because
`the homologues present different bactericidal activity,z it is
`sometimes necessary to determine not only the total amount
`of BAC but also the ratio of its homologues in the formula-
`tions. Among the European,3 British,4 and United States1
`pharmacopoeias, only the USP specifies the percentage of
`individual homologues: (1) the content of the n-C12H25 ho-
`mologue is not <40.0%, (2) the content of the n-Cl,Hzs
`homologue is not <20.0%, and (3) the total content of the
`C1zHz5 and C,,H,,
`homologues comprise together not
`
`= GH17
`
`c18H37
`
`L
`
`-
`
`< 70.0% of the total alkylbenzyldimethylammonium chloride
`content .1
`A quick and easy method for the determination of both the
`total and relative homologue ratio is desirable. HPLC with
`U V detection is a desirable technique because of its separa-
`tion capability and suitability for automated routine analysis.
`For ophthalmic preparations, such determinations are not
`trivial because BAC is usually present in very low concen-
`trations (0.007-0.01%, w/v), requiring low W detection
`wavelengths (210-215 nm) for good sensitivity. Because other
`excipients are usually present in much higher concentrations,
`interference at these low wavelengths is a common problem.
`Over the years, various specific and nonspecific methods
`have been developed for the determination of BAC. These
`have included extraction by complexing BAC with various
`dyes,5-9 titration of quarternary ammonium compounds with
`iodatelo or tetraphenylboron,ll pyrolysis and subsequent gas
`chromatography,lz chemical ionization mass spectrometry,l3
`and HPLC of simple aqueous solutions.1616 In addition, the
`USP monograph describes a reversed-phase HPLC method to
`determine the homologue ratio of BAC raw material at a
`relatively high concentration (4 mg/mL) and a titration
`method to determine the total content of BAC in solution
`based on potassium iodate equivalents.1 None of these meth-
`ods could be directly used for the analysis of BAC in complex
`ophthalmic solutions because either they do not have the
`required specificity and sensitivity or they can not completely
`separate BAC from the matrix. Interferences have been
`observed by the presence of polymeric material, suspended
`particles, and active ingredients. These kinds of samples are
`not suitable for direct HPLC injection and therefore require
`some kind of sample preparation prior to HPLC.
`The purpose of this study was to develop an HPLC method
`appropriate for measuring BAC in an experimental oph-
`thalmic solution containing BAC (0.007%, w/v) and the
`polymeric material tyloxapol (0.25%, w/v). Tyloxapol (21, a
`polymeric alkyl aryl polyether alcohol commonly used as an
`emulsifier or surfactant, presented a problem for HPLC
`analysis of BAC because it produced a large solvent front that
`partially masked the BAC peaks. The USP HPLC method for
`BAC could not be used because of the interference of tyloxapol
`and the lack of sensitivity at 254 nm. To solve this problem,
`a combination solid-phase extraction (SPE)/HPLC procedure
`was developed. The combination of SPE sample clean-up with
`the resolving capability of HPLC provided a powerful tool for
`the routine analysis of complex ophthalmic solutions. This
`paper describes a SPE/HPLC method suitable for the deter-
`mination of the total BAC content as well as each homologue
`ratio in an experimental ophthalmic solution containing
`tyloxapol. This technology should be applicable to other types
`of complex formulations.
`
`11 72 I Journal of Pharmaceutical Sciences
`Vol. 82, No. 17, November 1993
`
`0022-3549/93/1100- 1 172$02.50/0
`0 1993, American Pharmaceutical Association
`
`LUPIN EX 1055
`
`Page 1 of 3
`
`
`
`Experimental Section
`Apparatus-An HPLC system that consisted of a Hewlett-Packard
`1090 quaternary pump (Hewlet-Packard, Fullerton, CA), a Waters
`Associates (Waters, Milford, MA) WISP 710B autoinjector, 490
`programmable multiwavelength detector, and a Spectra-Physics
`ChromJet Integrator (Spectra-Physics, San Jose, CA) was used. All
`HPLC separations were performed isocratically on a 5 pm (150 x 4.6
`mm, id.) Ultrasphere cyano (nitrile-bonded silane, CN) column
`(Beckman, San Ramon, CA). A Burdick & Jackson, 12-port solid-
`phase extraction manifold (Burdick & Jackson, Muskegon, MI) was
`used for the sample extractions. The manifold was connected to a n
`in-house vacuum source, and a control valve was used to regulate the
`magnitude of vacuum applied. The container itself was large enough
`to allow 12 10-mL volumetric flasks to be attached to the rack at the
`same time for sample collection. A stop valve was also provided on
`each port for individual flow stoppage. All extractions were performed
`with Supelcoclean (Supelco, Bellefonte, PA) disposable cyan0 SPE
`columns with 1-mL capacity.
`reagents and solvents were reagent
`Reagents and Solutions-All
`or HPLC grade and purchased from J. T. Baker (Phillipsburg, NJ).
`The phosphate solution was prepared by dissolving 6 mL of concen-
`trated phosphoric acid (reagent grade) in 1950 mL of distilled water.
`The pH was adjusted to 5.0 by the addition of 50% NaOH solution, and
`the total volume was adjusted to 2 L with distilled water. The mobile
`phase was acet0nitrile:phosphate solution (60:40, v/v), and the wash
`solvent was acet0nitrile:phosphate solution (30:70, v/v). An experi-
`mental ophthalmic formulation was used for this study that con-
`tained proprietary drug (0.1%), mannitol (4.7%), sodium citrate
`(0.04%), citric acid (0.02%), tyloxapol (0.25%), BAC (0.007%), and
`edetate disodium (0.01%; all w/v). BAC was deleted for validation
`purposes, making this formulation a BAC vehicle.
`Sample Preparation-Test
`solutions were prepared by the addi-
`tion of appropriate amounts of BAC to an ophthalmic solution BAC
`vehicle (an ophthalmic solution containing all ingredients except
`BAC) that contained tyloxapol(O.O25%, w/v) as one of the ingredients.
`A flow control valve was attached to each SPE column, and the whole
`unit was placed onto the female her fitting of the vacuum manifold.
`Reduced pressure (-10 mmHg) was applied to the manifold with an
`in-house vacuum line. The SPE columns were conditioned with
`acetonitrile (2 mL) followed by distilled water (2 mL). When the level
`of the distilled water had reached - 1 mm above the top of the column
`packing, slow addition of the test sample (4 mL) was initiated.
`(Caution was taken not to disturb or dry out the column packing bed).
`After the sample had passed through, the column was washed with
`wash solvent (2 mL). The vacuum was disconnected, a 10-mL
`volumetric flask was placed under each SPE column, then the reduced
`pressure was applied again. The retained BAC was eluted from the
`column with mobile phase (5 mL), the vacuum was then disconnected,
`and the flasks were removed and diluted to volume with distilled
`water. These samples were directly analyzed by HPLC.
`HPLC Assay Procedure-The mobile phase was mixed and
`filtered before use. The chromatographic system employed a flow rate
`of 2 mL/min, 100-pL injection volume, 10-min run time, W detection
`(210 nm) at 0.01 AUFS, a recorder attenuation of x8, and a chart
`speed of 0.5 cndmin. After a stable baseline was established, replicate
`standards were injected to ensure reproducibility prior to sample
`analysis. System suitability criteria were established: relative stan-
`dard deviation of six replicate injections, 52.0%; resolution between
`the C,, and C,, peaks, 22; tailing factor for the C,, peak, 52; and
`number of theoretical plates, 13000 plates/column. A standard was
`inserted between every six samples. The BAC homologues were
`quantitated by the calculation described in the USP HPLC method,'
`taking into consideration the molecular weight of each homologue.
`The percentage of each BAC homologue and the percent recovery of
`total BAC were calculated as follows: %of each homologue = 100 AB,
`and % recovery = 100 B,, 5standard,
`where A i G the product of the
`area obtained from each Lomologue multiplied by its molecular
`weight and B is the sum of all of these products. The molecular
`weights of the Clo, C,,, C,,, C,,, and C,, BAC homologues (most
`common) are 312, 340, 368, 396, and 424, respectively.
`
`polymeric material tyloxapol. The low concentration of BAC
`in this experimental ophthalmic formulation (0.007%, w/v)
`necessitated using low UV wavelength (210 nm) detection for
`increased sensitivity. However, this low UV wavelength
`magnified interference problems encountered with direct
`RPLC analysis: tyloxapol eluted as a large peak after the
`solvent front, making quantitation of the BAC C,, homologue
`difficult and the BAC Clo homologue impossible (Figure 1).
`Solid-phase extraction was employed prior to HPLC to remove
`most of the interference by tyloxapol and, thereby, reduce
`excessively long run times (Figure 2). Only SPE columns from
`one manufacturer were used to obtain the data herein because
`vendor-to-vendor variability in SPE columns has been previ-
`ously reported for cyano cartridges.17
`Validation data were generated for this method with the
`experimental formulation. Linearity was satisfactory (Table
`I). Three six-point vehicle standard curves (duplicate samples
`at three different concentrations) were generated for the
`experimental ophthalmic formulation with concentrations of
`BAC ranging from 50 to 150% Label [0.007% (w/v) BAC =
`100% Label; the concentration range for injected samples was
`0.014-4.042 mg/mL]. The curves obtained were linear (r2 =
`0.999) and the y-intercepts, ranging from 1.2 to 3.1%, were
`small enough to justify the use of a single-point standard
`(Table I). Total recoveries were acceptable and in the range
`97-103% (Table I). The precision was also satisfactory (Table
`11). The injection of three sets of six vehicle standard repli-
`cates [0.007% (w/v) BAC] gave acceptable values for relative
`standard deviations (ranging from 0.74 to 1.52%).
`Though good results were obtained with this method most
`of the time, spurious results were infrequently observed:
`occasionally, an unexpected value (low or high by -2%) was
`obtained for BAC. A significant amount of effort was expended
`trying to track down spurious data that might have been the
`
`a b
`
`d
`
`1
`0
`
`I
`6
`
`1
`10
`
`Results and Discussion
`A combination SPE/HPLC method was developed for de-
`termination of BAC in ophthalmic solutions containing the
`
`tlmo
`(minutor)
`typical HPLC chromatogram of BAC sample prior to
`Figure 1-A
`extraction: (a) drug; (b) tyloxapol; (c) BAC Clo; (d) BAC C12; (e) BAC C,,;
`(9 BAC Cte; (9) BAC Cie.
`
`Journal of Pharmaceutical Sciences / 1 173
`Vol. 82, No. 11, November 1993
`
`Page 2 of 3
`
`
`
`b
`
`c
`
`C
`
`:
`0 1 a
`
`(minutoo)
`Tim.
`typical HPLC chromatogram of BAC sample after SPE
`Flgure 2-A
`extraction: (a) BAC Clo; (b) BAC C12; (c) BAC C14; (d) BAC C16; (e) BAC
`ClE?
`
`~
`
`Concentration,
`mglmL
`
`Table I-EtAC Vehicle Standard Curves'
`Area Counts (Recovery %)
`Standard
`Standard
`Standard
`Curve 2
`Curve 3
`Curve 1
`0.014
`128 352 (98) 128 372 (98) 123 153 (98)
`0.01 4
`128 641 (98) 125 067 (97) 124 270 (98)
`263 479 (98) 264 238 (101) 258 765 (100)
`0.028
`263 836 (98) 269 979 (1 00) 263 762 (1 02)
`0.028
`393 348 (97) 394 036 (1 03) 397 71 0 (1 03)
`0.042
`397 215 (98) 389 692 (100) 392 404 (101)
`0.042
`P
`0.999
`0.998
`0.999
`Avg. Recovery, % 98
`100
`100
`aSamples ranging from 50% to 150% of the target concentration
`(0.028 mglml) were prepared by solid-phase extraction and were
`analyzed by high-performance liquid chromatography.
`
`result of one or more possibilities; for example, allowing the
`SPE column to dry after conditioning and SPE column-to-
`column variability. The drying of a column after conditioning
`probably resulted in desolvation of the column packing and,
`hence, variable adsorption characteristics. 18 Also, batch-to-
`batch variation has been previously reported for disposable
`SPE columns for basic drugs on cyan0 (Bakerbondlg) or C18
`(Bakerbond's or Polymer Institute2O brands), and catharan-
`thus alkaloids on diol (Analytichemzl) cartridges, suggesting
`the possibility of poor quality control of the SPE column
`packing process.
`In conclusion, this method was proven to be sensitive,
`specific, precise, and accurate for the SPE/HPLC analysis of
`BAC in an experimental ophthalmic solution containing
`
`1 174 I Journal of Pharmaceutical Sciences
`Vol. 82, No. 1 7 , November 1993
`
`Table II-BAC Vehicle Standard Replicates'
`Concentration,
`Area Counts (Recovery %)
`mglmL
`Replicate Set 1 Replicate Set 2 Replicate Set 3
`0.028
`260 543 (101) 259 993 (101) 258 516 (100)
`259 225 (101) 260 055 (101) 256 329 (100)
`0.028
`0.028
`256 120 (99) 256 736 (100) 253 007 (98)
`0.028
`260 979 (101) 260 260 (101) 263 153 (102)
`0.028
`256 638 (100) 257 514 (100) 263 088 (102)
`257 997 (100) 261 950 (102) 258 469 (100)
`0.028
`Rel. Std. Dev., % 0.78
`0.74
`1.52
`Avg. Recovery, % 100
`101
`100
`a Samples of 100% target concentration (0.028 mg1mL) were prepared
`by SPE and were analyzed by HPLC (see Experimental Section for
`conditions).
`
`tyloxapol. The sample clean-up step (i.e., SPE extraction) and
`lower wavelength detection represent improvements over
`existing methods (e.g., the USP HPLC BAC method) that
`allow for analysis of BAC at low concentrations in tyloxapol-
`containing formulations. The method should be easily
`adapted to other solutions and suspensions containing poly-
`meric material. The disadvantage of an infrequent spurious
`result was easily remedied by reassay of the suspect sample.
`It is thought that the occasional lack of precision was a result
`of variability between the SPE columns. None of these
`problems was deemed significant enough to preclude the use
`of this method because the magnitude of error for total BAC
`content was seldom >2%. Improvements in the commercially
`available SPE columns or alterations in the HPLC conditions
`may render this technique even more reliable, but until then,
`it is still an acceptable method for the analysis of BAC in
`complex ophthalmic solutions.
`
`References and Notes
`1. United States Pharmacopeia, 22nd rev.; U.S. Pharmacopeial
`Convention: Rockville, MD, 1990; p 1905.
`2. Giles, R.; Daoud, N. N.; Gilbert, P.; Dickson, N. A. J. Pharm.
`Pharmacol. 1983, 34fsuppl.), 110.
`3. European Pharmacopeial 2nd. ed.; Council of European: France,
`1985, Part 11-9, p 371.
`4. British PharmacoDeia. vol. 1: British PharmacoDeia Commission:
`U.K., 1988; p 63.' '
`5. Auerbach M. E. Anal. Chem. 1943,15,492.
`6. Colichman, E. L. Anal. Chem. 1947,19,430.
`7. Ballard, C. W.; Isaacs, J.; Scott, P. G. W. J. Pharm. Pharmacol.
`1954, 6, 971.
`8. Chatten, L. G.; Okamura, K. 0. J. Pharm. Sci. 1973, 62, 328.
`9. Marsh, D. F.; Takahashi, L. T. J. Pharm. Sci. 1983, 72, 521.
`10. Brown, E. R. J. Pharm. Pharmacol. 1963,15,379.
`11. Metcalfe, L. D.; Martin, R. J.; Schmitz, A. A. J. Am. Oil Chem.
`SOC. 1966,43, 355.
`12. Jenninw. E . C.: Mitchner. H. J. Pharm. Sci. 1967.56. 590.
`13. Daoud,%. N.; Crooks, P. A.; Speak, R.; Gilbert, P. i. Pharm. Sci.
`1983, 72, 290.
`14. Meyer, R. C. J. Pharm. Sci. 1980, 69, 148.
`15. Ambrus, G.; Takahashi, L. T.; Marty, P. A. J. Pharm. Sci. 1987,
`76, 174.
`16. Comez-Gomar, A.; Gonzalez-Aubert, M. M.; Garces-Torrents, J.;
`Costa-Segarra, J. J. Pharm. Biomed. Anal. 1990,8,871.
`17. Moors, M.; Massart, D. L. Anal. Chim. Acta 1992,262, 135.
`18. Van Horne, K. C. Sorbent Extraction Technology; Analytichem
`International: Harbor City, CA, 1985; pp 1616.
`19. Moors, M.; Massart, D. L. J. Pharm. Biomed. Anal. 1991,9, 129.
`20. Marko, V.; Radova, K.; Novak, I. J. Liq. Chromatogr. 1991, 14,
`1659.
`21. Vendrig, D. E. M. M.; Holthuis, J. J. M. J. Chromatogr. 1987,
`414, 91.
`
`Page 3 of 3
`
`