Journal o_f Chromatography, 353 (1986) 309- 318
`Elsevier Science Publishers B.V. , Amsterdam - Printed in The Netherlands
`
`CHROMSYMP. 742
`
`DETERMINATION OF THE ANTILEUKEMIA AGENTS CYTARABINE
`AND AZACITIDINE AND THEIR RESPECTIVE DEGRADATION PROD(cid:173)
`UCTS BY HIGH-PERFORMANCE LIQUID CHROMATOGRAPHY
`
`LARRY D. KISSINGER* and NICKL. STEMM
`Control Research and Development, The Upjohn Company. Kalamazoo, MI 49001 ( U.S.A .)
`
`SUMMARY
`
`A reversed-phase high-performance liquid chromatography (HPLC) system
`was developed for the determination of the antineoplastic agents cytarabine and
`azacitidine. Separations were performed on an octadecylsilane column with a mobile
`phase of methanol-phosphate buffer pH 7.0 (5:95). The assay methods are suitable
`for bulk drugs and sterile powder formulations of the agents. Specificity in the pres(cid:173)
`ence of analogues and decomposition products was demonstrated. UV spectra of the
`components of interest were obtained in the HPLC effluent, and appropriate wave(cid:173)
`lengths were employed for the various analytes. Samples of azacitidine in various
`solutions were analyzed as a function of time by HPLC to determine the three
`first-order rate constants associated with its decomposition.
`
`INTRODUCTION
`
`Cytarabine (Ara-C; cytosine arabinoside; 1-P-o-arabinofuranosyl cytosine)
`and azaci ti dine ( 5-AC; 5-azacytidine; 4-amino- l-/J-o-ribofuranosyl-s-triazin-2( 1 H)(cid:173)
`one) are nucleoside analogues which have antitumor activity. Cytarabine is formu(cid:173)
`lated as a freeze-dried powder and is marketed for induction and maintenance of
`remission in acute mye1ocytic leukemia (Cytosar-U®, The Upjohn Company). A new
`drug application was made in the U.S.A. for a freeze-dried powder containing equal
`amounts of 5-AC and mannitol (Mylosar®, The Upjohn Company) for induction of
`remission in acute non-lymphocytic leukemia.
`This report describes a reversed-phase high-performance liquid chromato(cid:173)
`graphic (HPLC) method for quantitative determination of these antineoplastic agents
`in bulk drugs and pharmaceutical formulations. The sample preparations and chro(cid:173)
`matographic conditions are simple and rapid. The determinations are specific for
`these agents in the presence of analogues and products of hydrolytic decomposition.
`Although Ara-C and 5-AC are stable in the solid state, they are degraded by
`hydrolysis in aqueous solutions. Hydrolytic deamination of Ara-C (I) results in the
`elimination of ammonia and the formation of uracil arabinoside (II), as shown in
`Fig. 11 •2 . In common infusion solutions of Ara-C, degradation is less than 1 % in five
`days 3 • In neutral and basic solution, the hydrolysis of 5-AC (III) occurs via nucleo-
`CELGENE 2019
`APOTEX v. CELGENE
`IPR2023-00512
`
`© 1986 Elsevier Science Publishers B.V.
`
`0021-9673/86/$03 .50
`
`

`

`310
`
`.;; ~N)
`
`HOCa•o
`H HO
`H
`H
`OH H
`
`L. D. KISSINGER, N. L. STEMM
`
`-
`
`--½
`
`O~N)J
`2
`
`HOC~O ·
`H H
`H
`
`
`
`H
`
`OH H
`II
`Fig. I. Hydrolytic deamination of cytarabine (I) to uracil arabinoside (II).
`
`philic attack, opening the triazine ring at the 5,6-position to form N-(formylamini(cid:173)
`ne)-N'-P-o-ribofuranosylurea (IV), as shown in Fig. 24 • 5 . The N-formyl group of IV
`is eliminated to form 1-P-o-ribofuranosyl-3-guanylurea (V). In strongly acidic solu(cid:173)
`tions, the glycoside bond is hydrolyzed to produce 5-azacytosine and o-ribose. Hy(cid:173)
`drolysis of 5-AC in infusion solutions is much more rapid than that of cytarabine.
`Over 10% of 5-AC is degraded in common infusion solutions within 4 h 6
`• The ki(cid:173)
`netics of 5-AC decomposition have been studied previously by UV spectroscopy4,
`HPLC6 , and NMR spectroscopy 7 .
`
`NH2
`
`NH,
`
`NH2
`
`HNAN
`
`Hto Ao
`
`HOCH 2 HN
`
`H2N~N
`
`Ao
`
`N~N
`
`l..NAo
`HOQ
`
`HO OH
`
`Ill
`
`k12
`
`k,, o~
`HOQ
`
`HO OH
`
`HO OH
`
`IV
`
`V
`
`Fig. 2. Stepwise hydrolysis of azacitidine (III) to N-(formylaminine)-N'-2P-ribofuranosylurea (IV), then
`I -{1-o-ribofuranosyl-3-guanylurea (V).
`
`EXPERIMENT AL
`
`Chromatographic system
`The mobile phase was delivered by an Altex Model ll0A pump (Altex Scien(cid:173)
`tific, Berkeley, CA, U.S.A.). Sample injection was performed with a WISP 710B
`autosampler (Waters Assoc., Milford, MA, U.S.A.) or a manual Model 7010 loop
`injector (Rheodyne, Cotati, CA, U.S.A.). Detection was generally performed at 254
`nm with a LDC Model 1203 detector (Laboratory Data Control, Riviera Beach, FL,
`U.S.A.). A Tracor (Austin, TX, U .S.A.) Model 970A variable-wavelength detector
`was also employed. Some full UV chromatograms were obtained with HP Model
`1040A diode-array detector (Hewlett-Packard, Palo Alto, CA, U.S.A.). Separations
`were performed on a 30 cm x 3.9 mm I.D. µBondapak C 18 column (Waters Assoc).
`The mobile phase, water-methanol (95:5) containing 1.34 g disodium hydrogen phos(cid:173)
`phate heptahydrate, 0. 71 g sodium dihydrogen phosphate monohydrate per liter, was
`pumped at a rate of 1 ml/min. The apparent pH of the mobile phase was 7.0.
`
`

`

`HPLC OF CYTARABINE AND AZACTTJDINE
`
`311
`
`Reagents
`Methanol was distilled-in-glass grade (Burdick & Jackson, Muskegon, MI,
`U.S.A.). Disodium hydrogen phosphate heptahydrate and sodium dihydrogen phos(cid:173)
`phate monohydrate were analytical-reagent grade. The p-toluic acid was obtained
`from Crescent Chemical (Hauppauge, NY, U.S.A.). Cytosine, D-( - )-ribose, and D(cid:173)
`( - )-arabinose were obtained from Sigma (St. Louis, MO, U.S.A.). A sample of 5-
`azacytosine was provided by Ash-Stevens (Detroit, MI, U.S.A.).
`
`Cytarabine preparations and procedure
`Internal standard solution. A 1.4-mg/ml solution of p-toluic acid was prepared
`in methanol.
`Standard preparation. A 0.02-mg/ml solution of uracil arabinoside in water
`was prepared. Approximately 3 mg of Ara-C was accurately weighed and 5.0 ml of
`the uracil arabinoside solution, 5.0 ml of internal standard solution, and 20 ml of
`mobile phase were added.
`Bulk drug preparation. Cytarabine (3 mg) was accurately weighed and 5.0 ml
`of internal standard solution and 25 ml of mobile phase were added.
`CYTOSAR- U® sterile powder preparation. The contents of the vial were quan(cid:173)
`titatively diluted with water to prepare a 1-mg/ml solution of Ara-C. A 3.0-ml portion
`of this Ara-C solution was combined with 5.0 ml of internal standard solution and
`20 ml of mobile phase.
`Procedure. Portions of 10 µ1 of the preparations were injected. The cytarabine
`and uracil arabinoside content of the samples were calculated by comparing the ratio
`of the peak response relative to the internal standard to the ratio of the standards.
`
`Azacitidine preparations and procedure
`Internal standard solution. A 2-mg/ml solution of p-toluic acid was prepared
`in water-methanol (20:80).
`Standard and bulk drug preparations. A 1-mg/ml solution of 5-AC was prepared
`in internal standard solution.
`MYLOSAR® sterile powder preparation. The contents of the vial were quan(cid:173)
`titatively diluted with internal standard solution to prepare a 1-mg/ml solution of
`5-AC.
`Procedure. Exactly 15 min after addition of the internal standard solution to
`the 5-AC, a 2-µl portion of the preparation was chromatographed. The 5-AC content
`of the sample was calculated by comparing the ratio of the peak responses to the
`internal standard to the ratio of the standards.
`
`Azacitidine decomposition
`Decomposition of 5-AC in several solutions was monitored by performing
`from 50-120 HPLC assays as a function of time. Large•volume parenteral (L VP)
`solutions in 1-1 glass bottles and plastic bags were obtained from Travenol Labs.
`(Deerfield, IL, U.S.A.) and Abbott Labs. (Chicago, IL, U.S.A.). The pH of the so(cid:173)
`lutions was adiusted to the desired value with hydrochloric acid or sodium hydroxide.
`Solution admmistration sets with particulate filters were obtained from Travenol
`Labs. Direct injection with a 50-µl injection loop and no internal standard was used
`to analyze the dilute solutions. The time range monitored was two to three days,
`except for at 4°C, where the samp]e was monitored for 3 weeks.
`
`

`

`312
`
`RESULTS AND DISCUSSION
`
`L. D. KISSINGER, N. L. STEMM
`
`Since Ara-C and 5-AC are polar molecules, reversed-phase HPLC must be
`performed with a mobile phase containing a low concentration of organic modifier.
`Although 5-AC has been assayed by reverse-phase HPLC without any organic mod(cid:173)
`ifier in the mobile phase 5 , the addition of 5% methanol to the aqueous mobile phase
`improved the reproducibility of retention times.
`Tailing of some of the peaks, e.g. of 5-AC, was observed with unbuffered and
`acidic mobile phases. The irreversible adsorption which resulted in peak tailing is
`assumed to be the result of bonding of the primary amines of the analytes with
`residual silanol group on the surface of the stationary phase. The neutral pH of the
`phosphate buffer in the mobile phase was an effective compromise to assure sym(cid:173)
`metrical peaks and resonable column life. Retention times of the analytes varied from
`column to column and slowly decreased with time. Relative retention behavior of the
`components of interest were reproducible on all columns. With this mobile phase, a
`small, approximately 2 mm, void would form at the head of the column and result
`in reduced chromatographic efficiency. To maintain acceptable chromatographic per(cid:173)
`formance, it was necessary to repack the head of the column after ca. 40 h of opera(cid:173)
`tion. The mobile phase was not bacteriostatic, and the column would be ruined if
`stored with the mobile phase for extended time periods.
`
`Chromatographic specficity
`Selectivity of the chromatographic system is demonstrated for Ara-C, 5-AC,
`
`D
`
`254nm
`
`E
`
`200nm
`
`2 mAU
`
`A
`
`C
`
`8
`
`F
`
`0
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`10
`
`Time (min)
`Fig. 3. Chromatograms recorded at 200 and 254 nm for cytarabine, azacitidine, and their analogues.
`Peaks: A = 0.45 µg V; B = 0.20 µg cyclocytidine; C = 0.90 µg IV; D = 2.6 µg azacitidine; E = 2.0 Jig
`cytarabine; F = 0.20 µg uracil arabinoside. See Figs. 1 and 2 for full names and structures.
`
`

`

`HPLC OF CYT A RABINE AND AZACITIDINE
`
`313
`
`and their analogues in Fig. 3. A pro-drug, cyclocytidine, which is hydrolyzed to form
`Ara-C 8 , is separated from Ara-C and uracil arabinoside. Cleavage of the glycosidic
`bond in Ara-C and 5-AC would result in the formation of o~arabinose and o-ribose,
`respectively. These sugars elute at the column void volume. The bases which result
`from cleavage of the glycosidic bond elute earlier than the corresponding nucleoside
`analogues, with retention times of 4.1 and 3.6 min for cytosine and 5-azacytosine,
`respectively. None of the analogues or products of decomposition interfered with the
`peaks for Ara-C or 5-AC.
`A chromatogram of the analysis of a Cytosar-U sterile powder sample is shown
`in Fig. 4. Although Ara-Chas an absorbance maximum at wavelengths higher than
`254 nm, the common HPLC detection wavelength of 254 nm was employed. The
`additional sensitivity which could be obtained at the wavelength of maximum ab(cid:173)
`sorbance was not necessary for pharmaceutical samples. Extrapolation from elevated
`temperatures yields a decomposition rate constant at 25°C of 2.2 - 10- 5 h - i (ref. 9).
`Thus, only 0.05% of the Ara-C will be degraded in the sample preparation in one
`day. A small amount of uracil arabinoside is expected to form in the product during
`the freeze-drying process. A small peak for uracil arabinoside is present in the chro(cid:173)
`matogram of Cytosar-U sterile powder that corresponds to 0.05% of the Ara-C
`content.
`Chromatograms at 230 and 254 nm are shown in Fig. 5 for a sample prepared
`from Mylosar sterile powder. No detector response is observed for mannitol, which
`is present at an amount equal to the 5-AC in Mylosar, because it does not absorb
`UV radiation at these wavelengths. Small amounts of the primary decomposition
`
`A
`
`B
`
`254nm
`
`i
`50mAU l
`
`D 254nm
`
`C
`
`50mAU
`
`l
`l
`
`B
`
`~ l )~
`
`0
`
`10
`
`0
`
`10
`
`Time (min)
`Time (min)
`Fig. 4. Chromatogram recorded at 254 nm of a 100-mg Cytosar-U sterile powder sample. Peaks: A =
`cytarabine; B = p-toluic acid.
`Fig. 5. Chromatograms recorded at 230 and 254 nm of a 100-mg Mylosar sterile powder sample. Peaks:
`A = V; B = IV; C = azacitidine; D = p-toluic acid.
`
`

`

`314
`
`L. D. KISSINGER, N . L. STEMM
`
`product (IV, Fig. 2) and the secondary decomposition product (V, Fig. 2) can be
`detected in the Mylosar sample. The decomposition products of 5-AC are formed
`during the freeze-drying process and in the sample preparation, prior to injection.
`The peak for the primary decomposition product corresponds to 2.1 % of the 5-AC
`content, of which 1.7% is present in the freeze-dried cake and 0.4% formed in the
`sample preparation. To minimize decomposition of 5-AC in the sample and standard
`preparations, methanol- water (80:20) is used as a solvent for dilution of the prepa(cid:173)
`rations. Hydrolysis rates in the aqueous alcohol are much slower than in purely
`aqueous solution. The preparations are injected 15 min after addition of the alcoholic
`internal standard solution to reproduce the extent of decomposition. Use of the lower
`wavelength (230 nm) was necessary to obtain a response for the secondary decom(cid:173)
`position product.
`
`UV spectra
`UV spectra obtained with the diode-array detector for Ara-C, 5~AC, and their
`analogues are shown in Fig. 6. The spectra were obtained as the analytes eluted from
`the chromatograph. In the neutral pH of the mobile phase,,Ara-C absorbs light from
`200 through 300 nm with a small maximum at about 275 nm. Cytosine, the base
`portion of Ara-C, has an absorbance spectra that closely parallels Ara-C. The sugar
`
`90%
`(mAU)
`
`(a )
`
`0
`
`- 10L.....L---L......L....L--'--L.....1.---L......L..,-J---'--L.....1.---L......L....L-......... L.....L~,....._ ......... .L-'
`200
`250
`300
`
`90%
`(mAU)
`
`Wavelength (nm)
`
`(b)
`
`0
`-10'--'___.___._.......__.__..._..___.___._.......__.__ .......... ___.__._...__.__ ........... _,,....._._...__
`200
`250
`300
`
`Wavelength (nm}
`Fig. 6. Spectra of (a) cytarabine and analogues, and (b) azacitidine and analogues. (a)--, D-arabinose
`(Amax = 200 nm, max. abs. = 30 mAU); ---, cytosine (A.max = 200 nm, max. abs. = 137 mAU); ---;
`cytarabine (Amax = 200 nm, max. abs. = 314 mAU); - • - •-, uracil arabinoside (Amax = 262 nm, max.
`- , V (,l,mu = 200 run, max. abs. = 115 mAU); ----, 5-azacytosine (Amax = 200
`abs. = 12 mAU). (b) -
`nm, max. abs. = 321 mAU); ----, IV (Amax = 242 nm, max. abs. = 216 mAU); - . - . - . -, azacitidine
`0-max = 202 nm, max. abs. = 506 mAU).
`
`

`

`HPLC OF CYTARABINE AND AZACITIDINE
`
`315
`
`portion of the molecule, o-( - )-arabinose, only shows a weak end absorbance above
`200 nm. The primary decomposition product of Ara-C, uracil arabinoside, has a
`broad peak with a maximum at 270 nm. Except for the sugar, the common 254 nm
`HPLC detectors provide good sensitivity for the analysis of these components.
`The UV spectra of 5-AC and 5-azacytosine have broad bands, which are sim(cid:173)
`ilar to those of Ara-C and cytosine. The primary decomposition product of 5-AC
`has a maximum at ca. 254 nm. Maximal absorptivity for the second decomposition
`product of 5-AC (V) is at lower wavelengths, and sensitivity for it can be increased
`by changing from the normal analytical wavelength of 254 nm to 230 run. Ribose,
`the sugar portion of the 5-AC molecule, and mannitol, the sugar which is present in
`an amount equal to the 5-AC in the Mylosar formulation, are analogous to o-( - )(cid:173)
`arabinose in that they contain no significant ultraviolet chromophore.
`
`Linearity and precision
`Linearity of response was investigated by injecting nine samples of Ara-C,
`ranging from 0.3 to 1.5 µg. These amounts correspond to 30-150% of the amount
`specified in the assay procedure. Linear regression analysis of the aµiount determined
`by peak height and peak area versus the amount added resulted in slopes of 1.006
`and 1.012, respectively, and intercepts which were less than the assay variation. Cor(cid:173)
`relation coefficients (r) for these lines were greater than 0.9999. An Ara-C sample
`was assayed a total of eight times on six different days with a relative standard
`deviation (R.S.D.) of 1.1 and 0.9% by peak height and peak area, respectively.
`Nineteen spiked sample preparations of Ara-C were prepared with 0-15% of
`its decomposition product, uracil arabinoside. Regression analysis of the uracil ar(cid:173)
`abinoside amount determined by peak height versus the amount added resulted in a
`slope of 0.975 with r > 0.99. The amounts of uracil arabinoside determined in the
`presence of large amounts of Ara-Care slightly below theory (2.5%), because of the
`limited resolution from the major component. Peak-area quantitation tended to be
`less accurate than peak-height quantitation of uracil arabinoside on the tailing side
`of the Ara-C peak. A total of eight assays on six different days were performed on
`a sample that contained 0.5% uracil arabinoside with a resulting R.S.D. of 3.7% by
`peak height.
`Linearity for 5-AC was established by injecting seven samples ranging from
`0.8 to 2.8 µg. These amounts correspond to 40-140% of amount specified in the
`assay procedure. Peak-height and peak-area responses were linear over this range
`with r > 0.999 and intercepts which were less than the assay variation. Precision of
`the procedure has been established by a relative standard deviation of 1.0% for
`twenty determinations performed on a solid sample over a period of a year.
`
`Azacitidine decomposition kinetics
`The kinetics of the two-step decomposition route of 5-AC can be fit to the
`following equation: [III] = Ae - a1 + Be-Pt (refs. 7 and 9). Iterative best-fit calcula(cid:173)
`tions were performed on [III] versus time (t) profiles to determine the coefficients (A,
`B) and exponential terms (a, (3). From the terms of the biexponential equation, the
`respective first-order kinetic rate constants were calculated10.
`The values for the three first-order rate constants associated with the decom(cid:173)
`position of 5-AC are listed in Tables I and II. Kinetic rate constants were not sig-
`
`

`

`316
`
`L. D. KISSINGER, N. L. STEMM
`
`TABLE I
`FIRST-ORDER DECOMPOSITION RATE CONSTANTS FOR 0.1 mg/ml AZACITIDINE SOLU(cid:173)
`TIONS
`
`Solvent
`
`Temperature
`(° C)
`
`Water
`Water
`Phosphate (0.0067 M)
`Phosphate (0.067 M)
`Phosphate (0.2 M)
`
`25
`30
`30
`30
`30
`
`k11
`(min- 1 )
`
`0.000692
`0.00120
`0.00135
`0.00214
`0.00562
`
`k12
`(min- 1 )
`
`0.00180
`0.00312
`0.00333
`0.00558
`0.0139
`
`k13
`(min- 1 )
`
`0.000336
`0.000542
`0.000842
`0.00185
`0.00366
`
`nificantly different for solutions made from the b"ulk drug and Mylosar formulation,
`indicating that the small amount of mannitol does not affect the solution stability of
`5-AC.
`An example of the profiles of 5-AC and its two decomposition products as a
`function of time in a lactated Ringer's injection solution USP is shown in Fig. 7.
`These profiles were calculated from the values of the biexponential equation which
`was fit to the experimental data for 5-AC. Within the 1 % R.S.D. of the HPLC assay,
`values for 5-AC which were calculated in an iterative fashion from the first-order
`kinetic rate constants equaled the experimental values.
`The thermal dependence of the three first-order rate constants fit the Arrhenius
`
`TABLE II
`FIRST-ORDER DECOMPOSITION RATE CONSTANTS FOR AZACITIDINE (0.1 mg/ml) IN VARIOUS
`THERAPEUTIC SOLUTIONS OF MYLOSAR
`Solvents: water = purified water USP; BWFI = bacteriostatic water for injection; LRI = lactated Ringer's injection
`USP; dextrose = 5% dextrose injection USP; saline = 0.9% sodium chloride injection USP. Suppliers of solvents:
`A = laboratory purified water USP; B = The Upjohn Company, C = Travenol Labs., D = Abbott Labs.
`
`Solvent
`
`Solvent
`supplier
`
`pH
`
`Temperature Container
`r·cJ
`type
`
`k11
`(min - 1 )
`
`k12
`(min- 1 )
`
`k13
`( min- 1)
`
`IV
`set
`
`Water
`A
`Water
`A
`BWFI
`B
`LRI
`C
`LRI
`C
`LRI
`C
`LRI
`C
`LRI
`C
`LRI
`C
`LRI
`D
`LRI
`C
`Dextrose D
`Dextrose C
`Saline
`C
`Saline
`C
`Saline
`D
`
`7.0
`7.0
`5.7
`7.0
`6.0
`6.5
`7.0
`7.0
`7.0
`7.0
`7.5
`4.0
`6.0
`4.5
`4.5
`5.8
`
`30
`40
`25
`4
`25
`25
`25
`25
`25
`25
`25
`25
`25
`25
`25
`25
`
`Glass
`Glass
`Glass
`Glass
`Glass
`Glass
`Glass
`Plastic
`Plastic
`Plastic
`Glass
`Plastic
`Glass
`Glass
`Plastic
`Plastic
`
`0.00121
`0.00343
`0.000526
`0.0000619
`0.00149
`0.00100
`0.000878
`0.000769
`0.000897
`0.000730
`0.000991
`0.00304
`0.000608
`0.00296
`0.00296
`0.00172
`
`0.00334
`0.00938
`0.00170
`0. 000 13 6
`0.00413
`0.00258
`0.002501
`· 0.00178
`0.00224
`0.00181
`0.00268
`0.00829
`0.00115
`0.0106
`0.00810
`0.00451
`
`0.000498
`0.00157
`0.000385
`0.0000158
`0.000355
`0.000341
`0.000462
`0.000348
`0.000370
`0.000354
`0.000428
`0.00131
`0.00147
`0.00168
`0.000436
`0.00332
`
`Yes
`
`Yes
`
`Yes
`
`Yes
`
`

`

`HPLC OF CYTARABINE AND AZACITIDINE
`
`317
`
`120
`
`100
`
`80
`
`60
`
`40
`
`'-.... ------
`"
`
`20
`
`--------------------------------------------------------------
`
`0 o
`
`1000
`
`2000
`
`Minutes
`
`3000
`
`4000
`
`Fig. 7. Calculated profiles for azacitidine and its decomposition products in a solution of Lactated Ringer's
`injection USP, pH 6.5, 25°C.
`
`equation for neutral solutions (pH 7) over the temperature range of 4-40°C. The
`calculated activation energies for k 11 , k 12 , and k 13 were 8.33, 8.76, and 9.58 kcal/mol,
`respectively. For pharmaceutical preparations of 5-AC, the stability is greatly en(cid:173)
`hanced by reducing the temperature. After IO h, ca. 20% decomposition occurs at
`25°C, while ca. 4% decomposition occurs at 4°C.
`The rate constants in Table I show the decrease in stability of 5-AC with
`increasing ionic strength of pH 7 phosphate buffered solutions. The negative effect
`of ionic strength on the stability of 5-AC has been observed by other researchers4 .
`Infusion solutions of low ionic strength provide optimal stability for administration
`of Mylosar®. Of the pharmaceutical solutions, degradation of 5-AC is slowest in the
`bacteriostatic water for injection USP solution. Electrolytes in each of the L VP sol(cid:173)
`vents have a destabilizing effect on the 5-AC. Because of the lower pH of the saline
`a· i dextrose solutions, the decomposition of the drug is much more rapid than the
`nearly neutral lactated Ringer's injection USP solution. Even within the specified
`range of 6.0- 7.5 for the pH of lactated Ringer's injection USP, differences in the
`decomposition rates are detectable. The maxiµmm for stability in this infusion so(cid:173)
`lution is obtained at pH 7.0. Use of an intravenous (i.v.)-solution-administration set
`with particulate filter did not affect the decomposition kinetics.
`Without isolating the decomposition products of 5-AC, IV and V, their re(cid:173)
`sponse factors can be calculated from the observed responses and the kinetic factors.
`Responses for the decomposition products in solution are a result of the initial level
`of the component and the amount formed as a function of time. After subtracting
`the initial response, the response factor can be calculated from the amount predicted
`from the kinetic factors. Reproducible values for the response factors were observed
`for the time window from 700 to 4200 min. Response factors calculated from early
`time points were not reproducible because only a small amount of the decomposition
`products had formed. For data from three different solutions using the 254-nm mer(cid:173)
`cury line source detector, an average response factor for IV versus 5-AC on a mass
`basis was 2.14, with an R.S.D. of 3%. For data from three different solutions using
`a variable-wavelength detector set at 254 nm, the factor was 2.20, with an R.S.D. of
`
`

`

`318
`
`L. D. KISSINGER, N. L. STEMM
`
`15%. Better precision for the relative response factor is expected with the line source
`detector, because of the reproducibility of the source wavelength. With the photo(cid:173)
`diode-array detector, a relative response factor for V versus 5-AC on a mass basis at
`230 nm was found to be 0.56, with an R.S.D. of 7%.
`
`REFERENCES
`
`1 R. E. Notari, J. Pharm. Sci., 56 (1967) 804-809.
`2 R. E. Notari, M. L. Chin and A. Cardoni, J. Pharm. Sci., 59 (1970) 28-32.
`3 Y. Cheung, B. R. Vishnuvajjala and K. P. Flora, Am. J. Hos. Pharm., 41 (1984) 1802-1806.
`4 R. E. Notari and J. L. DeYoung, J. Pharm. Sci., 64 (1975) 1148-1157.
`5 J. A. Beisler, J. Med. Chem., 21 (1978) 204-208.
`6 Y. Cheung, B. R. Vishnuvajjala, N. L. Morris and K. L. Flora, Am. J. Hos. Pharm., 41 (1984)
`1156-1159.
`7 K. K. Chan, D. D. Giannini, J. A. Staroscik and W. Sadee, J. Pharm. Sci., 68 (1979) 807-812.
`8 L. E. Kirsch and R. E. Notari, J. Pharm. Sci. , 73 (1984) 896--902.
`9 R. E. Notari, M . L. 'Chin and R . Wittebort, J. Pharm. Sci., 61 (1972) 1189-1196.
`IO M. Gibaldi and D. Perrier, Pharmacokinetics, Marcel Dekker, New York, 1975, p. 88.
`
`