(5) K. B. Bruce and W. R. Maynard, Jr.. Anal. Chem.. 41, 977
`(1969).
`( 6 ) L. J. Dombrowski, P. M. Comi. and E. L. Pratt, J . Pharm. Sci.. 62,
`I761 (1973).
`(7) G . K. Wilkinson, Anal. Lett., 3, 289 (1970).
`(8) T. Walk and I{. Ehrsson. Acta Pharm. Suec., 8.27 (1971).
`(9) D. J. Edwards and K. Blau, Anal. Biochem., 45,387 ( I 972).
`(10) J. C. Lhuguenot and B. F. Maume, J . Chromarogr. Sci., 12, 41 I
`(1974).
`( I I ) 1.. Neelakantan and H. B. Kostenbaudcr, J . Pharm. Sci., 65, 740
`( 1976).
`
`( 1 2 ) t i . Kinsun, M. A . Moulin, and E C. Savini. J Pharm. Sci.. 67, 1 I8
`(1978).
`(13) W. I). Mason and E. N . Amick,J. Pharm. Sci.. 70,707 (1981).
`(14) B. 34. Farrell and T. M. Jefferies, J. Chromarogr., 272, I 1 I
`(1983).
`(I S) F. T. Noggle, Jr.. J . .4ssoc. Of$ .4nal. Chem , 6 5 7 0 2 (1980).
`
`ACKNOWLEDGMENTS
`
`We thank Dr. Pieter Bonscn for critically reviewing this manuscript and
`Ms. Rose Wright for her extensive literature research.
`
`Determination of Sodium Levothyroxine in Bulk, Tablet, and
`Injection Formulations by High-Performance Liquid
`Chromatography
`
`JAMES F. BROWERX, DUCKHEE Y. TOLER, and JOHN C. REEPMEYER
`Received July 15, 1983, from the National Center for Drug Analysis. Food and Drug Administration. Sr. Louis, MO 63/01.
`September 28. 1983.
`
`Accepted for publication
`
`Abstract 0 Sodium levothyroxine was determined in bulk drugs, tablets, and
`injections by high-performance liquid chromatography (HPLC). Levothy-
`roxine was separated from excipients and impurities on a 10-pm cyanoalkyl
`column using an acetonitrile-water-phosphoric acid mobile pha.e. The HPLC
`method is shown to be linear, accurate, and precise, and the results obtained
`by the HPLC and USP XX methods are compared.
`Keyphrases 0 Sodium levothyroxine-IiPLC,
`determination of bulk, tablet.
`and injection formulations 0 HPLC- sodium levothyroxine. determination
`of bulk, tablet, and injection formulations
`
`In a survey] of sodium levothyroxine products and formu-
`lations, 63 samples of tablets, representing 20 formulations
`from 5 manufacturers, 9 samples of injections from 2 manu-
`facturers, and 6 samples of bulk sodium levothyroxine from
`5 manufacturers, were analyzed in this labroatory. The pur-
`pose of the survey was to evaluate the quality of sodium levo-
`thyroxine products on the market and the adequacy of present
`compendial standards and methods. Methodology was de-
`veloped for content uniformity analysis of sodium levothy-
`roxine by high-performance liquid chromatography (HPLC).
`This method, unlike the official compendial method (I), dif-
`ferentiates levothyroxine from iodinated impurities and deg-
`radation products.
`Previously developed methods for the determination of so-
`dium levothyroxine (2-1 3) were evaluated and tested. A
`modification of the HPLC procedure described by Garnick
`et al. (1 3), using a cyanoalkyl bonded phase column, was se-
`lected. This method offers advantages over those already in
`the literature by avoiding buffers in the mobile phase, no
`sample derivatization, faster analysis times, greater sensitivity
`due to shorter retention times, lower flow rates, and 229 nm
`detection. A sample solvent was selected that readily dissolved
`sodium levothyroxine without degradation from tablet for-
`mulations. This method and the results obtained on the survey
`sample are reported here.
`
`I This study was a national survey for the Food and Drug Administration.
`
`~~
`
`~
`
`EXPERIMENTAL SECTION
`Apparatus A modular high-performance liquid chromatograph2 (HP1.C)
`was equipped with a fixed-wavelength (229 nm) cadmium lamp UV detector’.
`an automated injector4, a microprocessor controller5, and a recorder-inte-
`grator6. A stainless steel column (3.9 mm X 30 cm) was packed with irregular
`10-pm silica particles to which a layer of cyanoalkyl silane was chemically
`bonded’.
`The mobile phase consisted of acetonitrile-water
`650:l). This solution was passed through a 0.45-pm
`then pumped through the HPLC system at a rate of
`
`)hosphoric acid (350:
`ilter*, deaerated, and
`mL/min.
`
`0
`
`m
`MINUTE8
`Figure 1 -Chromatogram of sodium iecorhyroxine bulk drug decomposed
`by heating in air; derecror 01 229 nm and 0.02 AUKS“. Key: ( 1 ) sodium Iei:o-
`thyroxine at a iecel of -I00 pg/niL.
`
`Model A1.C 204; Waters Associates.
`Model 44 I ; Watcrs Associates.
`Model 710B W I S P Waters Associates.
`Model 720 System Controllcr; Waters Associates.
`Model 730 Data Module; Waters Associates.
`F-BondapakKN; Waters Associates.
`Durapore UVLPO4700; Millipore Corp, Bcdford. Mass.
`
`0022-3549J84J 0900- 13 15$0 1.OOJO
`@ 1984, American Pharmaceutical Association
`
`Journal of Pharmaceutical Sciences J 13 15
`Vol. 73. No. 9, September 1984
`
`Mylan Ex 1018, Page 1
`
`

`
`Table I-Sodium Levothvroxine Determined bv HPLC and LSP XX Procedures
`1.abcl Claim, %
`Composite Assay
`
`Manu-
`facturcrO
`
`Dosage
`Form, mg
`
`C.U.h
`
`IIPLC
`
`LISP
`
`Manu-
`facturer"
`
`Dosage
`Form, mg
`
`Label Claim, %
`Composite Assay
`
`C.L'.h
`
`HPLC
`
`USP
`
`Tablet, 0.025
`Tablct. 0.025
`Tablet, 0.025
`Tablet. 0.025
`Tablet. 0.025
`Tablet, 0.050
`Tablet, 0.025
`Tablet. 0.050
`Tablet. 0.050
`Tablet. 0.100
`Tablet, 0.050
`Tablet. 0.100
`Tablet. 0. I00
`Tablet, 0.050
`Ta blet. 0.100
`Tablet. 0.050
`Tablet. 0.1 25
`Tablet, 0.100
`Tablet. 0.100
`Tablet, 0.125
`Tablet. 0.100
`Tablct,O.l5O
`Tablet. 0.100
`Tablet, 0.:50
`Ta blct, 0.100
`Tablet, 0.150
`Tablet. 0.100
`Tablct.O.175
`Tablet. 0.150
`Tablet, 0.200
`Tablet. 0.150
`Tablet. 0.200
`Tablet, 0.150
`Tablet: 0.200
`Tablet, 0.150
`Tablet, 0.200
`Tablet, 0.150
`Tablet, 0.300
`Tablet, 0. I50
`Tablet, 0.300
`Tablet. 0.200
`Bulk drug
`Tablet. 0.050
`Tablet, 0.200
`Tablet, 0.200
`Tablet; 0.100
`Tablet, 0.200
`Tablet, 0.175
`Tablet, 0.200
`Tablet. 0.300
`Tablet, 0.200
`Tablet. 0. I00
`Tablet, 0.200
`Tablet. 0.100
`Tablet, 0.300
`Tablet, 0.200
`lnj., 0.100
`Bulk drug
`Inj., 0.100
`Inj., 0.100
`Tablet. 0.100
`Inj., 0.200
`Tablet, 0.100
`lnj., 0.200
`113.6
`97.4
`Tablet, 0.200
`Inj., 0.200
`Tablet, 0.200
`I 11.7
`95.6
`lnj., 0.500
`Bulk drug
`96.4
`100.7
`Inj., 0.500
`Bulk drug
`Tablet, 0. I00
`IOI.8
`62.3
`Bulk drug
`Tablet, 0.300
`104.3
`101.7
`F
`Bulk drug
`97. I
`98. I
`83.8 (9.4) 3
`Inj., 0.500
`( A ) A m o u r Pharmaceutical Co.. Scottsdale. Arit.: (B) Chelsea Laboratories Inc., Inwood. N.Y.; ( C ) Generic Pharmacculical. Palisndes Park. N.J.: (D) Western Research
`Laboralorieh. Denver. Cola; (I!) Travenwl Laboratories. Deerfield. 111.: (I:) Carter-(ilogau I.aboralories. Glendale. A r m * Content uniformity by HPLC. mean; KSD in parenlheses;
`number of units tested.
`Reagents -Methanol9, acctonitrileY, phosphoric acid"', and sodium hy-
`droxidc" were used as receivcd. Deionized water was supplied by a com-
`mcrcialI2 water system. Samplcs and standards were prepared in 0.01 ;M so-
`dium hydroxide in 50% methanol.
`I0 mg of USP lcvothyroxine reference
`Standard Preparation-About
`standard was accurately weighed into a 50-ml. volumetric flask, dissolved.
`and diluted to volume with 0.01 M mcthanolic sodium hydroxide. (This stock
`solution is stable for several weeks when refrigerated.) Two milliliters of this
`bolution was pipetted into a 100-mL volumetric flask and diluted to volume
`with 0.01 M methanolic sodium hqdroxide.
`Sample Preparation- Tablets -One tablet was placed in a glass-stoppered
`flask and accurately diluted with 0.01 M mcthanolic sodium hydroxide to -4
`pg/niL. The flask was placed in an ultrasonic bathI3 until the tablet disinte-
`grated; the tablet was then mechanically shaken14 for 30 min. The solution
`was then filteredI5 into a vial for injection.
`Powdered Composites and Lyophilized Injections-An accurately weighed
`portion of the powdered sample, equivalent to -400 p g of sodium levothy-
`roxine, was placed in a 100-mL volumctric flask. Fifty milliliters of 0.01 M
`methanolic sodium hydroxide was added, and the flask was placed in an ul-
`trasonic bath for 1 min. The flask was shaken for 5 min; the contents were
`diluted to volume with 0.01 M methanolic sodium hydroxide and mixed well.
`The solution was then filteredI5 into a vial for injection.
`sample of bulk drug substance was prepared
`Bulk Drug Substances-A
`in the same manner as the standard.
`
`L.:
`
`99.7 (3.7) 10
`99. I (2.9) 10
`97.8 (2.7) 10
`97.8 (2.1 j 10
`02.0 (2.1) 10
`01.5 0 . 7 ) 10
`06.1 (i.4j i0
`02.6 (2.3) 10
`94.5 (1.4) 10
`04.3 (2.3) 10
`05.1 (2.7j 10
`04.3 ( 1.9) 10
`04.8 (2.0) 10
`05.5 (2.0) 10
`02.1 (1.5) 10
`02.9 ( I .2) 10
`05.6 (1.2) 10
`03.7 ( I .4) 10
`05.2 (1.6) 10
`04.4(1.8) 10
`88.7 (1.5) 10
`06.3 (3.0) 10
`04.4 (2.7) 10
`05.8 (2.3j 10
`03.6 (3.3) I0
`05.0 (2.7) 10
`05.1 (].xi 1 0
`102.0(1.8) 10
`84.1(1.5) 5
`94.2 (2.9) 5
`89.0 (4.8) 5
`94.4 (4.7) 5
`95.5 (2.8) 5
`97.5 (1.3) 5
`95.5 iosj 5
`93.4 (3.6) 5
`
`108.2
`106.6
`
`86.5
`
`94.4
`
`97.8
`99.2
`
`95.4
`98.0
`
`Determination--- Fifty microliters each of the standard and sample solutions
`wcrc injected into the liquid chromatograph, and the chromatograms were
`rccordcd. Sodium lcvothyroxine was calculated on the basis of peak areas.
`
`RESULTS AND DISCUSSION
`
`series of validation tests were
`Linearity, Reproducibility, and Recovery-A
`performed on thc HPLC method. A linear response wasobtaincd when four
`standard solutions containing from 1 to 8 & n L were tested. Placebo samples,
`based on the batch formulation of the manufacturers, were spiked with various
`levels of the standard. The recovcries ranged from 99.5 to 100.6%. The re-
`producibility of the method was determined by consecutively injccting 10
`aliquots of standard solution. The KSD was 0.2%.
`Decomposition and Stability Studies-In
`the course of the survey, it was
`found that the bulk drug substances were sensitive to the conditions of the USP
`drying procedure (60°C in a vacuum over P2O5). I f the vacuum was not
`maintained below I0 mm tig, sodium levothyroxine decomposed rapidly
`(10- 15% in 4 h). Undried bulk drug samples were uxd in the analyses because
`of this problem, and corrections were made for moisture content. Figure 1
`shows a chromatogram of sodium levothyroxine bulk drug dried in the pres-
`ence of air. Some decomposition products can be seen in the chromatogram;
`these were not observed in the chromatograms of an unheated sample of so-
`dium levothyroxine measured at the same sensitivity. In addition, there are
`probably other decomposition products in the heated sample which are not
`eluted from the column. The identity of these decomposition products will be
`investigated at a later time.
`To test the stability of sodium levothyroxine in the sample solvent (0.01 M
`methanolic sodium hydroxide), solutions of the bulk drug samples were stored
`at 5OC and assayed periodically (cvrsus freshly prepared reference standards)
`over a six month period. The assay values remained constant over the entire
`tcsting period, indicating little or no decomposition.
`
`A
`
`B
`
`C
`
`D
`
`96.7 ( I .2) 30
`97.6 ( I .6) 30
`90.8 ( I .8) 30
`94.3 (2.1) 30
`96.7 ( 1.8) 30
`95.8 ( I .3) 30
`07.2 ( I .9) 30
`97.5 (2.5) 30
`97.6 ( I .7) 30
`98.4 (0.8j 30
`97.8 (2.0) 30
`104.4 (2.0) 10
`95.0 (1.8) 10
`95.0 ( I .8) 30
`97.4 (1.3) 10
`100.0 (2.5) 30
`I 00.6 i I .9j 30
`98.5 (1.7) I0
`102.0(1.4) 10
`102.5 (2.6) 10
`
`80.7 ( I .2) 10
`99.1 (1.7) 10
`88.3 (2.2) 10
`90.9 (1.9) I0
`92.4 (2.6) 10
`89.S (6.7) 10
`99.1 (5.2) 10
`
`99.4 (3.2) 20
`99.6 (2.1) 20
`94.7 ( I .8) 10
`90.2 (2.5) 20
`
`63.9 (2.5) 20
`100.8 (1.5) 10
`
`104.9
`96.5
`109.2
`97.8
`100.8
`106.3
`98.4
`100.2
`Past expiration date
`Past expiration date
`Past expiration date
`Past expiration date
`94.3
`97.4
`95.7
`96.6
`102.5
`105.2
`100.9
`96.8
`
`Mathcson, Coleman and Bell. Cincinnati, Ohio.
`l o NF Grade: Mallinckrodt Inc.. St. Louis. M o .
`II A R Grade; Mallinckrodt Inc.
`l 2 Milli-Q; Millipore Corp.
`l 3 Model SC400T; Randall M f g . Co.. Inc.. Hillside. h.J.
`l 4 Oscillating shaker: Eberbach Corp.. Ann Arbor. Mich.
`I s Nylon-66. 13-mm diameter. 0.45-pm pore size; Rainin Instrument Co.. Woburn.
`Mass
`
`1316 I Journal of Pharmaceutical Sciences
`Vol. 73, No. 9, September 1984
`
`Mylan Ex 1018, Page 2
`
`

`
`Low assay values were a problem experienced by most manufacturers; this
`problem would not be recognized if the assays were based only on the results
`from the USP method. Some tablet composite samples gave suitable assay
`values for total iodine by the USP method, but gave extremely low assays for
`sodium levothyroxine by HPLC. All samples which gave low assay values by
`HPLC gave suitable assays by the USP method. This fact indicates that the
`problems of low assays of marketed sodium levothyroxine are, in all proba-
`bility, attributable to sodium lcvothyroxine instability.
`
`REFERENCES
`( 1 ) “The United States Pharmacopeia XX,” U.S. Pharmacopeial Con-
`vention, Rockville. Md., 1980. pp. 446-447.
`(2) B. L. Karger and S. C. Su, J. Chromatogr. Sci., 12,678 (1974).
`(3) F. Nachtmann. G. Knapp. and H. Spitzy. J . Chromaiogr., 149,693
`( 1978).
`(4) M. T. W. Hearn, W. S. Hancock, and C. A. Bishop, J . Chromatogr.,
`157,337 (1978).
`(5) N. M. Alexander and M. Nishimoto, Clin. Chem., 25, 1757
`( 1979).
`(6) R. S. Rapaka, P. W. Knight. V. P. Shah, and V. K. Prasad. Anal.
`Lett., 12, I201 (1979).
`(7) E. P. Lankmayr and K. W. Budna, J. Chromarogr.. 198, 471
`(1980).
`(8) D. J. Smith and J . H. Graham, J . Assoc. 0). Anal. Chem., 62,818
`( 1 980).
`(9) B. R. Hepler, S. G. Weber, and W. C. Purdy, Anal. Chim. Acta. 113,
`269 (1 980).
`(10) D. J. Smith, M. Biesmeycr, and C. Yaciw,J. C‘hromafogr. Sci., 19,
`72(1981).
`( I I ) E. P. Lankmayr, B. Maichin, and G . Knapp, J . Chromarogr., 224,
`239(1981).
`(12) I. D. Hay, T. M. Annesley, N. S. Jiang, and C. A. Gorman, J .
`Chromatogr., 226,383 ( I 98 I ).
`( 1 3) R. L. Garnick, G. F. Burt, F. R. Borger, J. P. Aldrcd, J . W. Bastian,
`and D. A. Long, in “Hormone Drugs, Proceedings of the FDA-USP Work-
`shop on Drug and Reference Standards for Insulins, Somatropins, and Thy-
`roid-Axis Hormones,” U.S. Pharmacopeial Convention, Rockville, Md., 1983.
`pp. 504-5 16.
`
`ACKNOWLEMMENTS
`The authors thank Dr. Thomas P. Layloff and Donald P. Page for assistance
`during the survey.
`
`I
`I2
`
`0
`
`6
`MINUTES
`Figure 2-Chromatogram ojsodium levothyroxine and sodium liothyronine.
`detector at 229 nm and 0.02 AUFS. Key: ( I ) sodium liothyronine: (2) sodium
`Ievoihproxine, each at a level of-4 pg/mL.
`The chromatographic system parameters were adequate to separate sodium
`levothyroxine from sodium liothyronine for testing the bulk drug substances
`for impurities; Fig. 2 shows a chromatogram of this separation. Liathyronine
`was found at levels of 0.04-0.96% in five sodium levothyroxine bulk drug
`samples analyzed by HPLC.
`1 lists the results obtained from survey by the
`Sample Analysis-Table
`HPLC procedure. A comparison of results for composite samples, obtained
`with the HPLC and the USP X X methods. shows that the latter gave a higher
`result in practically every case. This is not surprising since the USP X X assay
`is nonspecific for levothyroxine and measures total iodine content. However,
`the difference in assay values could not be totally accounted for by a total area
`summation of HPLC peaks. The major impurities and degradation products
`probably are not eluted from the column with this mobile phase.
`
`Determination of Isosorbide 5-Mononitrate in Human
`Plasma by Capillary Column Gas Chromatography
`
`P. STRAEHL and R. L. GALEAZZIx
`Received February 22. 1983, from the L‘nioersity of Bern, Depariment of Medicine. Inselspiial, ClI-3010 Bern, Swiizerland.
`publication August 25. 1983.
`
`Accepted for
`
`Abstract 0 An electron-capture gas chromatographic method for the deter-
`mination of isosorbidc 5-mononitrate in human plasma using a capillary
`column is described. lsosorbide 5-mononitrate and the internal standard
`(isosorbide dinitrate) are extracted from the alkalinized plasma with ether.
`The lower limit of detection for isosorbide 5-mononitrate is 1 ng/mL of
`plasma.
`5-mononitrate, human
`Keyphrases 0 Gas chromatography-isosorbide
`plasma 0 lsosorbidc 5-mononitrate-GC, human plasma
`
`[ 1,4:3,6-dianhydro-~-glucitol
`lsosorbide 5-mononitrate
`5-nitrate (I)] is the primary metabolite of isosorbide dinitrate
`[ I ,4:3,6-dianhydro-~-glucitol dinitrate (Il)] which has been
`
`used for many years in the treatment of angina pectoris and
`congestive heart failure. Recent studies of the hernodynamic
`effect of the mononitrate ( 1 ) indicate that, after acute ad-
`ministration, cardiac work load decreases at rest and during
`exercise. Pharmacokinetic studies (2-4) showed that the
`mononitrate is rapidly and completely absorbed from the G1
`tract without undergoing first-pass elimination. The maximum
`concentrations were reached within 1 h after oral adminis-
`tration, and the substance was eliminated with a half-life of
`-4 h. Thus, the mononitrate has a half-life which is at least
`four times as long as the half-life of the dinitrate (5). Published
`gas chromatography (GC) assays (2,6, 7), using older column
`
`0022-35491 8 4 0900- 13 17$0 1.001 0
`@ 1984, American Pharmaceutical Association
`
`Journal of Pharmaceutical Sciences I 13 17
`Vol. 73, No. 9, September 1984
`
`Mylan Ex 1018, Page 3