Assay of para-Aminobenzoic Acid Formed by Hydrolysis of Procaine in CP 1 B Solution 2000, 52, 115-119 Ph. Galais ~ / C. Dauphin 2 / D. Pradeau 2./A. Chevallie? 1 H6pital Broussais, Service de Pharmacie, 96, Rue Didot, 75014 Paris, France 2 Pharmacie Centrale des H6pitaux (PCH) Assistance Publique-H6pitaux de Paris, Laboratoire d'analyses physico-chimiques, 7, Rue du Fer,5 Moulin, 75005 Paris, France KeyWords Thin-layer chromatography Cardioplegia solution Procaine p-Aminobenzoic acid Summary CP1 is a cardioplegia solution containing procaine, a substance that is unstable at pH > 7. For this reason the product it is supplied as two solutions that are mixed immediately before use. para-Aminobenzoic acid (PABA) was separated by horizontal high-performance thin-layer chromatography (HPTLC) on silica gel 60 nanoplates with fluorescence indicator. This method is specific, repeatable, and reproducible. The autoclave sterilization conditions currently used are the least aggressive towards pro- caine. Under normal storage conditions, the concentration of procaine in CP1B complies with standards (4Z5 - 52.5 mmol L 1) for a period of four years. The shelf-life of 2 years is thus satis- factory Introduction Cardioplegia solutions, or cardioplegics, are drugs used in heart surgery to stop the heart from beating, thereby facilitating the work of surgeons. Candidate drugs must reduce the energy requirements of the myocardium, create a favorable envir- onment for the resumption of energy pro- duction and combat the harmful effects of ischemia [1]. Although there are many formula- tions, there is no unanimous consensus among surgical teams [2]. Five solutions are currently used for cardioplegia in the Paris Hospital System (Assistance Publi- que des H6pitaux de Paris; AP-HP): CP1 | (PCH) or Fabiani's solution, CP9 | (PCH) or Menasche's solution, SLF 103 | (Aguet- tant), Plegisol | (Abbott), and SLF 11 | (Fresenius), equivalent to Bretchneider's solution. The compositions of these pre- parations are listed in Table I. CP1 ac- counted for 90% of cardioplegia solutions purchased from the Pharmacie Centrale des H6pitaux de Paris in 1996, of which 30% were used by the AP-HP. The classic composition of the formulations includes potassium, the plegia active ingredient, calcium, which blocks muscle contraction, and other ingredients intended to main- tain pH, osmolality, and protect cell inte- grity [1]. The role of procaine, found in CP1 only, is essentially to stabilize the cell membrane [3, 41. Procaine is the ester ofpara-aminoben- zoic acid (PABA) and diethylaminoetha- nol. It can be hydrolyzed, a process which is accelerated by alkaline pH [5], yielding PABA and diethylaminoethanol accord- ing to the breakdown reaction: Procaine 0 [/CH3 H~,N~ o~NVCHs ~./CO OH + HO_CFI/_CH~_/2H, H2N ~1~2H S PABA diethylaminoethanol Because CP1 solution is buffered at pH 8 with Tris(hydroxymethyl)aminomethane (THAM), it is packaged as two solutions to be mixed just before use the vehicle CP1A (1000 mL) containing THAM, and concentrated CP1B (10mL) which con- tains magnesium chloride and the pro- caine. The rate of hydrolysis is also consider- ably accelerated by heat [6]. Cardioplegia solutions are, however, injected directly into the heart of the patient and must thus be sterile. CP1B ampoules are currently sterilized by autoclaving for 30min at l l0~ Hydrolysis of procaine yields para-aminobenzoic acid [6], which colors the solution yellow; its systemic toxicity at low doses has not been clearly established [7]. It is thus important to quantify pro- caine breakdown to determine the extent of exposure of patients to this breakdown product and to verify the compliance of Original Chromatographia Vol. 52, No. 1/2, July 2000 115 0009-5893/00/02 115- 05 $ 03.00/0 (cid:14)9 2000 Friedr. Vieweg & Sohn Verlagsgesellschaft mbH
`
`MYLAN ET AL. - EXHIBIT 1022
`
`

`
`Table I. Composition (mmol L L) of the cardioplegia solutions available at the PCH. CPI CP9 SLF 103 PLEGISOL SLF 11 Na + 144 0 147 110 12 K + 20 16 20 16 l0 Ca 2+ 4 0 2 1.2 0 Mg 2+ 16 3 16 16 2 CF 201 22 203 160 34 THAM 2.8 0 0 0 0 HC1 qs pH = 8 0 0 0 0 Procaine 1 0 0 0 0 procaine with standards, both after auto- claving and after storage. The aim of this work was to determine the stability of procaine in CP1B as a function of sterilization conditions and sample storage times. After validation of the analytical method, PABA (procaine breakdown product) was assayed in sam- ples of the same batch of CP1B subjected to four different sterilization cycles, and in 32 batches of CP 1B manufactured over 5 years and stored at room temperature protected from light. Experimental Products and Equipment Reagents All starting materials used were analytical grade, supplied by Prolabo (Paris, France) and Sigma (St Louis, MO, USA). Fluo- rescent silica gel nanoplates for thin-layer chromatography (HPTLC plates; 100 mrn x 200mm, silica gel 60F2s4) were sup- plied by Merck (Darmstadt, Germany). Standard Solution Six aqueous standard solutions of PABA were prepared for the calibration ranges (50.5, 58.0, 58.3, 59.7 and 63.5mgL -1) and the control (79mgL 1). They con- tained 1.36gL -t procaine, 60gL -1 po- tassium chloride and 145gL -1 magne- sium chloride, hexahydrate. This compo- sition was thus very close to that of CP 1B. Chromatographic development of the standards was comparable with that of the samples. These solutions were stored at + 4 ~ protected from light. Sample The study was conducted with drug sam- ples stored at room temperature and pro- tected from light. All batches (32 in total) manufactured in a five-year period were examined; one was rejected because its procaine concentration was non-compli- ant. A special batch of 2500 ampoules of CP1B was manufactured by the PCH pro- duction unit for determination of the ef- fect of sterilization conditions on procaine hydrolysis. The batch was divided into five sub-batches, corresponding to differ- ent sterilization cycles: non-sterilized, 30 rain at 110 ~ 45 min at 110 ~ 20 min at 120 ~ and 30 min at 120 ~ Manufac- turing and sterilization conditions were chosen to enable work with methods vali- dated according to Good Laboratory Practice 8. Equipment An automatic sample applicator (AS30 TLC applicator) and the densitometer for quantitative analysis (CD60 Densito- meter) were obtained from Desaga (Hei- delberg, Germany). The horizontal linear development chamber was obtained from Camag (Muttenz, Switzerland). Methods Procaine Assay Procaine was assayed by UV spectropho- tometry at 290 nm, a method that has been validated for the quality control of manu- factured batches. In compliance with the decree of December 9, 1992 9, the pro- caine concentration was _+ 5% of the target value, or 50 + 2.5 mmol procaine per liter CP lB. The results of procaine assays con- ducted when each batch was validated are found in the corresponding analysis certi- ficates. Because this assay is not specific for procaine, the results are actually the total quantities of procaine and PABA in ampoules of CP 1B. PABA Assay PABA was separated by high-performance thin-layer chromatography (HPTLC) on fluorescent silica gel nanoplates. The as- say was conducted as described in the test for related substances in the monograph of the European Pharmacopoeia, 3rd edi- tion 10. Fluorescence inhibition was de- termined by densitometry at 275 nm. This wavelength was selected after acquisition of a solid-phase absorption spectrum be- tween 230 and 350 nm. At 275 nm, absor- bance by fresh plates was low whereas ab- sorption by PABA was close to the maxi- mum. Sample Application The automatic applicator used enabled quantification of PABA, because of the precision of the volume applied and the homogeneity of the resulting applications. Sample (2 gL) bands were 5 mm long; the first band was 15 mm from the edge of the plate and adjacent applications were sepa- rated by 15mm. The spraying rate was 10 gLmin 1. If several applications were made at the same site the interval between applications was 10 s. The 100 mm x 200mm silica gel plates used enabled 12 applications, or four range points and 4 samples in duplicate, to be made on oppo- site edges of the plate. Among the samples applied, a control enabled the assay to be validated. The calibration range was prepared with the automatic applicator, using one of the five standard PABA solutions de- scribed above. The quantities applied for the four range points were between 100 and 510 ng PABA. Development Chromatographic development was per- formed horizontally (Camag system) in a ventilated cabinet. Runs took less than 15 rain, with samples applied in parallel on each side of the plate. The mobile phase was freshly prepared acetic acid-di- butyl ether-hexane, 4:80:16 (v/v). Under these conditions, procaine remained at the origin of the plate and PABA migrated ca 12 mm. The plate was dried in a ventilated cabinet and the migration distance of PABA was measured under a UV lamp (254nm) for precise location of PABA spots to enable programming of the den- sitometer. Densitometric Assay Densitometry measures the fluorescence extinction of spots with reference to the fresh plate. Peak areas were calculated by means of the CD60 software. A linear re- gression line not passing through the ori- 1 16 Chromatographia Vol. 52, No. 1121 July 2000 Original
`
`

`
`1 700 600 50D 400 300 200 100 0 -100 il r ii Figure 1 Chromatogram (PABA). of para-aminobenzoic acid gin (a peculiarity of thin-layer chromato- graphy) was then calculated. This function determines the quantity of PABA applied as a function of the measured peak area. Each sample was applied in duplicate and the quantity of PABA was calculated from the mean area of the two peaks. The coefficient of variation between two mea- surements of the same sample did not ex- ceed 5%. The control was applied in duplicate in all assay series. The series assayed was re- garded as validated if the control did not differ from its target value by more than 10%, and if the coefficient of variation be- tween the two measured values did not ex- ceed 5%. The results enabled the molar concen- trations of PABA and procaine in am- poules of CP1B to be calculated by use of the molar masses of procaine and PABA, 272.8 and 137.1 g mol 1, respectively. Analysis of the Results After development, typical chromato- grams contained two fluorescence extinc- tion bands that could be seen under UV light at 254 nm. These bands are described in the section Test for Related Substances in the procaine monograph of the Eur- opean Pharmacopoeia, 3rd edition 10. Under the conditions used, procaine did not migrate, remaining at the site of appli- cation (origin of the plate). PABA, the only breakdown product described in the procaine monograph, migrated 12 mm on average under the conditions used. The densitometric output from a typical chro- matogram is depicted in Figure 1. Table II. Technical validation of the HPTLC assay of PABA. Linearity y = b.x + a Reproducibility Repeatability Range Assay RF Assay 50.5mgL -1 b = 1.56 a = 94.23 r = 0.996 59.7 mgL I b = 1.62 a = 110.98 r = 0.995 58mgL -1 b= 1.95 a = 42.72 r = 0.994 58,3 mgL 1 b = 1.44 a = 2.45 r = 0.995 Standard All the range Standard solution points of the solution at 316 ng linearity study per application Results from Results from Results from thirty assays: sixty assays: six assays: m = 302 ng m = 0.316 CV= 4.8% CV= 3% CV= 3% 63,5mgL 1 b= 1.28 Recovery a = 51.92 / 95.60% r = 0.996 m, mean; CV, coefficient of variation; r, correlation coefficient. Results Validation of the Method Validation of an analytical procedure en- sures the quality of a result by scientific study of the reliability of the method by use of a statistical tool. At present there is no validation method suited to quantita- tive thin-layer chromatography. Despite this, the general principle for validation of an analytical method applied to the assay of a medicinal substance 11 was used as a reference in this work. The validation of assay methods using high-performance li- quid chromatography described by the SFSTP 12 is a practical basis for the method used. The validation results are summarized in Table II. Specificity The specificity of an HPTLC separation comes from the choice of the plate and mobile phase; these enable the separation of the components of the sample assayed to be optimized. PABA, the main break- down product of procaine, contains the aminobenzoic ring that is responsible for the spectral properties of the parent com- pound, and thus absorbs at the same wa- velength. Procaine is thus the only sub- stance in CP1 B that can interfere with the assay of PABA. The method used in this work enabled PABA to be separated from procaine, as described in the procaine monograph of the European Pharmaco~ poeia, 3rd edition 10. It is thus specific for PABA. Sensitivity The limits of detection and quantification of the method were determined by use of a PABA concentration range of eight points from 29 to 580 ng. The limit of detection is the smallest quantity of substance that can be detected. It was defined as the quantity de PABA that furnished a signal- to-noise ratio 3. The limit of detection was found to be 29 ng PABA. The limit of quantification is the smal- lest value that can be considered as valid for an assay. It was arbitrarily set at 3.3 times the limit of detection. The limit of quantification was 100 ng PABA. Linearily The limit of linearity was determined gra- phically by use of the calibration plot pre- pared for the PABA range prepared for study of sensitivity and was found to be 510 ng PABA. Five calibration ranges of four points each were prepared by means of five PABA standard solutions (50.5, 58.0, 58.3, 59.7 and 63.5mgL 1). The four range points, equivalent to 2, 4, 6, and 8 gL standard solution, were applied in triplicate to each plate. The regression coefficient for each plot, taken individu- ally, was consistently > 0.994. Reliabilily Repeatability. The same standard solution was assayed six times on the same plate. The coefficient of variation was 4.8% and so the method can be regarded as repeata- ble. Original Chromatographia Vol. 52, No. 1/2, July 2000 117
`
`

`
`Table III. Effects of sterilization conditions on the hydrolysis of procaine in CPI B. Sterilization Procaine PABA PABA conditions (mmolL -1) (mmolL -1) (mmolL 1) PABA (%) Non-sterilized 50 < LOD < LOD 0 110 ~ 30min 50 0.067 0.49 1 110 ~ 45 min 50 0.1 0.73 1.5 120 ~ 20 min 50 0.114 0.83 1.7 120 ~ 30min 50 0.123 0.90 1.8 LOD, limit of detection. Table IV. HPTLC assay of PABA in different batches of CP 1B. Analysis certificates Procaine lm PABA assays Storage time (years) (mmol L -1 year l) PABAm (% year 1) Procaine 2m (mmol L -1 year -1) 1 49.8 2.1 48.8 2 49.4 3 47.9 3 50.3 3.2 48.7 4 50.3 3.5 48.6 5 49.5 4.3 47.3 Procaine lm, mean initial procaine concentration on release of batch; Procaine 2m, mean procaine concentration calculated from amount of breakdown into PABA; PABAm, mean amount of PABA. Reproducibility. Reproducibility of chromatographic development can be as- sessed by calculating the retention factor (RF) which for each spot is the distance migrated by the solute divided by the dis- tance migrated by the solvent. The mean Rv calculated from the 60 range points used in the linearity study was 0.316 (CV = 3%). The RF can thus be regarded as reproducible. The concentration of the standard PABA solution (79mgL -1) was deter- mined 30 times (4 gL per application), in- cluding 16 times to validate different ser- ies of assays. The mean value found was 302 ng PABA with a CVof 3%. The meth- od can thus be regarded as reproducible. AccurQcy The accuracy of the assay was determined by use of results obtained from the stan- dard solution in the reproducibility study. The mean value obtained for this solution (302 ng PABA) differed from the target value by 4.4%, and the maximum varia- tion compared with the target value was 10%. The method is thus at least 90% ac- curate. Effects of Sterilization on Procaine Hydrolysis The concentration of PABA in CPIB be- fore sterilization was lower than the limit of detection of the method. PABA is thus formed as a result of breakdown of pro- caine during steam sterilization. PABA was assayed in each solution sterilized by one of the four sterilization cycles tested. The measured concentration remained consistently lower than 1 mmol PABA per liter CP1B (Table III). The quantity of procaine remaining in the CP1B solution was, therefore, consistently higher than the lower acceptable limit of 47.5 mmol L l, and the batch was always com- pliant. The results show that the extent of breakdown of procaine into PABA in- creased with autoclaving time and tem- perature. Sterilization conditions cur- rently applied to CP1B, 30 min at 110 ~ are the least aggressive towards procaine. Assay of PABA in CPIB The maximum concentration of PABA measured in one of the batches of CP1B after five years Of storage was 0.314 g L -1, i. e. 2.3 mmol L -l. Results from the study of the stability of procaine as a function of storage time are listed in Table IV. Discussion The PABA assayed is a procaine break- down product and not an active ingredi- ent, and is present in very small quantities. The precision of the method used is suffi- cient for assay of an identified breakdown product. The limit tests in the monograph of the European Pharmacopoeia 3rd edi- tion call for a calculated maximum PABA content of 500 gg per gram procaine base 10. The use of densitometric plate read- ing as method of detection enabled the limit of quantification to be reduced to 100 ng PABA. When performing linearity studies, the SFSTP recommends preparing six differ- ent ranges, each containing five range points assayed six times 12. In our work, five different ranges were prepared, each containing four points applied three times. The resulting calibration curves had different slopes (between 1.28 and 1.95; CV = 16%) and Y-intercept values (between 2.45 and 111; CV = 71%). The calibration curves are thus not reproduci- ble. This difference between the equations shows that the results are highly depen- dent on operating conditions, in particu- lar the quality of chromatography plates that are not perfectly homogeneous. These results show the necessity of prepar- ing a validated calibration range for qual- ity control for each assay series on the same plate. The reproducibility of RF values shows the homogeneous particle size distribu- tion of the silica gel nanoplates and the ex- cellent reproducibility of development. The C V of repeatability (4.8%) was ab- normally higher than that of reproducibil- ity (3%). This can be explained statisti- cally - the standard was assayed 30 times for reproducibility, and only six times for repeatability. Nevertheless, the CVvalues for repeatability and reproducibility are satisfactory (< 5%). The same is true for the minimum accuracy of the method (90%). The maximum expected error (10%) is acceptable for assay of an identi- fied impurity present in very small quanti- ties. Procaine breakdown detected in the study of the sterilization of CP1B was found to be temperature-dependent, but remained relatively low, not exceeding 2.5% (Table III). The procaine concen- tration thus remained within acceptable levels in all the samples (47.5 to 52.5 mmol L-l). Steam sterilization under the conditions described does not mas- sively destroy procaine as might have been expected. An alternative sterilization pro- cedure is thus not necessary. The plot of a hydrolysis curve (Figure 2) shows the ef- fect of sterilization conditions on procaine hydrolysis. The results in Table IV show that the concentration of PABA increased with time. Procaine is thus unstable in solution. After a given storage time, its residual le- vel in CP1B depends on its initial level. 118 Chromatographia Vol. 52, No. 1/2, July 2000 Original
`
`

`
`.~ 0.5 10 0 rain ~110~ 45 min 110~ 30 min Stetiliza~on co~itlol~ Figure 2 Effect of sterilization conditions on the appear- ance of PABA formed by hydrolysis of pro- caine in CP1B. After five years of storage at room tem- perature protected from light, the pro- caine concentration in one batch only was found to be lower than standards (46.4mmolL -1) and this batch was re- jected. Two other batches were at the low- er limit of acceptability. The shelf life of two years after the date of manufacture is thus totally satisfactory. After two years of storage, mean procaine breakdown was 3%, a level that remained almost constant during the third and fourth years. Conclusion As might have been expected, there was no evidence of considerable breakdown of procaine during sterilization of manufac- tured batches of CP1B. The concentration of procaine in the finished product re- mained compliant with the standards set by the European Pharmacopoeia 3rd edi- tion. The PABA formed is a breakdown product, the concentration of which must be determined, justifying this work. On the other hand, and at the current state of scientific knowledge, it is not a real public health risk - it is the active ingredient in other pharmaceutical preparations admi- nistered systemically. References 1 Donnelly, A.J.; Djuric, M. AJHP 1991, 48, 2444. 2 Hearse, D.J.; O'Brien, K.; Braimbridge, M.V. Ji Thorac. Cardiovasc. Surg. 1981, 81,873. 3 Bixler, T.J.; Gardner, T.J.; Flaherty, J.T.; Goldman, R.A.; Gott, V.L.J. Thorac. Cardiovase. Surg. 1978, 75(6), 886. 4 Nishi, T.; Guilmette, J.E.; Wakabayashi, A. Ann. Thorac. Surg. 1980, 30(4), 349. 5 Higuchi, T.; Havinga, A.; Busse, L.W.J. Am. Pharm. Assoc. 1950, 39, 405. 6 Martindale. The Extra Pharmacopoeia 1989, 29th edn, The Pharmaceutical Press, London. 7 Richard, J.; Lewis, S.R. Sax's Dangerous Properties of Industrial Materials 1996, Van Nostrand-Reinhold, New York. 8 Bonnes pratiques de fabrication (Good Manufacturing Practices), arr6t6 du 10 mai 1995 (Decree of May 10, 1995) Jour- nal Officiel de la Rdpublique Franfaise 1995, May 13. 9 Arr~t6du9 ddcembre 1996 (Decree of De- cember 9, 1996) Journal Officiel de la Rd- publique Franfaise 1996, December 29. 10 European Pharmacopoeia 1997, 3rd edn, Council of Europe. 11 Touratier, S.; Pradeau, D. Validation d'une mdthode analytique appliqu6e au dosage du m6dicament. In: Analyse pra- tique du mddicament 1992, 115. E. M. In- ter. Cachan. 12 Caporal-Gauthier, J.; Nivet, J.M. STP Pharma Prat. 1992, 2(4), 205. Received: Nov. 8, 1999 Accepted: Jan 24, 2000 Original Chromatographia Vol. 52, No. 1/2, July 2000 119