6 DEVELOPMENTS IN PHARMACAUTICAL ANALYSIS Proc. Analyt. Div. Chern. Soc. Developments in Pharmaceutical Analysis The following are summaries of nine of the papers presented at a Meeting of the Analytical Division organised by the Scottish Region and in conjunction with the Joint Pharmaceutical Analysis Group held on September 16th and 17th, 1975, and reported in the October issue of Proceedings (p.264). Recent Advances in Pharmaceutical Analysis- Do We Really Need Them ? C. A. Johnson British Phavmacopoeia Commission, 8 Bulstvode Stveet, Louzdon, W1 M 5FT A meeting of the International Pharmaceutical Association was held recently in Dublin, and after that meeting I met one of the delegates attending a convention of American morticians. On hearing that 1 was an analytical chemist concerned with the purity of medicines, he asked me what I thought was the most important thing about the formalin which he used in large volumes. My answer that it should be of maximum possible strength was not correct and he told me that the formalin he used on his cadavers should be absolutely free from toxic impurities, the reason being that in any examination following an exhumation he wanted to be able to prove that any arsenic or similar substance found in one of his clients was not.present in his formalin! He also gave me a salutary reminder that analytical chemists are now “finding smaller and smaller amounts of things that are hardly even there,’’ and our encounter reminded me to discuss in this paper impurity levels and sensitivities of methods. Referring to the title of this meeting and of my contribution to it, I decided that “recent” should cover a period of 21 years, as true “advances,” rather than continuing “develop- ments,” are made relatively rarely, and also the concept of ensuring that a medicine is suitable for the patient has changed radically during that period.At the beginning of the period under review, the foundations of gas chromatography had recently been laid by James and Martin.Although the relevance of this technique to pharma- ceutical analysis seemed doubtful then, only 6 years later, when gas - liquid chromatography was being applied to the separation of certain oestrogenic steroids, the vital importance of gas chromatography as a tool in the pharmaceutical laboratory and not just as one of value to lipid and petroleum chemists had been recognised.Also, near the beginning of this period, non-aqueous titration, the Schoniger oxygen-flask combustion technique and the use of metal indicators in Schwarzenbach’s titration procedure for hardness of water, which was transformed into an accurate, precise and simple method for determining most metal ions, were notable breakthroughs. Soon, thin-layer chromatography, introduced by Stahl, became established as one of the most extensively used methods in pharmaceutical analysis.In those days before the “Guide to Good Manufacturing Practice,” quality control as we know it today was in its infancy, and the adequacy of any precautions taken depended largely on the experience of the production manager.Often incidents had to occur before specific dangers were recognised and, as diverse incidents came to be recognised as stemming frorn the same type of general failing, general codes of practice were developed. Probably the greatest impetus to the establishment of the pharmaceutical analyst as a key figure in ensuring over-all quality came in the aftermath of the thalidomide disaster.The growth of registration authorities, requiring to know the details of the structure, composition and analysis of new compounds proposed for medicinal use, has elevated the pharmaceutical analyst to a position of importance he could not have aspired to 21 years ago. ,4n incident involving defective thyroid tablets, where biological assessment showed that the activity of a suspect batch of tablets was only one fifth of the expected value, although the total iodine content of defective and satisfactory tablets were identical, reminded pharmaceutical analysts that results obtained from purely chemical and physical measurements may sometimes belie the actual worth of the material being examined.Januavy, 1976 DEVELOPMENTS I N PHARMACEUTICAL ANALYSIS 7 Is it possible that the tremendous upsurge of interest in the analysis and control of pharma- ceuticals has led in some instances to over-reaction and the loss of a sense of proportion and reality? The BP 1953 contained in its specifications very few tests to limit impurities, and these were restricted to named impurities that could be sought by classical procedures.However, the advent of gas and thin-layer chromatography transformed the situation, as can be seen in the BP 1963 and the BP 1968. Sensitive detection systems such as fluorescence quenching on thin-layer plates, the use of the flame-ionisation detector and multiple detection systems, has enabled the concept of homogeneity of drug substances to be exploited.However, although qualitative information on impurities can be obtained, precise quantitative information can be obtained only if the impurities are isolated, characterised and quantified. It has even been suggested that they should then be synthesised for pharmacological and toxicological testing. Is this justified as routine practice? In my view, to demand such procedures in every instance is an excessive and costly use of scientific resources.Perhaps we, as analytical chemists, are in danger of losing sight of why we carry out analyses and are rather pursuing them for their own sake, in order to achieve greater and greater sensitivity. Of course, my views are based on my concern with standards and specifications for drug substances and medicines, and more sensi tive methods are essential in clinical and metabolic studies, where further advances will always be necessary.Nonetheless, restraint and commonsense in applying these advances are needed. Bioavailability, because of the difficulty of assessing the clinical response to a given dose of a drug, tends to be measured in terms of blood levels, which in my view may not lead to accurate assessments of clinical response.Even when the pharmaceutical analyst chooses the most favourable conditions when carrying out laboratory tests with carefully selected animals he must accept limits of errorthat would often be unacceptable in a chemical or physical context. The clinician, however, has none of these advantages as the variability between his patients is great.In addition, different subjects metabolise drugs at different rates, and it seems an almost impossible task to predict the biological response in a patient selected at random. This is not a recommendation to abandon the collection of information on blood levels of drugs in patients, but rather a plea to regard such information from a perspective of common sense and not to invest the precise results that can now be achieved on blood levels with more importance than they can bear.This, of course is a difficult plea to make as there are many commercial interests with various attitudes and there is an expansion of “research” units that collect more and more data without concern for their relevance in real life. In my opinion there is no logic in insisting, because a pharmacopoeia1 monograph could not be expected to require that in vivo blood levels be determined in humans, that all standards be abandoned.I t still seems useful t o require confirmation that the right drug, of adequate purity, has been uniformly distributed in the product. The solution rate test is also inadequate for predicting clinical response, but it does provide information on a physical attribute that could be useful in ensuring that there is no significant variation from batch to batch. It should always be borne in mind that the job of the pharmaceutical analyst is not to demonstrate the marvels of different analytical techniques, but to protect the patient from failure of medicines and to protect his employer from the consequences of human and mechanical errors.Maximum effort should be put into evaluating and compensating for real clinical differences that have been observed in practice. In this context it is my view that, as a low-dose presentation of a rather insoluble but highly potent drug, the tablet is un- satisfactory; it should be abolished and replaced with a more effective means of dosage in the very few instances where it could mean a matter of life or death. In considering microbial contamination also there is a need for a balanced perspective of the significance of different situations.In some instances it is now virtually impossible to obtain some materials of a quality considered to be acceptable, e.g., tragacanth. Is this proscription really necessary, and is the major emphasis of effort sometimes misdirected? The answer, of course, must be “yes,” on every front.We need greater sensitivity of methods, particularly for metabolic studies, for solution rate tests, for determination of uniformity of content in low- dosage preparations and for forensic studies, but we must not demand determinations of smaller and smaller amounts of trace impurities unless there is a demonstrable reason why Let us consider the attitude to impurities over the last 20 years.Do we, then, really need advances in pharmaceutical analysis?8 DEVELOPMENTS IN PHARMACECTICAL ANALYSIS PYOC. Agzalyt. Dia. Chewz. SOC. we should. We must never allow our consideration of acceptable criteria to b: conditioned by the sensitivity of the methods available to us.We need methods capable of automation so as to permit more frequent, less time-consuming analyses free from human error. We must develop newer methods to deal with new problems, and must adopt techniques currently used in other fields, e.g., enzymatic, fluorimetric and electrophoretic methods. We need more penetrating studies of the containers and presentation of pharmaceutical products. In all of these aspects, the advances in our knowledge and ability should not automatically be coupled with more stringent requirements.Our Coat or Arms depicts a balance, evenly poised, and that is the situation to which we must aspire. Automation in the Analysis of Drugs J. W. Murfin Analytical Development, Reseavch, The Boots Company Limited, NottiNgham, NG2 3 A A Automated analysis is necessary in the assay of unit small-dosage preparations, with which the chance of variation of dosage from unit to unit is high; it is desirable in the analysis of single- dose preparations, in which the content of active ingredient is larger, particularly when long manufacturing runs are made.The automation of the measurement of dissolution rates and dissolution profiles of tablets can save much operator time.The two main systems of fully automated analysis are the continuous-flow system and the discrete system. The continuous-flow system has fewer moving parts, and can handle greater dilutions and reproduce more manual techniques than the discrete system. It is, however, much slower. The discrete system can handle up to 300 samples per hour, but requires more instrumentation : in particular, a series of syringe pumps that are capable of delivering precise volumes over a long period.There is a good choice of automated equipment available now, except, in the author’s opinion, for sampling units. A sampler that will accept the solid dosage form and that can be programmed to dissolve it in an accurate volume of solvent as low as 10 ml and at least as high as 100 ml is needed.Filtration in high-speed systems poses problems which do not seem to have been properly solved yet. Discrete systems handling 300 samples per hour require high filtration speeds for short periods. To obtain the necessary speed a filter having a larger dead volume than is desirable may be required. Filters that are satisfactory at low speed may clog or pass fine particles at higher speeds.There is also the problem of changing the filter automatically between samples. Many types of detector have been used. The author has found that Vitatron colorimeters for the visible range and the Cecil 272 spectrophotometer for the ultraviolet range are reliable. Fluorimeters can be adapted to take flow-through cells, and other specialised techniques such as flame photometry, atomic-absorption spectrometry and polarography have been used automatically.A reaction which seems imprecise when used manually may be acceptable under automated conditions, e.g., the reaction of salicylic acid or paracetamol with iron(II1) chloride. Many manual operations can be automated. Serial dilution is simple in continuous-flow analysis, as is extraction into an immiscible solvent and re-extraction into an aqueous medium.Heating can be accommodated up to a point, also the insertion of delay coils in order to enable a slow reaction to take place. In general, however, the longer a system, the greater is the chance of loss of sample integrity. Aqueous solvents run most quietly in continuous-flow systems, but other solvents, except acetone and esters, can be used provided that attention is paid to the bubble pattern. Most tertiary bases in pharmaceutical formulations can be assayed by extraction into chloro- form of the ion pair formed with certain dyes, followed by measurement of the colour of the chloroform layer.For extraction the author prefers the method described by Kuze1,l using a capillary coil of PTFE or polyethylene tubing.Kuzel also used bromocresol purple to form the ion pair, which has been found generally satisfactory. Digoxin, in tablets that contain 0.25 mg each, can be assayed by oxidising the digitoxose portion of the molecule to malonaldehyde with potassium periodate, followed by reaction with thiobarbituric acid to give a red colour.Jamavy, 1976 DEVELOPMENTS IN PHARMACEUTICAL ANALYSIS 9 The uniformity of dosage of tablets of thyroxine containing 50 or 100 ,ug of thyroxine can be determined, but the method is not accurate enough to be used for assay purposes.Thyroxine, among other iodine compounds, catalyses the reduction of cerium(1V) salts with sodium arsenite. The calibration graph is curved, i.e., Beer's law is not obeyed, and results of about 95 per cent.of the theoretical are obtained. It is said that because an automated reaction takes place under exactly similar conditions every time, it does not need to go to completion in order to achieve satisfactory results. Two colorimetric reactions were used to check this statement. One was the assay of digoxin, already mentioned, the other the assay of ammonium salts with hypochlorite and alkaline phenol.In the digoxin assay, a 40-ft delay coil heated at 75 "C was inserted in the system at the colour formation stage to enable the reaction to go to completion. The heating tempera- ture was varied from 30 to 80 "C at 10 "C intervals so that the extent of reaction varied from 9 to 100 per cent. In the ammonium assay, which took place at room temperature, the system was shortened successively at the colour formation stage to give reaction extents from 100 to 8 per cent.The sampling time was varied in both methods, and the mean absorbance of ten replicates compared with the steady-state reading. Assuming 1 per cent. to be a reasonable upper limit for the relative standard deviation of a colorimetric assay, this criterion was satisfied in the digoxin method for all reaction extents from 9 to 100 per cent., when the absorbance was at least 90 per cent.of the steady-state reading. The ammonium assay was satisfactory when the extent of reaction exceeded 45 per cent. and the absorbance was again at least 90 per cent. of the steady-state reading.This work has been described in detail else- where.2 It is possible to obtain good repeatability in automated analysis, even though the reaction be far from complete. The reaction to be used should be checked, however, before accepting this statement generally. The application of automated analysis in the pharmaceutical industry has been gaining impetus over the last few years, and has become another useful tool in the hands of the analyst.References The reduction in colour is measured. 1. 2. Kuzel, N., J . Phavm. Sci., 1968, 57, 852. Murfin, J . W., UV Spectvovuz. GvouP Bull., 1974, No. 2, 21. The Fluorirnetric Analysis of Qestrogen in Oral Contraceptive Preparations J. H. McB. Miller and P. Duguid Medicines Testing Labovatovy, Phavnzacezztical Society of Gveat Bvitain, 36 Yovk Place, Edivzbuvgh, EH1 3H U At the present time there are three types of oral contraception available : the sequential type whereby oestrogen is administered alone for the first week, followed by a lesser dosage of the oestrogen in conjunction with a progestogen for the remainder of the course; the combined type in which both an oestrogen and a progestogen are present in the tablets; and the progesto- gen type in which a progestogen alone is administered.The work reported is confined to the first two types. The most commonly encountered oestrogens are ethinyloestradiol and mestranol, which are present at a very low dosage level (30-100 pg) in combination with progestogen, which is present at a level of from 5 to 1000 times that of the oestrogen in the preparation.Thus, an extremely specific method, which is unaffected by a large excess of progestogen, is required for the analysis of these oestrogens. Such a method should also be capable of being applied to con- tent uniformity tests for individual tablets as such tests are being increasingly applied as a measure of good manufacturing practice by regulatory agencies throughout the world.Many methods are available for the determination of these synthetic oestrogens but they may lack specificity or sufficient sensitivity for single-tablet assays. Ethinyloestradiol can be determined by its ultraviolet absorption but an Ei:k value of 71 precludes its applicability as a method for content uniformity tests. Titration with silver nitratel is obviously not possible but greater sensitivity can be achieved by utilising the reaction of the silver ions with the ethinyl group a t position 17 on the D ring of the steroid nucleus so that unreacted silver ions can be deter-10 DEVELOPMENTS IN PHARMACEUTICAL ANALYSIS Proc.Analyt. Div. Chewz. SOC. mined colorimetrically2 after complexation with dithizone or by atomic-absorption spectroscopy.Unfortunately, amounts of less than 1 mg cannot be analysed by this procedure as below this level stoicheiometric conditions cannot be achieved. Various c o l ~ r i m e t r i c ~ ~ ~ methods have been applied, mainly based on the Kober reaction5 by which a pink colour is generated on reaction of the oestrogen with sulphuric acid in ethanol.Such methods have been successfully applied to the bulk analyses of ethinyloestradiol and mestranol tablets3y4 but the method is at the limit of its sensitivity for single-tablet analysis and subject to interference from progestogens. A recent paper6 described the use of the re- action of diazotised 5-chloro-2,4-dinitroaniline with phenols in the determination of ethinyl- oestradiol in the presence of norethinodrone.However, an amount of tablet powder equivalent to 500 pg of ethinyloestradiol was required for the determination and, of course, the method could not be applied to the assay of mestranol as there is a methoxy group at position 3 on this molecule. Gas chromatography is sufficiently sensitive, and is highly selective, for the analysis of ethinyloestradiol or mestrano17js but suffers from the disadvantage that if applied to the analysis of individual tablets the procedure would be laborious and time consuming.This is especially true when silylation procedures are required. Fluorimetry, with its great sensitivity, thus appears to have potential in this area. Ethinyl- oestradiol and mestranol both exhibit a native fluorescence which is the basis of the official method of analysis of ethinyloestradiol tablets.9 Fluorescence can also be induced in oestro- gens by a number of reagents, including an ethanolic solution of sulphuric acid,1° phosphoric acid,ll sulphuric acid plus acetic acid12 and a mixture of chloroform and acetic anhydride.13 A number of fluorimetric procedures have been compared.Obviously, as the native fluorescence of ethinyloestradiol has been applied by the British Pharmacopoeia this was considered as a possible method for the analysis of oral contraceptive tablets.The method, is, however, inapplicable to oral contraceptive tablets because of quenching of the native fluorescence owing to the presence of progestogens (Fig. 1). For example, norethisterone in solution with ethinyloestradiol at the levels to be expected in commercial formulations results in quenching.n I 1 1 I n 3 d Quencher concentrationimg per tablet Fig. 1. Reduction of the native fluorescence of ethinyloestradiol in the presence of norethisterone ace- tate in the amounts likely to be present in oral contraceptive tablets. Norethisterone absorbs light at 240 nm and shows no absorption at 280 nm, the wavelength of -m=I-ujmum ~bsm-pt~c-~ cf et*~2-~jllce,.t~~adiol, so that t3e quexchivg is like2j- to be due to cu& sional processes.On the other hand, another progestogen, megestrol acetate, absorbs light at about 280 nm so that a t the levels likely to be present in solution most of the incident light will be absorbed by the megestrol acetate molecules so that quenching, in this instance, occurs by a “trivial” process.January, 1976 DEVELOPMENTS I N PHARMACEUTICAL ANALYSIS TABLE I 11 THE EFFECT OF THE DEGREE OF MIXING IN AN ULTRASONIC BATH OF ETHINYLOESTRADIOL RESIDUES IN CONTACT WITH SULPHURIC ACID - ETHANOL (70 + 30) Excitation wavelength, 542 nm ; emission wavelength, 560 nm.Time/min fluorescence intensity variation, per cent.Mean relative Coefficient of 0.5 2.0 5.0 42.4 49.8 70.0 30.0 18.9 8.1 The induction of fluorescence in the molecule with sulphuric acid was unsatisfactory owing to the low solubility of ethinyloestradiol in the sulphuric acid - ethanol reagent. The degree of mixing in an ultrasonic bath had a marked effect on both the relative fluorescence intensity and on the reproducibility of that fluorescence response (Table I).In further tests quenching also occurred with some progestogens, e.g., megestrol acetate, and an increase in fluorescence was observed with norethynodrel and norgestrel as these compounds also fluoresced under these conditions at wavelengths similar to those for ethinyloestradiol and mestranol. Partition of the fluorescent moiety between the reagent and an organic solvent seemed to have advantages over the above method.This procedure was first used by Ittrich14 in order to overcome interferences from residual organic compounds co-extracted from blood and urine in clinical chemistry. Similar problems can occur by “carry-over” of tablet excipients in extraction procedures in pharmaceutical analysis.Ittrich extracted the sulphuric acid induced fluophor of oestrogen with p-nitrophenol in chloroform. Later, he added hydro- quinone to the reagent to improve the stability of the fluorescence. Re-examination of the method for its possible application to pharmaceutical analysis was undertaken. The main reagent used throughout the work was 70 per cent. sulphuric acid in ethanol and the main solvent employed was dichloromethane, which was shown by Ittrich15 to give good stability and precision when applied to his system.However, it should be stressed at this point that the reagent must be recently prepared, the dichloromethane must be freshly purified and redistilled,l6 the glassware must be cleaned in chromic acid1’ and grease on glass joints and rubber stoppers must be avoided.Provided these precautions are observed excellent results will be obtained. Ittrich usedp-nitrophenol in the solvent and hydroquinone in the reagent in order to stabilise the fluorescence of the oestrogen. Fig. 2 shows the stability of the fluophor of ethinyloestradiol with various solvent - reagent combinations. It is clear that the reagent containing hydroquinone results in the most stable fluorescence.Table I1 shows the effect of the addition of the two stabilisers, p-nitrophenol and hydroquinone, on the precision of the results, fromwhichit is clear that the addition of P-nitrophenol causes a great increase in the co- efficient of variation and therefore this solvent system was considered to be unsuitable. Various concentrations of hydroquinone in the reagent were examined and optimum results were obtained by using 0-2 per cent.m/m hydroquinone in the reagent. Variation in shaking time with the reagent in contact with solvent resulted in variation in the relative fluorescence intensity. A shaking time of 5 min was chosen in order to ensure complete mixing with the TABLE I1 PRECISION OF THE FLUORESCENCE RESPONSE OF STANDARD SOLUTIONS OF ETHINYLOESTRADIOL WITH VARIOUS SOLVENT AND REAGENT SYSTEMS Mean relative fluorescence Coefficient of Solvent system Reagent intensity variation, per cent.(n = 5) 2% y / V p-nitrophenol in dichloromethane 0-2 % yn/ V hydroquinone in sulphuric acid - ethanol (70 + 30) 54.3 15-1 Dichloromethane 0.2% yn/ V hydroquinone in sulphuric acid - ethanol (70 + 30) 50.9 ethanol (70 + 30) 72.1 Dichloromethane Sulphuric acid - 0-35 2.912 DEVELOPMENTS IN PHARMACEUTICAL ANALYSIS PYOC. Analyt.Div. Chewz. SOC. L-_L--L 1 ‘ I ___ 0 1 2 3 4 T irne/li Fig. 2. The stability of the fluo- phor, induced by a number of re- agents, in various solvents: (A) sulphuric acid - ethanol (70 + 30) ; (B) solvent dichloromethane and reagent sulphuric acid - ethanol (70 + 30) ; (C) solvent dichloro- methane and reagent sulphuric acid - ethanol (70 + 30) with 0.2 per cent.mlm of hydroquinone; (D) solvent dichloromethane with 2 per cent. of p-nitrophenol and reagent sulphuric acid - ethanol (70 + 30) with 0.2 per cent. m/m of hydroquinone. minimum breakdown of the fluophor. Self-quenching was not found below concentrations of 1 pg ml-l of ethinyloestradiol in dichloromethane.The usual working concentrations were 0.2 pg ml-l of ethinyloestradiol for the samples and 0.1-0-4 pg ml-l of ethinyloestradiol for the standards . Procedure for Analysis of Ethinyloestradiol in Tablets Place a tablet in a 150-ml separator containing 20 ml of water acidified by the addition of 5 ml of dilute hydrochloric acid and allow the tablet to disintegrate.Extract with four 20-ml aliquots of dichloromethane and pass each extract through a plug of anhydrous sodium sulphate into a 100-ml calibrated flask. Dilute this solu- tion with dichloromethane, if necessary, to obtain a final concentration of 0.2 pg ml-1 of ethinyloestradiol. Pipette 15 ml of this solution into a 25-ml calibrated flask to which is then added 3 ml of reagent [0.2 per cent.wz/m hydroquinone in sulphuric acid - ethanol (70 -t 30)]. Shake for 5 min. Transfer the upper dichloromethane layer to a fluorescence cell and measure the fluorescence using an excitation wavelength of 542 nm (emission wavelength, 560 nm). At the same time, add 3 ml of reagent to 15 ml of a standard solution of ethinyloestradiol in dichloromethane (0-2 pg ml-l) and subsequently treat as the sample.A number of commercial ethinyloestradiol tablets were then analysed by the proposed fluorimetric procedure and by other methods, including ultraviolet spectroscopy and native fluorescence, the results of which are presented in Table 111. The agreement between the results obtained by the various methods was excellent and the precision of the proposed method was generally superior.The proposed method was then applied to a number of oral contra- ceptive tablets of varying strengths of ethinyloestradiol and containing various progestogens. For tablets containing between 30 and 100 pg with megestrol acetate, norethisterone acetate, d,Z-norgestrel and d-norgestrel the ethinyloestradiol content found was 90.7-97.9, 96-8, 97.5 and 100-5-104-7 per cent.of the stated content, respectively, and the single-tablet assay results showed excellent uniformities of content, the variation being less than 10 per cent. A large number of oral contraceptive tablets have been examined in this laboratory and all have exhibited good uniformity of content patterns. Make up to volume with dichloromethane.Jaizuary, 19’16 DEVELOPMENTS I N PHARMACEUTICAL ANALYSIS TABLE I11 13 COMPARISON OF RESULTS OBTAINED BY DIFFERENT METHODS FOR THE ANALYSIS OF ETHINYLOESTRADIOL TABLETS AND MESTRANOL TABLETS Content of oestrogen, per cent.of stated amount ---------------h----------- - Ultraviolet spectroscopy Native fluorescence Induced fluorescence Amount r----h---- 7 r---h--7 r--__h- 7 stated Found Standard Found Standard Found Standard Tablet on label/pg deviation deviation deviation Ethinyloestradiol 100 94.3 1.80 94.0 2.59 98.6 1-17 Ethinyloestradiol 50 52.4 2.20 - - 52-0 1.03 Mestranol 100 98.8 2-38 - - 98.9 3.64 Although the proposed method is applicable to a large number of oral contraceptive prepara- tions, there are a number which cannot be analysed in this way.There are some progestogens that cause interference in this method by either increasing the fluorescence response, for example, dimethisterone, or by quenching, for example, ethisterone. Some means of treating such samples must be found. Firstly, by the addition of the interfering substance to the standards at the concentration expected in the sample. This has the disadvantage that should the quenching be great the sensitivity of the method may be sufficiently reduced to render the method inapplicable or that the calibration graph is not rectilinear.Secondly, chemical modification of an interfering progestogen has been employed.l* Norethynodrel and mestranol fluoresce at the same wavelength when reacted with sulphuric acid - acetic acid reagent.Reaction with sodium borohydride reduces the keto group in position 3 in the A ring of norethynodrel to a hydroxy group and subsequent treatment with l~ydrochloric acid changes the position of the unsaturated bond to the 3,4-position. The resulting compound, norethinodiol, fluoresces at a lower wavelength and the interference is obviated. Cullen et ~ 1 . ~ 9 showed that the addition of water to the sulphuric acid - ethanol reagent, after reaction with a steroid, may result in a shift of the excitation wavelength to shorter wavelength.For example, the norethisterone fluorescence excitation maximum is at 465 nm but on addition of water it shifts to 406 nm. Similarly, ethisterone shifts from a maximum at 465 nm to one at 415 nm. However, oestrogens do not show this phenomenon.This is another possible way of overcoming interferences from progestogens present in oral contraceptive tablets. Finally, various high-pressure liquid chromatographic methods have been published for the separation of oestrogens from proge~togens~o-~~ but these methods cannot be applied until a suitable fluorimetric detector is available. There are a number of approaches.References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. “British Pharmacopoeia 1968,” The Pharmaceutical Press, London, 1968, p. 395. Kizk, M., Vallon, J . J., and Badinand, A., Analytica Chirn. Acta, 1973, 65, 220. W u , J. Y . P., J . A s s . Off. Analyt. Chern., 1974, 57, 747. Wu, J. Y . P., J . A s s . Off. Analyt. Chem., 1975, 58, 75.Kober, S., Biochem. J., 1938, 32, 357. Eldawy, M. A., Tawfik, A. S., and Elshabouri, S. R., J. Phavm. Sci., 1975, 64, 1221. Talmage, J. M., and Penner, M. H., J . Pharm. Sci., 1967, 56, 657. Shroff, A. P., and Grodski, J., J. Phavm. Sci., 1967, 56, 460. “British Pharmacopoeia 1973,” I3.M. Stationery Office, London, 1973, p. 193. Silber, R. H., Busch, R. D., and Oslapas, R., Clin.Chem., 1958, 4, 287. Ichu, S., Forchielli, E., Perloff, W. H., and Dorfman, R. J., Analyt. Chem., 1963, 5, 422. Kalant, H., Biochern. J . , 1958, 69, 79. James, T., J. Pharm. Sci., 1972, 61, 1306. Ittrich, G. Z., Hoppe-Seyler’s 2. Physiol. Chern., 1958, 312, 1. Ittrich, G. Z., Acta Endow., Copenh., 1960, 35, 34. Spencer-Peet, J., Daly, J . R., and Smith, V., J. Endocv., 1965, 31, 235.Mattingly, D., J . Clin. Path., 1962, 15, 374. Pastor, J., Pauli, A. M., and Papoccia, N., Trav. Soc. Phavm. Montpelliev, 1973, 33, 411. Cullen, L. F., Rutgers, J. G., Lucchesi, P. A., and Papariello, G. J., J . Phavm. Sci., 1968, 57, 1857. Siggia, S., and Dishman, R. A., Analyt. Chem., 1970, 42, 1223. Bailey, F., and Brittain, P. N., J . Chvomat., 1973, 83, 431. Cavina, G., Moretti, G., Mollica, A., and Antonini, R., J .Chromat., 1971, 60, 179. Henry, R. A., Schmit, J . A., and Dieckman, J - F., J . Chrornat. Sci., 1971, 9, 513.14 DEVELOPMENTS IN PHARMACEUTICAL ANALYSIS Proc. Artalyt. Div. Chem. SOC. Recent Applications of High-performance Liquid Chroma t og rap h y i n P ha r m aceu t i ca I An a I ysi s John H. Knox, Jadwiga Jurand and Andrew Pryde Wolfson Liquid Chromatography Unit, Department of Chemistry, University of Edinburgh, West Maim Road, Edinburgh, EH9 3 J J The development of high-performance liquid chromatography (HPLC) since the late 1960s is in many ways similar to that of gas chromatography. The technique is now about 8 years old and is a t about the same stage of development as gas chromatography was in 1958 when capillary columns and ionisation detectors appeared.In HPLC the validity of small-particle packings is now fully established and increasing use is being made of bonded packing materials, i.e., packings in which the surface of an adsorbent is drastically modified by chemical reaction. TO date there has been no dramatic breakthrough in methods of detection but already advances towards more universal detectors are being made, for example by the use of HPLC - mass spectrometric interfaces. At the same time detectors of high specificity are being developed, for example, those based on fluorescence and chemical reaction prior to optical detection.The expansion of the literature on HPLC also mirrors that on gas chromatography. The rate of production of papers in major journals is about 300 per annum and the rate is doubling roughly every two years.We can expect this to continue for between 5 and 8 years more. Roughly one third of the papers are devoted to the technique of HPLC per se, another third to general applications, and the final third to pharmaceutical or closely related biochemical problems. It is therefore evident that by far the most important single area of application is pharma- ceutical chemistry and biochemistry.It is unlikely that this bias will be significantly altered in the future. We simply note here that typical column-inlet pressures are in the range 30-300 atm, flow-rates are from 0-3 to 3 cm3 min-l, sample volumes are from 1 to 10 p1 and columns are from 100 to 500 mm in length and of 2-5-mm bore; packing materials are of particle size from 5 to 20 pm.Detection sensitivities for favourable samples and according to the detector used are: refractive index monitor, 1 in lo6; ultraviolet photometer, 1 in 109; fluorescence monitor, 1 in lolo; wire transport - flame ionisation detection systems, 1 in 106. Table I shows that gas chromatography and HPLC now have very comparable performances in terms of speed, resolution, sensitivity and precision but that the cost of HPLC equipment is still two to three times that of gas-chromatographic equipment.Equipment for HPLC has been described many times elsewhere. TABLE I COMPARISON OF PERFORMANCE OF GAS CHROMATOGRAPHY AND HPLC Performance HPLC Gas chromatography Range of compounds All soluble compounds especially involatile ones.Relative molecular mass > 100 Minimum sample 10-9-10-6 g Precision 0.5-3% PIate efficiency 1000-100000 Cost @OOO-10000 Analysis time 1-30 min Data handling Recorder, integrator and computer All compounds volatile without decomposition. Relative molecular mass < 400 10-12-10-9 g 0.5-3 1000-1 000000 1-30 min i1000-3000 Recorder, integrator and computer Although HPLC is most analogous to gas chromatography, it is clearly likely to be used for many analyses currently carried out by thin-layer chromatography (TLC) .Comparison of HPLC with TLC reveals that while separation times may be similar the resolution of well re- tained solutes in TLC is much poorer than in elution chromatography (HPLC). This is because such components traverse only a fraction of the plate or column in TLC, whereas they traverse the entire column in HPLC.It is also clear that to produce the same analysis times in HPLC and TLC the linear velocity of eluting agent in HPLC must be 5-20 times that inJanuavy, 1976 DEVELOPMENTS I N PHARMACEUTICAL ANALYSIS 15 TLC. As shown below this requires the use of very small particles for column packing and high pressures across the column.In order to appreciate how the operating conditions are chosen for HPLC it is important to recognise that two independent factors determine any chromatographic system. To obtain resolution of two solutes the disengagement of the peak maxima must exceed the mean peak width (or some chosen multiple of it).Peak disengagement is governed exclusively by the equilibrium distributions of the solutes between the moving and stationary zones in the column and increases linearly with distance migrated, while peak width in well behaved systems in- creases as the square root of the distance migrated. In this way resolution increases with the root of the distance migrated. The peak width is largely governed by kinetic factors and of these the most important is the rate of mass transfer of solute between the mobile and station- ary zones.This rate is maximised by using particles of very small size. However, if reduc- tion in particle size is carried too far, band spreading due to flow and axial diffusion becomes important, and the pressure drop for a given plate performance eventually rises.Thus, a compromise must be struck with optimum values for particle size and column dimensions. With the current pressure capabilities of up to 100 or 200 atm the optimum particle size is from 3 to 10 pm; the optimum column dimensions are from 100 to 500 mm (length) and from 5 to 8 mm (bore size), with the larger particles being used with the wider and longer columns.One important consequence of the theory of peak migration is that the degree of retention of a solute must be strictly independent of its actual concentration in the mobile or stationary zones. If this is not so peak tailing or fronting occurs, the peak width no longer increases as the root of distance migrated, and resolution is grossly impaired. In adsorption chromato- graphy this arises especially with solutes whose molecules are both rigid and contain a large number of functional groups.The tetracyclines provide a good example of this ‘(template effect ,” whereby a surface which is uniform with regard to single-site adsorption appears heterogeneous for multi-site adsorption. To overcome this problem the surface of the ad- sorbent may be chemically modified but it is necessary to do this very thoroughly in order to avoid residual adsorption on to unreacted silanol groups.With efficient reaction, excellent separations of rigid polar molecules can then be carried out, as shown by Fig. 1. 15 10 5 0 T ime/m in Fig. 1. High-efficiency separation of tetra- cyclines on reverse-phase SAS-silica (Wolfson Liquid Chromatography Unit, Preparation No.354). Column, 125 x 5 mm; particle size, 6 pm; eluting agent, acetonitrile - water (35 + 65 V / V ) made 0.1 M in HCIO,; detection, Cecil ultraviolet photometer, 280 nm. Order of elu- tion: 1, epitetracycline; 2, tetracycline; 3, im- purity ; 4, chlortetracycline : 5 , epianhydrotetra- cycline ; 6 , anhydrotetracycline.16 DEVELOPMENTS IN PHARMACEUTICAL ANALYSIS Proc.Analyt. Div. Chem. SOC. In general, adsorption chromatography using silica or alumina is well suited to the analysis of not too polar solutes, for example aromatic hydrocarbons, simple phenothiazines and bromostilbenes, but is less satisfactory for more polar and more highly functionalised com- pounds. Such compounds are best separated on reverse-phase packings, i.e., adsorbents modified so as to possess a non-polar surface.Reference has already been made to the separation of the tetracyclines. The morphine alkaloids are also well separated on such a material, using aqueous methanol buffered with ammonia. Ionic compounds such as nucleotides and penicillins are generally separated by ion- exchange chromatography. Classically this is carried out on ion-exchange resin beads that are either homogeneous or pellicular (surface layered on to glass beads).In both instances mass transfer in and out of the resin is slow and wide peaks result unless very low elution velocities are used. Much better performance is obtained by bonding an ion-exchanging sur- face on to silica gel, for this gives a porous material with an ion-exchanging surface of high area and accessibility.Such ion-exchange silicas are ideally suited to the rapid analysis of nucleotides. Weak ion exchangers (for example pyridyl - alumina, or aminopropyl - silica) can also be used as polar reverse-phase adsorbents. Examples of their use in this way are the separations of barbiturates on pyridyl - alumina using a borate buffer at pH 7, and of paraquat and diquat in urine using aqueous methanol at an apparent pH of 3.3.A particularly fruitful ap- proach is to use reverse-phase partition chromatography with an eluting agent containing a small concentration of a surface-active agent. In the separation of B vitamins, for example, an acidic aqueous eluting agent was used that contained 0.1 per cent. of sodium lauryl sulphate.The packing was reverse-phase SAS silica gel (hydrocarbon surface). We believe that lauryl sulphate anion is adsorbed by the hydrocarbon surface and acts as a dynamic ion exchanger for amine cations. This technique has been used successfully in a number of other applications such as the separation of tricyclic tranquillisers and of catecholamines. This summary of some of our more recent work on the applications of HPLC to pharma- ceuticals and related biochemicals is no more than an illustration of the types of separations that can be achieved and of how one can set about choosing column packings and eluting agents.The range of applications will undoubtedly expand enormously but we can now be reasonably confident that no classes of compound will be found that will not be amenable to elution as sharp peaks from modern HPLC columns.Ion-pair formation can also be used as an aid in separation. A Review of Applications of Derivative Formation in the Quantitative Gas-chromatographic Analysis of Pharmaceuticals J. Nicholson Mediciws Testing Laboratovy, Pharmaceutical Society of Great Britain, 36 Yovk Place, Edi-uzbuvgh, EH1 3H U There are a number of reasons for preparing the derivative of a compound prior to chroma- tography rather than attempting to chromatograph the parent compound.These can be summarised as follows : lack of stability of the compound; non-volatility of the compound; to increase the sensitivity of detection, e.g., by electron-capture detection; to separate com- pounds that are not readily separated without derivative formation; to aid the rapid con- firmation of identity of a chromatographic peak ; to overcome adsorption problems during chromatography; and to produce compounds more suitable for gas - liquid chromatography - mass spectrometry.Having accepted the need to prepare a derivative, there are certain criteria which, ideally, it should meet. The following are some of the main properties of an ideal derivative : it must satisfy the reasons for making it; it must possess thermal stability; it should be formed rapidly; its preparation should require the minimum of manipulation; it should be stable; it should be formed quantitatively or at least reproducibly for quantitative analyses; its formation should be predictable with no side reactions or structural changes; it, or excess of reactants, should not react with the stationary phase ; when electron-capture detection is being used the derivative should be readily separable from excess of reagent; and the derivative should not be capable of being confused with other, possibly similar, compounds.All of these criteria are rarely met and a compromise has to be made.Januavy, 1976 DEVELOPMENTS I N PHARMACEUTICAL ANALYSIS Acylation 17 Drugs containing amine and phenolic groupings can readily form derivatives by acylation with acyl chlorides or anhydrides.Such derivatives have a greater relative molecular mass than the parent compounds and this property usually results in increased retention times. These increases can be minimised by the use of fluorinated reagents, of which one of the most popular is heptafluorobutyric anhydride (HFBA).Derivative formation is achieved by direct interaction of the drug and HFBA under an- hydrous conditions. Excess of reagent can be removed easily under a current of dry nitrogen before the derivative is dissolved in a suitable solvent for gas - liquid chromatography. The choice of solvent is important as hydrolysis of, for example, the derivatives of pentazocinel and oxyphenbutazone2in a much favoured solvent, ethyl acetate, has been noted.It has been suggested that benzene may be a better ~ h o i c e . ~ However, by maintaining the derivatives in excess of HFBA until just prior to gas - liquid chromatography their fidelity is ensured. In addition to improving the chromatographic be- haviour of the drugs, conversion into their heptafluorobutyrates also allows them to be detected with high sensitivity by electron-capture detection. Prediction of the response of electron- capture detection to halogenated acyl derivatives is difficult.Thus, the HFBA derivative of methamphetamine allows more sensitive detection than does the pentafluorobenzoyl deriva- t i ~ e .~ In some instances warming may be necessary. With thymol the situation is r e ~ e r s e d . ~ Methylation with Diazomethane An ethereal solution of diazomethane will methylate rapidly, for example, carboxylic acids and barbiturates, on its addition to the solid drug. Methylation is usually smooth and pre- dictable although unsaturated dicarboxylic acids such as maleic acid or fumaric acid have been reported to give anomalous results.6 An elegant technique has been reported for the rapid determination of salicylic acid in aspirin using diazomethane in tetrahydr~furan.~ For different drugs that give the same methyl derivative, e.g., theophylline and theobromine, use can be made of diazoethane in order to achieve separation.8 Esterification of Acids with Methanol A common reagent for the preparation of methyl esters of fatty acids is a 3 per cent. m/m solution of hydrogen chloride in methanol.Derivative formation occurs in about 15 min when the acid is heated with the reagent in a beaker of hot water for about 15 min. A similar reaction occurs when using about 14 per cent. m/V of boron trifluoride or 12 per cent.m/V of boron trichloride in methanol. In a boiling water bath, reaction is complete in about 3 min. Alkylation Acidic drugs such as barbiturates, glutetliimide and phenytoin can be methylated with dimethyl sulphate in a medium of aqueous methanol saturated with potassium carbonate. The reaction is complete in 5-10 min at 70 "C. With phenytoin, care has to be exercised with the potassium carbonate concentration, which, if too high, will allow the unstable 1,3-dimethyl- phenytoin to form in addition to the normal 3-methylphenyt0in.~ Under similar conditions, pentafluorobenzyl bromide will alkylate acidic compounds, such as flurbiprofen,1° to give derivatives capable of being detected with high sensitivity by electron- capture detection. However, reaction is relatively slow, i.e., about 90 min for completion at 60 "C.This time can be significantly reduced by using the same reagent in a technique known as extractive alkylation. By this means pentazocinell and ketobemidonell can be quantita- tively alkylated in 10 min. Flash Alkylation Flash alkylation is a technique that involves derivative formation by the interaction of the drug and a quaternary ammonium hydroxide to form a salt, which is then pyrolysed in the heated injection port (250-360 "C) of a chromatograph in order to produce an a k j d derivative.Derivative formation merely involves dissolving a convenient amount of the drug in excess of18 DEVELOPMENTS IN PHARMACEUTICAL ANALYSIS PYOC. Auzalyt. Div. Clzem. SOC. reagent, e.g., 0.1 M tetramethylammonium or trimethylanilinium hydroxide, and injecting this solution directly into a chromatographic column, when the methyl derivative will be formed.Typical drugs that will undergo derivatisation include barbiturates, xanthines, phenytoin and similar compounds. It has been reported12 that the phenylbarbiturates do undergo decomposition during derivative formation.Thus, phenobarbitone will form, in addition to the expected NN'-dimethylphenobarbitone, N-methyl-a-phenylbutyramide and a small amount of NN-dimethyl-x-phenylbutyramide. Silylation Silylation is now a common technique for producing stable and volatile derivatives of, for example, alcohols, phenols, carboxylic acids and amines. One of the most popular reagents for producing such derivatives is NO-bis(trimethylsily1)acetamide (BSA), which replaces re- active hydrogen atoms with trimethylsilyl groups.Reaction can be achieved by using pure BSA or a solution of BSA in, for example, pyridine and/or dimethylformamide. N-Trimethylsilylimidazole (TSIM) is one of the most powerful silylating agents known but reacts only with hydroxyl groups. In admixture with BSA and trimethylchlorosilane, a re- agent is produced that is capable of silylating all hydroxyl groups whether sterically hindered or not .I3 However, both they and the reagents are susceptible to hydrolysis and must be protected from moisture.Trimethylsilyl derivatives possess high volatility and chromatograph well. Carbamate Formation Carbamate formation is a reaction that has been used to prepare derivatives of tertiary amines that can be detected with high sensitivity by flame-ionisation or electron-capture detection.The amine is made to react with, for example, ethyl chloroformate (for flame- ionisation detection) or pentafluorobenzyl chloroformate (for electron-capture detection) in the presence of a base catalyst to give an NN-disubstituted carbamate.The subsequent removal of excess of reagents does present problems if electron-capture detection is to be used. Thus, amitriptyline will undergo carbamate formation by the elimination of an N-methyl group,14 whereas NN-dimethyldibenzo jb,f]- thiepin-10-methylamine (DTM) forms an NN-dimethylcarbarnate after the elimination of the major part of the molecule.l5 Forecasting the reaction product is difficult.1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. References Pittman, K. A., and Davison, C., J . Phavm. Sci., 1973, 62, 765. Bruce, R. B., Maynard, W. R., and Dunning, L. K., J . Phavrn. Sci., 1974, 63, 446. Ehrson, H., Walle, T., and Brotell, H., Acta Phavm. Suec., 1971, 8, 319. Driscoll, R . C., Barr, F. S., Gragg, B. J., and hloore, G. W., J . Phavm.Sci., 1971, 60, 1492. McCallum, N. I<., and Armstrong, R. J., J . Chvornat., 1973, 78, 303. Mlejnek, O., J . Chrornat., 1972, 70, 59. Watson, J . R., Matsui, F., McConnell, R. M. J . , and Lawrence, R. C., J . Phavm. Sci., 1972, 61, 929. Ferry, D. G., Ferry, D. M., Moller, P. W., and McQueen, E. G., J . Chvornat., 1974, 89, 110. Sabih, I<., and Sabih, K., J . Phavm. Sci., 1971, 60, 1216.Kaiser, D. G., Shaw, S. R., and Vangiersen, G. J., J . Phavm. Sci., 1974, 63, 567. Brotell, H., Ehrson, H., and Gyllenhall, O., J . Chvomat., 1973, 78, 293. Osiewicz, R., J . Chvomat., 1974, 88, 157. Chambaz, E. M., and Horning, E. C., Analyt. Lett., 1968, 1, 201. Hartvig, P., and Vessman, J., Acta Phavm. Suec., 1974, 11, 115. Degon, P. H., and Riess, W., J . Chromat., 1973, 85, 53.Recent Developments in Electrophoretic Analysis P. H. Corran iNational Institute f o v Biological Stavzdavds avzd Contvol, Holly Hill, Haunpstead, London, N W3 6RR Electrophoresis is simple, both in theory and in practice, and affords a cheap and rapid way of separating a wide variety of ions; it has not, however, found much favour in pharmaceuticalJanuary, 1976 DEVELOPMENTS I N PHARMACEUTICAL ANALYSIS 19 analysis.In this paper some new analytical requirements for which electrophoresis is proving useful are described briefly, and a technical development that seems to be of potential use in pharmaceutical analysis is also described. For general reviews of this subject, see references 1 and 2. Four kinds of electrophoresis can be distinguished : moving boundary and zone electro- phoresis, isotachophoresis and isoelectric focusing; they have in common the movement of ions in an applied electric field, but this paper will be confined almost entirely to zone electro- phoresis and isotachophoresis.Zone Electrophoresis1,2 In the classical model of electrophoresis an ion accelerates in an electric field to a limiting velocity at which the viscous drag of the solvent balances the electrostatic force; the mobility (the velocity in unit field) of an ion is then dependent directly on the charge, and inversely proportional to its Stokes radius (approximately cc 31/wT).As most ions of pharmaceutical interest are weak acids and/or bases, a careful choice of pH is important for optimum separa- tion.The supports used vary from paper or cellulose, for ions of low relative molecular mass (<loo), to polyacrylamide or agarose gels for those of high relative molecular mass (>3 000). Five of the six electrophoretic tests specified in the British Pharmacopoeia (BP) are for zone electrophoresis, mostly to limit the levels of impurities. Particularly of interest is the recent introduction of “mapping” of the peptides produced by digestion of the synthetic ACTH analogue tetracosactrin with trypsin as a test for identity.The mixture of peptides is separated first by electrophoresis, then the cellulose plate is turned through 90” and sub- jected to thin-layer chromatography. The pattern of spots revealed by staining is character- istic of tetracosactrin, and the “map” or “fingerprint” can be compared with a standard “map” prepared in parallel; the test is simple, fairly rapid and of obvious usefulness as more and more polypeptides enter pharmaceutical use.Polyacrylamide Gel Electrophoresis (PAGE) Polyacrylamide gel was introduced as a support in electrophoresis by Ornstein and Davis in 1963; the gels have proved simple to prepare and convenient to use.The average pore size of the gel can be adjusted between 1 and 6 nm by a suitable choice of acrylamide and NN’- methylenebisacrylamide concentrations, thus allowing a controlled molecular sieving effect to be added to the electrophoretic one. Much effort has been expended on the technique3 and on supplying a theoretical ba~kground.~ PAGE was introduced into the BP in the 1975 Addendum, where it is specified as a test to limit desamidoglucagon and foreign peptides in glucagon. The intensities of the bands due to the impurities that are seen after staining are compared with the glucagon bands at loadings of from 5 pg to 1 mg of the sample per gel.The intensities of bands caused by foreign peptides at the highest loading must not be greater than the intensity of the glucagon band at the lowest loading, which must be visible for the test to be valid.The sensitivity for insulin (a possible contaminant) is 0.2%. Another example in which PAGE may prove to be of value is in the examination of insulins that have been specially purified in order to remove the low levels of proinsulin (which co- crystallises with insulin), arginine insulin and aggregates, which are present in normal corn- mercial samples ; all of these contaminants, and the desamido insulins, are resolved from insulin itself by use of PAGE.Many biological substances in the BP are controlled by use of bioassay and little else (for instance, protamine) ; for these substances as well as for the sub- stantial number of synthetic and natural polypeptide hormones which are likely to be of therapeutic use in the near future, PAGE seems to be an appropriate means of analysis.Control of more complex mixtures of proteins is less simple; perhaps a method such as Laurell cross-electrophoresis, in which after gel electrophoresis in one direction the proteins are quantified by electrophoresis at right-angles into a gel containing antibody to the mixture, may be the solution. The result of using the Laurell method is certainly striking and of enhanced resolution.It is also possible that isotachophoresis may be a suitable method.20 DEVELOPMENTS IN PHARMACEUTICAL ANALYSIS Proc. Analyt. Div. Chewz. SOC. Isotachoph~resis~ When a less mobile ion (B-) is migrating electrophoretically behind a more mobile ion (A-) with a common counter ion (Rt) the equilibrium concentration of B depends simply on the concentration of A, on the charges of A and B and on the mobilities of the three components; the boundary between A and 13 is self-maintaining.For a series of zones of successively lower mobility, the equilibrium concentration can again be shown to depend only on the concen- tration of the leading ion A for a given system, and the length of each zone (in a container of constant cross-section) depends on the amount of the relevant components present.This phenomenon is illustrated in Fig. 1, with the output from a thermocouple shown below; if a separation is carried out at constant current (as is usual) the heat generated in each zone is inversely proportional to the mobility of the component ion, and the thermal output shows steps as each boundary passes the detector.BeIow these lines the voltage gradient is shown, together with the concentrations of the zones assuming equal charge. A pH gradient can also be included. Direction of electrophoresis c Differential thermal i t i \ I \ - Time - Distance Fig. 1. in an isotachophoretic separation. Diagrammatic representation of equilibrium conditions This property of the equilibrium concentrations makes the method applicable to the study of metabolites (as little as 1 pmol of a substance absorbing ultraviolet light strongly can be quantified satisfactorily) although a wide range of ions of all kinds can be separated and determined quantitatively.The major drawback is, perhaps, the lack of a sufficiently sen- sitive universal detector (the ultraviolet detector can only be used at wavelengths above 254 nm) but the technique is still in its early stages and future development will be consider- able.Januavy, 1976 DEVELOPMENTS I N PHARMACEUTICAL ANALYSIS 21 Isoelectric Focusing1 Isoelectric focusing is more expensive than PAGE but may be of use in detecting minor components, which form much more concentrated zones than they do in PAGE.However, staining of zones of small polypeptide is difficult because of interference by the ampholytes that are used to create the pH gradient. References 1. 2 . 3 Strickland, R. D., Analyt. Cheruz., 1974, 46, 95R. Huetteman, R. E., Cotter, M. L., Shaw, C. J., Janicki, C.A., Almond, H. K., Moyer, E. S., Stroff, Gordon, A. H., in Work, T. S., and Work, E., Editovs, “Laboratory Techniques in Biochemistry Chrambach, A., and Rodbard, D., Science, N.Y., 1971, 172, 440. Haglund, H., Sci. Tools, 1970, 17, 2. A. P., and Vestano, F., Analyt. Chem., 1975, 47, 233R. and Molecular Biology,” Volume 1, North Holland Publishing Co., London, 1969, pp. 7-149.4. 5. Problems with Polymorphism of Medicinal Compounds J. A. Clements Depavtmevzt of Pharmacy, Heviot- Watt Univevsity, 79 Gvassmavket, Edinburgh, EH1 2HJ The incidence of polymorphism amongst natural and synthetic materials is high and embraces a wide range of substances. Among groups of medicinal compounds that have been studied systematically, it is common to find over half of them to be polymorphic.The existence of metastable crystalline forms of pharmaceutical materials has led to difficulties in formulation (particularly of suspensions), in non-equivalence of biological activity and in analytical pro- cedures used in tests of chemical identity. Causes of Polymorphism Polymorphism can arise from1: differences in bond type, as occur in allotropes of carbon and tin; transitions between an ideal and a defect structure; changes in immediate bond co- ordinates; and changes involving minor differences in bond distances and angles but with no significant change in co-ordinates.This last type of change between forms can be either constructive, involving breakage of bonds, or displacive. Oxalic acid and resorcinol are examples of dimorphic compounds for which a transition between the a- and p-forms involves disruption of intermolecular hydrogen bonds and is a constructive change.2 Identification of Polymorphs For two materials to be referred to as polymorphic they must first be shown by suitable techniques (such as nuclear magnetic resonance spectroscopy or solution-phase infrared spectroscopy) to be chemically identical.Polymorphism is then detected by differences in a number of properties of the solid state. Those methods most widely applicable are: examina- tion of crystal properties ; melting-point measurement ; X-ray diffraction ; infrared spectroscopy; differential thermal analysis and differential scanning calorimetry; and solubility measure- ments. Crystal Properties Microscopic identification of polymorphism in samples of medicinal compounds is usually not possible because the material is usually in the form of crystallites or fine powder, and the crystal form or habit is variable.Refractive indices can be measured microscopically using the Becke test1 and this can serve to distinguish polymorphs. The refractive indices of the four crystalline forms of sulphanil- amide have been shown to differ from one another along the various crystal axes.22 DEVELOPMENTS IN PHARMACEUTICAL ANALYSIS PYOC.Analyt. Div. Chem. SOC. Melting-point Measurement Although melting-points of polymorphs usually differ, commercial samples may contain variable or non-quantifiable amounts of impurities that make the method unreliable in deter- mining the polymorphic f ~ r m .~ Impurities, including breakdown products, can form during crystallisation, or during heating in a melting-point apparatus, so that variable melting-point among samples is not necessarily due to polymorphism. This may, for example, be the cause of the apparent polymorphism of a ~ p i r i n . ~ A reversion of form can occur during heating so that the observed melting-point is then that of the newly formed polymorph.These changes can involve a liquid melt, but the reversion can occur entirely in the solid state. For example, on heating form XI11 of phenobarbitone it undergoes a change to form 111. At 150 O C , a liquid melt is seen to form but not at about 70 "C where a solid-state transition O C C U ~ S . ~ X-ray Diffraction X-ray diffraction performed on single crystals gives the distance (d) between planes along each axis and thus the dimensions of the unit cell.From symmetry considerations the crystal system can be determined. Callow and Kennard6 reported the crystal properties, including the unit cell dimensions, of polymorphs of cortisone acetate. From the volume of the unit cell and the number of mole- cules within it, the density of packing of form I can be shown to be greater than in the other forms and is consistent with the observation that this is the stable form.X-ray diffraction measurements have been reported for three polymorphs of ~ulphanilamide.~-~~ In each instance, intermolecular hydrogen bonding takes place between the amide nitrogen or the amine nitrogen and the sulphone group in an adjacent molecule of the sulphonamide (Fig.l), but bond lengths differ among the polymorphs. \ - . - \ , 3.06 Fig. 1. Molecular packing and proposed hydrogen- bond network viewed along an axis of a-sulphanilamide crystal. 9January, 1976 DEVELOPMENTS I N PHARMACEUTICAL ANALYSIS 23 Although powder X-ray diffraction methods provide less information than single-crystal studies, they serve to characterise polymorphic forms.This technique of powder diffracto- metryll is quicker and more convenient than the earlier powder photography method. The diff ractograms of polymorphic forms usually show several differences. However, differences in diffractograms of samples of a. monomorphic material can arise if the powder samples do not present a random selection of crystal faces from which Bragg reflections can occur.Reducing the particle size, or mixing with 2 per cent. of carbon black,12 are recommended in order to avoid any preferred orientation of crystals in the powder bed. Powder X-ray diffraction can also be used to estimate the composition of mixtures of polymorphs in a sample by making use of the peaks characteristic of each form.ll Infrared Spectroscopy The infrared spectra of polymorphs in the solid state usually show several small differences and infrared spectroscopy is a useful technique for the detection of polymorphism in a com- pound.It is applied quantitatively in the limit test for polymorph A in samples of Chloram- phenicol Palmitate Suspension BPC. The infrared spectra may identify the differences in intermolecular bonding between the forms, as the spectra are often altered at those frequencies that correspond to groups partici- pating in hydrogen bonding.Thus, the characteristic stretching frequencies due to hydroxy and amino groups are decreased when bonding to an adjacent electronegative centre takes place.13 The extent of the spectral shift is inversely related to the length of the hydrogen bond and this has been confirmed for three polymorphs of s~1phanilamide.l~ The N-H stretching frequencies (two peaks) due to the primary amino group in sulphamethoxydiazine are reported15 as follows: for the compound in solution, 3500 and 3400 cm-l; for polymorph I, 3458 and 3345 cm-l; and for polymorph 11, 3365 and 3275 cm-l. These results show that the spectral shift differs between polymorphs.Other spectral shifts due to participating groups such as those involving N-H bending and S-0 stretching can be observed but are less distinct than those involving N-H stretching. Polymorphism and Qualitative Analysis Whereas the crystallographer makes use of physical differences to detect polymorphism, these differences can be a source of confusion to the analyst.For example, if an unknown chemical is examined in the solid state and its infrared spectrum differs from that produced by an authentic sample of the material it is thought to be, this may be due to a difference in the crystalline form of the two samples. A series of papersf6-18 from the Laboratory of the Govern- ment Chemist has examined the extent of this problem for steroids, sulphonamides and bar- bit urat es. However, where solu- bility considerations preclude solution-phase infrared spectroscopy, the recommended pro- cedure is t o recrystallise the sample and the authentic specimen from a solvent that has been found consistently to produce a single polymorphic form.Solvents for this purpose are specified in some instances in the information accompanying an Authentic Specimen or British Chemical Reference Substance.This difficulty does not arise if samples are examined in solution. Form Reversions One polymorphic form, metastable with respect to another under a given set of conditions, will tend to revert to a more stable form. However, the rate of reversioninthesolidphasecan vary greatly, with half-lives measured in seconds, days or years.Most problems in the use of metastable polymorphs in pharmaceutical formulations arise from those of intermediate stability. Grinding a crystalline sample can promote a change of polymorphic form, and the suscepti- bility of a metastable form is greatest in crystals with a small lattice energy. These reversions can occur during sample preparation, for example for infrared spectroscopy.The changes arising from various periods of grinding of cc-naphthaleneacetamide with potassium iodide in the preparation of the disc give rise to the appearance of new absorption peaks that are con- sistent with a polymorphic change. Similar changes during the preparation of Nujol mulls from barbiturates have been demonstrated.lS In the preparation of samples of metastable24 DEVELOPMENTS IN PHARMACEUTICAL ANALYSIS PYOC.Analyt. Div. Cheutt. SOC. crystalline forms, the minimum amount of grinding consistent with satisfactory sample dispersion should be used. Changes of crystalline form proceed more rapidly in contact with a solvent than in the solid state, and presumably involve the solution phase.These reversions are often encountered during dissolution studies of a metastable form. Fig. 2 shows the change in concentration of meprobamate during dissolution of metastable form 11. The solution becomes supersaturated with respect to the stable form I, which crystallises out from solution. The solubilities of the two forms can be determined from the graph, and from measurements made at other tempera- tures a van’t Hoff graph of the logarithm of the solubility veysus the reciprocal of the absolute temperature can be plotted.From these data, the thermodynamic parameters of the change from metastable to 7 1- 6 - ? - : 5 r E stable polymorph can be elucidated.20 - I Tim e/m in Tiine/h Fig. 2. meprobamate in distilled water a t 26 oC.20 Dissolution of form 1 (A) and form 11 (B) of It has been suggested21 that the biological activity of a compound is related to its thermo- dynamic activity.The free energy difference between polymorphs of meprobamate (402 cal mol-l) is intermediate between that of the polymorphs of mefenamic acid (251 cal mol-l) and of chloramphenicol palmitate (744 cal mol-1). A difference in the biological activities of the polymorphs when administered as oral preparations has been demonstrated for chloramphenicol palmitate but not for mefenamic acid.Further, the period of stability of the metastable form of meprobamate in contact with water is greater than 24 h, which suggests that an increased biological activity of this crystalline form would be observed after oral administration.In using in vitvo dissolution behaviour to predict the in vivo behaviour of two polymorphs, use of an inappropriate solvent can lead to an erroneous prediction. For example, Ravin et aLZ2 have reported a 2-fold difference in dissolution rates of polymorphic forms of an antiviral compound when tested in aqueous alcohol, but no difference when simulated gastric juice was used.Change in particle size distribution in suspensions is a troublesome phenomenon in their formulation. Dramatic increases in crystal dimensions are observed, for example, with metastable polymorphs of cortisone acetate in aqueous ~uspension.~~ This effect is additional to Ostwald “ripening,” which occurs much more slowly. Satisfactory suspensions are only formed from the polymorph which is stable in contact with the suspension medium.Although certain surface-active materials such as lyophilic colloids are capable of decreasing the crystal- lisation rate, it is not usually possible to prepare suspensions of a metastable polymorph with an acceptable shelf-life by this means.January, 1976 DEVELOPMENTS I N PHARMACEUTICAL ANALYSIS 25 With one or two exceptions, the biological activity shown by a metastable polymorph has not been significantly greater than that of the stable form.Polymorphism is in general an undesirable property in a medicinal compound. However, because a metastable polymorph with enhanced biological activity could arise in commercial production of an established com- pound, it is necessary to draw up quality control specifications for the raw material so as to prevent its use as solid in the preparation of a formulated product.In many instances the melting-point requirement is sufficient. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. References Hartshorne, N. H., and Stuart, A., “Crystals and the Polarising Microscope,” Fourth Edition, Edward Evans, R.C., “An Introduction to Crystal Chemistry,” Cambridge University Press, Cambridge, 1966, Aguiar, A. J., Krc, J . , Kinkel, A. W., and Samyn, J . C., J . Pharm. Sci., 1967, 56, 847. Bundgaard, H., J . Pharm. Pharvnac., 1974, 26, 535. Clements, J. A., and Stanski, D., Can. J . Pharm. Sci., 1971, 6, 9. Callow, R. K., and Kennard, O., J . Phavm. Pharmac., 1961, 13, 723.Allkaume, M., and Decap, J., Acta CrystaZZogv., 1965, 18, 731. Allkaume, M., and Decap, J., Acta Crystallogr., 1965, 19, 934. O’Connor, B. H., and Maslen, E. N., Acta CrystaZZogr., 1965, 18, 363. O’Connell, A. M., and Maslen, E. N., Acta Crystallogr., 1967, 22, 134. Shell, J . W., J . Pharm. Sci., 1963, 52, 24. Christ, C. L., Barnes, R. B., and Williams, E. F., Aqzalyt. Chem., 1948, 20, 789.Rellamy, J., “The IR Spectra of Complex Molecules,” Methuen, London, 1964, pp. 99 and 252. Novak, A., Lascombe, J., and Joshen, J . L., J . Phys., Paris, SuppZ., 1966, 27, 381. Moustafa, M. A., Ebian, A. R., Khalil, S. A., and Motawi, M. M., J . Phavm. Phavvnac., 1971, 23, 869. Mesley, R. J., and Johnson, C. A., J . Pharm. Phavmac., 1965, 17, 329. Mesley, R. J., and Houghton, E.E., J . Pharm. Phavmac., 1967, 19, 295. Mesley, R. J., and Clements, R. L., J . Phavm. Pharmac., 1968, 20, 341. Cleverley, B., and Williams, P. P., Chemy Ind., 1959, 49. Clements, J. A., and Popli, S. D., Can. J . Pharm. Sci., 1973, 8, 88. Higuchi, T., J . Am. Phavm. Ass., Sci. E d n , 1958, 47, 657. Ravin, L. J., Shami, E. G., and Rattie, E., J . Pharm. Sci., 1970, 59, 1290.Collard, R. E., Pharm. J . , 1961, 186, 113. Arnold, London, 1970, p. 21. p. 366. Quality Control of Plastics Used in Pharmaceutical Packaging J. E. Pentelow Quality Contvol, The Boots Company Limited, ATottinghav?a, NG2 3AA Major Plastics Used The major plastics used in the manufacture of containers for injection solutions are high- and low-density polyethylene, polypropylene and plasticised poly(viny1 chloride).Additives With polyethylene no additives need be used. Polypropylene contains a tertiary carbon atom and therefore must have an antioxidant added to prevent degradation during manu- facture. On break- down of the polymer, hydrochloric acid is given off and it is necessary to use a hydrochloric acid absorbing material. Heavy metal salts, such as lead, cadmium and barium stearates, and organotin salts are commonly used as stabilisers.For pharmaceutical use, it is necessary to use calcium or calcium plus zinc salts as stabilisers; a 2% level of stabiliser is adequate. Further, poly(viny1 chloride) is rigid and in order to make it flexible about 20-30y0 of a plasticiser must be incorporated, e.g., di-2-ethylhexyl phthalate. A level of 0.05% of antioxidant is normal.Poly(viny1 chloride) is unstable during fabrication unless a stabiliser is present. The Fundamentals of Testing In general, pharmaceutical manufacturers carry out development trials over a minimum of 3 months on formulations produced under carefully controlled conditions and contained26 DEVELOPMENTS IN PHARMACEUTICAL ANALYSIS PYOC. Analyt.Div. Chcm. SOC. in packs similar in essential features to those envisaged for sale. This testing is expensive, and as improvements or supply difficulties require slight modifications to either the formula- tion or the pack, such tests are continually being repeated. The cost of such testing is high, but it is essential. Formulations are required to have a shelf-life commensurate with the nature of the product, and the only way to substantiate such a shelf-life is to store and test the formulation.The foundation on which marketing takes place rests on the assurance that formula and pack are the same as those tested. This is an important aspect in the control of pharmaceutical packaging, for in the event of a slight modification to the pack the conclusion drawn from the initial tests on shelf-life may be invalidated.Development Testing of Plastics Extraction and Toxicity Tests Plastics are known to contain a diversity of additives and this has led to fears that these additives may be extracted. Because of this, all pharmacopoeias publish requirements for extractability and toxicity. The bases of these tests are: certain biological tests on auto- claved extracts of plastics; certain chemical tests on autoclaved extracts of plastics ; also, about half of the pharmacopoeias include chemical tests on the plastic itself.This approach has the disadvantage that analysts faced with imported samples have no tests to carry out on those samples. The British Pharmacopoeia has recognised this drawback and laid down requirements for the final pack.Cooper,l in his comprehensive and valuable booklet “Plastics Containers for Pharmaceuticals-Testing and Control,” has brought together English translations of the appropriate parts of the pharmacopoeias of the world. Biological tests On reviewing these tests it can be seen that the most popular biological tests are the acute toxicity test, the pyrogeii test, the intracutaneous test and the haemolysis test.The British Pharmacopoeia is alone in requiring solutions to comply with a biological test, performed on a cat, in which an electrocardiograph tracing, the blood pressure and the rate and depth of respiration are used as parameters for measurement. The future use of tissue culture methods is being considered.Chemical tests Minor variations do occur between the pharmacopoeias in respect of the time of sterilisation and the ratio of the area of the plastic to the volume of water used in the extraction. Tests found in all pharmacopoeias are those for non-volatile matter, heavy metals and pH. A considerable variety of other tests are used and, in practice, tests are chosen with reference to knowledge of additives likely to be found in the plastic concerned.The conditions of extraction are essentially those of steam sterilisation. Other Tests are also of great importance. out. Stvess cvack tests The principal disadvantage of plastic is that it is more prone to cracking than glass or metal and such cracking will lead to non-sterility. Certain substances enhance cracking, surface- active agents and silicones being examples. It is vital to pay attention to packing and stacking as moderate stresses applied over prolonged periods of time are often the cause of cracking.Impact and print tests It is essential that the container should not easily crack on impact and that the print, which will, of course, include instructions, should not be easily removed.The extraction and toxicity tests must be put into perspective, because physical weaknesses Other important tests on the container are therefore carried Tests on containers for the above properties are given in BS 4839.2Januavy, 1976 DEVELOPMENTS I N PHARMACEUTICAL ANALYSIS 27 Quality Control of the Plastic Quality Control departments must consider how to test the plastic in routine production in order to ensure that it is the same as the plastic tested a t the development stage.Chemical analysis is part of the testing used to give assurances in this respect. Such testing has been reviewed by Lefaux3 and analytical methods are available in the books by Haslam et ~ l . , ~ which is particularly useful for infrared data, Crompton5 and van der Heide,6 which is parti- cularly useful for thin-layer chromatographic work.Polyethylene and Polypropylene As the only additives in these materials are antioxidants, the only other variable that needs to be considered is the difference in polymer chain length, which is related to relative molecular mass and is measured by means of the melt flow index (MFI). This index is described in BS 27827 and is the mass (g) passing through a defined orifice under closely controlled conditions in 10 min.Such a value will be related inversely to the relative mole- cular mass. It has a particular significance in that plastics of low MFI are less prone to cracking; however, as they are also less easy to mould it is necessary to reach a compromise with the moulder.Antioxidants are normally determined by extraction followed by thin-layer cliromato- graphic analysis. If a polyolefin that contains no antioxidant is used, then absence of antioxidant is a valuable assurance of quality. An infrared analysis is useful in confirming identity. Hence, the assurance of quality of the plastic is built on the use of certain extraction tests followed by general chemical and biological tests, infrared analysis, melt flow index and antioxidant analysis.Poly(viny1 chloride) Differences in chain length in poly(viny1 chloride) have not led to problems and assurance of quality depends on general extraction tests, infrared analysis, the quantitative and qualita- tive analysis of the stabiliser and plasticiser and a limit test for vinyl chloride monomer content, which can be determined by gas - liquid chromatography.Conclusion All of the testing described must be supported by close collaboration between the fabricator and the pharmaceutical manufacturer. Cleanliness and good housekeeping are an essential part of all pharmaceutical manufacture. The fabricator must clearly understand tliat no part of his manufacturing procedure must be changed without prior consultation.Further, he must not change his source of supply of any material, even if it is sold as an equivalent material. The tests now in use were first put forward 15 years ago and have been found to be adequate in most instances. Over those 15 years we have seen a proliferation o f these tests, or slight modifications of them, throughout the pharmacopoeias of the world.I t is to be hoped that the next few years will see a rationalisation of these tests. References 1. 2. RS 4839: Part I : 1972. 3. 4. 5. 6. 5 . BS 2782: 1970. Cooper, J., “Plastic Containers for Pharmaceuticals-‘Testing and Control,” World Health Organisa- Lefaux, R., Annls Pharm. Fv., 1972, 30, 673. Haslam, J., Willis, H. A., and Squirrell, D.C. M., “Identification and Analysis of Plastics,” Second Edition, Iliffe Publishers, London, 1972. Crompton, T. R., “Chcmical Analysis of Additives in Plastics,” Pergamon Press, Oxford, 1971. van der Heide, R. F., “The Safety for Health of Plastics Food Packaging Materials. tion, Geneva, 1974. Principlcs and Chemical Methods,” Kemink en Zoon, Utrecht, 1964.28 DEVELOPMENTS IN PHARMACEUTICAL ANALYSIS Proc.AnaZyt. Div. Chefiz. SOC. Control of Low-level Cross-contamination by Means of Thin - layer Chromatography J. D. Edmond Medicines Testing Labovatory, Phavmaceutical Society of Great Britain, 36 York Place, Edinbuvgh, EHl 3H U The analyst in the quality control laboratory of a pharmaceutical factory is familiar with his own factory, products and likely sources of contamination.However, the analyst examining the samples in the Medicines Testing Laboratory is dependent on the Inspector, from the Medicines Inspectorate of the Department of Health and Social Security, for this type of information. He is completely unfamiliar with the manufacturer, his product range and the surroundings from which the sample has been taken. An analyst can, given time, devise techniques for the identification of most, if not all, un- known components of a mixture.Such a procedure is time consuming and the policy at this laboratory is to identify the contaminant that the Inspector suspects might be present. If additional contaminants are present they are usually only classified according to the group, i.e., steroids, phenols, alkaloids, etc.; sometimes, however, these additional compounds can be readily identified, in which event added information is available to the Inspector when re- visiting the factory. The majority of suspected cross-contaminations investigated at the Medicines Testing Laboratory are resolved by means of thin-layer chromatography, most of the remainder being checked by microbiological methods.In theory, the analyst has simply to select a suitable system for the separation of the contaminant from the suspect material, locate the spots by means of a suitable detection agent and compare the spots with those of standards run at the same time, taking into account the spot size and intensity. Alternatively, these comparisons can be made by use of instrumental techniques, either on the plate or after elution from the adsorbent.In practice, the procedure is seldom as simple as this and, among other factors, the analyst has to consider: the sensitivity of the detection agent ; the similarity of solubility of the contam- inant with that of the active ingredient, or even excipient matter, of the product; and the complexity of the formulation and the possibility of losing the contaminant in multi-extraction procedures.Sensitivity of the Detection Agent It is fairly obvious that if contamination of a product has occurred then the more sensitive the detection agent the more likely the detection of the contamination. In Table I the minimum detectable levels of three different classes of compounds by four different techniques are summarised, from which it can be seen that a judicious choice of detection agent is neces- sary in order to detect a particular contaminant.TABLE I RELATIVE SENSITIVITY OF DIFFERENT DETECTION AGENTS Detection agent Minimum detectable amount/pg r-------iL------ -7 Ephedrine Mestranol Paracetamol Ultraviolet spectroscopy (254 nm) 10 10 0.1 10% H,SO, in ethanol 10 0.05 10 Ninh y drin 0.05 10 10 Potassium hexacyanoferrate (I 11) 10 1 0.1 Iron(lI1) chloride Similarity of Solubility Ideally the thin-layer chromatographic system should leave the contaminated material on or near the origin of the chromatogramand allow the contaminant to run faster so that it is available, in a reasonably pure state, for location and determination.Frequently this situation is difficult to achieve without agreat deal of effort, but often, by means of suitable solvent extrac- tion, the problem can be overcome.For instance, in checking for the presence of phenacetin and caffeine in aspirin tablets the simplest technique is to suspend the product in dilute sodium hydroxide solution and extract the contaminants with chloroform.Januavy, 1976 DEVELOPMENTS I N PHARMACEUTICAL ANALYSIS 29 Complexity of Formulation The extraction solvent should, if possible, extract only the contaminant and leave the suspect material behind.Again this situation may not be easy to achieve but, with ingenuity, the analyst can usually find an answer to a difficult extraction problem. For example, only a complex extraction procedure can be used to extract capsicum from an embrocation rub containing methyl salicylate.However, the methyl salicylate can be readily removed from the product by steam distillation, allowing the capsicum to be extracted from the resulting residue after hydrolysis of the ointment base. The following is an example of a typical cross- contamination investigation carried out at the Medicines Testing Laboratory.Quinidine (R, 0.40) was sought in ( a ) , Barbitone Tablets BPC, ( b ) , Effervescent Potassium Tablets BPC and (c), Phenobarbitone and Theobromine Tablets BPC. Powdered tablets were suspended in dilute sodium hydroxide solution prior to extraction with chloroform. The extract was concentrated just to dryness, the resulting residue was taken up in 1 ml of chloro- form and suitable aliquots of the solution were subjected to chromatography [adsorbent, silica gel F,,, (Merck) ; mobile phase, dichloromethane - diethyl ether - methanol - water (77 + 15 + 8 + 1.2)J.Detection was by means of ultraviolet spectroscopy (254 nm), the minimum detectable amount of quinidine being in the range 0.2-0.3 pg. The contamination level in preparations ( a ) , ( b ) and (c) was found to be 300, 30 and 30 p.p.m., respectively. These levels were con- firmed by ultraviolet spectrophotometry and fluorescence emission spectrophotometry and also by running chromatograms in benzene - diethyl ether - diethylamine (20 + 12 + 5) on silica gel. The analyst carrying out the cross-contamination investigations can frequently be completely misled by false positive reactions or by the presence of impurities naturally present in one of the active ingredients of the product. For this reason it is essential that the cross-contamina- tion should be confirmed either by a second thin-layer chromatographic system or by some additional instrumental technique after elution of the contaminant from the adsorbent. For example, it has been found that caffeine can be detected readily in chloroform extracts of Phenobarbitone and Theobromine Tablets RPC. The caffeine in this instance is not a contaminant but a natural impurity associated with theobromine. It is common to use certain types of detection sprays for particular classes of compounds, e.g., dilute potassium iodobismuthate solution (RP specification) for the location of alkaloids. It has been found that a positive reaction with this spray is not necessarily indicative of the presence of an alkaloid. For example, the use of chloroform to extract morphine from Squill Linctus, Opiate, BPC, also results in the extraction of anise oil, which contains up to 90 per cent. of anethole, an unsaturated phenolic ether; anethole gives a false positive reaction with dilute potassium iodobismuthate solution, presumably by addition of iodine to the double bond of the unsaturated side-chain in the anethole molecule. These two examples illustrate why the analyst must confirm his original findings, preferably by a second thin-layer chromatographic system, or by an alternative technique. In attempting to put into perspective the levels of contamination found in products investi- gated at the Medicines Testing Laboratory, there is little point in trying to relate the levels found to the “related substances tests” which are frequently used in pharmacopoeia1 mono- graphs because one is almost certainly dealing with spurious or “unrelated foreign substances.” The most striking way to set the contamination levels into perspective is to consider the two extremes shown in Table 11. TABLE I1 DOSE OF CONTAMINANT MEASURED AS A FRACTION OF THE DAILY DOSE I N A MAXIMUM DOSE OF CONTAMINATED MATERIAL Maximum Contaminant Equivalent Tablet Contamination daily in maximum of maximum Product mass/mg level, p.p.m. dosage, tablcts dosagc/pg dosage Paracetamol E thinyl- tablets 600 10 oestracliol tablets 60 100 8 48 1 5 30 8 x 10-630 OBITUARY YYOC. Annlyt. Div. Chem. SOC. If one considers patient A taking paracetamol tablets contaminated with 6 pg of ethinyl- oestradiol and patient B taking ethinyloestradiol tablets contaminated with 6 pg of para- cetamol, then the paracetamol tablets are the less heavily contaminated. This, however, is not the information in which the analyst in quality control is interested because, if one con- siders the maximum daily dosage needed to achieve the required therapeutic action, it can be readily seen that patient A would also be receiving 48 pg of ethinyloestradiol, which is equiva- lent to the maximum daily dosage. Patient B would be at no risk from the contaminant in the et hinyloestr adiol tablets . It may be poor manufacturing practice to produce tablets contaminated with relatively small amounts of relatively innocuous contaminants, but the consequences are not serious. It is, however, not only poor manufacturing practice to produce tablets with small amounts of highly potent contaminants, but also, in many instances, highly dangerous. In conclusion, the manufacturer should assess his own situation along with the possible hazards and act accordingly, in order to ensure that the product is as free from harmful con- tamination as is possible. Each manufacturer will have his own special problems and it is up to the manufacturer to overcome these difficulties. This is probably the crux of the whole problem of cross-contamination.