626 Analyst, July, 1979, Vol. 104, pp. 626-636 Determination of Vitamin D2 in Multivitamin Tablets by High-performance Liquid Chromatography* C hristeen M ackay Department of Chemistry, Loughborough University of Tec.hnology, Loughborough, Leicestershire, L E l l 3 T U J. Tillman7 Fisons Limited, Pharmaceutical Division, Research and .Development Laboratories, Bakewell Road, Lough- borough, Leicestershire, L E l l OQY and D. Thorburn Burns Department of Analytical Chemistry, The Queen’s Universi:ty of Belfast, Belfast, BT9 5AG, Northern Ireland A procedure is described for the determination of vitamin D, in multi- vitamin tablets. The vitamin is released from the tablets by enzymic digestion, which is followed by solvent extraction and chromatography on a microparticulate silica column.Other fat-soluble vitamins do not interfere with the assay. The problem of defining and quanti:fying vitamin D, in order to correlate results with biological potency is discussed. The procedure is suitable for stability studies. Keywords : Vitamin D, ; pre-vitamin D, ; multivitamin tablets ; high- performame liquid chromatography Vitamins D, (ergocalciferol) and D, are the most active of the antirachitic vitamins and are the forms most commonly used in pharmaceutical preparations. They are usually incorpora- ted in tablets in the form of starch-coated, gelatin-protected beadlets. The assay of these vitamins in multivitamin preparations is difficult for three main reasons: firstly, vitamin D is a heat-labile molecule, which, in solution, readily isomerises to its precursor, pre-vitamin D ; secondly, vitamin D is usually present in micro-amounts relative to other fat-soluble vitamins, such as vitamin A and vitamin E, which affects the assay; and thirdly, complex matrices contain various excijpients and antioxidants, which may also have an effect.Laborious clean-up procedures prior to determination are usually required, as many of the chemical methods of assay are non-specific.lS2 This difficulty of accurate measurement of vitamins D, and D, is illustrated by the variety of techniques examined and by the increasing number of papers being published on the subject. The original method for the determination of vitamin D is a biological assay, which evaluates the curative response to vitamin D when administered to rachitic rats., Physico- chemical methods have also been reported.Direct ultraviolet spectrophotometry is only applicable to relatively pure, concentrated preparations. The most commonly used chemical procedure involves reaction with antimony( 111) chloride, commonly known as Nield’s reagent.4 This method is based on the ability of antimony(II1) chloride to form a transitory complex with the vitamin D C,-C,, seco-sterol dieine system. However, any other seco-sterol or unsaturated system will also react (e.g., vitamin A). Hence, extensive purification stages, such as saponification, solvent extraction and clerm-up on several columns, may be required prior to the colorimetric determination of vitarnin D.596 Thin-layer chromatography has also been employed, usually as a clean-up procedure prior to colorimetric Sheppard et aZ., have reviewed the application of gas - liquid chromatography to the deter- mination of fat-soluble vitamins, including vitamin D.Cyclisation of vitamin D and pre- vitamin D was a problem at the high column temperatures necessary for their elution, and this problem is now usually overcome by isomerisation and derivatisation after separation from other fat-soluble vitamins.1°-12 Determinations can take 2 d to complete ; however, the method does distinguish between vitamins D,; and D,. High-performance liquid chromatography (HPLC) provides an alternative to derivatisation gas - liquid chromatography for the separation of labile, heat-sensitive compounds, such as * Presented at the Fourth SAC Conference, Birmingham, July 17th to 22nd, 1977.t To whom correspondence should be addressed.MACKAY, TILLMAN AND BURNS 627 vitamins D, and D,. Reversed-phase chromatography has been used in order toseparate the D vitamins from other fat-soluble vitamins.l3-15 The separation of vitamin D, from D, is possible, especially in the presence of silver nitrate in the mobile phase.16 In general, adsorption chromatography will not separate vitamin D, from vitamin D,, but it has been used for the separation of the D vitamins from other fat-soluble vitamins.17-19 Adsorption chromatography has the advantage of providing a better separation of pre-vitamin D, from vitamin D3,l*-l9 compared with reversed-phase systems.15 Independent collaborative studies comparing the AOAC chemical method with HPLC and rat bioassay20,21 and also the AOAC method with HPLC and gas - liquid chromatography22 have recently been reported.It was recommended that HPLC should be adopted as the first official action for D, in resins and oils,21 with the reservation that thermal isomerisation to the pre-vitamin could occur in sample pre-treatment. This aspect of vitamin D chemistry is important and it implies that any quantitative analytical procedure for the determination of vitamin D must be performed under conditions where the equilibrium ratio is not altered. This paper describes a reliable, quantitative chromatographic method for the determination of vitamin D and pre-vitamin D, and a sample pre-treatment procedure that takes cognisance of, and resolves, this problem of the thermal isomerisation of vitamin D,.Experimental Reagents and Stock Solutions All reagents are of analytical-reagent grade unless otherwise stated. Cyclohexane (laboratory grade). Isopro$yl alcohol. n-Hexane (spectroscopic grade). Trypsin with 85% lactose as diluent. 4-Hydroxybiphenyl. Vitamin D,, pure, crystalline. Dilute sodium hydroxide solution (BP). Phosphate bufler (pH 7). Obtained from BDH Chemicals Ltd. Obtained from BDH Chemicals Ltd. Obtained from Koch-Light Laboratories. A 5% m/V solution in water. Dissolve 8.28 g of sodium dihydrogen phosphate (NaH,PO,.H,O) and 19.88 g of disodium hydrogen phosphate (Na,HPO,) (anhydrous) in distilled water and dilute to 11. Dissolve 20 mg of 4-hydroxybiphenyl in 5 ml of ethyl alcohol, then dilute to 100 ml with cyclohexane.Dilute 10 ml of stock solution to 100 ml with cyclo- hexane. Dissolve 50 mg of vitamin D, in 5 ml of ethyl alcohol and dilute the solution to 100 ml with cyclohexane. Dilute 2-, 4-, 6-, 8- and 10-ml aliquots of the stock vitamin D, standard solution, together with 5 ml of stock internal standard solution, to 50 ml with cyclohexane. These solutions cover the range 20-100 p g ml-l of vitamin D,. Internal standard stock solution. Dilute internal standard solution. Vitamin D, standard stock solution. Working standard solutions. Apparatus The apparatus used consists of a liquid chromatograph equipped with a constant-flow pump (Waters Associates, Model M 6000), an untraviolet absorbance detector (Waters Associates, Model 440) and a syringe - loop injection system (Waters Associates, Model U6K). Chromatographic Conditions Column.Stationary phase. Mobile phase. Flow-rate. 0.8 ml min-l. Detector. Chart speed. 0.2 cm min-l. Stainless-steel, 250 x 4.6 mm i.d. Microparticulate silica, 10 pm (Partisil 10). Cyclohexane containing 1.25% V/V isopropyl alcohol. An ultraviolet detector operated at 254 nm, with sensitivity of 0.02 a.u.f.s. Procedure Calibration graph Allow sufficient time for the column to equilibrate with the solvent in order to obtain a628 MACKAY et d.: DETERMINATION OF VITAMIN D, IN MULTIVITAMIN Analyst, VOl. 104 steady base line (15-30min). Then inject 5-pl aliquots of the dilute working standard solutions. Calculate the peak-height ratio for th.e vitamin D, peak (retention time approxi- mately 20 min) relative to the internal standard peak (retention time approximately 24 min) and construct a calibration graph of peak height ratio against vitamin D, concen- tration (in pg ml-1).Procedure for tablets Sample preparation. Weigh 20 tablets and grind them to a fine powder in a pestle and mortar. Accurately weigh duplicate portions, each equivalent to 50 pg of vitamin D,, into 100-ml stoppered conical flasks. Digestion process. To the sample add 500 mg of trypsin and 50 ml of phosphate buffer and digest in a water-bath at 37 "C for 1 h, swirling occasionally to release the vitamins from the tablet matrix. Transfer the digestion mixture into a 250-ml separating funnel, washing it in with two 10-ml portions of water, then two 5-nd portions of ethyl alcohol.Finally, rinse with two 40-ml portions of hexane, collecting a'U of the extracts in the separating funnel. Shake the funnel vigorously for 1 min, allow the two phases to separate and run off the lower, aqueous phase into a second separating funnel. Extract the aqueous phase with a further three 40-ml portions of hexane, combining each hexane extract with the extract in the first separating funnel. Next wash the bulked hexane extracts with 20ml of distilled water containing 2 ml of dilute sodium hydroxide solution. A white gelatinous precipitate (from stearate excipients) might be fonned; this should be run off and rejected with this alkaline solution. Wash the hexane extract with two 20-ml volumes of distilled water (or until it is free from alkali), then dry it by shaking with about 5 g of anhydrous sodium sulphate.Filter the extract into a 250-ml round-bottomed flask and evaporate it under vacuum to a low volume in a rotary evaporator a t 37 "C. Transfer the residual liquid quantitatively into a 25-ml pear-shaped flask and continue the evaporation to dryness. Dissolve the residue in 2 ml of dilute internal standard solution. Use 5 pl of this extract for chromatography under the same conditions as in the calibration procedure and calculate the peak-height ratio of vitamin D, to internal standard. From the calibration graph, determine the concentration of vitamin D, in the sample solution. Calculation Extraction. M Micrograms of vitamin D, per tablet = X x V x - m where X is the concentration of vitamin D, in the extract (pg ml-l), V the volume of extract (normally 2 ml), M the mean mass of the tablet (g) and m the amount taken (g).Investigation of Experimental Variables Mobile phase Laboratory-grade cyclohexane was chosen as the solvent as it is inexpensive and readily available and was shown to have negligible absorlbance at 254 nm. Cyclohexane itself offers no de-activation of the silica and hence the vitamin D is strongly absorbed on to the column, leading to long retention times and broad peaks. The addition of small volumes of iso- propyl alcohol to the mobile phase produces dramatic de-activation of the silica, and by increasing the concentration of isopropyl alcohol the retention of vitamin D is reduced (Fig. 1) and sharp peaks are obtained. However, if excessive amounts of isopropyl alcohol are added then column efficiency is reduced and separation of vitamin D from vitamins A and E is not achieved.It is recommended that the concentration of isopropyl alcohol necessary to produce sharp peaks and adequate separation from interfering vitamins be determined for each column. With the column used in this work 1.25% V/V was the optimum con- centration of isopropyl alcohol in the cyclohexane that was necessary to give a reasonably sharp vitamin D, peak and adequate separation from pre-vitamin D, and from vitamins A and E. Vitamins A acetate and E acetate were eluted at about 8min, pre-vitamin D, at about 15 min and vitamin D, at about 20 min. A further advantage of this system is that vitamin A alcohol (a possible degradation product of vitamin A acetate) is eluted after vitamin D,.Jab, 1979 TABLETS BY HIGH-PERFORMANCE LIQUID CHROMATOGRAPHY 60 50 C .- 5 4 0 - - 30- .- E 0 .- + : 20- c.' U 10 , - - 0 629 0 0.5 1 .o 1.5 2.0 Concentration of isopropyl alcohol in cyclohexane, % V/V Fig.1. Effect of isopropyl alcohol concentrations on retention time of vitamin D,. Choice of internal standard For quantitative work, using syringe injection, an internal standard is desirable. A suitable compound was found to be 4-hydroxybiphenyl, which has a retention time of about 25 min. A base-line separation from vitamin D, was obtained and there was also no inter- ference from the pre-vitamin or vitamin A alcohol. A typical chromatogram is shown in Fig. 2. 4 5 1 L - Time/min Fig. 2. Chromatogram of vitamin mixture in tablet pro- portions.1, Vitamin A acetate (17.2pg) andvitamin E (250pg); 2, unknown; 3, pre-vitamin D,; 4, vitamin D, (0.125 pg); 5, 4- hydroxybiphenyl; and 6, vit- amin A alcohol. Conditions as in Procedure.630 Linearity of response working range 0-0.5 pg of vitamin D,. Reproducibility The two main factors that affect precision in quantitative liquid chromatography are reproducibility of the peak height and reproducibility of the retention time. A series of eight replicate 5-p1 injections of a standard solution containing 40 pg ml-l of vitamin D, gave a mean peak-height ratio of 0.525 with a coefficient of variation of 0.93%. A greater problem with de-activated silica columns is the: reproducibility of retention time from day to day.The absolute retention times of the pre-vitamin D,, vitamin D, and the internal standard varied, but the relative retention times remained constant. MACKAY et ai. : DETERMINATION OF VITAMIN D, I N MULTIVITAMIN Analyst, voi. 104 The measurement of peak-height ratios gave (consistent linear calibration graphs over the Digestion and extraction conditions Many procedures involve digestion of the vitamin from the beadlets with water or dimethyl sulphoxide17 at 60-70 "C or saponification with alcoholic potassium hydroxide solution20 at a similar temperature. It was considered essential to avoid the risk of formation of pre- vitamin D, at these elevated temperatures and an alternative procedure was sought. Digestion with trypsin in a phosphate buffer at 37 "C for 1 h was shown to be a most effective procedure for breaking up the gelatin beadlets.This procedure also avoided the formation of vitamin A alcohol and hence reduced the analysis time. Hexane was chosen as the solvent with which to extract the fat-soluble vitamins from the aqueous digest as it gave minimum problems with emulsion formation. However, four extractions were required in order to obtain a quantitative recovery from aqueous so'lutions, as shown. Volume of hexane Recovery of vitamin D,, yo 1 x 100ml + 1 x 50ml 82 1 x 100ml + 2 x 50ml 92 1 x 8 0 m l + 3 x 40ml 96 Efect of heating vitamin D, solutions at 37 "C In solution, vitamin D, forms an equilibrium mixture with the pre-vitamin and the ratio, at equilibrium, is solely temperature dependent. The proportion of pre-vitamin increases with increasing temperat~re,,~ hence the need to :maintain low-temperature reaction conditions in order to minimise the formation of pre-vitamin during the analytical procedure. The total time at 37 "C for the analytical procedure described is of the order of 1+2 h.A standard solution of vitamin D,, with internal standard, was therefore maintained at 37 "C and 5-pl injections made after various periods of time. There was no appreciable change in the peak-height ratio over a period of at least 4 h and no increase in the small (not greater than 5%) pre-vitamin peak in the initial sample. After 3 d at 37 "C the pre-vitamin peak had increased and the change in peak-height ratio of vitamin D, indicated a loss of 10%. These results gave confidence that no additional pre-vitamin would be formed during the analytical procedure.Recovery from gelatin beadlets Recovery experiments, using the recommended procedure, were then performed on the vitamin D, beadlet concentrate and in the presence of tablet excipients, including the theoretical proportion of other vitamins. Vitamin D, was first assayed in the beadlets by use of ultraviolet absorption (due allowance being made for the presence of the antioxidants butylated hydroxyanisole and butylated hydroxytoluene) . The sample size taken for the ultraviolet assay was 100 mg, in an attempt to obtain as precise a result as possible, against which to compare the HPLC results. Ten 2-mg samples were taken for the HPLC assay, which corresponded to the assay requirement of about 50pg of vitamin D,.The results are shown in Table I. The slightly low recovery in the presence of tablet components was believed to be due to the slight emulsion problems that were encountered because of the bulk of the insoluble matter present ; however, it was considered to be acceptable for a micro-assay of this nature.J d y , 1979 TABLETS BY HIGH-PERFORMANCE LIQUID CHROMATOGRAPHY TABLE I 63 Z RECOVERY OF VITAMIN D, FROM BEADLET CONCENTRATE HPLC assay r A \ Ultraviolet Beadlets + tablet assay Beadlets components Mean vitamin D,/mg g-l . . .. .. . . 26.5 26.7 25.2 Recovery with respect to ultraviolet assay, % . . 100.5 95.1 Relative standard deviation, % . . .. . . 0.8 2.4 3.5 A fialysis of multivitamirt tablets A sample of 20 tablets containing theoretically 500 I.U.of vitamin D, per tablet was ground up and four replicate analyses were performed. A mean result of 513 I.U. per tablet was obtained, with a range of 478-532 I.U. per tablet. A typical chromatogram is illustrated in Fig. 3. It will be observed that there is some interference with the pre-vitamin D, peak from degradation products of vitamin A or vitamin E. In this instance it would be necessary to determine the pre-vitamin content by reference to the vitamin D beadlet concentrate as described later under Discussion of Potency Measurement. 1 3 - Tirne/min Fig. 3. Chromatogram of multivitamin tablet. 1, Vitamins A acetate and E; 2, vitamin D,; 3, internal standard; and X, retention of pre-vitamin D,. Conditions as in Pro- cedure. Relative detector Yesfionse for vitamin D, and $re-vitamin D, In order to quantify any pre-vitamin that may be present in beadlets or concentrate, the detector response to the pre-vitamin needs to be determined, as this has a much lower absorbance at 254 nm than vitamin D, itself.In the absence of a pure sample of pre-vitamin the detector response relative to vitamin D, was determined in the following manner. A 100-ml volume of a cyclohexane solution containing 10 mg of pure vitamin D, and 2.5 mg632 MACKAY et al. DETERMINATION OF VITAMIN D, I N MULTIVITAMIN Analyst, VOl. 104 of internal standard was isomerised by refluxing at 80 "C for 2 h. factor (R,) for the pre-vitamin is given by The relative response RP (isom.) R - - RD (init.) - RD (isom.) where RP (isom.) = peak-height ratio of pre-vitamin to internal standard in isomerised solution; RD (init.) = peak-height ratio of vitamin D, to internal standard in initial solution; and RD (isom.) = peak-height ratio of vitamin D, to internal standard in isomerised solution.Reproducible, but different, values were obtained for two different detectors: 0.797 for detector A (Varian) and 0.65 for detector B (Waters Associates, Model 440). This variation in response between detectors is contrary to the findings of Hofsass et aL21 and emphasises the need to calibrate each detector. Stability Aspects In order to determine whether or not the procedure would be suitable for stability studies of formulations, various samples of pure crystalline vitamin D, were subjected to exaggerated conditions designed to cause degradation.Efect of ultraviolet light A solution of vitamin D, (0.1 mg ml-1) in cyclohexane and pure solid vitamin D, were sealed in glass ampoules and placed in the sample compartment of a xenon-arc fade tester (Xenotest). The solution and solid had discoloured after 5 d. The solution showed a loss of 86% of the vitamin while the solid assayed at 53%. Efect of heat (on solid material) Assay after 1 week gave the following results: Pure crystalline vitamin D, was heated at 60 and 100 "C in sealed and unsealed ampoules. Vitamin remaining, yo Temperature/"C Sealed ampoule Unsealed ampoule 60 41 24 100 5 4 For samples stored at both temperatures there was considerable darkening and the solid had partly melted to form a resin-like mass.A11 samples showed the presence of pre-vitamin as well as other more polar degradation peaks. Typical chromatograms are shown in Fig. 4. Saponijcation A sample of vitamin D, concentrate in beadlet form was saponified by refluxing on a water-bath, using 5 ml of 50% m/V aqueous potassium hydroxide solution in 70 ml of 70% V/V aqueous ethanol in the presence of sodium ascorbate as antioxidant. A considerable amount of pre-vitamin (about 15%) is formed during this procedure (Fig. 5). Stability of vitamin D, in solution at elevated tem$eratuures Solutions of pure vitamin D, and internal standard were retained at various temperatures for different periods of time in order to determine the rate of formation of pre-vitamin: solutions in cyclohexane were stored at 20 and 4.0 "C; a solution in cyclohexane was refluxed (at 80 "C); and a solution in toluene was refluxed (at 102 "C).In all instances the only degradation peak to be produced was that due to the pre-vitamin (see Fig. 6 for an example) and an equilibrium was attained. It was therefore possible to determine the equilibrium ratio of vitamin to pre-vitamin and the time taken to reach equilibrium. These results are summarised in Table I1 and compared with the figures calculated by Keverling Buisman et aL2, for vitamin D,. Previous workers' have indicated that equilibrium and isomerisation rates of vitamin D, and D, are similar, as is confirmed by the above results.July, 1979 TABLETS BY HIGH-PERFORMANCE LIQUID CHROMATOGRAPHY 2 I - Time/min - Time/min Fig. 4. Chromatograms showing degradation of vitamin D,: (a) 60 OC and (b) 100 "C, after 1 week in unsealed ampoules.1, Pre-vitamin D,; 2, vitamin D,; and 3, internal standard. Other peaks unidentified. Discussion of Potency Measurement - Time/min Fig. 5. Chromatogram of vitamin D, concentrate in beadlet form after saponifica- tion. 1, Pre-vitamin D,; and 2, vitamin D,. 633 Attempts to correlate bioassay results with physico-chemical assay results have led to confusion in defining the vitamin D content of a product. The biological assay measures I Time/min - Fig. 6. Chromatograms showing the rate of formation of pre-vitamin D, from solutions in cyclohexane of pure vitamin D,, refluxed at 80 "C for different periods of time: (a), initial; (b) after 15 min; and (c) after 2 h. 1, Pre-vitamin D,; 2, vitamin D,; and 3, internal standard.634 MACKAY et al.: DETERMINATION OF VITAMIN D, I N MULTIVITAMIN Analyst, VOl.104 TABLE I1 RATE OF FORMATION OF PRE-VITAMIN D, AT DIFFERENT TEMPERATURES Temperature/"C A r \ 20 410 80 ---+-- +- - Theoreti- Theoreti- Theoreti- cal* HPLC cal HPLC cal HPLC pre-vitamin D, . . 93:7 95:5 89: 11 92:s 78:22 78:22 Ratio vitamin D, : Time to reach equilibrium . . .. 30d 14d 3.5d 2 d 2.4 h 2 h * According to Keverling Buisman et aLe3 102 - Theoreti- cal HPLC 7 2 ~ 2 8 74:26 30min 10min the vitamin D in terms of units of potency, the International Unit (I.U.), which was defined by the World Health Organization as being equivalent to 0.025 pg of pure, crystalline vitamin D, or D,; this is the amount of material required to produce a minimum prophylactic response in rachitic rats or chicks.The major problem in all assay procedures arises from the fact that vitamin D readily isomerises in solution to form the pre-vitamin. The rate of isomerisation and the equi- librium ratio appear to be solely temperature dependent and unaffected by solvent, light or catalysts. 24 Vitamin D2 Pre-vitamin D2 Pre-vitamin D has about 3540% of the biological activity of vitamin D itself. This corresponds to the percentage formation of vitamin D from the pre-vitamin at 40 "C in 9 h, as deduced from kinetic studies on the pre-vitamin - vitamin isomerisation. It can therefore be concluded that the apparent biological activity of the pre-vitamin is a function of its in vivo conversion to vitamin D at body temper(ature.25 Thus, any assay method that measures actual vitamin D only, and that causes conversion to the pre-vitamin during the procedure, will yield falsely low results.In order to over- come this problem, Keverling Buisman et aL2, and Mulder et aLZ4 have defined two terms: Actual vitamin D Potential vitamin D = vitamin D + pre-vitamin D = vitanlin D only It was shown that in many analytical procedures the actual vitamin D content could vary owing to formation of the pre-vitamin, but that the potential vitamin D content remained unaltered. It has therefore been suggested that the potential vitamin D content is the only usable measure of vitamin D activity. Under specified conditions colorimetric methods, thin-layer chromatography, gas - liquid Chromatography and HPLC procedures are all capable of yielding the potential vitamin D content of a preparation.Three approaches have been suggested.27July, 1979 TABLETS BY HIGH-PERFORMANCE LIQUID CHROMATOGRAPHY 635 Purification without separating vitamin and pre-vitamin and determination by a reaction that gives the same response to both compounds (e.g., Nield's reagent). Purification and separation of the vitamin and pre-vitamin followed by (a), analysing the fractions separately and adding the results, or ( b ) , re-uniting the fractions and analysing them together (e.g., thin-layer chromatography followed by extraction and Nield reaction on the separate or combined fractions). 3. Comparison of a measurable property of a purified sample solution and a standard solution after equilibration of both under identical conditions. These approaches would give no indication of the pre-vitamin content of the original sample because the vitamin and pre-vitamin are either measured in total or the ratio is altered during the analytical procedure.As the pre-vitamin is less active than the actual vitamin a false indication of the potency of a preparation could be obtained. These approaches would only be valid for preparations containing no pre-vitamin initially. The only universally valid approach to obtaining good correlation with biological potency is to quantify the vitamin and pre-vitamin separately and not to alter the proportion of each during the process. This implies that saponification or digestion at elevated temperatures prior to measurement would render a method invalid for preparations that contain an appreciable amount of pre-vitamin.The concept of potency measurement has been extended by Tartivita et aZ.18 and Vanhaelen- Fastre and Vanhaelen19 to make allowance for the lower biological activity of the pre- vitamin by inclusion of a bioactivity factor. Vitamin D potency = actual vitamin D content + BF x pre-vitamin content where BF is a factor to allow for the bioactivity of pre-vitamin D relative to actual vitamin D. Based on the work of Hanewald et ~ 1 . ~ ~ this can be taken to be 0.35, but it is obviously empirical and difficult to measure precisely and there may be a species variation. In the work reported in this paper, sample pre-treatment is carried out at less than 40 "C and it has been shown that no isomerisation of vitamin D occurs during the analysis time.Isomerisation of vitamin D in the solid state does not occur and is unlikely in solid multi- vitamin formulations. Thus, if the proportion of pre-vitamin present in the beadlets formed as a result of the manufacturing process is measured, then by the determination of the actual vitamin content of the tablet, allowance can be made for the pre-vitamin content and a valid determination of the vitamin D potency of the formulation obtained. With solution formulations of vitamin D, where isomerisation is a possibility, it would be necessary to quantify both the vitamin D and the pre-vitamin peaks separately. Conclusions In order to obtain good correlation between chemical assays and biological potency it is essential that actual vitamin D and pre-vitamin D should be measured separately and that the ratio should not be altered during the analytical procedure; allowance can then be made for the lower activity of the pre-vitamin.The proposed procedure for the determination of vitamin D in multivitamin tablets has several advantages and has proved to be satisfactory in routine use. Saponification is avoided by means of an enzymic digestion, and solvent extraction is carried out at low temperatures. This procedure eliminates the risk of pre-vitamin formation during the assay, as would ultrasonic vibration.19 The formation of the pre-vitamin is unlikely in tablet formulations at ambient temperature, so that the actual vitamin D content is also equal to the vitamin D potency (provided that pre-vitamin is excluded or limited in the beadlet concentrate incorporated in the tablet).For liquid formulations, where isomerisation is a possibility, it is also necessary to quantify the pre-vitamin content. The chromatographic conditions are such that other fat-soluble vitamins do not interfere ; this avoids laborious column clean-up procedures and allows at least four assays per day to be completed. The procedure was developed using vitamin D,, but should be equally applicable to preparations containing vitamin D, as their chromatographic and thermal properties are almost identical. The method is also suitable for stability studies providing there is no interference from degradation products of other fat-soluble vitamins.The authors thank Dr. A. G. Fogg, Loughborough University, for kind liaison and helpful discussions during the course of this work. 1. 2. Thus, the following term can be defined636 MACKAY, TILLMAN AND BURNS References Freed, M., “Methods of Vitamin Analysis,’’ Interscience, New York, 1966. Gyorgy, P., and Pearson, W. N., “The Vitamins,” Academic Press, New York, 1967. “Official Methods of Analysis of the Association of Official Analytical Chemists,” Twelfth Nield, C. H., Russell, W. C., and Zimmerli, A., J . BioZ. Chem., 1940, 136, 73. “United States Pharmacopeia XIX,” USP Convention Inc., Rockville, Md., 1975, p. 633. Mulder, F. J., and de Vries, E., J . Ass. Ofi. Analyt. Chem.. 1974, 57, 1349. Hanewald, K. H., Mulder, F. J., and Keuning, K. J., J . Pharm. Sci., 1968, 57, 1308. Johnson, G. W., and Vickers, C., Analyst, 1973, 98, 257. Sheppard, A. J., Prosser, A. R., and Hubbard, W. D., J . Am. Oil. Chem. Soc., 1972, 49, 619. Murray, T. K., Erdody, P., and Panalaks, T., J . Ass. 08. Analyt. Chem., 1968, 51, 839. Bell, J. G., and Christie, A. A., Analyst, 1973, 98, 268. Bell, J. G., and Christie, A. A., Analyst, 1974, 9’9, 385. Williams, R. C., Schmit, J. A., and Henry, R. A., J . Chromat. Sci., 1972, 10, 494. Osadca, M., and Araujo, M., J . Ass. OH. Analyt. Chem., 1977, 60, 993. Wiggins, R. A., Chemy Ind., 1977, 841. Tscherne, R. J., and Capitano, G., J . Chromat., 1977, 136, 337. Tompkins, D. F., and Tscherne, R. J., Analyt. (?hem., 1974, 46, 1602. Tartivita, K. A., Sciarello, J. P., and Rudy, B. C., J . Pharm. Sci., 1976, 65, 1024. Vanhaelen-Fastre, R., and Vanhaelen, M., J . Chromat., 1978, 153, 219. Hofsass, H., Grant, A., Alicino, N. J., and Greenbaum, S. B., J . Ass. 08. Analyt. Chem., 1976, 59, Hofsass, H., Alicino, N. J., Hirsch, A. L., Ameilta, L., and Smith, L. D., J . Ass. OH. Analyt. Chem. Borsje, B., Craenen, H. A. H., Esser, R. J. E., Milder, F. J., and de Vries, E. J., J . Ass. Ofl. Analyt. Keverling Buisman, J. A., Hanewald, K. H., Mulder, F. J., Roborgh, J. R., and Keuning, K. J., Mulder, F. J., de Vries, E. J., and Borsje, B., J . Ass. Off. Analyt. Chem., 1971, 54, 1168. Hanewald, K. H., Rappoldt, M. P., and Roborgh, J. R., Recl Trav. Chim. Pays-Bas Belg., 1961, Edition, Association of Official Analytical Chemists, Washington, D.C. , 1975, Section 43.166. 251. 1978, 61, 735. Chem., 1978, 61, 122. J . Pharm. Sci., 1968, 57, 1326. 80, 1003. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. Received December 13th, 1978 Accepted February 7th, 1979