192 Awalyst, March, 1967, Vol. 92, $9. 192-195 The Determination of Warfarin in Animal Relicta BY F. B. FISHWICK AND A. TAYLOR (Infestation Control Laboratory, Ministry of Agriculture, Fisheries and Food, Tolworth, Surrey) A method is described for the quantitative analysis of warfarin in animal relicta. Basically the method consists in the extraction of relicta, column clean-up of the extract, separation of the warfarin by thin-layer chromato- graphy, elution from the layer and spectrophotometric assay at 305 mp. WARFARIN, the recommended British Standard name for 3- (acetonylbenzyl)-4-hydroxy- coumarin, has been in use both as a rodenticide and for human therapy for many years. The need for a sensitive and specific method for detecting warfarin in biological material has long been apparent.The assay of warfarin present in rodenticidal baits was first reported by La Clair,l but in this and subsequent papers2 9 3 9 4 no general method was completely satisfactory. Garner5 y 6 has described a method for detecting the presence of warfarin in post-mortem material, but like Pyorala7 we have been unable to apply this quantitatively. Similarly, the method suggested by Wanntorps has been found reliable only with fresh relicta, otherwise high blank values are encountered, and these cannot be allowed for by the method described. O’Reilly, Aggeler, Hoag and Leongg have described a method for determining warfarin in human plasma but this has not been found suitable for low levels of warfarin owing to difficulties with emulsions. No sensitive colour reactions specific for warfarin appear to be known.The 2,4-dinitrophenylhydrazone can be prepared from warfarin, but once a fairly pure tissue extract has been prepared there is no advantage over a direct spectrophotometric determination. Investigation of the reaction between the carbonyl group and m-dinitro- benzene were not pursued as the colour produced was not significantly different from the reagent blanks. The oxidation of warfarin with alkaline potassium permanganate gives a carboxylic acid, 3-( cc-carboxybenzyl)-4-hydroxycoumarin, but in such poor yield as to be of low value analytically. The gas - liquid chromatography of warfarin has been describedlO but the method is rather insensitive. EXPERIMENTAL It was considered that the best results would be obtained from a method dependent on the isolation of warfarin followed by spectrophotometric measurement, rather than on a more chemical approach.Various ways of isolating warfarin from biological tissues by using ion-exchange resins, paper electrophoresis and column and thin-layer chromatography have been investigated. ION-EXCHANGE RESINS- Warfarin is taken up by De Acidite FF in the OH-, C1-, HC0,- and HS03- forms; De Acidite G in the OH- and C1- forms; and De Acidite E in the OH- form. It was found to be extremely difficult to elute warfarin from any of these, except De Acidite FF, in which case it was found that boiling with a 1 per cent. solution of sodium hydroxide was effective. Elution with methanolic ammonium chloride, as described in British Patent specification 881,855, gave variable recoveries and introduced the additional problem of isolating the warfarin from a large volume of eluate.PAPER ELECTROPHORESIS- Warfarin will move 5 cm in 4 hours in a borate buffer of pH 9 with a potential gradient of 10 volts per cm with Whatman 3MM paper. It is then possible, by viewing the paper under ultraviolet light of 254mp, to cut out the band containing the warfarin. This paper strip is then pulped and the warfarin eluted with a 1 per cent. sodium pyrophosphate solution. After acidification with 5 N hydrochloric acid and extraction with chloroform, the compound is recovered by removing the solvent on a steam-bath, and after dissolution in isopropyl alcohol containing 1 per cent. of acetic acid the extinction is determined at 305 mp.This solvent,FISHWICK AND TAYLOR 193 first suggested by Armstrong,ll has been found to be more specific and subject to less inter- ference than either sodium hydroxide or sodium pyrophosphate solutions. One of the difficul- ties of this method is in deciding on the width of the band to be removed from the electro- phoretogram; the wider the band the more interfering substances are removed; the narrower the band the lower the recovery of warfarin. Recoveries from pig's liver have been in the range of 65 to 85 per cent., and subject to wide variation. COLUMN CHROMATOGRAPHY- Florisil showed a batch-to-batch variation even when the pre-treatment was standardised, and alumina gave consistently low recoveries. Untreated silica gel gave little clean-up, but it was found that by activating the gel at 120" C for 16 hours and subsequently adding 15 per cent. w/w of water, a product was obtained that gave sufficient clean-up to avoid emulsion formation at the later stages of the analysis without reducing the recovery.A further reduction in the proportion of total solids in the extract is brought about by extraction of the eluate from the column with a 1 per cent. sodium pyrophosphate solution, which is then acidified and the warfarin transferred into chloroform. This solution is suitable for the separation of warfarin by thin-layer chromatography on kieselgel plates. Recovery experiments of pure warfarin from thin-layer plates are given in Table I. Several absorbents for column clean-up of extracted material have been tried.TABLE I RECOVERY OF WARFARIN FROM THIN-LAYER PLATES Warfarin added, Warfarin recovered, Mean recovery, 126 117 92.6 86.6 82-3 81.4 94.5 56.3 49.1 48.5 86.9 30.6 26.2 27-2 87.4 18.1 14.9 17.8 90.1 Pg PQ per cent. METHOD The method consists in the extraction of relicta by ether, column clean-up of the extract, purification of the eluate by chloroform partition, separation of the warfarin by thin-layer chromatography, elution from the layer and spectrophotometric assay at 305 mp. APPARATUS- M.S.E. Homogeniser-Catalogue No. 7700. Quick$ and Quartz liquid - liquid extraction unit-Catalogue No. EX9/33. This is fitted Glass tubes f o r chromatography-Aimer. Catalogue No. AGG/type F of l-cm diameter Spectrophotometev-Hilger Uvispek, type 700H. Ultraviolet light source-Hanovia Chromatolite, 254 mp.Thin-layer spreading apparatus-Suitable for the preparation of thin-layer plates Pasteur long- form pipettes. All materials should be of analytical-reagent grade. Ethyl acetate. Diethyl ether, peroxide free-Diethyl ether is washed with concentrated iron(I1) sulphate solution, followed by 0-5 per cent. potassium permanganate solution and 2 N sodium hydroxide solution. I t is then dried over calcium chloride and redistilled. This is fitted with a 100-ml beaker. with sintered-glass distributor EX9/30S. and 30cm long. (10 x 20 cm) with a layer thickness about 250 p. REAGENTS- Isopropyl alcohol containing 1 per cent. of acetic acid. Hydrochloric acid, 5 N. Sodium Pyrophosphate solution, 1 per cent. in water. Celite 545-Obtainable from Johns-Manville Co.Ltd. Silica gel-Hopkins and Williams, M.F.C. grade. Prepared by heating at 120" C for 16 hours, then adding 15 per cent. w/w of water and shaking the mixture for 3 hours.194 FISHWICK AND TAYLOR DETERMINATION [Analyst, Vol. 92 Sodium sdphate, anhydrous. Chloroform. Kieselguhr GF254-Available from E. Merck, Darmstadt. Mobile solvent for thin-layer chromatography-Diethylether - hexane - acetic acid, 75 + 25 + 1 (v/v). PROCEDURE- Macerate 6 g of tissue with 30 ml of water and transfer the suspension with washing to the liquid-liquid extraction unit. Add 5 ml of 5 N hydrochloric acid and extract the mixture with ether for 45 hours. Fill a chromatographic column with ether and add 1 g of Celite 545. Allow the mixture to settle on the sinter and pour a suspension of 9 g of the prepared silica gel in the minimum volume of ether on to the top of the column and assist the gel to settle down by tapping the column.Cover the silica gel layer with a 3-cm layer of anhydrous sodium sulphate and allow the ether to drain to the upper surface of the column packing. Transfer the ether extract, reduced by distillation to not more than 5 ml, to the top of the prepared column and allow it to percolate into the column packing. Wash the distillation flask thoroughly with successive 10-ml volumes of ether, combine the washings and add them to the column; wash the column with sufficient ether to bring the total volume to 150ml. Transfer the effluent to a separating funnel and extract with 15-ml and 10-ml volumes of sodium pyrophosphate solution.Combine the alkaline extracts acidified with 3 ml of 5 N hydrochloric acid and extract with two 10-ml aliquots of chloroform. Combine these extracts and evaporate them on a water-bath to less than 1 ml. Prepare thin-layer chromatographic plates by suspending 30 g of Kieselguhr GF254 in 63 ml of water, shake the mixture vigorously for 30 seconds and apply to 10 x 20-cm plates in a conventional way to give a layer about 250-p thick. Allow the layer to set and dry in an oven at 110" C for 30 minutes. Apply the concentrated chloroform extract to the layer as a streak (great care is needed at this stage to avoid losses). Develop the plate in the mobile solvent to 12 cm from the origin in a lined and sealed tank at ambient temperature.Allow it to dry, and locate the warfarin under ultraviolet light of wavelength 254 mp; the RF value is about 0.46. Firmly plug a Pasteur pipette with a small piece of cotton-wool and rinse it with ethyl acetate. Transfer the warfarin band to the pipette and elute the warfarin with not less than 8 ml or more than 10 ml of ethyl acetate. Collect the eluate in a test-tube and evaporate to dryness over a water-bath. Dissolve the residue in 5 ml of isopropyl alcohol - acetic acid mixture and read this solution on the spectrophotometer at 305 mp in a l-cm cell against a solvent blank. The EiZ; value for warfarin in this solvent at a wavelength of 305 mp is 361. RESULTS In preliminary investigations with the liver from several animal species the interfering substances from pig's liver have caused more trouble than any other.As most of our field samples have involved this animal, it was selected for determining recovery values. Blank TABLE I1 RECOVERY OF WARFARIN FROM LIVER Warfarin added, Warfarin recovered, Mean recovery, 128 125 96.9 Pg tG per cent. 123 124 56.3 44.6 78.2 41.7 45.7 44.3 30-6 22.7 79.8 24.1 24.5 26.3 15.6 14-7 16.1 14.8 18.1 84.0March, 19671 OF WARFARIN I N ANIMAL RELICTA 195 values on reagents are equivalent to 0.48 p.p.m., expressed as warfarin. Blank values have been determined on three different samples of liver, and 12 determinations gave a mean optical density of 0.040 and standard deviation of &0.014, equivalent to 0.92 p.p.m. of warfarin. Recovery experiments at four levels are given in Table 11. These values are corrected for a liver blank of 0.44 and a reagent blank of 0.48 p.p.m. of warfarin. We thank Dr. H. G. Dickinson, of Ward, Blenkinsop & Co. Ltd., for several interesting discussions, and Miss M. R. Boulton for invaluable technical assistance. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. REFERENCES La Clair, J. B., J . Ass. Off. Agvic. Chem., 1952, 35, 372. -, Ibid., 1953, 36, 373. -, Ibid., 1954, 37, 634. -, Ibid., 1955, 38, 299. Garner, R. J., Novd. VetMed., 1956, 8, 514. -- , Ibid., 1957, 9, 464. Pyorala, K., Annls Med. E X ~ . Biol. Fenn., 1965, 43, Suppl. No. 3. Wanntorp, H., Acta Phavmac. Tox., 1959, 16, No. 2. O’Reilly, R. A., Aggeler, P. M., Hoag, M. S., and Leong, L., Thromb. Biath. Haenzorvh., 1962, Kazyak, L., and Knoblock, E. C., Analyt. Chem., 1963, 35, 1448. Armstrong, W., Chem. & Ind., 1959, 154. 8, 82. Received June 21st, 1966