Analjst, July, 1968, Vol. 93, $9. 445452 446 A One-step Extraction and Clean-up Procedure before Gas = Liquid Chromatographic Determination of Some Organochlorine Pesticide Residues in Blood BY G. CZEGLI?DI-JANKO* AND V. CIELESZKY (Institute of Nutrition, Budapest I X , Gydli ut 3/a, Hungary) An apparatus and a method are described for the extraction and clean-up of organochlorine pesticides, e.g., DDT, DDE, a- and y-BHC (lindane) and dieldrin, from heparin-treated and lyophilised blood samples for gas-chromato- graphic determinations. Extraction and clean-up are camed out in one step on a suitable column with amounts of solvent as small as 25 to 30 ml. Two types of column are used according to the type of organochlorine pesticide to be determined; for acid-stable substances a sulphuric acid - diatomaceous earth column is used, and for alkali-stable pesticides an alkaline column, on which saponifi- cation of fat takes place simultaneously with extraction.The method has been applied successfully in worker as well as in general population surveys. THE accumulation and storage of organochlorine pesticides, particularly DDT, in the human organism is well established.l~2s3s4 Investigation of this phenomenon, however, has been carried out mainly on human fatty tissue and milk samples. Methods for the identification and determination of organochlorine residues are exten- sively reviewed by Beynon and Elgar.6 In general, extraction and clean-up of the active agent are carried out separately. A one-step procedure for the two operations on a conditioned Florisil column was recently described by Langlois, Stemp and Liska.6 Experimental results recording the occurrence and level of organochlorine residues in human blood are scarce.Moreover, the few investigations of this kind that have been carried out were concerned primarily with alkali-stable active agents in relation to occupationally exposed p e ~ p l e . ~ . ~ However, DDT and DDE could not be determined separately by the methods used in these investigations. For the assessment of organochlorine pesticides in the blood of experimental animals, a method was published by Jain, Fontan and Kirk,g and hexane-extractable organochlorine insecticides in human-blood samples were determined by Dale, Curley and Cueto.lo In these tests, as well as in the experiments of Radomski and Fiserova-Bergerova,ll no clean-up was used before analysis. For the determination of organochlorine pesticide residues in blood by gas - liquid chromatography we tried to develop a procedure with which substantially lower pesticide levels in blood originating from nutritional intake could be assessed, not only in occupationally exposed people but also in the general population. For this purpose, efficient clean-up of the blood extracts seemed necessary, particularly as the sensitivity of the tritium-foil electron- capture detector, which was also used in our investigations, becomes seriously impaired by lipid co-extractives present in blood extracts.12 The present paper describes an apparatus and a method by which the extraction of small amounts of organochlorine (DDT, DDE, E- and y-BHC and dieldrin) residues in heparin- treated and lyophilised human-blood samples, and the clean-up of the extract, can be achieved in a single step by using only 25 to 30 ml of solvent.The resulting extract is suitable for gas - liquid chromatographic analysis. The addition of the anti-coagulant to the blood samples became necessary because in this method, even when only incipient clotting occurs, the extractability of the organo- chlorine pesticides considerably decreases and, generally, these compounds cannot be extracted at all from clotted blood. Lyophilisation in pesticide residue analysis has been recently mentioned by Branden- berger and Miiller.13 Also, in the course of other investigations in this Institute, good results * Present address : State Institute of Hygiene, Department of Disinfection, Budapest IX, GyAli ut 2-4, Hungary.0 SAC and the authors.446 [Analyst, Vol. 93 have been obtained by using lyophilised samples in various types of analysis. The procedure had already been successfully applied to the investigation of organochlorine residues in foods, as well as to the determination of other agents, e.g., phosphate esters. EXPERIMENTAL REAGENTS- CZEGL~DI- J A N K ~ AND CIELESZKY : GAS-CHROMATOGRAPHIC All materials should be of analytical-reagent grade, unless otherwise stated. Florisil, 60 to 100 mesh-This was washed with light petroleum, air-dried and condi- tioned by the method of Langlois, Stemp and Liska.6 It was kept for 10 to 12 hours at 140" C, then mixed with 5 per cent. of water; it may be stored for 1 week in a glass-stoppered bottle.Sodium sulphate, anhydrous. Diatomaceous earth , commercial grade. Sulphuric acid mixture-Prepare by mixing equal volumes of concentrated sulphuric Potassium hydroxide. Alumina (Merck)-This was kept at 450" C for 3 hours and then mixed with 10 per cent. Sodium chloride. Heparin solution. Saline (sodium chloride) solution, 0.9 per cent. Hexane, re-distilled, boiling-range 62" to 64" C. Light petroleum, re-distilled, boiling-range 32" to 37" C. Methylene chloride, re-distilled, boiling-point 42" C. Benzene, re-distilled, boiling-range 80" to 81" C. Olive oil. acid and fuming sulphuric acid containing 20 per cent. of sulphur trioxide. of water. LYOPHILISATION OF SAMPLE- Freshly drawn blood (2 to 10 ml), depending on the expected amount of pesticide, is placed into a glass tube containing 1 drop of heparin solution.By centrifuging the blood sample, plasma and red blood cells can be examined separately. For this purpose the cellular components must be washed thoroughly with saline and the washings added to the plasma. After mixing the sample with 50 per cent. of air-dried Florisil, lyophilisation occurs. In our experiments, an apparatus (supplied by Labor Co., Budapest, Hungary) operated by an oil vacuum pump was used, the schematic representation of which can be seen in Fig. 1. Hg vacuum A = Flask B = Receivers C = Freezing mixture at -70" C Fig. 1. Schematic representation of the apparatus for the lyophilisation of blood samples The Florisil- blood mixture is spread on the inner wall of flask A so that a uniform layer is formed.The flask is then immersed in a freezing mixture consisting of solid carbon dioxide and light petroleum until the layer becomes frozen. During this procedure a thin layer of ice is formed on the outer wall of the flask. For subsequent lyophilisation, flask AJuly, 19681 DETERMINATION OF ORGANOCHLORINE PESTICIDE RESIDUES IN BLOOD 447 is placed into the apparatus and the vacuum pump started. Vapours of sublimating ice are condensed in receiver B, which is cooled to -70" C by the freezing mixture, and thus do not reach the vacuum pump. Lyophilisation is completed when the layer of ice on the outer wall of flask A melts away. The water-free layer adhering to the inner wall of the flask is now scraped off and pulverised.mm of mercury. Under these con- ditions, no organochlorine residues, except lindane (see Table I), could be extracted in recovery experiments by methylene chloride (which is used generally in this laboratory in water analysis14), in determinable amounts from the ice accumulated in the receiver after thawing, indicating that DDT, DDE and dieldrin do not escape from blood during lyophilisa- tion. In some instances, however, when a higher vacuum is needed, it may become essential to re-check the lyophilisation procedure for eventual loss of pesticide residues. EXTRACTION AND CLEAN-UP OF SAMPLE- adsorbent (Fig. 3). The vacuum used in this experiment was 5 x For extraction and clean-up, column A, in the apparatus shown in Fig.2, is filled with D A = Column containing adsorbent B = Flak Fig. 2. C = Condenser D =ilnsulated tube Apparatus for extraction of organochlorine pesticides and clean-up of the extract With acid-stable active agents, such as DDT and its metabolite DDE, or BHC, a modified Davidow column15 (Fig. 3-Acid) can be used, and with alkali-resistant dieldrin an alkaline column (Fig. 3-Alkaline), developed by the authors, on which saponification of fat takes place simultaneously with extraction. An alkaline column of different composition has been described by Albert .16 In the apparatus shown in Fig. 2 a small amount of light petroleum, containing 20 per cent. of methylene chloride for dieldrin, circulates through the lyophilised sample placed on top of the adsorbent in column A.The amount of solvent used (25 to 30 ml) should be such that, after percolating through the column, about 6 to 8ml would accumulate in flask B. Water is then circulated in condenser C. The water jacket around column A, and condenser C (Fig. 2), prevents the vapour pressure of the low-boiling solvent from disrupting the column, which would counteract the continuous circulation of the solvent. The solvent is made to circulate by gently warming flask B on a suitably regulated water-bath. The solvent vapours enter through insulated tube D into condenser C, where condensation occurs, and the con- densed solvent again percolates through the sample and column. The warming of flask B is regulated so that above the sample there should always be a solvent layer of 0.5 to 1.0 cm.With correct column filling the solvent flow-rate is about 3 ml per minute.448 AND CIELESZKY : GAS-CHROMATOGRAPHIC [A ?Za&St, VOl. 93 Acid Alkaline Fig. 3. Acid and alkaline columns: A, lyophilised blood sample; B, sulphuric acid - diatomaceous earth mixture (3.5 ml of fuming sulphuric acid and 4 g of diatomaceous earth) ; C, 3 g of pulverised potassium hydroxide, wetted with methylene chloride saturated with water; D, 0.6 g of diatomaceous earth; E, 0.5 g of alumina; F, 3 g of Florisil, conditioned; G, 0.5 g of sodium sulphate; H, glass-wool After 3 hours' circulation the contents of flask B are transferred into a conical centrifuge tube aad the solvent evaporated. The residue is dissolved in 0.25 to 1.0 ml of hexane and is ready for gas - liquid chromatographic analysis.With alkaline column filling the extract in flask B is transferred into a separating funnel, washed twice with water and dried over anhydrous sodium sulphate. GAS - LIQUID CHROMATOGRAPHIC ANALYSIS OF PURIFIED EXTRACTS- A Perkin-Elmer gas chromatograph, Model 452, was used, and a Pyrex-glass column, 2 feet long and $-inch diameter, filled.with Chromosorb W, 60 to 100 mesh, with 2-5 per cent. of Apiezon L and 0.75 per cent. of Epikote resin 1001. The conditions used were: carrier gas, nitrogen; inlet pressure, 1.4 kg per cm2; flow-rate, 200 ml per minute; column temperature, 170" C; injection temperature, about 230" C; detector, electron capture and applied potential, 45 volts; amplifier gain, &th to Ath; and injection syringe, Hamilton microlitre syringe, 701-N.In experiments reported here, 10-pl aliquots of the purified and dissolved extract were injected into the gas chromatograph and chromatograms obtained by selecting the appropriate amplifier gain according to active-agent content. Insecticides were identified by comparing their retention times with those of compounds used for reference.* Thin-layer chromatography, as described by K o v ~ c s , ~ ~ was used as a complementary technique. The quantitative determination of the pesticide was carried out by multiplying the peak height in the chromatogram by the peak width at half the peak height. Calibration graphs were prepared by injecting 1 to 10 p1 of the standard solutions into the gas chromatograph containing 0.5 to 1.0 pg of pure pesticide per ml of solvent.Amounts of blood extracts and reference substances must be chosen so that comparative measurements can be carried out with the same amplifier gain. For DDT, with both test and reference substances, the ratio of the peak heights of DDT, and of DDD originating from a minor decomposition of DDT under the working conditions already described, must be considered. * All organochlorine compounds used for reference were recrystallised several times and controlled by their melting-points.July, 19681 DETERMINATION OF ORGANOCHLORINE PESTICIDE RESIDUES IN BLOOD 449 RECOVERY FROM COLUMNS- The recovery of known amounts of organochlorine pesticides in pure solution from the columns described has been checked. It was found that on both types of column, acidic as well as alkaline, 94 to 96 per cent.of the active agents added in amounts of 0.05 pg (equiva- lent to 0-01 p.p.m. of agent in 5ml of blood) was recovered after 10 minutes’ circulation. After 20 minutes the recovery was 98 to 100 per cent. EXTRACTION TIME- To determine the extraction time necessary to obtain the maximum of extractable residual material, extraction times of 1 to 4 hours were carried out. It was found that whereas 2 hours were insufficient for optimum results, the highest residue yields could be achieved with certainty in 3 hours. With the solvent flow-rate already described, extraction for 3 hours with only 25 to 30 ml of solvent corresponds to a conventional extraction carried out with 550 to 600 ml of solvent.RECOVERY OF ADDED PESTICIDES- Because of lyophilisation, it was necessary to devise special techniques of adding various known amounts of pesticides to blood. The reference substances could be added to blood only in those solvents that do not interfere with the freezing of the Florisil- blood mixture before lyophilisation. In addition, the complete incorporation of the added material into the sample was necessary to avoid. substantial evaporation under vacuum. All of these requirements could be met by dissolving the reference substances in benzene and by mixing known portions of the benzene solution with olive oil containing (from previous experiments) no appreciable amounts of organochlorine residues. One microgram of each pesticide, viz., DDT, DDE, lindane and dieldrin, dissolved in 1 ml of benzene containing 5 drops of olive oil, was added to 10ml of blood (equivalent to 0.1 p.p.m.of pesticide in the blood). The mixture was vigorously shaken and then allowed to stand for a few hours, during which period the shaking was frequently repeated. Subse- quently, an amount of Florisil equal to 50 per cent. w/w of the blood sample was added and the mixture lyophilised. Recoveries are shown in Table I. TABLE I RECOVERY OF ORGANOCHLORINE PESTICIDES ADDED TO BLOOD Amount of pesticide in blood after adding 0.1 p.p.m. of reference Amount of pesticide Pesticide p.p.m. p.p.m. 0.105 DDT .. .. .. 0.0 104 0.107 0.105 0.119 DDE .. .. .. 0.0172 0.1 16 0-113 0.025 Lindane .. .. .. 0-0061 0.025 0.017 0.094 Dieldrin . . .. ..<0-0001 0.097 0.098 in blood, substance, Recovery, per cent. 96.2 97.1 94.7 102.0 98.8 96-3 20.0 20.5 11.9 94.5 97-2 98-3 It is noted that although the recovery of added DDT, DDE and dieldrin was almost complete, that of lindane was only partial. This might arise from the considerable solubility in water and volatility of lindane, in consequence of which most of this substance, when added to blood samples, escapes during lyophilisation and can be recovered in receiver B. Of the lindane originally present, however, no trace could be detected in the receiver, and we found no explanation for this, but the assumption of Dale, Curley and Hayes,ls according to which organochlorine pesticide residues in blood are probably bound to proteins, deserves attention. Therefore, the results must be evaluated with respect to recovery experiments.450 CZEGL~DI- J A N K ~ AND CIELESZKY: GAS-CHROMATOGRAPHIC [Analyst, Vol.93 RESULTS AND DISCUSSION With the apparatus described, extraction of organochlorine pesticides (DDT, DDE, dieldrin, lindane and a-BHC) from blood, and purification of the extract before gas - liquid chromatographic examination can be performed in one single step with a minimum of solvent. The clean-up of extracts containing acid-stable or alkali-stable pesticides is carried out by using columns with different fillings. With an alkaline column, a partial decomposition of methylene chloride present in the solvent mixture can occur. Our experiments, however, proved that neither did interfering peaks appear on the chromatogram, nor did interference with quantitative extraction or recovery of added reference substance occur as a consequence of this decomposition. Under these experimental conditions, clear-cut peaks appeared in the course of gas - liquid chromatographic determinations separately carried out with DDE, dieldrin, lindane and a - B H C , indicating that no decomposition occurred.DDT, however, became slightly decomposed to DDD in the gas - liquid chromatographic procedure, thus giving two peaks. Identification of the small DDD peak was achieved by using DDD as a reference sub- stance. When establishing the DDT calibration graph, the peak area of DDT and the small one of the DDT breakdown product were considered. In measurements of the peak areas of the sample, DDT and DDD were compared with corresponding peak areas in the chromato- gram of pure DDT solutions used for the preparation of the calibration graph. The decomposition of DDT depends also on the condition of the gas-chromatographic column.With a freshly filled column increased decomposition was observed, and the rate of decomposition is also affected by the absolute amount of DDT present, for instance, a larger proportion of 1 ng than of 10 to 15 ng is decomposed. Therefore, comparison of sample and reference DDT must be made as far as possible with similar amounts, with the same amplifier gain and also by preparing sample and reference chromatograms in rapid succession. In Fig. 4 the gas chromatogram of the acid-treated blood extract of a healthy person, with only the usual environmental exposure to insecticides, is shown, together with the thin-layer chromatogram of the same extract and of reference compounds. Although the presence of #I-BHC is revealed by the thin-layer chromatographic plate, the peak of this compound cannot be identified in the gas chromatogram, because, under the conditions used for these measurements, the peak of P-BHC and other unidentified peaks partially overlap.It is noteworthy that the small, unidentified peak after the DDE peak can be seen only in the chromatogram of acid-treated blood extracts and was not noticed, for instance, in fat extracts. In the chromatogram of blood samples other peaks also appear, the identities of which are not yet established. Fig. 5 shows the gas and thin-layer chromatograms of the blood extract of a worker who was employed on a plant that produced aldrin-treated fertiliser.The blood extract was purified by an alkaline column before gas - liquid chromatography, and the thin-layer chromatogram clearly shows two spots corresponding to dieldrin and DDE, the latter including also DDT converted into DDE by the alkaline treatment. Dieldrin is a metabolite of aldrin formed in the organism. The method has also been used in general population surveys for the determination of the DDT, DDE, a- and y-BHC (lindane) and dieldrin contents of blood, and in monitoring the dieldrin concentration of the blood of workers engaged in the production of aldrin-treated fertilisers. TABLE I1 ORGANOCHLORINE PESTICIDE RESIDUE LEVELS IN THE BLOOD OF THE GENERAL Results are expressed in p.p.m.number DDT DDE a-BHC (lindane) Dieldrin POPULATION IN BUDAPEST, HUNGARY, 1967 Serial y-BHC 1 0.0 104 0.0172 0.0004 0-0051 <0~0001 2 0.0085 0.0061 0*0007 0-0063 <0-0001 3 0.0081 0.0052 0.0003 0.0041 < 0.000 1 4 0.0124 0.0462 0~0002 0-0032 <0-0001 5 0.0255 0.0253 0~0001 0.0035 t0~0001July, 19681 DETERMINATION OF ORGANOCHLORINE PESTICIDE RESIDUES IN BLOOD 451 Table I1 shows results from material continuously collected by us from among the general population; the presence of a considerable amount of lindane is remarkable. It is pointed out by Hayesls that the presence of lindane in blood is generally a consequence of some unusual exposure to this substance. Our findings, however, show that with the method 0 oDDE ODieldrii. 0 0 DDE (a) 0 0 DDT 0 c a - B H C 0 0 O y - B H C B - B H C DDT DDE DDD a 7 Blood Start -extract X DE Fig.4. (a) Thin-layer chromatogram and (b) gas chromatogram of an acid-treated blood extract (general population blood sample). X = unidentified peak. On gas chromatogram, section A was recorded a t 30 inches per hour, with an amplifier gain of &th and section B a t 15 inches per hour with an amplifier gain of &th. The diagram is reduced to one third of the original size Dieldrin 1 X 1 I Fig. 5. (a) Thin-layer chromatogram and (b) gas chromatogram of an alkali-treated blood extract (aldrin-exposed worker's blood), X = unidentified peak. The gas chromatogram was recorded a t 15 inches per hour, with an amplifier gain of h t h . The diagram is reduced to one third of the original size452 CZEGL~DI- J A N K ~ AND CIELESZKY described above, 0.01 p.p.m.of lindane can be detected regularly in the blood of the general population at present in Hungary. In Table I11 some representative results are shown, as selected from material collected among workers on a plant producing aldrin-treated fertiliser, corresponding to two different levels of exposure. The first three values relate to transport personnel working in the store- house of the plant, the other three to workers operating the mixing machine inside the plant. It can be seen that the difference between heavy and moderate exposures also reveals itself in the dieldrin levels in blood. TABLE I11 DIELDRIN LEVELS IN THE BLOOD OF WORKERS EMPLOYED IN A PLANT PRODUCING ALDRIN-TREATED FERTILISER Serial number Dieldrin in blood, p.p.m.0.017 1 Transport workers 0.024 0.020 0.195 5 Workers in the mixing room 0.103 6 0-123 In Table IV some results are presented relating to the distribution of DDT and DDE in plasma and erythrocytes in the blood of the general population. TABLE IV DDT AND DDE CONTENTS IN PLASMA AND RED BLOOD CELLS OF PEOPLE NOT OCCUPATIONALLY EXPOSED TO PESTICIDES, BUDAPEST, 1966 I In plasma p.p.m.* Serial & number DDT DDE 1 0.005 0-016 2 0.008 0.026 3 0.008 0.039 4 0.007 0.030 5 0.003 0.018 2 3 In whole blood, as calculated from 1 and 2, In red blood cells, p.p.m.* p.p.m. & - DDT DDE DDT DDE 0.002 0.007 0-007 0.023 0.004 0.009 0.012 0.034 0.003 0.011 0.011 0.050 0.003 0,016 0.010 0,046 0.003 0-007 0.006 0.025 * Calculated to whole blood.4 In whole blood, as determined, p.p.m. DDT DDE 0.006 0.021 0.010 0.032 0.010 0.046 0.009 0-051 0-006 0.028 The results reported in this paper show the suitability of the method for the purposes outlined above. Results and evaluation of our monitoring work will be published elsewhere. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. REFERENCES Quinby, G. E., Hayes, W. J., Armstrong, J. F., and Durham, W. F., J . Amer. Med. Ass., 1965, Maier-Bode, H., Medna Exp., 1960, 1, 146. DCnes, A., Die Nahrung, 1962, 6, 48. Robinson, J., Richardson, A., Hunter, C. G., Crabtree, A. N., and Rees, H. J., Brit. J . Ind. Med., Beynon, K. I., and Elgar, K. E., Analyst, 1966, 91, 143. Langlois, B. E., Stemp, A. R., and Liska, B. J.. J . Agric. Fd Chem., 1964, 12, 243. Brown, V. K. H., Hunter, C. G., and Richardson, A., Brit. J . Ind. Med., 1964, 21, 283. Richardson, A.. Robinson, J.. Bush, B., and Davies, J. M., Archs Enuir. Hlth, 1967, 14, 703. Jain, N. C., Fontan, C. R., and Kirk, P. L., J . Pharm. Pharmac., 1965, 17, 362. Dale, W. E., Curley, A., and Cueto, C . , jun., Life Sciences, 1966, 5, 47. Radomski, J. L., and Fiserova-Bergerova, V., Ind. Med. Surg., 1965, 34, 934. McCully, K. A., and McKinley, W. P., J . Ass. Off. Agric. Chem., 1964, 47, 652. Brandenberger, H.. and Muller, S., Mit. Geb. Lebensmittelunters. u. Hyg., 1965, 56, 281. Cieleszky, V., Egdszsbgtudomdny, 1967, 11 , 93. Davidow, B. J., J . Ass. Ofl. Agric. Chem., 1950, 33, 130. Albert, R. A., Ibid., 1964, 47, 659. Kovacs, M. F., Ibid., 1963, 46, 884. Dale, W. E., Curley, A., and Hayes, W. J., jun., Ind. Med. Surg., 1967, 36, 275. Hayes, W. J., jun., “Scientific Aspects of Pest Control,” Publication No. 1402. National Academy First received November 17th, 1966 Amended January 26th, 1968 191, 175. 1965, 22, 220. of Sciences, National Research Council, Washington, D.C., 1966, p. 314.