A gas chromatographic method of measuring residual water in freeze-dried smallpox vaccine*

Bull. Org. mond. Sante 1972, 47, 7-11Bull. Wld Hlth Org. A gas chromatographic method of measuring residual water in freeze-dried smallpox vaccine* L. C. ROBINSON 1 A new method for the accurate determination of residual water infreeze-dried vaccines is described. Tests may be carried out on very small samples in sealed ampoules. The method is based on the adsorption of water from the vaccine with benzene and its estima- tion by gas chromatography. It appears that the residual water and total nitrogen content are important factors in the stability of the vaccine. The freeze-drying of virus and bacterial suspen- sions for the production of vaccines is usually car- ried out with little or no regard to the amount of water remaining in the product. In the few experi- ments in which residual water has been measured, either the Karl Fischer method (Greiff & Rightsel, 1968, 1969) or a gravimetric technique (Namura et al., 1956) was used. The Karl Fischer technique requires all reagents to be carefully dried and is not suitable for small samples in sealed ampoules. The gravimetric method, which relies on drying to con- stant weight, is highly inaccurate since the samples are small and extremely hygroscopic. The method described here is based on the adsorption of water from the sample with benzene and its subsequent estimation by gas chromatography. The method is accurate, rapid, and applicable to small samples in sealed glass ampoules. Although it was developed primarily to facilitate the production and quality control of Lister Institute tissue-culture smallpox vaccine, the method has been successfully applied to other freeze-dried virus and bacterial vaccines and to plasma fractions. MATERIALS AND METHODS Gas chromatography A grade of benzene suitable for gas chromato- graphy was used. It was dried by refluxing over phosphorus(V) oxide for 2-3 hours and was then * A patent application for the method has been filed under No. 12895/71. 1 Biochemist, Virus Vaccine Unit, Lister Institute of Preventive Medicine, Elstree, England. distilled over freshly drawn sodium wire 0.5 mm in diameter. The benzene was stored in nonactinic glass bottles over sodium wire and freshly activated molecular sieve, Linde type 4A. A Perkin-Elmer model Fl1 gas chromatograph with a hot-wire detector was used. The dual col- umns were 1-m stainless steel tubes 1 mm diameter packed with Chromosorb 102. The carrier gas was high-purity hydrogen at a pressure of 2.8 kgf/cm2 flowing at 80 ml/min. The oven temperature was 1 70^C, the detector temperature 200°C, and the injection port temperature 250°C. The attenuation range was 2 and the detector supply setting was 7. The recorder scale was 2.5 mV and the chart speed 4 mm/min. Calibration curve for water in benzene The calibration curve (Fig. 1) was constructed by preparing known solutions of water in benzene; pentane (20 ,l/ml) was added as an internal stan- dard that was used in all determinations, and 2.0-1l volumes of the benzene/water solutions were injected. The ratio of the peak height of water over that of pentane (hw/hp) was plotted against the amount of water expressed as ,g/ml. Treatment of the vaccine container Rubber-capped vials. These vials were kept over- night in a desiccator over phosphorus(V) oxide, dried, and then weighed; 1 ml of dry benzene was injected into each vial. After equilibration for 30 min, a 2.0-pl sample from the ampoule was injected into the injection port of the gas chromato- 2867 -7 8 L. C. ROBINSON 60 50 40 30 20- 20 10 0 50 100 200 300 400 pg of water/ ml of benzene Fig. 1. Calibration curve for water in benzene. graph. After the contents had been sampled, the rubber cap was removed from the vial and both the vial and the cap were rinsed in acetone, washed in water, and dried to constant weight. The dry weight of vaccine could then be determined. Sealed glass ampoules. A special cell was devised to permit the ampoule to be opened in a sealed, dry atmosphere. The cell consists of two Quickfit screw thread joints, thread size 28 (catalogue No. SQ28),1 fused together (Fig. 2). The base con- tains a short length of copper tube in which the ampoule is held. The side-arm is fitted with a piece of mild steel rod that is free to slide over a distance of about 2.0 cm through a rubber sleeve. The top of the cell is fitted with a silicone rubber membrane 1 mm thick. In use, the cell was dried overnight in the oven at 100°C. Ampoules to be tested were kept overnight in a desiccator over phosphorus(V) oxide; they were then scratched with a file at the weak point, weighed, and placed in the cell. The mild steel rod was tapped lightly until the ampoule frac- tured; 1 ml of benzene was then injected through the membrane into the ampoule. Samples (2 ,ul) taken after a 30-min equilibration period were injected into the chromatograph. After sampling, the ampoule and all fragments in the cell were rinsed with acetone and then with water, and dried in the oven at 80°C. The fragments were weighed and the weight of the vaccine was determined. I Manufactured by Quickfit & Quartz, Ltd., Stone, Staffordshire, England. Determination of residual water Residual water was expressed as a percentage of the weight (wt %) of vaccine. The method was checked by means of two experimental models. (I) Determination of water lost by sodium sulfate decahydrate at its transition point. A known weight (approximately 10 mg) of Na2SO4.10H2O was heated with 25 ml of dry benzene to 32-34°C under reflux, the condenser being fitted with a molecular sieve water trap. Samples were removed and injected into the column. A typical result was 0.0228 wt% cal- culated and 0.0250 wt% found by experiment. Plastic screw cap(OC28/9) Silicone rubber disk Steel rod r2mm, uibbersleeve 25 - n = Side-ann . e Copper tube Silicone rubber seal Copper disk Fig. 2. Cell for opening ampoules in a sealed, dry at- mosphere. (2) Determination of water adsorbed on glass beads by the chromatographic method, and comparison of the result with that obtained by a gravimetric method. About 20-30 g of 4-mm diameter glass beads were weighed, heated at 1000C for 6 hours, and allowed to cool in a vacuum desiccator; they were then weighed again. This procedure was repeated until a constant weight was obtained. An equal weight of the beads (from the same bottle on the same day) METHOD FOR MEASURING RESIDUAL WATER IN FREEZE-DRIED VACCINE Fig. 3. Typical chromatogram of residual water in smallpox vaccine. a, Air; b, benzene; p, pentane; w, water; the arrows indicate injections. 9 L. C. ROBINSON was washed with 2.0 ml of dry benzene and analysed by gas chromatography. A typical result for the weight of water adsorbed on to glass beads was 0.172 wt% by the gravimetric method and 0.169 wt% by gas chromatography. Smallpox vaccine Experiments were carried out on batches of Lister smallpox vaccine prepared by a conventional method Table 1. Stability and residual water and total nitrogen content of smallpox vaccines derived from animal skins Potency (logio pock-forming units/ml Serial Dried at Water *Total No. Dried at -10C and (wt%) nitrogen Predried -10 C stored at (wt%) 37'C for 4 weeks Batches failing to meet WHO potency requirements 1 8.3 7.4 7.3 0.318 23.6 2 8.3 7.5 7.1 0.550 26.5 3 8.3 7.8 7.2 0.572 30.0 4 8.5 8.1 7.3 >0.200 34.0 5 8.2 7.4 7.3 0.485 45.0 6 8.3 8.1 7.3 0.527 32.4 7 8.3 7.6 7.3 0.514 24.8 8 8.5 8.2 7.3 0.424 26.5 Batches passing the WHO heat-resistance test 9 8.4 8.2 8.1 1.525 54.8 10 8.3 8.2 8.1 0.900 67.0 11 8.4 8.1 8.1 0.860 75.0 12 8.6 8.2 8.2 1.500 50.0 13 8.8 8.1 8.1 1.520 47.3 14 8.6 8.2 8.1 0.620 50.5 15 8.7 8.8 8.3 0.758 48.5 16 8.7 8.8 8.3 0.797 43.4 (Kaplan & Murray, 1962). Total nitrogen was deter- mined by the direct estimation of ammonia in total micro-Kjeldahl digests by the indophenol reaction (Jones, 1967). Virus infectivity was assayed by the method of pock counts in the chorioallantoic mem- branes of 12-day-old chick embryos (Westwood et al., 1957). RESULTS The results of these experiments are shown in Table 1. A typical chromatogram is shown in Fig. 3. Five of the vaccines (1, 2, 3, 5, and 7) that failed to satisfy the WHO potency requirements (Table 1) lost much of their infectivity after freeze-drying, but were relatively stable at 37°C. In these batches the residual water content of the freeze-dried vaccine was less than 0.6 wt%, and the total nitrogen con- tent before freeze-drying was less than 45.0 wt%. The heat resistance test (WHO Expert Group on Requirements for Biological Substances, 1966) re- quires that after 4 weeks' storage at 37°C a smallpox vaccine shall have a titre exceeding 1.0x 108 pock- forming units per ml, and that at least one-tenth of the virus titre shall be retained. The vaccines that passed this test (Table 1) all had a total nitrogen content above 45.0 wt% before freeze-drying and a residual water content above 0.6 wt% after freeze- drying. DISCUSSION Several authors have commented on the dangers of overdrying (Fry & Greaves, 1951; Hilleman et al., 1951; Hutton et al., 1951; Naylor & Smith, 1946). Merryman (1966), on the other hand, stated that for maximum stability the residual water should be minimal. Results obtained with the method described here indicate that there is no simple cor- relation between the residual water or total nitrogen of a vaccine and its respective stability at 37°C. The data presented in Table 1 are too few to enable definite conclusions to be drawn but they suggest that a vaccine with a low nitrogen content may be relatively unstable if the residual water content is very low. ACKNOWLEDGEMENTS The author thanks the World Health Organization for a grant for the purchase of equipment, and acknowledges the valuable technical assistance given by Mr D. J. Walker. 10 METHOD FOR MEASURING RESIDUAL WATER IN FREEZE-DRIED VACCINE 11 RESUME UNE METHODE DE CHROMATOGRAPHIE GAZEUSE DESTINEE A MESURER L'EAU RESIDUELLE DANS LES VACCINS ANTIVARIOLIQUES LYOPHILISES L'auteur decrit un proce&d permettant de determiner la teneur en eau residuelle des vaccins lyophilis6s. La methode, basee sur l'adsorption de 1'eau de l'6chantillon par le benzene et sa mesure par chromatographie gazeuse, est pr&cise, reproductible et ne requiert qu'un appareillage relativement simple. Destin6e a l'origine a contr6ler la qualite des vaccins antivarioliques, elle est applicable a d'autres vaccins viraux ou bacteriens et aux fractions du plasma. Le r6le de la teneur en eau residuelle et en azote total sur la stabilit6 d'une preparation vaccinale est brievement examine. REFERENCES Fry, R. M. & Greaves, R. I. N. (1951) J. Hyg. (Lond.), 49, 220-246 Greiff, D. & Rightsel, W. A. (1968) Appl. Microbiol., 16, 835-840 Greiff, D. & Rightsel, W. A. (1969) Appl. Microbiol., 17, 830-835 Hilleman, M. R. et al. (1951) Publ. Hlth Rep. (Wash.), 66, 1195-1203 Hutton, R. S. et al. (1951) J. Bact., 61, 309-319 Jones, C. R. (1967) Lab. Pract., 16, 1486 Kaplan, C. & Murray, H. G. S. (1962) In: International Symposium on Smallpox Vaccination, Lyon, Institut Merieux, pp. 150-154 Merryman, H. T. (1966) In: Cryobiology, London, Academic Press, pp. 610-654 Namura, M. et al. (1965) Jap. J. med. Sci. Biol., 18, 249-256 Naylor, H. B. & Smith, P. A. (1946) J. Bact., 52, 565- 573 Westwood, J. C. N. et al. (1957) J. Hyg. (Lond.), 55, 123-139 WHO Expert Group on Requirements for Biological Substances (1966) Wld Hlth Org. techn. Rep. Ser., No. 323