Comparison in the efficiences of the tests for pyrogens in various pharmacopoeias

Abstract

Pyrogen test method in four kinds of pharmacopoeiae (Japanese 7th edition, U. S. 16th edition, British 1958 edition, and International 2nd edition) was comparatively examined by the statistical treatment of its efficiency from the standard of assay by the Monte Carlo method. The variance (σ2=0.0179), obtained by the administration of glucose injection, was found to be a good measure having sensitive discriminatory power. The British method was especially good but it required the largest number of animals per unit test. The variance (σ2=0.0835), obtained by the administration of a pyrogenic substance, showed a fair difference in the efficiency among the four methods and the American method was found to be the most rational, there being less chance of making false judgement of positive result than by the Japanese or International method, although there was no likelihood of over-looking the pyrogenic substance in any of the methods. The British method had a greater chance of missing the pyrogenic substance than the other methods, in spite of the use of a larger number of animals per unit test. It was concluded that, if the American pharmacopeal method is followed, individual difference of animals tested for the rise of body temperature must be below ca. 0.09 in variance, and there would be an increasing danger of over-looking the contamination of minimum pyrogenic substance of 0.6° by the British method if this variance of the individual difference is over ca. 0.02.