Technical and bacteriological aspects of a continuous survey of drug resistance in tuberculosis in South Africa

The existence of several methods for determining in vitro drug resistance of M. tuberculosis, as well as the differences in definitions of drug resistance, prompted the authors to use a discrimination approach described by Mitchison in 1969. This method was employed to analyse a continuous survey of drug resistance in South Africa undertaken between 1964 and 1970.A retrospective analysis of 16,511 strains on record was made, using electronic data processing, to measure the critical level of drug concentration. The strains were isolated from the sputum of African patients in hospital. Results from a research and a routine laboratory were compared, the former using the absolute concentration method and the latter the vertical diffusion method.The patients' history was used to differentiate the strains into `predominantly sensitive' and `predominantly resistant' groups. The definition of resistance was obtained at the drug concentration that best distinguished between the two groups.The procedure was also suitable for judging the quality of in vitro tests. The recent trend to consider growth at a low concentration of major drugs as the critical level of resistance was confirmed as bacteriologically correct. The critical levels for most drugs remained for both methods fairly constant during the 6 years.It is unnecessary to use high concentrations of PAS. The vertical diffusion method was shown to be inaccurate for certain drugs.