Cases may be found by searching for them in the community (in the course of a prevalence survey or in the context of surveillance) or they may be detected among those individuals who present themselves to the health services with symptoms (passive case finding). The latter is the recommended method of routinely identifying cases.
Isolation of the micro organism by culture
As tuberculosis is a disease caused by the organism Mycobacterium tuberculosis, its definitive isolation by bacteriological culture is the sine qua non of tuberculosis. Specially prepared media are required for the isolation of the organism and standardized methods have been outlined for such cultures47. Strict quality assurance methods must be applied to bacteriological services to ensure that the results are valid; a disturbing proportion of positive cultures in routine laboratories have been demonstrated to be due to contamination48 49.
The proportion of cases among those treated for tuberculosis who have the micro organism isolated varies according to the site and extent of the disease. Among 5,307 cases of tuberculosis notified in Britsh Columbia, Canada from 1975 to 1989, 74.1% of them were confirmed by culture. This varied from 27.5% of those with primary tuberculosis to 76.5% of those with minimal pulmonary tuberculosis and 94.2% of those with far advanced pulmonary tuberculosis.
It also varies according to the examination methods used; among post primary pulmonary cases in the same study, only 68.0% of those diagnosed by a screening examination were culture positive as compared with 84.7% for cases routinely detected by passive case-finding.
Test characteristics
Using the information provided above, assuming 9 suspects examined for every case and a contamination rate for cultures of 1 per cent, it is possible to calculate the test characteristics for cultures. For passive case finding: sensitivity 87%, specificity 99%, positive predictive value 92% and negative predictive value 98%. If the culture is used in a survey and the prevalence of disease in the community is 50 per 100,000 persons, the values are: positive predictive value 4% and negative predictive value greater than 99%.
Identification of micro organisms by morphology
In addition to isolation of the micro organism by bacteriological culture, it is also possible to identify the organism morphologically by means of direct microscopy of biological specimens, usually sputum samples. Standardized methods have also been outlined for such procedures50. This is the recommended method of confirmation of cases in low income countries.
Test characteristics
If the smear examination is used for passive case finding of tuberculosis, the specificity is between 96% (under field conditions)51 and 98% (under research conditions)52; sensitivity of the test is approximately 50%. Positive predictive value is between 89% and 95%; negative predictive value is 76%. The reproducibility of the test is, under the experimental conditions noted above, 92% and, under field conditions, 87%.
The validity of the test varies according to the epidemiological situation. The most frequent cause of a positive sputum smear which does not represent Mycobacterium tuberculosis is the presence of environmental Mycobacteria species in the sputum which occurs in patients with the Acquired Immune Deficiency Syndrome (AIDS) in areas where tuberculosis is relatively uncommon53. Where tuberculosis is relatively common, a high proportion of specimens with positive smears yield cultures positive for Mycobacterium tuberculosis; the isolation of environmental mycobacteria under these conditions is exceedingly rare. A study in the National Tuberculosis Program in Singapore54 indicated that, of 500 patients positive on at least one of two sputum smears, 483 (97%) were confirmed with at least one positive of two cultures performed. In the same study, 115 of the 500 cases were positive on only one of the two sputum smears of whom 103 (90%) were confirmed on culture.
If the smear examination is used in a survey of the population, the proportion of cases culture positive who will additionally be smear positive may be low (22% in a population with regular screening examinations)55, thus considerably reducing the sensitivity of the test.
Test characteristics of smear examination used for detection of contagious tuberculosis can be estimated from studies of contacts of active cases of tuberculosis. In a study of 13,472 contacts56, test characteristics were: sensitivity 96%; specificity 35%; positive predictive value 17%; negative predictive value 99%. Positive predictive value and specificity are low as many infected individuals had been infected by others than the individual contact in the study. While the smear examination is not so efficient in detecting cases, it is highly sensitive in identifying those who are likely to infect others, while being inefficient in identifying all sources of infection in the community.
Molecular genetic techniques
Molecular methods can identify precise strains of Mycobacterium tuberculosis, using restriction fragment length polymorphism (RFLP), and allow the tracking of pathogens to enhance the understanding of transmission57. Using RFLP, point source outbreaks of M tuberculosis have been investigated in a variety of (primarily institutional) settings. In addition, a "library" of strains has been prepared to identify epidemiologically related strains58 and to track the transmission dynamics of multi drug resistant strains. Finally, by comparing initial isolates with those isolated in patients who develop tuberculosis disease again permits the identification of reinfection59. Finally, RFLP can be used to trace contamination of bacteriological cultures.
