THE EPIDEMIOLOGICAL BASIS OF
TUBERCULOSIS CONTROL
D.A. Enarson and Annik Rouillon
2.1 THE PROBLEM OF TUBERCULOSIS:
EXTENT AND MEASUREMENT
2.1.1 THE EXTENT OF THE PROBLEM
Tuberculosis continues to be a very major problem throughout the world: in the early 1990s as many as 16 million cases of tuberculosis have been reported with, each year, 8 million new, cases (one-half of which are infectious) and 3 million deaths due to the disease [1]. These estimates are likely to be high; a more reasonable estimate of the total number of new cases each year is 5.5 million, of which 74% occur in Asia and another 12% in Africa. As many as one billion (one thousand million) people throughout the world may be infected with tuberculosis. Cases of tuberculosis arise either from this large pool of infected people or from those with previous disease, now inactive. The only way to prevent tuberculosis entirely is to stop the transmission of this infection in the community. However immediate cessation of transmission would be accompanied by only a gradual disappearance, because of disease arising from previously infected individuals who carry a diminishing risk of developing disease throughout their lives.
In spite of the ubiquitous nature of tuberculosis and the 'vicious cycle' that it entails, the bacillus is remarkably inefficient and this inefficiency is the key to its elimination. To become infected requires contact (usually prolonged and intimate contact) with an infectious case [2]. In industrialized countries, the number of contacts per case of infectious tuberculosis appears to be declining; in the middle 1970s, in Canada, the number of contacts per case was seven; at present there are only three contacts per case, only one being under 20 years of age.
2.1.2 MEASURING TUBERCULOSIS IN THE COMMUNITY
When tuberculosis was common and specific treatment was not available, the mortality rate was a good indicator of the size of the tuberculosis problem, as there was a fairly constant relationship between incidence and mortality. Rates have been declining ever since the first records were kept. The rate of decline prior to the introduction of chemotherapy was steady at approximately 4% per annum. Around 1945, the rate in industrialized countries plummeted and then began to stabilize again around 1960 following the introduction of streptomycin in 1946, para-aminosalicylic acid (PAS) in 1948 and isoniazie (H) in 1952.
A second measure of tuberculosis in a community is the tuberculin skin test developed by Koch, which indicates the presence of infection with tuberculosis independent of clinical disease. Much of the earliest epidemiological use of this test was in Scandinavia[3]. Since these early studies, the tuberculin test has been used to determine the prevalence of infection in samples of the general population of many countries. The skin testing of new recruits to the military service of the Netherlands provided the data from which the determination of the average annual risk of tuberculosis infection was developed [4], although the concept was first described in the study of other infectious diseases (notably smallpox). It indicates the probability of becoming infected within a given year, either estimated from the age-specific prevalence of tuberculin skin sensitivity or directly measured by repeating the test in the same population at several points in time and calculating the rate of development (incidence) of tuberculin skin sensitivity (indicating infection with tuberculosis). The calculation of the estimate from a single tuberculin survey is based on an algebraic formula ARI = 1 - (1 - p)1/a where ARI is the average annual risk of infection, p is the prevalence of tuberculin sensitivity in the sample and a is the average age of the group tested. This estimation gives the average of the experience over the period of time represented by the age (reflecting the average experience at the mid point of the average life span of the tested individuals) and therefore cannot measure short-term changes.
Prevalence rate, from representative samples of the general population, has been determined in some countries, indicating the number of cases of active tuberculosis within the community at a given point in time. The protocol for such work has been developed superbly in Japan and has been undertaken in a number of countries, especially in East Asia. These surveys can identify all infectious sources (the smear-positive cases) at a point in time. Under stable conditions, without intervention, the ratio of prevalence to incidence in a community is 2:1 (that is, the average duration of a case is 2 years). The introduction of specific chemotherapy of tuberculosis has resulted in a disruption of the ratio of prevalence to incidence, by shortening the duration of cases.
In countries where notification of tuberculosis cases is quite thorough, the incidence of active tuberculosis can be approximated by the notification rate of cases, and has been used to monitor the epidemiological situation. This measure requires the supervision and review of all notifications by tuberculosis experts to ensure the validity and completeness of the registration of cases. The incidence of all active cases of tuberculosis is approximately twice that of smear-positive pulmonary cases. An annual risk of infection of 1% is approximately equal to an annual incidence of all active cases of 100 per 100 000 [5] where there is no HIV infection; where there is a significant overlap of HIV and tuberculosis infection, this ratio is no longer valid.
For purposes of discussion, various 'bench-marks' of tuberculosis (in rates per 100 000 per year) have been proposed as follows [6].
1000 above this rate, tuberculosis can be said to be 'epidemic.'
100 above this rate, groups can be defined as at 'high risk' for tuberculosis.
10 or below this rate, groups can be defined as at 'low risk' for tuberculosis.
1 below this rate, the tuberculosis programmes are entering the elimination phase.
0.1 at this level, tuberculosis can be said to be eliminated.
2.2 THE NATURAL HISTORY OF TUBERCULOSIS
2.2.1 IN THE INDIVIDUAL
In order to be infectious, a patient must have a sufficient concentration of micro-organisms in the sputum to create the floating, infecting dust particles when coughing, sneezing or singing. Tuberculosis patients, to be infectious, must have pulmonary tuberculosis, with the sputum containing acid-fast bacilli visible on direct microscopy [7].
