Indian Pediatrics 2000;37: 9-18

Tuberculosis Control, Without Protection From BCG

The 15- year follow up of the famous Indian Council of Medical Research (ICMR) BCG trial in Chingleput district, for protection from tuberculosis, has recently been reported in the Indian Journal of Medical Research(1). The paper closed with the following statements: "In conclusion, the Chingleput trial, which was designed to evaluate BCG as a public health measure, has shown that BCG offers no protec-tion against adult type bacillary tuberculosis. Consequently, BCG cannot be expected to reduce the transmission of tuberculosis"(1). Where do we go from Chingleput?

In this essay I shall briefly review and summarize the ICMR BCG trial and results, and revisit the National Tuberculosis Programme (NTP, 1962) and the Revised National Tuberculosis Control Programme (RNTCP, 1992) in the light of the final outcome of the trial. There are clear indications of the inade-quacies of the NTP/RNTCP strategy and tactics of TB control in India. Therefore, the main purpose here is to propose a rational, revised and realistic strategy for the control of tuberculosis.

The ICMR BCG Trial

The objective of this ambitious trial was to measure the protective effect of BCG vaccination against pulmonary tuberculosis (TB) in the population. Two vaccine strains (Danish and French) and two doses (high, 0.1 mg and low, 0.01 mg) were compared. The study started in 1968 and recruitment was completed in 1971. Follow up was sequential in 6 intervals of 2.5 years each, for a total of 15  years. Among a total of 281,161 persons aged 1 month or above, two-thirds were given BCG and a third were given placebo in randomized blind allocation. In the pre-vaccination stage, everyone beyond infancy was tested with PPD-S and PPD-B (mycobactin of Battey strain of M. avium, representing environmental non-TB mycobacteria). Also, everyone above 10 years was X-rayed. Those below 10 were assumed to be free from TB, unless ill. The subsets of population analyzed for BCG efficacy were those with <8 mm PPD-S reaction and normal chest X-ray at recruitment and those who were infants at that time. They were defined as uninfected by Mycobacterium tuberculosis at recruitment.

Follow up included miniature X-ray chest of all subjects over 5 years of age every 30 months. Those with symptoms or X-ray abnormality had one spot sputum and one over-night sputum specimen collected and cultured for M. tuberculosis. Between surveys, anyone with symptoms was also similarly examined. The overall prevalence of bacillary (i.e., open pulmonary) TB (in all subjects above 10 years, irrespective of PPD reaction) fell gradually from 10.55/1000 to 10, 10.37, 9.53, 8.68, 8.59 per round to follow up, and finally to 7.75/1000 after 15 years. Except for the prevalence at the initial survey, the subsequent rates are not equal to the prevalence in the general population (in which new persons are added each year), but are specific only for the cohort of population recruited into the study. Therefore we cannot assume that the prevalence in the general population would have declined at all. In the study cohort, a total of 67.47 persons among 1000 persons had bacillary disease within just 15 years. The overall annual incidence of bacillary disease was 3.83/1000 at the start and it declined to 2.3/1000 at the last round (male, 5.76 to 3.61; female, 1.87 to 1.02). This decline was in the study cohort, and it should not be interpreted to indicate a fall in incidence in the general population, as already pointed out.

There were 35,708 persons over 5 years, with > 11 mm of PPD-S reaction (hence defined as already infected wiht M. tuberculosis) at start, and 21,418 among them were still available at the last follow up. Their culture-positive TB incidence declined from 5.54/1000 at entry to 4.45, 3.56, 3.56 and 4.05 and 4.25 at subsequent rounds of follow up. On the other hand, among those with <12 mm PPD-S reaction (including those with <8 mm reaction) the incidence increased from 0.35/1000 at entry to 0.55, 0.43, 0.72, 0.73 and 1.09, during the subsequent rounds of follow up. At recruitment, 94% of bacillary cases were among those who had already been infected with M. tuberculosis. During follow up, 83% to 89% of cases continued to be among them. Even after 15 years, at the final round of follow up, 80% of cases were still among those who had been found infected at the starting point. These data suggest that the predominant pathogenesis of pulmonary bacillary TB is reactivation rather than re-infection (see below).

The segment of population who were defined not infected with M. tuberculosis at entry, by virtue of their PPD-S reaction being <8 mm, or being infants at that time, ranged from 227,321 (first round) to 142,680 (final round). Their annual incidence of bacillary TB was 0.5 and 0.51 per 1000 in those given high and low dose of BCG respectively and 0.48/1000 in those given placebo, thus showing no protection by BCG. In the age group of 1 month to 9 years (at recruitment), there was 27% (for high dose BCG) and 21% (for low dose BCG) apparent protective efficacy. But, the 95% confidence intervals were wide and ranged from below zero, thus showing no statistical significance. In all other age groups, the annual incidence rates were marginally more among the BCG-vaccinated, but again without statis-tical significance. The two strains of BCG showed no difference in performance. Earlier contact with environmental mycobacteria did not offer obvious protection against tuber-culosis, nor did it improve the efficacy of BCG vaccination. Since BCG offered no protection, the subsets of population who received BCG or placebo are epidemiologically similar. As this population advanced in age, the incidence of bacillary TB increased from 0.136/1000 at entry to 0.327, 0.354, 0.628, 0.764 and finally 1.044/1000 during the follow up rounds. These incidence rates do not apply to the total popula-tion of the community, since they would include those who had >7 mm PPD-S reaction at the start and also the new entrants by birth.

The 7.5-year follow up had been reported in 1980(2). The 15 years of follow up was over in September 1987, but there has been an inordinate delay in publishing the final results, until 1999. The earlier report also had concluded that BCG vaccination has no role in the prevention of spread of TB(2). The present final report confirms the earlier conclusion with the full data set(1). Those who designed and conducted this truly monumental study deserve our gratitude and appreciation for a difficult task done exceedingly well.

BCG and Childhood (Primary) Tuberculosis

The purpose of this trial was to determine if BCG vaccination would reduce the trans-mission of M. tuberculosis. Sputum positive TB patients are the primary source of transmission. Only if BCG vaccination reduces the incidence of such cases, has it a role in the control of tuberculosis. The impact of BCG vaccination on childhood forms of TB was not investigated in this trial.

The WHO launched the Expanded Program on Immunization (EPI), including routine infant BCG vaccination, only in 1974 and India adopted it in 1977-1978. In other words, the Chingleput trial, which started in 1968 was not meant to give any guidance regarding the role of BCG vaccination on childhood TB. There-fore it is not appropriate to examine the useful-ness of BCG as part of EPI in the context of this trial. We must use evidence independent of this trial for deciding whether BCG vaccination in infancy does any good or not. And the bulk of evidence is that BCG vaccination is highly effective for protection of children from pro-gressive primary (non-pulmonary) TB, particularly TB meningitis(3). Other forms of progressive primary TB include bone and miliary TB and scrofula.

There is no evidence in the Chingleput study, that BCG will protect children against infection per se with M. tuberculosis, or against the later development of pulmonary bacillary TB as a result of either reactivation or re-infection. On the other hand, it is clear that the incidence of pulmonary TB had steadily increased with time among those who were previously (at the start of the study) not infected with M. tuberculosis, but had been vaccinated with BCG (or given placebo). This is evidence for continued spread of infection in the community inspite of active case detection and treatment, and BCG vaccination. In short, BCG did not protect against primary infection or its progression to secondary TB. As long as transmission of M. tuberculosis is rampant in the community, we must continue with BCG vaccination in EPI as it is the only intervention we have to reduce the risk of progressive primary TB.

For the sake of completeness and trans-parency, it must be pointed out that the annual incidence of TB was higher in the vaccine group than in the placebo group in the first 5 years after BCG vaccination. However, it was in those  vaccinated when under 10 years, that there was the 27% (for high dose) and 21% (for low dose) apparent efficacy, confirming that BCG had no adverse epidemiological effect on child health. In the next 7.5 years after vaccination, the annual incidence was higher in the placebo group than in the vaccine group. At the final round, the incidence was once again higher in the vaccine group than in the placebo group. As mentioned earlier, over the total period of observation, the average annual incidence was equal in the vaccinated and the placebo groups.

The National Tuberculosis Program


The NTP was established in 1962 and BCG was included in it as a public health tool on faith, and this trial was to give a clear answer one way or the other regarding its role in NTP. So the Chingleput trial was designed in mid-1960's and commenced in 1968. The long-term objective of NTP was to control TB until it is no longer a public health problem. Control was defined as the occurrence of less than one new infection for every case of bacillary TB and also as the reduction of the cumulative prevalence of infection (detected by PPD test) to less than 1% by age 14 years(4,5). The operational objectives were to diagnose and treat all TB cases that came to medical attention and to vaccinate all infants with BCG.

It makes good sense to detect bacillary cases very early so that treatment could be started immediately both to make the person well and to arrest further transmission to fresh suscep-tibles. It was expected that people with symp-toms of TB would attend health care institutions and that they would be investigated for TB. Two decades after the start of NTP it was estimated that nationally the case finding rate and treat-ment completion rate were below 40%(4). The primary health care system did not succeed in India for various reasons and people sought treatment outside the system. The tactic of case detection had to be redesigned.

The Revised National TB Control Program (RNTCP)

The Government of India, the WHO and the World Bank together redesigned the TB control efforts into the RNTCP and launched it in 1992. The RNTCP defined certain operational targets. It was postulated that, if 70% of actual bacillary TB cases were detected through sputum smear examination (Zeihl-Neelsen staining and micro-scopy to detect acid fast bacilli, AFB) and if 85% of them were rapidly cured, then trans-mission could be reduced to less than one fresh infection per every sputum positive case(4,5). Case detection is again passive, relying upon primary health care, but now under public sector and also private sector. From 1994, treatment is through supervised short course chemo-therapy regimen, known as directly observed therapy (DoT). The program's administrative structure is vertical from the National level (Deputy Director General of Health Service for TB) through State and District (district tuber-culosis centers, DTC with facilities for chest X-ray and sputum microscopy) and sub-district levels. The care delivery is through the primary and referral health care system(4,5). The area under cover of RNTCP started with a few districts (2.35 million population covered in 1993) and it has been expanded over time (15.83 million covered in 1995) until the entire country will come under the programme at a future day(4,5).

The Operational Guidelines for TB Control is an exhaustive manual containing all necessary information for the operation of the RNTCP(6). In an ideal world, or in even an established and moderately disciplined health care system, the guidelines would work well, but in the real world situation of India, the motions of processes may be gone through, and perhaps  even the process targets acheived on paper, but the goal will not be reached. My reasons are given below.

Implications of TB Epidemiology in Chingleput for RNTCP

In the Chingleput study, the case detection was active and passive. "These methods were expected to maintain a continuous surveillance and pick up almost all the cases occurring in the study area. Each round of examination was concluded only after obtaining at least 90% coverage"(1). The investigators estimated 95% success in case detection(1). This community had better TB control program, except for DoT, than what was envisaged even in the RNTCP. So we can have a foretaste of what could be expected if we implemented the RNTCP as if under military-like discipline. During the 15 years of follow up, the incidence of bacillary TB declined only in the previously infected segment of population, but it increased steadily among the previously uninfected segment. Obviously, they continued to be exposed to infection and those who got infected behaved like those in the study cohort found to have been infected at the start of the study. Had re-infection been the major reason for bacillary TB in the previously infected, we would expect the disease incidence among them to increase or at least remain steady. Since the incidence rate among them declined with time, we can surmise that reactivation (with gradual exhaustion of infected stock) was the more likely mechanism of bacillary disease among them, rather than re-infection. It is fairly obvious that even under the best of circumstances TB control under the RNTCP will not achieve results any better than what was achieved in Chingleput. And, what was found in Chingleput after achieving 95% case detection and treatment, was an increase in the incidence of sputum positive TB in those who had not been infected with M. tuberculosis at the beginning of the study, not a decrease.

The continued transmission inspite of early case detection and treatment has important but negative implications on the strategy and tactics of RNTCP.

It is obvious that we cannot control TB by the current methods prescribed under the NTP and RNTCP. The NTP had defined TB control status to fulfill two criteria: one `case' infects less than one person, and, the prevalence of infection declines to less than 1% in children below 14 years. We must examine the technical adequacy of the strategy of control, its management rigour and its ethical standards. We must rethink our TB control strategy.

Where Do We Go from Chingleput? Rethinking TB Control

Technical Adequacy

For human-to-human transmitted infections, on the average one infected person must infect one susceptible person in order to maintain its prevalence. In other words, the effective reproductive rate of an infection, R, must be 1 for a steady state(7). If R is >1, then an epidemic is on; if <1, then incidence is declining. This is the premise of the criterion of less than one infection per case, set for defining control status. The criterion is technically faulty. It is the bacillary case that acts as the source of infection for others. Since every infected person does not progress to bacillary disease, each case must be infecting more than one person, among whom one develops disease, for a steady incidence over time. If the life-time risk of disease is 10% among the infected, then a case is infecting 10 persons now, one of whom will develop disease and R will remain 1 for disease. Since 10 infected persons will result in 10 new infected persons, R is 1 for infection as well. If one case infects less than 10 susceptibles, then the incidence of infection and of disease will decline. If one case infects less than one person, the infection is being eliminated. Thus, the unrealistically ambitious elimination criterion has been applied to control programme, which is untenable from a technical viewpoint. The criterion should be less than one new case for every case that has occurred.

When does a case infect others? While it is true that infectiousness of the index case declines rapidly upon start of chemotherapy, it is not widely appreciated that some of the spread of infection to contacts have already occurred even before disease is diagnosed by detecting AFB in Ziehl-Neelsen stained sputum smear(8). Therefore, case detection (target 70%) and chemotherapy (target 85%), for effective cure of 60% cases (70 times 85% = 59.5%), may not control TB. It is well known that the sputum of a TB patient is first culture positive even as the smear examination is negative; diagnosis by smear examination, although it is the only feasible method currently at the community level, is inadequate for control purposes, unless made more efficient and effective. In Chingle-put, case detection was active, not merely passive; and diagnosis was by culture, not smear examination. Yet, case detection and chemo-therapy did not control infection to control level.

The second criterion of <1% prevalence of infection by 14 years of age requires PPD-S  testing of children from time to time to know where on the path of TB epidemiology we are at. This is not being done in India. From a technical viewpoint, the magnitude of the problem of TB must be looked at as the pool of infected individuals, because it is a proportion among them that will later become infectious cases. The cross sectional rates of prevalence of positive PPD-S reactions have been reported by two groups of investigators(9,10). Below 15 years the rate was 1% or 2%. By 15 years, the rates were 15% to 17%. The rates rose steeply up to about 35 years and reached 79% in one study(9) and 47% in the other(10). Such is the magnitude of the transmission of infection. Un-fortunately, both these studies were conducted over 2 decades ago. We need to assess system-atically the rates of transmission, by age, in all communities, if we take control objectives seriously. If we allow the infected persons to go through the natural history of infection, them some of them will develop adult type bacillary disease and disseminate the microbes in the environment resulting in new infections. In the original NTP strategy, BCG was expected to reduce the risk of bacillary TB, but it does not do so. The only alternative method to reduce the risk of bacillary TB is to give `preventive chemotherapy' to those who have been infected(11-14). In the classical study of pre-ventive therapy in Alaska, subsequent progres-sion to bacillary disease was prevented in 90% and the reduction was sustained over 20 years(11). In other words, only early detection of infection and preventive treatment will achieve what was originally expected of BCG for the reduction of the proportion of infected persons from progressing to infectious form of TB.

Management Principles and RNTCP

In any venture, objecties are defined in such a way that they can be measured for evaluating the efficiency of inputs and effectiveness of  achieving the expected outcome. How can we measure if 70% of all bacillary TB cases have been detected unless we have a fair idea of the total burden of TB in every community? How will we assess if the secondary infection rate is declining from 10 per case, unless methods are designed to test and are applied from time to time? How can we know if infection prevalence at 14 years of age has declined to less than 1% unless it is measured periodically?

Obviously, the international experts who were mainly responsible for setting these criteria were not unaware of these management principles. I believe that they expected that basic primary health care is available in India, in which case all criteria of control and operational objectives could be and would be assessed either regularly or at least periodically. The criteria are meant to be applied, for which a reasonable level of primary health care and referral services must be available. The rest of the world thinks that we have provided basic primary health care, which is the foundation upon which the TB control programme has been envisioned.

The Ethics of TB Control Methods and Criteria

One major ethical issue that must be discussed is whether we should set goals for case detection (target 70%) and treatment (target 85% cure) solely for the public health purpose of disease control, or, if diagnosis and treatment should be made available to all citizens (target 100%) as a matter of right of the individual. Secondly, providing TB control in selected populations and not in every part of the country is unethical, unless primary health care and reasonable services even against TB were available everywhere, but an additional vertical TB control programme was being superimposed on it.

It is not merely for ideological and ethical considerations alone that the above statements are made, but also for very practical reasons. If people have the confidence that the primary and referral health care system will help them in their need of any and all illnesses, then and then only will we be able to diagnose TB in its early phase of symptoms. Passive case detection will not improve until sick people have the confidence to attend our clinics for their needs, not merely for our purpose of disease control by our priorities. Public health measures such as safe water supply and sewage disposal or vector control can be made effective irrespective of the quality of primary health care. On the other hand, for public health measures involving case detection and treatment, such as for the control of TB, malaria, leprosy, sexually transmitted diseases, human immunodeficiency virus (HIV) infection, etc. people's participation and cooperation are essential. For this end, there should be accountability for the local area health care establishment for the health care of the population under its jurisdiction. The referral system must be made accountable so that problematic cases may be promptly attended to by the secondary and tertiary health care institutions nearest to the community. Where private sector institutions are given specific tasks, coordination must be established. India has de-linked public health from health care, but we must link them back together if we must succeed in any specific disease control programme. India remains one of the few countries that have the competence and capability to deliver good primary health care, but has not given it due importance. If anyone has any symptom of illness, not merely cough for over 3 weeks, that person deserves medical attention. In our country wages are still low and investment in health care delivery will increase employment and improve our economy(15).

Improving Sputum Smear Examination

The current process of sputum smear examination by Ziehl-Neelsen staining methods needs revision. It is laborious and it could be improved upon. Although culture may be more sensitive, it is not realistic to expect culture facilities everywhere at the present time. On the other hand, technology has advanced, and sputum smear examination by fluorescence microscopy technicque can be made available at the level of primary health care(16). It will improve the sensitivity and quality of sputum examination. Eventually, sooner than later, every district hospital should have a microbiology laboratory and the capability to culture mycobacteria (and all other pathogens as well). If we do not accept these steps as realistic in the new century, we are resigning ourselves to the despair of unfulfilled plans and programmes in our health sector.

Detecting and Treating Infection

There should be a systematic approach for determining the prevalence and incidence of M. tuberculosis infection among children and adults(11-14). This has to become the corner stone of TB control and of assessing the change in the trend of infection over time, in the community, until it reaches the required less than 1% level in children below 14 years, as defined for control status. We must develop a national consensus as to when and how children and adults should be systematically tested with PPD-S. At the first point in time when PPD-S reaction of a pre-determined size is detected, preventive therapy must be instituted. The exact formulation of PPD-S, the size of reaction for preventive therapy and the choice of drug(s), dose and duration should also be evolved by consensus. Such treatment will reduce the incidence of future bacillary disease among the infected, and also reduce the transmission frequency. This is not only good public health but also good clinical medicine(11-14). As seen earlier, case management alone does not give us the confidence to prevent the spread of infection.

New Inputs for TB Control

India must also upscale the use of information technology in primary health care, disease control programmes and public health in general. Patients diagnosed to have pulmo-nary TB must be registered on computers, irrespective of whether diagnosis was made in public sector or private sector institution. The registry should be managed at the district level. For every patient, the health worker or agency responsible for treatment and follow up must be made known both to the patient and to the health worker or agency. The treatment status, clinical progress and laboratory test results of every patient must be reviewed at regular intervals by computer programming.

A Unique Opportunity

The control of TB offers India a unique opportunity to construct a model of primary health care linked to public health. The control of TB is a barometer to measure the capacity of the health care and public health systems to offer the rural and underprivileged people primary health care as a basic human right. To those who might say that it will not be possible to deliver a basic level of primary health care, as the foun-dation for TB control, let me say that we must then declare this to the international and United Nations agencies and to the public at large.

We must urgently redesign our primary health care infrastructure, if we want to succeed in building primary health care and to link health care to public health. Only on such a backbone will we be able to superimpose the vertical pro-grammes of control of specific diseases such as TB or malaria, of sustainable universal child-hood immunization or of reduction of neonatal and infant mortality rates. We now have the opportunity to bring modern management tech-niques and information technology to our advantage in our health care and public health systems.

The Extreme Urgency of the Situation

HIV infection reached India prior to 1986 and has been relentlessly spreading to new areas and to more people. TB is the commonest serious seconary disease amogn HIV infected subjects. The risk of progression to TB disease is said to be 10% per year in contrast to the life-time risk of 10% among non-HIV infected people. Thus, each year we delay the control of TB, quantitatively the more difficult it will become to achieve control. Moreover, the preva-lence of drug resistance is on the rise(17,18) and it has been succinctly stated:" unknowingly we are transforming an eminently treatable disease into one which is life-threatening and exhor-bitantly expensive to treat"(19).

T. Jacob John,
2/91 E2, Kamalakshipuram,
Vellore, Tamil Nadu 632 002,
E-mail: tjjohn@md4.vsnl.net.in


1. Tuberculosis Research Centre (ICMR), Chennai, Fifteen-year follow up of trial of BCG vaccine in south India for tuberculosis prevention. Indian J Med Res 1999; 110: 56-69.

2. Tuberculosis prevention trial, Madras. Trial of BCG vaccines in south India for tuberculosis prevention. Indian J Med Res 1980; 72 (Suppl): 1-74.

3. Zodpey SP, Maldhure BR, Shrikhande SN, Tiwari RR. Effectiveness of BCG vaccination against tuberculous meningitis: A case-control study. J Indian Med Assoc 1996; 94: 338-340.

4. Park K. Epidemiology of Communicable Diseases. Tuberculosis. In: Park's Textbook of Preventive and Social Medicine, 15th edn. Jabalpur, Banarsidas Bhanot Publishers, 1997; pp 138-151.

5. Khatri GR. National Tuberculosis Control Programme. J Indian Med Assoc 1996; 94: 370-375.

6. Anonymous. Operational Guidelines for Tuberculosis Control. Central TB Division, Directorate General of Health Services, New Delhi, May 1997.

7. Anderson RM. Directly transmitted viral and bacterial infections of man. In: Population Dynamics of Infectious Diseases. Theory and Applications. Ed. Anderson RM. London, Chapman and Hall, 1982; pp 1-37.

8. Kamat SR, Dawson SJY, Devadatta S. A controlled study of the influence of segregation of tuberculosis patients for one year on the attack rate of tuberculosis in close family contacts in south India. Bull WHO 1966; 34: 577-632.

9. John TJ, Frimodt-Moller J, Feldman RA, Jeyabal P, Kamath KR. Infection and disease in a group of south Indian families. Part 13. Skin sensitivity to 6 mycobacterial antigens. Indian J Med Res 1971; 59: 1727-1736.

10. National Tuberculosis Institute, Bangalore. Tuberculosis in a rural population of south India: A five-year epidemiological study. Bull WHO 1974; 51: 473-488.

11. Comstock GW, Baum C, Snider GE Jr. Isoniazid prophylaxis among Alaskan Eskimos: A final report of the Bethel isoniazid studies. Amer Rev Respir Dis 1979; 119: 827-830.

12. Dash LA, Comstock GW, Flynn JPJ. Isoniazid preventive therapy. Retrospect and prospect. Amer Rev Respir Dis 1980; 121: 1039-1044.

13. American Thoracic Society, Centers for Disease Control and Prevention. Treatment of tubercu-losis and tuberculosis infection in adults and children. Amer J Respir Critic Care Med 1994; 149: 1359-1374.

14. Centres for Disease Control. The use of preventive treatment for tuberculosis in the US. Recommendations of the Advisory Committee for eliminaton of tuberculosis. Morb Mort Weekly Rep 1990; 39 (No. RR 8): 9-12.

15. Sen A. Health in development. Bull WHO 1999; 77: 619-623.

16. Toman K. Tuberculosis. Case finding and chemotherapy. World Health Organization, Geneva and Jaypee Brothers, New Delhi, 1989; pp 1-72.

17. Chandrasekaran S, Jagota P, Choudhuri K. Initial drug resistance to antitubercular drugs in urban and rural district tuberculosis programme. Indian J Tubercul 1992; 39: 171-175.

18. Jain RF. Faulty prescription, an avoidable cause of multidrug resistance in tuberculosis. J Indian Med Assoc 1996; 94: 385-388.

19. Jain NK. Drug resistance in India. A tragedy in the making. Indian J Tubercul 1992; 92: 145-148.

Key Messages

1. BCG vaccination does not protect against infection by Mycobacterium tuberculosis, or against the development of secondary pulmonary cavitory (bacillary) tuberculosis (TB). Since bacillary disease is the source of infection to new subjects, BCG offers no contribution towards the control of TB.

2. As human immunodeficiency virus infection increases several-fold the risk of developing bacillary TB, and as multi-drug resistance of M. tuberculosis is on the increase, TB control must be given high priority and urgent attention. Delay may make TB as uncontrollable as AIDS is today.

3. Alternate tactic is necessary to reduce the risk of developing bacillary disease, in order to help control TB. The only available intervention is to give preventive chemotherapy to those with inactive infection. This method has to be included in the national TB control programme. It must also become standard medical practice.

4. The target approach of the national TB control programme must be balanced with the provision of total primary health care to all people, in order to attract all sick persons who may have TB or other diseases of public health importance. The sputum smear examination for acid fast bacilli must be made more efficient and sensitive. Anti-TB treatment must become a routine part of primary health care, which should be supported by modern information technology.

5. BCG vaccination of infants/children must continue since TB is uncontrolled and since it is the only available intervention to reduce the risk of primary infection progressing to disease at distant sites such as meningitis, miliary, bone TB and scrofula. The Chingleput BCG trial did not investigate this issue, nor was it intended.