Dr. PERKINS. I am on the board in Chicago, and we are trying to increase the use of BCG but confine its use to those particularly exposed. We feel it should be used more extensively among those particular groups.

Dr. LONG. In the list I read of desirable research I mentioned a search for a better vaccine than BCG. BCG, I think we will all agree, does some good, but everything that has been said so far in regard to the two groups-and, incidentally, it was a fine study in connection with the Indians-everything indicates we do not know everything we should know about vaccination, and there have been many suggestions that we can get something better than BCG. The National Tuberculosis Association has an investigation in which a somewhat similar organism derived from a human being rather than a cow offers a chance to get a better vaccine than BCG.

The CHAIRMAN. I had not been aware there was such a vaccine as BCG. I thought I had had every kind of vaccine there was, but I had never heard of this one.

Mr. CARLYLE. Mr. Chairman.

The CHAIRMAN. Mr. Carlyle.

Mr. CARLYLE. This BCG vaccine is comparatively new, is it not? Dr. PERKINS. No. It is old. It was first used in humans about 1920. Mr. CARLYLE. It is used extensively all over this country? Dr. PERKINS. It is used in practically every part of the country, but confined pretty much to these people undergoing unusual exposure. (The following material was submitted for the record regarding BCG vaccine:)

(By Carroll E. Palmer and Lawrence W. Shaw')

At the annual meeting of the National Tuberculosis Association 3 years ago, Dr. Robert Anderson' discussed the attitude of the Public Health Service toward BCG vaccination. He pointed out the wide differences of opinion on many questions about BCG and its use; the recurrent theme throughout his talk was how little we really knew about BCG-how little solid evidence there is amid all the claims and counterclaims of its value.

Because of this, Dr. Anderson said the Public Health Service could recommend the use of BCG only for selected groups with known high risk of exposure to tuberculous infection. Tuberculosis had been declining sharply for many years and there was no reason to think that it would not continue to decline-perhaps even more rapidly. Moreover, we were not convinced that the advantages of using BCG would outweigh the disadvantages of losing the tuberculin test as a diagnostic and case-finding tool. We could find no justification for mass vaccination programs in this country, except for the purpose of making carefully documented studies to determine the value of BCG in tuberculosis control.*

That is not a recommendation one makes lightly, in the face of strong popular belief in BCG and its widespread use abroad. It is not a recommendation one makes without feeling an obligation to make known any indication that it should be changed. But about the only reason for changing the recommendation would be decisive evidence of the effectiveness of BCG in human beings. A number of groups in different parts of the world, including the Public Health Service, are actively trying to find such evidence. It is our purpose in this paper to describe very briefly the several large-scale evaluation studies undertaken by the Service and to present the first preliminary results of these studies.

1 Read at the annual meeting of the National Tuberculosis Association, May 20, 1953, Los Angeles, Calif.

Medical Director and Statistician, respectively, from Field Service Branch, Division of Chronic Disease and Tuberculosis.

Anderson. Robert J.. and Palmer, Carroll E.: BCG, Journal of American Medical Association, 1950, 143, pp. 1.048-1,051.

Anderson. Robert J.: Editorial on Licensure of BCG Vaccine, Public Health Reports, August 4, 1950.

In planning the studies we had to make decisions about several very critical issues.

In the first place, we decided that it was absolutely necessary to have control studies, and by that we meant we would deliberately not vaccinate a randomly selected part of the population found eligible for vaccination by the tuberculin test. By so doing, we would have two comparable groups: one vaccinated, the other not vaccinated; and we would measure the effectiveness of BCG by comparing the amount of tuberculosis that developed in the two groups. We decided we could not compromise on that issue. We would not, as some have tried before, vaccinate all who were willing to be vaccinated and use those who refused as controls. Nor would we compromise and vaccinate all who were eligible and expect to measure the effect of vaccination by the change in tuberculosis morbidity and mortality.

Another important decision concerned the size of the studies. With the steady drop in the prevalence of tuberculosis-only with large studies could we expect enough cases of tuberculosis to indicate what BCG could do on a mass basis for general population groups.

This aspect led to a third-the matter of evaluation. Very few people nowadays have the time, the staff, or the money to do regular, periodic examinations of large population groups. Evaluation therefore would have to depend primarily on the established systems of reporting tuberculosis deaths and cases.

Finally, in order to determine the impact of the BCG program on the total tuberculosis problem, it would be necessary to know about the tuberculosis occurring among not only the vaccinated and controls but also among the tuberculin reactors who were ineligible to receive BCG.

Thus, we agreed that the main characteristics of the studies should be: first, that they would be control-studies (some people would be vaccinated, some not). Second, that they would be very large-scale studies. Third, that followup would be made an integral part of the usual system of notification of tuberculosis cases and deaths. Fourth, that both those eligible for vaccination and those not eligible would be included in the followup. Such studies, we fully recognized, might be criticized because the evaluation depends so heavily on mortality and morbidity reporting. However, we are dependent on just such information to measure all our efforts in tuberculosis control. After all, if BCG vaccination in a population is not effective enough to be recognized through reasonably efficient reporting of tuberculosis cases and deaths, its widespread use can hardly be advocated.

The first study was started in 1947 in Muscogee County, Ga.-a community with a population of about 100,000 persons. In the spring of 1947, the entire school population was tuberculin tested, and half of the nonreactors were vaccinated. Three years later, in 1950, the BCG program was made part of a mass survey of the whole population who were invited to have an X-ray, tuberculin test, and BCG vaccination. About 10,000 school children were included in the early program; about 60,000 of the general population were added in 1950. Muscogee County is a reasonably typical community with a tuberculosis mortality rate somewhat below the rest of the country. The program there is directed by a full-time Public Health Service physician who, in cooperation with the practicing physicians, is able to insure accurate diagnosis and efficient reporting of cases and deaths from tuberculosis.

Our second program is among the American Indians, where the prevalence of tuberculosis is many times higher than in the rest of the country. The field work of this cooperative undertaking was carried out in 1949 by the Bureau of Indian Affairs and included a high proportion of the Indian children attending Federal and mission schools in the continental United States. The number of children under study is 27,000.

The third program is in Puerto Rico where the field work of testing and vaccinating was begun in the fall of 1949 and completed in the spring of 1951. It is a cooperative undertaking with the insular government and was offered to all of the estimated 400,000 children attending school on the island. Although the field teams visited almost every school on the island, only about one-third of the expected number actually participated. Even so, we expected that the high tuberculosis mortality rate in Puerto Rico would give, fairly soon, an indication of what BCG could do. The total number of children under study is 165,000, including about 18,000 preschool children.

A fourth study was undertaken in a special group-20,000 mental patients of all ages in State institutions in Ohio.

Table 1 shows how the population in each study was subdivided into three groups; the tuberculin reactors, the vaccinated, and the controls. In the first

three studies, about half of the populations were tuberculin reactors and not eligible for vaccination, leaving the other half to be divided into the vaccinated and control groups. The division, however, was different in each study: in Georgia every other person was vaccinated, in Puerto Rico 2 out of 3 were vaccinated, and for Indians 3 out of 4. Of the 20,000 Ohio mental patients, three-fourths were already reactors, leaving just over 5,000 eligible for vaccination. In this group 6 out of 7 were vaccinated.

The total number of persons included in the four studies is 282,000: 87,500 vaccinated 50,500 controls, and 144,000 tuberculin reactors. The nonreactors were subdivided into vaccinated and control groups by strict unbiased methods which insure that the two groups are entirely comparable. The stage has thus been set for several large-scale tests of what BCG could accomplish as a broad public-health measure.

You can well believe that it has been no small task to carry the studies this far-to tuberculin test and vaccinate as carefully and precisely as we knew how, to make accurate individual records, and to set up the necessary statistical files or rosters. The followup phase of the studies now going on consists of matching current tuberculosis mortality and morbidity reports with the cards in the rosters and, insofar as possible, to check and verify the accuracy of these reports.

Let us turn now to what we hopefully expected would be, by 1953, a fairly clear indication of the value of using BCG in a number of different situations and populations. The results, in some respects, are both striking and informative. All four studies show that a very large part of the tuberculosis appearing in the study groups comes from among those who were already infected—the tuberculin reactors who could not be vaccinated. In other respects, the results are inconclusive. Little tuberculosis has appeared in those eligible for vaccination, whether or not they had been vaccinated.

The material we have collected is summarized in table 2.

In Muscogee County there have been only three deaths from tuberculosis in the study population-all of them among the tuberculin reactors. The morbidity has also been remarkably low-only 25 new cases of tuberculosis have occurred— 20 in the 3 years of followup of the older group and 5 in the 6 years of followup of the original school group. The noteworthy finding is that the bulk of the new cases has occurred among those not eligible for vaccination-very little tuberculosis has appeared among the nonreactors. The five cases among the vaccinees and controls have occurred in the large group who entered the study in 1950there have been no cases of tuberculosis in 6 years among either the vaccinated or control-school children.

Results from the Indian study are limited to information on deaths. There have been 13 tuberculosis deaths among the 12,000 tuberculin reactors. Only one tuberculosis death has been reported for the 3,500 controls, and, if this rate were applied to the vaccinated, we would have expected 3 deaths, and actually there was none. No one can draw conclusions on so few cases, but the results do not disagree with the view that vaccination is of value.

In Puerto Rico we also find a considerable concentration of tuberculosis mortality in the tuberculin reactors-24 deaths during the first few years of the program. In this study, however, we have a little stronger indication that BCG may be of value. There have been 4 deaths in the controls, and only 1 in the vaccinated group which contained twice as many children. If we apply to the vaccinated the mortality rate of the controls, 8 deaths would be expected but only 1 occurred. The numbers, of course, are very small, but they do suggest that the vaccinated are faring better than the controls.

Tuberculosis morbidity statistics in Puerto Rico are somewhat deficient in the completeness of the reporting and the accuracy of diagnosis, but we have no reason to suspect any bias related to the fact of BCG vaccination. For the present paper we have only a very limited, but presumably unbiased, batch of morbidity reports. They show 6 new cases of tuberculosis among the vaccinated, 2 among the controls, and 75 among tuberculin reactors. Here again we find most of the cases in the reactors. Unlike the mortality data, the morbidity material does not show signs of the vaccinated being benefited by their vaccination.

Among Ohio mental patients we find essentially no difference between the vaccinated and controls with respect to tuberculosis mortality: 1 death among the 700 controls and 4 among the 4,500 that had been vaccinated. As in the other studies, the great majority of the deaths from tuberculosis have occurred among the tuberculin reactors.

New cases of tuberculosis in the Ohio study numbered 43 in the vaccinated and 6 in the controls. If the morbidity rate among the vaccinated had been the

same as in the controls, 36 cases would be expected in the vaccinated; thus the morbidity rate in the vaccinated is actually somewhat higher than in the controls. Again the bulk of the reported cases has occurred among the tuberculin reactors, 286 out of the total of 335 cases reported so far.

What conclusions can be drawn now from all of this work?

First, and all four studies agree on the point, a very large proportion of the tuberculosis that appears in a population during the first few years after a vaccination program evidently occurs in the group who would not be vaccinated because they had already been infected. This implies that mass vaccination, even if BCG is effective, cannot be expected to have a large and immediate influence on tuberculosis mortality and morbidity rates.

The second major point more or less follows, and indicates why we are still so uncertain about the effectiveness of BCG: so little tuberculosis appears among those eligible for vaccination, whether or not they are vaccinated. It is possible that in the Indian and Puerto Rican mortality data we may be seeing the early signs that BCG, under certain circumstances, could be useful in tuberculosis control. On the other hand, it is also true that none of the morbidity comparisons shows any evidence of a beneficial effect of BCG.

The third point is that our studies do not indicate that tuberculosis in this country would be more effectively controlled by adding mass vaccination programs. Muscogee County in Georgia may not be entirely representative of the whole country, but since the effect of the vaccination program there is imperceptible, we have little reason to expect very different results in most of the rest of the country.

There has been time today only to tell very briefly about the Public Health Service studies. Neither these results, nor those in recent reports by others, indicate a need for changing the Public Health Service recommendations of 3 years ago.

TABLE 1.-Location and size of the BCG control studies of the Public Health Service

TABLE 2.-Tuberculosis cases and deaths reported by December 1952, among persons in Public Health Service BCG control studies

(Editorial reprinted from Public Health Reports, August 4, 1950, by the Department of Health, Education, and Welfare, Public Health Service, Washington 25, D. C.)

On July 12, the Public Health Service licensed the Research Foundation and the University of Illinois for "manufacture, exportation, importation, and sale"

of BCG. Until now, licensure of the product has awaited manufacture in accordance with certain requirements. In view of the divergence of opinion about this biological product, it seems in order to consider the significance of such action. It means that the vaccine produced by the licensed laboratory has been found safe by trial with animals, that it is free from contaminating substances, and that it will produce a satisfactory immediate reaction in animals and human beings when used within the prescribed time limit. Thus, the vaccine may enter interstate commerce and will be available to health officers and clinicians who wish to use it as a protective measure against tuberculosis.

In those places of the world where tuberculosis is a national emergency and where prosecution of the usual control methods is impossible, it is understandable that BCG has been given extensive application. In this country, where we are not faced with the same deficiencies, the medical, profession for the most part has not advocated the widespread usage of the vaccine. The Council on the Management and Treatment of Diseases of the Chest, reporting for the American College of Chest Physicians, has recommended that the use of BCG vaccine be restricted to controlled studies (1). The American Trudeau Society (2) recommends that the use of BCG be limited to groups especially exposed to the risk of tuberculous infection.

The Public Health Service, like others concerned about tuberculosis, would welcome any agent which would prevent the personal tragedy and public health problem of tuberculosis. But it has not yet been conclusively demonstrated that BCG is such an agent. Moreover, efforts to find more stable and suitable immunizing agents are going forward. Indiscriminate use of BCG here could, we believe, not only negate its potential future application but might divert attention from the control activities which are serving the Nation well and which, under the circumstances prevailing in the United States, could lead to the virtual eradication of tuberculosis. It is our feeling that we must be very careful not to imperil the gains we are making with proved control methods and must not relax in any area the pursuit of case finding and treatment to care for the sick and to protect the well.

If the use of BCG in the United States is to contribute more information than has been gained in almost 30 years of use elsewhere, vaccination programs must be carefully planned. It would be desirable if State and local health departments which are immediately responsible for tuberculosis control were to develop plans for the use of the vaccine in their jurisdictions and keep records of those who are vaccinated. A beginning has been made in Wisconsin where the State health department has reviewed all requests for the vaccine desired from research laboratories, and in New York where the State department of health has manufactured BCG vaccine and has kept records of persons in the State who were vaccinated.

We feel that mass BCG vaccination campaigns are not indicated in this country where tuberculosis morbidity and mortality rates are relatively low (3). It is our recommendation that vaccination be limited to those persons who are particularly vulnerable to exposure. These include:

1. Those physicians, nurses, laboratory workers, hospital employees, and others who are exposed by occupation.

2. Those individuals or groups exposed to continued contact with tuberculosis. 3. Patients, inmates, and employees of institutions, such as mental hospitals and prisons, in which case-finding programs indicate that exposure to tuberculosis is likely to be high.

ROBT. J. ANDERSON,

Medical Director, Chief, Division of Tuberculosis.

REFERENCES

(1) Joint meeting of Council on the Management and Treatment of Diseases of the Chest and Council on Public Health of American College of Chest Physicians, San Francisco, June 22, 1950.

(2) ATS Statement on BCG. Am. Rev. Tuberc., 60: 681-682 (1949).

(3) Anderson, Robt. J., and Palmer, Carroll E.: BCG. J. A. M. A., 143: 1048-1051 (July 22), 1950.

Mr. CARLYLE. Are the reported cases of tuberculosis in this country on the increase or decrease?

Dr. PERKINS. The reported cases are about uniform. I think the rate of development of tuberculosis is going down but with our im