Examining the BCG vaccine: a remarkable follow-up study

July 2004

For reasons that I don’t fully understand, I’ve always had more than a passing interest in bacille Calmette Guérin (BCG) vaccines, even beyond my interest in vaccines in general. Perhaps this started when I matriculated at what was then known as Western Reserve University School of Medicine in 1953.

Those concerned about medical student health at that time had a healthy respect for tuberculosis (TB) as an occupational hazard of physicians. One approach taken to minimize this hazard was to immunize all tuberculin-negative freshman medical students with BCG vaccine. I don’t know how widespread this practice was at U.S. medical schools at the time, but it was standard operating procedure at Western Reserve.

Those like me, who failed to convert their skin tests six weeks after immunization, were treated to a second dose of BCG; six weeks later I had still stubbornly failed to convert, at which point it was concluded that I must have some as yet unknown innate immunity.

TB control in the United States has been based on case detection, prompt treatment of clinical cases and preventive therapy of skin-test converters; clearly, it has generally worked superbly, and for that reason BCG vaccine has never played a significant role in TB control in this country. In many parts of the world, however, BCG vaccination of infants continues because of its demonstrated efficacy in preventing lethal systemic or central nervous system infections in children. One estimate is that 100 million doses each year are still given to infants for this purpose.

The occasion of this issue of Infectious Disease News, devoted to adult immunization issues, seems an especially opportune time to comment on a highly unusual follow-up study of BCG vaccine, published in the Journal of the American Medical Association (JAMA) on May 5, 2004. In setting the scene for this study, recall that there have been only a very limited number of controlled field trials of BCG vaccine, and that these have yielded widely divergent results.

There have been eight major controlled trials, which have fairly consistently shown efficacy in preventing miliary and meningeal TB in children. Efficacy in preventing pulmonary disease, however, has been quite variable. Most of the controlled trials were carried out between 1930 and 1960. Because of the divergent results of previous trials, one last controlled trial was undertaken, the Madras, India, trial, widely acclaimed as “the BCG trial to end all BCG trials.” (Fortunately, that will likely not prove true.) That trial failed to show efficacy in preventing pulmonary disease.

The first of the early studies was carried out in American Indians and Alaska natives by Townsend, Aronson and others and reported in the American Review of Tuberculosis in 1942. A 20-year follow-up of that study, published by Aronson et al in 1958, revealed an 82% efficacy in preventing TB mortality and a 75% efficacy in preventing x-ray–diagnosed TB. This was clearly one of the studies, probably the strongest one, to show efficacy in preventing pulmonary disease.

All the preceding is a lengthy introduction to the May 5, 2004, JAMA report, the senior author of which is the granddaughter of J.D. Aronson, DO, who carried out the original studies. This paper reports an astonishing 60-year follow-up of the original study population. That population, enrolled in 1935-1938, consisted of 3,025 American Indian and Alaska Native children and adults, aged 1 month to 20 years. The 60-year follow-up study investigators, working from 1992 to 1998, were able to obtain information on an astonishing 85% of study participants! Although efficacy appeared to wane somewhat from that reported in the 20-year follow-up, it was still 52%, with a 95% confidence interval of 27% to 69%. Thus, there was evidence of efficacy that persisted for 50 to 60 years, truly a long duration of protection.

The most astonishing aspect of this report for me is that it could be done at all. Any clinical investigator who has carried out follow-up studies, or tried to, can appreciate the enormous effort that must have gone into this study. It would be possible perhaps only in a unique population such as this that is generally less mobile than usual and has generally only a single central source of medical care. This report is a remarkable accomplishment.

As pointed out in an accompanying editorial by Christopher Dye, DPhil, of WHO, this is easily the longest reported follow-up of protection afforded by BCG vaccine. Publication of this report may have other, mostly salutary, effects as well.

Dye noted there likely will be a resumption of the debates about the reasons for the divergent results of the BCG field trials. Several explanations for the divergent results have been proposed in the past, including variation in vaccine strain potency, lack of protective effect against the “endemic” strain, the confounding effect of exposure to environmental mycobacteria, and that TB disease in the trial area was due more to exogenous reinfection than to endogenous reactivation. Each of these hypotheses may contain some kernel of truth, but none of them has emerged as a single dominant explanation.

Some years ago, the late Alvan Feinstein, MD, and colleagues from Yale University published a study (JAMA. 1983; 249:2362-2369) that ranked the several, field trials according to the authors’ assessment of the scientific and statistical quality of the study, freedom from biases, validity of the case ascertainment methodologies and the like. They concluded that the trials with the best methodological quality and the greatest statistical precision reported the highest efficacy and suggested bias or inadequate statistical power likely contributed to the divergent results of other trials. This report pleased few, but perhaps may have contributed in some way to the CDC’s decision to re-look at BCG vaccines more broadly by awarding a contract to the Harvard School of Public Health to carry out a meta-analysis of BCG vaccine trials. This report (JAMA. 1994:271;698-702) found an overall 51% efficacy against all forms of TB and a 71% efficacy against death due to TB.

Meanwhile, as we all are aware, the world of TB has changed dramatically in the last 20 years due to the emergence of the HIV/AIDS pandemic and multidrug resistance. The desirability of an effective immunologic approach to TB control is thus even greater now than ever before. There is much promising research ongoing to develop second-generation TB vaccines that might offer broad and durable protection. I believe the Aronson report will provide strong encouragement to that group of scientists. Although we may indeed have seen the end of BCG trials, I truly believe we are not anywhere close to the end of TB vaccine trials!