BCG vaccine weakens over time, researchers say

Weakening strains underscore need for new vaccine

Three years after a meta-analysis estimated that the only vaccine for tuberculosis, Bacille Calmette-Guerin (BCG), was on average 50% effective, a leading TB researcher suggests the various strains have weakened considerably to lower levels. While that is bad news to the many countries that depend on the vaccine for TB prevention, the reasons for the attenuation could offer insights into developing a new, more effective vaccine.

In plotting the efficacy and tuberculin reactivity of BCG strains used in clinical trials against their number of in vitro passages, researchers at Stanford University in San Francisco found that four out of five strains appear to have lost their efficacy while maintaining their reactivity, according to a letter in the Sept. 11 issue of Nature.1

The finding is not surprising, considering that developers of the original vaccine used a strain that had gone through 231 passages before it was weakened enough to be considered safe. "It is not surprising that protective efficacy of BCG waned over the next 1,000 passages," writes Peter Small, MD, assistant professor of medicine at Stanford University Medical Center and lead author of the letter. "We hypothesize that the attributes of the current BCG have resulted from subtle pressures to minimize adverse reactions and maintain tuberculin reactivity during vaccine development and testing."

The pressure to minimize adverse reactions grew as tuberculosis mortality decreased and the public became less accepting of typical BCG side effects, which often include scars from ulceration. Subsequent offspring of the strains differ in mycolic acid composition, protein transcription, and genotype, Small notes.

"If you look at efficacy as a function of passage number, then the efficacy in all the clinical trials decreases, and that is what the correspondence reports," Small tells TB Monitor.

George Comstock, MD, professor of medicine at Johns Hopkins University and a leading authority on BCG, agrees that the strains used in clinical trials do appear to have decreasing efficacy. "It is true. If you go down the trials and look at them in order, there is an indication of a downward trend in efficacy," he says. "We not only need a better vaccine, we need to seriously think about how we would know if we had a better one if we had it."

For years, researchers have disagreed over just how well BCG protects against TB, with estimates ranging from 0% to 80%. Most studies have shown that it offers some protection for children at least, particularly childhood TB meningitis. Many developing countries, too poor to afford adequate TB control and treatment measures, routinely vaccinate children. In Europe and other developed regions it has been limited to high-risk children and health care workers. For example, Sweden used to vaccinate 95% of all newborns. Routine vaccination was stopped in 1975, however, because of side effects and lower risk of tuberculosis exposure. Only 4% of Swedish-born children now receive BCG.2

Three years ago, a meta-analysis conducted by researchers at Harvard University came to the conclusion that current BCG strains conferred, on average, a 50% level of protection. Both Comstock and Small question the validity of the analysis. "I think it may not be coincidental that if you average a group of random numbers, you end up with 50% — that is the cynic in me," Small says.

Unfortunately, a meta-analysis may be the only approach to determining the actual benefit of BCG. The reason is an ethical one. Conducting a controlled, double-blind trial would be difficult to justify considering that health officials in most countries where such a trial would take place believe BCG offers some protection, Comstock says.

"You might be able to pull [a trial] off, but is it ethical to use a placebo when they believe that BCG is helpful?" he explains. "I think a lot of countries where the trial would be useful would say, ‘We aren’t going to let our kids be unvaccinated.’ So you are stuck comparing two vaccines."

An opportunity to learn about protection

While Small’s finding may be disillusioning and will only add mud to the pool of BCG studies, his hypothesis has several implications. First, he notes, efficacy estimates for BCG may be overstated, and rational vaccine policy dictates an accurate assessment of current strains. Second, "comparison within strains of recent and more ancestral forms (such as retention lots from vaccine trials) may provide insight into vaccine characteristics that correlate with efficacy and tuberculin reactivity."

And finally, he notes that until researchers learn more about the correlates of immunity and determinates of virulence, increased vaccine efficacy will likely result in increased adverse reactions.

While the differences in BCG strains may account for differences in efficacy, Comstock cautions that other theories exist. One of the more plausible ones is that non-tuberculosis organisms may interfere with its protection. "I don’t think that has been disproven by any means, but personally, I think strain differences are more tenable," he says, adding that "both hypotheses might prove to be true."

In a related study on the effects of BCG vaccination, researchers from Sweden have attempted to answer whether it can increase the risk of atopy or allergies in children. A recent study in the Lancet suggests that vaccination doesn’t have an impact on atopic disease before children reach school age.2

For unknown reasons, atopic disease is increasing in developed countries, reaching as high as 30% in children with asthma, atopic dermatitis, rhinoconjunctivitis, or food allergy. Because BCG modulates the immune response, it has been thought to play a part in the development of atopic diseases.

Researchers assessed the influence of BCG on the development of atopy in children with atopic heredity. Specifically, it evaluated 780 children born between 1989 and 1992 who received BCG before the age of six months. The study found that a positive history or clinical signs consistent with atopy were found in 36% of those children given BCG compared to 41% of the control group. The distribution of atopic symptoms in the two groups did not differ significantly, the researchers report.

Comstock says the findings fit with his observations that any association between BCG vaccination and atopy is small, if not insignificant.

The study shows that early BCG vaccination in infants with atopic heredity did not affect development of atopic disease before school age. If the findings hold true for older children and children without a predisposition to atopy, vaccination may not be an effective prevention strategy for atopy, the authors note.


1. Small P, Behr M. Has BCG attenuated to impotence? Nature 1997; 389:133-134.

2. Alm J, Lilja G, Pershagen G, et al. Early BCG vaccination and development of atopy. Lancet 1997; 350: 400-403.