A Booster for Tuberculosis Vaccines

In 1935, a cohort of Native Americans began to enroll in an experiment designed to address a topic about which everybody was unsure, if not entirely ignorant: the efficacy of BCG vaccination against tuberculosis (TB). In the early 1930s, 2 decades before the advent of combination chemotherapy, 65 of every 10 000 Alaska Natives and Eskimos died of TB each year.¹ For those unfortunate enough to develop active disease following infection, the principal remedies were fresh air and hope. With a case-fatality rate of about 50%, the prognosis for patients with TB was grim.

The BCG vaccine had been developed by 1921, and a collection of experiences in the 1920s suggested that it could markedly reduce both TB incidence and mortality. But doubts remained because no large, properly controlled prospective studies of its efficacy had been conducted. Given the early promise of BCG, the new series of immunization trials began with great expectations.

Between 1935 and 1938, Aronson and collaborators recruited about 3000 American Indians and Alaska Natives aged 1 month to 20 years to participate in a placebo-controlled trial carried out across 6 US states. Prospective case finding continued until 1947. With the latest efforts to trace participants reported in this issue of THE JOURNAL,² surviving members of the cohort have now been followed up for more than 60 years. The investigation has been handed down 2 generations, from Aronson grandfather to granddaughter. From 1948 to 1998, covering more than 100 000 person-years of observation, 36 cases of pulmonary and extrapulmonary TB occurred in the vaccinated group and 66 in the control group, giving respective incidence rates of 66 and 138 per 100 000 per year and a crude (unadjusted) vaccine efficacy of 52%.

This is easily the longest recorded period of significant protection conferred by BCG vaccine. The results appear to be a boost for TB vaccination in general, but with several important qualifications. It is reassuring that BCG can still provide durable protective immunity against both pulmonary and extrapulmonary TB. The caveats are about the conditions under which BCG, and perhaps other TB vaccines, can provide lasting protection. The fact that Aronson and colleagues now report a longer duration of protection than previously documented will reopen the debate about why BCG efficacy has been so variable.

The consensus from a large number of trials is that BCG has consistently high efficacy (mostly in the range of 65%-90%)^{3 - 4} in protecting against severe forms of TB in infants, principally meningitis and miliary disease; roughly, the 100 million doses given around the world each year prevent an estimated 50 000 cases of severe disease in children.⁴ By contrast, the efficacy against disease in older children and adults, evaluated for varying periods, has ranged from zero (or even potential harm) to more than 80%.³ Many explanations have been offered for these mixed results, including differences in trial methods, in vaccine strains and doses, in the risk of reinfection following vaccination, in the risk of disease from recent rather than historical infection, in the health and sex of vaccinees, and in the prevalence of coinfection with helminths or environmental mycobacteria.³ Infection with environmental mycobacteria may provide some protection against TB while diminishing the efficacy of BCG.⁵

Among this long list of possibilities are several factors that could help to explain why the efficacy reported in the current study by Aronson et al was not lower. First, the participants in the trial probably had relatively low exposure to environmental mycobacteria. In particular, the trial excluded children who responded to a strong dose of tuberculin, among whom would have been those infected with other mycobacteria. Second, although the trial used 2 different strains of BCG vaccine, neither of the strains appeared to have very low efficacy, which could have obscured a significant result overall. Third, the trial began when reinfection of vaccinees was much more frequent than it would be today, possibly boosting the protection initially obtained from BCG.

In contrast, at least 2 reasons help to explain why the long-term efficacy reported by Aronson et al was not higher. Efficacy showed some signs of waning in the later years of the trial, when TB incidence was lower and a higher proportion of cases would have arisen from reactivation of old infections rather than the rapid progression of new ones. It remains unclear whether BCG vaccine is less efficacious in preventing disease due to reactivation or whether protection against any form of disease simply weakens with time. Moreover, the findings suggest that the proportion of vaccinees protected would have been greater if the trial had included only women. The best estimate of efficacy for women was 70%, compared with 29% for men. This difference between men and women also remains unexplained.

While the results of this latest report are different from those obtained in previous studies, they are not inconsistent with current thinking about why BCG vaccine efficacy is so variable. The findings also leave a checklist of questions for those who are now trying to develop new and better TB vaccines. Among a growing list of new candidate vaccine antigens,⁶ 2 of the most promising are now undergoing phase 1 safety trials in humans. One is a live attenuated BCG bacterium (rBCG30) that overexpresses an antigen 85 protein and in guinea pigs provides greater protection than BCG alone.⁷ The other is a fusion protein of 2 different antigens that could be used with an adjuvant as a booster to either BCG or to rBCG30.⁸ As these early safety trials proceed, vaccinologists will be asking whether a new vaccine can prevent infection as well as the progression to disease, whether it can protect against pulmonary as well as extrapulmonary disease, whether it can work well in the presence of nontuberculous mycobacteria, whether it will need to be boosted naturally or artificially and how often, and whether efficacy will differ between children and adults and between men and women. There is and will continue to be active discussion about how TB vaccine trials can be designed.^{9 - 10} The trial by Aronson et al is a sharp reminder that, even if a new vaccine does have consistently high and durable efficacy, it could take years of field studies before its full potential is understood.

For every reason it might be difficult to make a new vaccine, there are more reasons to persist with the effort. The biggest advance in TB control since the start of the trial by Aronson et al has not been immunization but rather the introduction of combination chemotherapy. The vast majority of patients with active TB can still be cured with current drugs. During 2001-2002, approximately 1 million sputum smear–positive patients were either cured or completed treatment with short-course chemotherapy under the internationally recommended strategy for TB control, known as DOTS (directly observed therapy, short course).¹¹ However, these patients represent only one third of the increasing number of new TB cases that emerged worldwide during 2001.¹¹ These incident cases were identified with an ancient diagnostic technique (sputum smear microscopy) that predates BCG and all of these cases had to be treated for at least 6 months, during which 12% were lost to follow-up.¹¹ Outside of DOTS programs, default and death rates are typically higher and cure rates are lower. Moreover, as surveys continue to map the distribution of drug-resistant TB organisms,¹² most physicians and other health care workers are acutely aware that all of the principal antituberculosis drugs were introduced during the 1950s and 1960s. No new drug has been added to the first-line treatment regimen for TB for more than 30 years.

The remarkable results of the 60-year follow-up study by Aronson et al still leave plenty of questions about vaccination unanswered. Nonetheless, these findings are relevant to the well-being of millions who are still threatened by or who have TB and provide important lessons for future TB vaccine development.