sciencebriefs
13:00in productionCh. 1 · Weakening a killer, one culture at a time/ 13:00 · ceiling 15 min
Medicine

BCG vaccine

Albert Calmette and Camille Guerin spent thirteen years weakening a tuberculosis strain into the BCG vaccine, first given to a child in 1921, whose protection has since proved real but strikingly inconsistent from one country to the next.

Between 1908 and 1921, Albert Calmette and Camille Guerin at the Pasteur Institute repeatedly subcultured a virulent tuberculosis-related bacterium, 239 times over thirteen years, until it could no longer cause disease in laboratory animals, producing the strain known as Bacille Calmette-Guerin, or BCG. First given to a newborn in Paris in 1921, the vaccine survived a serious setback in 1930 when a contaminated batch in Lubeck, Germany caused tuberculosis in over a hundred infants and killed dozens, a disaster that weighed on Calmette until his death three years later. BCG remains the only licensed tuberculosis vaccine and is used across most of the world, but trials have found its protection ranges from strong and long-lasting in some countries to negligible in others, a variability still not fully explained.

Chapters & takeaways6
  1. 0:08
    Weakening a killer, one culture at a time

    Calmette and Guerin subcultured a virulent tuberculosis-related strain 239 times over thirteen years until it lost the ability to cause disease in test animals.

  2. 2:10
    The first vaccination, 1921

    The resulting BCG strain was given to a newborn infant in Paris in 1921, beginning its use in humans.

  3. 4:20
    A contamination that turned deadly

    In Lubeck in 1930, a batch of BCG vaccine contaminated with the virulent strain caused tuberculosis in over a hundred infants and killed dozens.

  4. 6:30
    Protection that will not settle on one number

    Trials have found BCG's protective effect ranging from roughly 60 to 80 percent in some countries to no measurable effect in others.

  5. 8:40
    Still the only vaccine of its kind

    Despite the inconsistency, BCG remains the sole licensed tuberculosis vaccine and is given routinely across most of the world.

  6. 10:50
    Why an imperfect vaccine still merits attention

    Worth understanding because it shows how a vaccine can be genuinely useful, genuinely dangerous once, and genuinely inconsistent, all at once.

Worth your time?

Yes. Study the whole thing.

4/ 5
What works
  • the concreteness of 239 subcultures over thirteen years as a measure of how deliberate the attenuation process was
  • the unflinching inclusion of the Lubeck disaster rather than skipping past it
  • the honest range of efficacy figures instead of a single reassuring number
What does not
  • it does not explain why BCG's protection varies so much by region, only that it does
  • it leaves open exactly how long protection lasts, with estimates ranging from a couple of decades to much longer in some populations
Study it if
  • readers who assume vaccine efficacy figures are fixed and universal
  • anyone curious about the Lubeck disaster and how public health recovered from it
  • readers interested in why a vaccine can work well in one country and poorly in another
Skip it if
  • readers wanting a settled explanation for the geographic variability in effectiveness
  • anyone after the immunology of how BCG stimulates the immune system
The written brief3 min read

Weakening a killer, one culture at a time

Starting in 1908 at the Pasteur Institute in Lille, the physician Albert Calmette and the veterinarian Camille Guerin set out to produce a strain of tuberculosis-related bacterium weak enough to be given safely to people yet still capable of training the immune system to recognise the disease. Their method was repetition: they grew a virulent strain of Mycobacterium bovis, related to the bacterium that causes human tuberculosis, on a bile-containing culture medium and subcultured it again and again, checking periodically whether it could still cause disease in test animals. This continued for thirteen years and 239 successive subcultures, through the disruption of the First World War, until by 1919 the strain, by then known as Bacille Calmette-Guerin, or BCG, had lost the ability to cause tuberculosis in the animals it was tested on.

The first vaccination, 1921

Calmette and Guerin moved to the Pasteur Institute in Paris in 1919, and in 1921 the attenuated strain was given to a newborn infant at the Hopital de la Charite, the first use of what would become BCG vaccination in humans. The approach relied on giving a live but weakened bacterium rather than a killed one or a purified fragment of it, on the reasoning, drawn partly from earlier work by the Norwegian researcher Kristian Feyer Andvord, that a living organism, even a weakened one, would provoke a more durable immune response than dead material. Early use expanded gradually through the 1920s as the vaccine’s apparent safety in the initial cases built confidence in the method.

A contamination that turned deadly

That confidence was badly shaken in 1930, when a vaccine batch administered in Lubeck, Germany turned out to have been contaminated with the original virulent tuberculosis strain, apparently through storage alongside it in the same incubator. Of 251 infants given the contaminated batch, 173 developed tuberculosis and 72 died, a disaster that became public knowledge and weighed heavily on Calmette, who died three years later. Production and safety practices were overhauled, and mass vaccination resumed in many countries by 1932, but the episode remains part of BCG’s history as a reminder that a live attenuated vaccine depends entirely on strict separation from the virulent organism it is derived from.

Protection that will not settle on one number

Once safety practices recovered, BCG’s protective effect turned out to be real but far from uniform. A 1994 review put its overall effect at roughly halving the risk of tuberculosis, but individual trials have found very different results depending on location: studies in the United Kingdom have consistently shown protection in the range of 60 to 80 percent, while trials conducted nearer the equator have sometimes found no measurable protective effect at all. How long protection lasts is similarly unsettled, with one long-running UK study finding it fell to under 60 percent after 15 years and to zero after 20, while a study of a Native American population found evidence of continuing protection six decades after vaccination.

Still the only vaccine of its kind

Despite that inconsistency, BCG remains the only licensed vaccine against tuberculosis a century after its introduction, and as of 2022 it was included in the vaccination schedules of 155 countries, generally those where tuberculosis remains a significant cause of illness and death. In countries such as the United States, where tuberculosis is comparatively rare, health authorities have instead relied on testing and treating latent infection rather than routine vaccination, a choice that reflects how differently BCG’s costs and benefits weigh depending on how common the disease already is in a given population. The vaccine’s continued use across most of the world, imperfect as it is, reflects the absence of any better alternative rather than confidence that the geographic variability has been resolved.

Why an imperfect vaccine still merits attention

This is worth understanding because BCG resists the tidy narrative that vaccines either clearly work or clearly do not: it has demonstrably protected people in some settings for decades, it once caused a documented, fatal disaster through contamination, and its effectiveness varies by region in ways researchers still cannot fully explain a century later. Sitting with all three facts at once, rather than reaching for whichever supports a preferred conclusion, gives a more honest picture of what a live attenuated vaccine can and cannot guarantee. An hour spent on BCG’s history is a useful corrective for anyone inclined to expect a single clean number from any vaccine.

Same field · Medicine4 of 88
Up next in Science

Turing test

Alan Turing · 9:04

The Turing test doesn’t ask whether machines think—it asks whether we’ve stopped noticing the difference.

9:04