BCG Against Coronavirus: Less Hype And More Evidence, Please

Close up of reaction of Bacillus Calmette Guerin or BCG vaccination infants.
Close up of reaction of Bacillus Calmette Guerin or BCG vaccination infants.

I was born and raised in India. On the day I was born, I am told I got my first jab, a vaccine called BCG (bacille Calmette-Guerin). Kids born in India today still get this vaccine.
This century-old vaccine is suddenly in the news, thanks to a flurry of ecological studies (pre-prints, at this stage) that claim a strong correlation between BCG vaccination and protection against COVID-19. These studies have used the BCG World Atlas that my team developed nearly a decade ago, and updated in 2017. The Atlas is neither perfect nor complete, but it is the only such database.
As a TB researcher, I would be thrilled if BCG worked against COVID-19. But, these ecological studies have serious limitations (which I will cover later) which most media reports have ignored. In fact, BCG has been talked up to be a silver bullet’ and a ‘game-changer’. We need to tone down the hype and focus on getting stronger trial evidence. The good news is that rigorous trials are getting underway, to settle the issue, one way or another.
A controversial vaccine with some utility
BCG is easily the most widely used vaccine worldwide. Next year will mark the 100th anniversary of the first human administration of this vaccine, a live, attenuated version of a virulent bovine strain of tubercle bacillus.
BCG is inexpensive and quite safe. But, it also the most mysterious and controversial vaccine, with at least 10 BCG substrains used across countries, variable vaccine efficacy, and huge variations in vaccination practices across the world (Figure below). In fact, TB scientists like me spend hours fighting over what exactly the vaccine does and how it works.
Take vaccine efficacy, for example. In clinical trials, the efficacy of the BCG vaccine against pulmonary TB in adults has been reported to be 0–80%. The largest vaccine trial was done in south India, and the efficacy of BCG was estimated to be 0%.
Despite this many countries, mostly low- and middle-income countries (LMICs) where TB rates are high, still offer BCG to infants. This because trials show that BCG can protect children from severe, extrapulmonary forms of active TB disease. But vaccine efficacy in adults is poor. So, current BCG practices do not substantially contribute to the control of the global TB epidemic (which affects 10 million people every year), as in most countries, the BCG vaccine is administered once, at birth, and its protection is unlikely to extend consistently into adolescence.
Interestingly enough, in contrast to the effectiveness of BCG against TB, BCG seems to be more protective against leprosy. Since TB and leprosy are both mycobacterial infections, this makes some sense. What is really strange is that BCG might have non-specific immune-boosting effects that might confer some protection against mortality by any cause, perhaps by preventing infections other than TB. BCG is also used as an immunotherapeutic agent for patients with bladder cancer. So, it is tempting to hypothesize that BCG could work against coronavirus.
Ecological studies on BCG and COVID-19
At least half a dozen studies now exist, all exploring the link between BCG vaccination policies in various countries, and incidence of COVID-19 cases and deaths. While data on BCG vaccination policies is derived from the BCG Atlas, data on COVID-19 cases are taken from WHO or other public databases tracking the pandemic.
The first such ecological study used COVID-19 data as of 21st March 2020, and concluded that “the correlation between the beginning of universal BCG vaccination and the protection against COVID-19 suggests that BCG might confer long-lasting protection against the current strain of coronavirus.” This study got lots of uncritical media attention, even before peer-review. Clearly, it inspired many other similar studies. While some studies have been careful and cautiousothers could have used some epidemiological input.
Limitations of ecological studies
First, ecological studies are inherently limited since they take aggregate data and try to make inferences at the individual level. For example, countries might be the unit of analysis, but inferences might be made about individuals living in the countries. It is easy to see the problem here. I could be living in a “low-income country” but be very wealthy (or vice-versa). So, a correlation seen at the country-level might not apply to me. Epidemiologists call this the ‘ecological fallacy’. Ecological fallacy arises from thinking that relationships observed for groups necessarily hold for individuals.
Second, timing really matters in this pandemic. Some of these analyses were done a month ago. Since then, COVID-19 cases and deaths have shot up in many LMICs. For example, India reported 195 cases on 21st March, while the number on 11th April was 8446. That is a 40-fold increase. In South Africa, cases have increased from 205 on March 21 to 2028 on 11th April, a 10-fold increase. 
In fact, the growth rate of confirmed coronavirus cases is now much faster in LMICs. So, if these early ecological analyses were to be repeated now, they might produce very different results. In fact, more recent ecological analyses have shown less optimistic results than the first one.
Third, many LMICs, including India, are seriously under-testing for COVID-19. Even the richest countries are struggling to ramp up testing. This means, reported case numbers from LMICs (which universally vaccinate with BCG) are seriously under-estimated, and deaths can also be under-reported, since COVID-19 symptoms overlap with many other respiratory infections and fevers. A recent ecological analysis which accounted for variation in testing rates suggests that BCG vaccination may not offer protection against COVID-19.
Fourth, confounding is another major issue. For example, the distribution of age in India is vastly different from Italy, for example. Europe and North America have ageing populations, while Asia, South America and Africa have younger populations. This matters because COVID-19 death rates are higher among elderly people. Only a few of the ecological analyses have adjusted for age as a confounder, and adjustment for age does indeed make the correlation much weaker.
Even if known confounders such as age are adjusted for, it is impossible for ecological analyses to adjust for all confounders. For example, countries that offer BCG routinely also offer many other vaccines at the same time. What if one of them is protective against COVID-19 and not BCG? Alternatively, countries that offer BCG routinely might have a much higher burden of many other infectious diseases and that might factor into protection against COVID-19. We don’t have any way of teasing such issues out in an ecological study. A randomized trial, on the other hand, can help avoid confounding.
Fifth, ecological studies gloss over some big inconsistencies. Countries such as China, Iran, South Korea, Singapore, and Japan give BCG at birth. And they have all seen COVID-19 outbreaks, quite early in the pandemic. Also, countries that no longer offer BCG vaccine did offer them in the past. For example, UK gave BCG routinely to school kids until 2005. Many other European countries (e.g. Italy, France, Germany, Spain) gave BCG vaccination in the past, and presumably have older individuals who are vaccinated. Apparently, this has not prevented outbreaks and deaths among older adults in these countries.
Sixth, we know BCG offers some protection to children and very little protection to adults for TB. By analogy, even if BCG does offer protection against COVID-19, it would make sense that such protection should be seen in children. The claim that BCG protects older adults from COVID-19 is a real stretch from the biological standpoint. Accounting for variations in BCG strains is another challenge - even in TB, the effect of strain variation is not well understood.
All of these limitations are already being discussed in blogs and pre-prints and should hopefully temper the media hype.
The only way to really test the BCG-COVID-19 link is to conduct randomized trials. Such trials are starting in Australiathe Netherlands, and the US. However, unlike the ecological studies that primarily looked at universal BCG vaccination policy at birth, these trials focus on adults (e.g. healthcare workers).
Until trial results are out, LMICs dealing with the COVID-19 pandemic must focus on interventions such as aggressive testing, isolation, contact tracing, and physical distancing (where feasible). They should also offer personal protection equipment to healthcare care providers and strengthen hospital capacity to deal with severely ill patients. The most vulnerable populations must receive cash and social benefits. It would be dangerous for policy makers to assume protection from BCG and not act. Researchers and journalists need to be responsible as well, and not raise false hopes based on weak evidence. And countries should not hoard BCG vaccines for COVID-19, since BCG supply chain is weak, and really necessary to protect children in LMICs from childhood TB.
In some ways, the hype around BCG is analogous to the hype and hope around choloroquine and hydroxychloroquine (HCQ) for COVID-19. The whole world is looking for good news and silver bullets. This is understandable. While this may change in future, currently, there is no evidence to support the clinical use of either BCG or HCQ, except in carefully controlled clinical trials.

No comments:

Powered by Blogger.