Written By | Health News Agencies
A landmark study of 59,000 infections stretching across 19 years has found that dirty air does not strike everyone the same way
Search
SCIENCE & TECH
Right now, as you read this, there is a decent chance you are carrying the organism that kills more small children than almost any other bacterium on Earth.
It is sitting quietly in your nose and throat. It is not making you ill. You have never noticed it, you have never been tested for it, and until this week nobody could tell you that the particular version of it you happen to be carrying might decide what a lungful of polluted air does to you.
That changed on Monday.
In a study published in Nature Microbiology on 14 September, an international team led by the Wellcome Sanger Institute in Britain, the National Institute for Communicable Diseases (NICD) in Johannesburg and the Barcelona Supercomputing Center did something nobody had managed before. They took nineteen years of hospital records — roughly 59,000 cases of invasive pneumococcal disease drawn from South Africa's national GERMS-SA surveillance programme — and laid them against nineteen years of air quality, temperature and humidity data. Then they asked a question that sounds simple and turns out to be anything but.
Does dirty air make everyone equally sick?
The answer, emphatically, is no. And the reason why is a piece of biology most people have never heard of.
THE PASSENGER ALMOST NOBODY KNOWS THEY HAVE
The bacterium is Streptococcus pneumoniae. The pneumococcus. It lives in the respiratory microbiome — the community of microbes that colonises the human nose and throat — in roughly one in five adults worldwide at any given moment. In South Africa, between 40 and 60 per cent of children are carrying it.
For the overwhelming majority of those people, it does absolutely nothing. It is a lodger, not an intruder. It sits in the nasopharynx, it competes with its neighbours, and it eventually moves on.
But when it slips downward into the lower respiratory tract, or sideways into the bloodstream, it becomes something else entirely. That is invasive pneumococcal disease — IPD — and it is the umbrella term for bacterial pneumonia, bloodstream sepsis and bacterial meningitis. It is one of the great infectious killers of the very young and the very old. Global estimates put pneumococcal deaths in children under five at somewhere in the region of 294,000 a year. South Africa logged more than 1,800 reported IPD cases in 2022 alone. In England, the UK Health Security Agency counts more than 5,000 cases annually, concentrated hard among the under-fives, the over-65s and people with underlying conditions.
And here is the crucial detail that makes this week's paper matter: there is no such thing as "the" pneumococcus. There are more than 100 distinct serotypes — subtypes distinguished by the sugar capsule wrapped around the bacterium like armour. They are related the way breeds of dog are related. They are not interchangeable.
SAME SMOG, DIFFERENT CLOCK
What the Sanger-NICD-Barcelona team found is that those serotypes respond to air pollution in startlingly different ways — not just in how badly, but in how fast.
Three subtypes stood out for sheer danger. Serotypes 14, 19A and 8 were linked to the greatest risk of invasive disease following air pollution exposure. If those are the strains circulating in your neighbourhood, a bad air week is a bigger threat than it would be somewhere else.
Then there is the timing, which is the genuinely eerie part. Across the dataset as a whole, the peak risk of invasive disease arrived roughly two weeks after the pollution exposure — a slow fuse, burning quietly while nobody connects the smog of a fortnight ago to the child now on a ward. But in areas where serotypes 4, 8, 23F and 19F were common, the pattern collapsed. Cases rose immediately, within the same week.
Same air. Same particulates. Two completely different clocks, set by which bacterium happened to be in residence.
The researchers also pulled apart the role of weather, and found it behaves in the same split-personality way. High temperatures were associated with an immediate jump in disease risk. Cold temperatures produced a delayed rise. The team's suggested explanation is that cold weather slows bacterial transmission, or that damage done by seasonal respiratory viruses opens the door for a bacterial infection to walk through a week or two later — the well-documented one-two punch in which a virus softens the tissue and a bacterium finishes the job.
Crucially, the analysis accounted for temperature, humidity and population density, so the pollution signal is not simply winter, or crowding, wearing a disguise.
WHY SOUTH AFRICA WAS THE PLACE TO ASK
There is a reason this work was done in South Africa rather than Stockholm.
South Africa has, in the GERMS-SA programme, one of the most sustained national laboratory surveillance systems for invasive bacterial disease anywhere in the developing world — nearly two decades of consistently collected, genomically characterised isolates. That is the raw material. Very few countries have it.
It also has an air pollution problem of a scale that makes the signal visible. The Highveld region and the industrial belts of Gauteng and Mpumalanga host the country's largest coal-fired power stations. Analyses of fine particulate pollution have attributed roughly 42,000 deaths in South Africa in a single recent year to PM2.5 exposure — including more than 1,200 among children under five, and amounting to something in the order of 8 per cent of all deaths in the country.
Put a world-class bacterial surveillance system on top of one of the world's dirtiest airsheds and you get a natural experiment that could not be run deliberately.
WHAT THE SCIENTISTS THEMSELVES SAY
Professor Anne von Gottberg, co-senior author at the NICD in Johannesburg, framed the finding as a direct argument for air quality policy.
"Invasive pneumococcal diseases are a significant public health issue in South Africa and globally," she said. "This research adds to the ongoing evidence that air quality greatly impacts our health and shows that it has different effects depending on the strain of bacteria that is found in an area. We need to focus on monitoring and improving air quality in areas with high disease risk to prevent or reduce some of these infections."
Professor Rachel Lowe, co-senior author at the Barcelona Supercomputing Center and the Catalan Institution for Research and Advanced Studies, placed it in a bigger and more uncomfortable frame.
"Climate change and rapid urbanization are fundamentally altering the environmental conditions that influence risk of infectious disease," she said. "By linking long-term climate and air quality monitoring with health outcomes, this work highlights how shifting environmental patterns can amplify public health risks. Protecting communities most at risk will require cross-sector collaboration and adaptive health strategies that can keep pace with a changing environment."
And Dr Sophie Belman — the paper's first author, formerly of the Wellcome Sanger Institute and the Barcelona Supercomputing Center and now an assistant professor at the Yale School of Public Health — delivered the line that should stop anyone who assumes environmental risk is a single number printed on a government dashboard.
"While we found that temperature and air pollution generally increase the risk of invasive pneumococcal diseases, such as bacterial meningitis, our research also suggests that it is the bacterial subtypes a person is carrying that modulate infection rates," Belman said. "The strain of bacteria impacts the timing of disease and who might be more at risk depending on their respiratory microbiome, meaning that the risk is not the same in every situation or for every person. By extending our findings to other parts of the globe, we will better understand who has the highest health risk from environmental exposures, and what factors need to be addressed in different regions, such as improving air quality in cities."
The risk is not the same for every person. That is the sentence.
THE VACCINE PROBLEM HIDING IN THE DATA
There is an uncomfortable subplot here, and it has to do with vaccines.
Pneumococcal conjugate vaccines — PCV10, PCV13, PCV15, PCV20 — do not protect against all 100-plus serotypes. They protect against a shortlist. Global surveillance work through the PSERENADE project has put pooled coverage of childhood invasive disease at roughly 40 per cent for PCV10, 50 per cent for PCV13, 51 per cent for PCV15 and 57 per cent for PCV20.
Which means that between two-fifths and three-fifths of the serotypes causing invasive disease in children are, depending on the vaccine in use, outside the net.
Worse, the pneumococcal population does not sit still. Squeeze out the vaccine-covered serotypes and others expand to fill the vacated space — a phenomenon known as serotype replacement. The most notorious example is serotype 19A, which rose globally after the first conjugate vaccines rolled out and which PSERENADE confirms remains one of the predominant causes of post-vaccine invasive disease worldwide.
Serotype 19A is also, per Monday's paper, one of the three subtypes most strongly linked to invasive disease after air pollution exposure.
That is not a coincidence anyone should find comfortable. It raises the possibility that in heavily polluted cities, the strains that survived the vaccine era are precisely the ones best positioned to exploit bad air. The study does not prove that. But it is exactly the question the authors are pointing at when they argue for combining environmental monitoring with genomic surveillance.
WHAT THIS STUDY CANNOT TELL YOU
Sober caveats matter here, and a responsible reading demands several.
This is an observational study. It links patterns in environmental exposure to patterns in recorded disease across a population; it did not put anyone in a chamber and pump in particulates. Association is not causation, and the authors do not claim otherwise.
It is also, at its core, an ecological analysis. Pollution exposure is assigned by area, not by individual. A person living in a high-PM2.5 district may spend their days elsewhere. Serotype distribution is likewise characterised at the level of circulating strains in a region rather than swabbed from each patient's own throat before they fell ill.
And it is South African. The country's serotype mix, vaccine schedule, HIV burden, housing stock, indoor cooking fuels and industrial geography are its own. The team is explicit that extending the findings to other parts of the globe is future work, not a finished result.
What the study does establish, robustly, is that treating "air pollution" and "bacterial infection" as two separate conversations is a mistake — and that the biological middleman between them has a name and a number.
WHAT THIS MEANS FOR YOU
You cannot find out which pneumococcal serotype is living in your throat. There is no home test, and asking your GP will not get you far. So the practical advice is not about strains. It is about the two levers you actually control.
The first is vaccination. Pneumococcal vaccines do not cover everything, but the serotypes they do cover include some of the deadliest. If you are over 65, if you have a child under five, if you or someone in your household has a chronic heart, lung, liver or kidney condition, diabetes, a suppressed immune system or no functioning spleen, the pneumococcal vaccine is not an optional extra. Check your status. Check your children's schedule. The flu vaccine belongs in the same conversation, because of that viral one-two punch: a respiratory virus is frequently the thing that opens the door for the bacterium.
The second is exposure. On high-pollution days — and most countries now publish a daily air quality index — the people who should actually change their behaviour are precisely the ones this study identified as most vulnerable: young children and older adults. That means shifting outdoor exercise indoors or to early morning, keeping windows shut on the side of the building facing traffic, and running a filter if you have one. If you live near a coal plant, a busy freight corridor or a heavy industrial zone, treat the air index the way you treat a weather forecast.
And there is a symptom lesson buried in the timing data. The study found disease risk peaking up to two weeks after exposure. Human beings do not naturally connect a fever today to smog a fortnight ago. So if a chest infection follows a bad air stretch, and particularly if it comes with confusion, a stiff neck, severe headache, a rapidly worsening cough, chest pain or difficulty breathing, do not wait it out. Invasive pneumococcal disease moves fast once it is invasive. Meningitis and sepsis are measured in hours, not days.
THE BIGGER PICTURE
For decades, the case against air pollution has been built almost entirely on chronic damage — the slow scarring of lungs, the hardening of arteries, the shaved years at the end of a life. That case is overwhelming, and it has still not been enough to close many of the plants doing the damage.
This week's paper adds something different and, for the public, considerably more visceral: a short-term, acute, infectious consequence. Not a statistic about life expectancy in 2050. A child in a hospital bed a fortnight after a bad air week.
It also hands public health officials something they have never had. If you know which serotypes are circulating in a district, and you can see a pollution event coming, you can in principle predict the shape and the timing of the surge that follows — same-week for some strains, two weeks out for others — and staff the wards accordingly. The authors argue precisely this: that integrating environmental monitoring with genomic surveillance could forecast infection spikes, sharpen vaccination strategy and let hospitals brace before the ambulances arrive.
That is a genuinely new capability, and it is being built out of data South Africa has been patiently collecting since before most of today's air quality apps existed.
The air is loading the gun. It always was. What nineteen years of South African hospital records have now shown is that something living in the back of your throat is holding the trigger — and that not every finger squeezes at the same speed.
