Fitter. Happier. Fewer early-life infections?


A person lucky enough to survive to the age of 20 in 1840s London could expect to live for a further 40 years, to the age of 60. In 2011, a 20-year-old Londoner would expect to live for a further 60 years, to the age of 80. That’s a 50% increase in remaining lifespan, in a little over 150 years. How has this happened?

In 19th century Europe, around 40% of children died of infectious diseases like smallpox, whooping cough and measles before they reached adulthood. However, it’s important to recognize that both of our 20-year olds survived to age 20, so the difference in their life expectancies has nothing to do with the fact that children in modern London are unlikely to catch dangerous infections. OR DOES IT?


(13) Scream_wikimedia commons

From Wikimedia Commons


Looking at patterns of mortality across time throws up some interesting results. For example, a study of 19th and 20th century populations in Britain and Sweden showed that the death rates of individuals were more closely related to their year of birth than to the year in which their mortality was assessed.


(13) Kermack table

Mortality rates per thousand individuals in England and Wales, 1845-1925. The diagonals show data for the same cohort of individuals followed across their lives, and it’s apparent that the mortality rate remains similar as the cohort ages. The best predictor of mortality is birth year, not calendar year. For example, the cohort born aged 10 in 1855 is the same group who are aged 60 in 1905 and their mortality rates are identical. Meanwhile, the cohort aged 10 in 1905 has a much lower mortality rate.


The authors concluded that:

The figures behave as if the expectation of life was determined by the conditions which existed during the child’s earlier years…the health of the child is determined by the environmental conditions existing during the years 0-15, and the health of the [adult] is determined preponderantly by the physical constitution which the child has built up.

 So what might these ‘conditions’ be? An influential paper published in 2004 suggested that the link between early-life conditions and later-life mortality might be due to infections experienced in childhood. Infections elicit inflammatory immune responses, which may persist at a chronically high level. Chronic inflammation is linked to risk of heart disease, stroke and cancer in later life. The ‘cohort morbidity phenotype’ hypothesis suggests that since childhood infections have become increasingly rare, so have chronic inflammation, their associated pathologies and early death. Thus, we live longer.


(13) Cohort morbidity phenotype

A simplified version of the ‘cohort morbidity phenotype’ hypothesis. Infections cause inflammatory responses, which lead to atherosclerosis (thickening of artery walls) and thrombosis (clotting), which are linked to heart disease, stroke and mortality. The full version includes a couple of added nuances!


This is an exciting (and controversial) idea, so we tested it, in a new paper published in Proceedings of the National Academy of Sciences of the USA. The data we used came from, as I usually can’t help blurting out when meeting people and explaining what I do, ‘some dead Finnish people’. We had data on births, marriages and deaths from church records for over 7,000 individuals, born between 1751 and 1850, in seven different populations across Finland.


(13) Finns

A Finnish family, pictured in the late 19th century. Photo courtesy of Virpi Lummaa.


OK, I admit it: a number of previous studies have tested for links between early and later mortality. But…. these studies have looked at how many children born in a given year survived infancy (the ‘cohort mortality rate’), and then correlated that with the survival rate of these individuals in later life. Instead of using data on child deaths from all causes, we used data on child deaths from infections.

For each year of our study, and in each parish, we knew how many children were alive, and how many of those children died of an infectious disease. Our measure of disease exposure for a given birth year was the number of children who died of infectious diseases, divided by the number of children alive. We calculated this measure of disease exposure for each of the first five years of a child’s life. We then went on to use statistical models to determine the association between early disease exposure and:

  1. Mortality risk in adulthood
  2. Risk of mortality from cardiovascular disease, stroke and cancer
  3. Reproductive success

We predicted that our measure of early disease exposure should be linked with higher mortality risk, a greater risk of mortality from cardiovascular disease, stroke and cancers, and lower reproductive performance. And we found…



Nothing. There was no link between early disease exposure and adult (after age 15) mortality risk. We did find the expected differences between social classes, with wealthy farm owners and merchants surviving better than poor crofters and labourers, and between the sexes, with women surviving better than men. However, higher disease exposure was predicted to increase mortality risk by a piddling, and statistically insignificant, 2%.

We also found no association between early disease exposure and deaths specifically from heart disease, stroke and cancer. The (nowhere near statistically significant) trend was for a lower probability of death from these causes with increasing early disease exposure. Men were more likely to die from these causes, but there was no difference between the social classes.

Lastly, we found that reproductive success was not affected by early-life disease exposure. To take any survival effect out of the equation (not essential, so it turned out…), we only analysed people who survived to age 50, and who therefore had almost certainly reached the end of their reproductive lives. Early disease exposure was not linked to age at first birth, lifetime children born, child survival rate, or lifetime children surviving to adulthood. Excellent.


(13) Listening to the results

As the results were revealed, the excitement of the audience was almost palpable.


Normally at this point, there’d be loads of cool graphs and stuff…but we didn’t make any. Instead, take a look at the paper and the enormous supplement for details on the results! Before the headline conclusions….some caveats:

We have no idea who was exposed to disease. The only records of infections were where someone died and the cause was recorded in the church register as being of an infectious disease. This meant we had to assume that, in years when lots of children died of infections, our study individuals were (on average) more likely to get disease. But at worst, it’s possible that none of our study individuals ever got sick as kids.

What doesn’t kill you makes you stronger. It’s possible that individuals who were exposed to disease responded in two ways. They could have been weakened by disease and so died earlier, or they could have been the crème de la crème: robust enough to survive and be awesome at everything. A balance between damaged individuals and robust survivors would lead to…no net effect. So we may have actually found TWO cool effects…but without any evidence for them.

But…we did do some good things. We used a measure of disease exposure based on death from infectious disease, rather than deaths of all causes; we tested for effects on specific causes of death; we found the same patterns in seven parishes across Finland; we looked at effects on reproduction. We also did some cool (and relatively straightforward!) stats to remove temporal trends…but this isn’t the place for that.

Overall, we found absolutely bugger all evidence for a link between early-life disease exposure and mortality risk, cause of death, or reproductive success in later life. The results challenge the idea that extended lifespan in modern populations is due to reduced childhood disease exposure, but they certainly do not disprove it. It does seem however, that in common with a few other recent studies, the early-life environment may have weaker effects on events occurring in adulthood than do the conditions experienced during adult life.

If you’re interested in the paper, but can’t access it, please do get in touch.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s