August 14, 2024

expert reaction to study looking at ultraviolet light exposure and mortality

A study published in Health & Place looks at UV light exposure and mortality rates. 

Prof Brian Diffey, British Association of Dermatologists, said:

“There are various limitations to the study, so people don’t need to make knee-jerk changes to their sun safety habits, and there’s no need for changes to public health advice.

“One of the major limitations is that the researchers used local UV levels to estimate how much sunlight people were exposed to.  Dosimetric studies, which measure the actual UV exposure people experience, repeatedly show that the amount of time people spend outside, and their activities have a much bigger impact on their sun exposure than the local UV levels.

“We know from previous evidence that sun safety is about balancing the benefits and risks of sun exposure and it is possible to follow current safety advice and still get many of the health benefits of sun exposure.  Notably, wearing sunscreen doesn’t prevent people from synthesising vitamin D.  Similarly, spending short periods in the sun without protection, for example, on a short shopping trip through town, can be beneficial¹.”

1 https://documents.manchester.ac.uk/display.aspx?DocID=44418

Rachel Richardson, Methods Support Unit Manager, Evidence Production and Methods Directorate, Cochrane, said:

“It is important to be aware that the participants in this study are not fully representative of the UK population, for example, the UK BioBank only includes people who were between the ages of 37 and 73 at recruitment.  These findings may not therefore apply to older and younger age-groups.  A second limitation is that some of the results are based on small numbers of participants, for example the findings related to mortality from melanoma.  This means that there is not enough information in the study to draw conclusions about whether there was an association between UV exposure and deaths from melanoma.”

Prof Kevin McConway, Emeritus Professor of Applied Statistics, Open University, said:

“This is an interesting study, though it needs some care in interpretation.  Rightly, the press release and the research paper don’t go as far as claiming that the research found that having more exposure to ultra-violet (UV) radiation (from the sun or otherwise) causes people in low-sunlight countries to live longer than would otherwise happen.

“The findings are consistent with that possibility, but they are consistent with other possibilities too.  That’s why the press release says that “UV rays may boost health in low sunlight countries”, and the paper says “The benefits of UV exposure may outweigh the risks in low-sunlight countries.”  (My emphases.)

“I’ll explain why there remains doubt below.  But first, I want to make a point about sunbed use.  A quick glance at the research paper or the press release might give someone the idea that the more one uses a solarium or sunbed, the better.  But in fact the researchers only compared people who said they used a sunbed at least once a year, with people who said they used a sunbed less than that, or not at all.  There is no comparison between different levels of sunbed use beyond that.  So, for instance, they didn’t compare people who used a sunbed just once a year with people who used it many more times a year.  There’s been previous research on harmful health effects of excessive use of sunbeds and sunlamps, and this new research doesn’t contradict those findings.

“So, why the doubt about cause and effect in the conclusions?  It’s because the research was observational.  The researchers didn’t tell people to live in sunny or less sunny parts of the UK, or to use or not use sunbeds.  They just recorded what the participants did, followed them up for about 13 years in average, and recorded which participants died.  For the 7 in every 100 participants who did die during the study, the researchers recorded what the cause of death was.  They were particularly interested in deaths from cardiovascular disease (CVD, such as heart disease and strokes), and from cancer.  And they found that, on average, deaths occurred rather later in life in people who lived in sunnier parts of the UK, or who had used a sunbed at least once a year.

“The snag here is that people who live in different parts of the country differ in many other ways than the amount of UV radiation they get from the sun, and same is true of people who use or don’t use sunbeds.  And the correlation between UV exposure and death rates could be caused by these other factors and not actually by the exposure to UV rays at all.

“Obviously the researchers were well aware of this.  It’s possible to make statistical adjustments to allow for some other factors – so-called potential confounding factors – and the researchers did that for several factors.  But the researchers rightly point out that there may be some relevant confounding that they did not adjust for – and it’s not possible to adjust for factors on which there is no data.  So some doubt has to remain about the extent to which the different UV exposures explain the differences in death rates, in terms of cause and effect.

“The researchers do try to justify what they did about confounding by using terms and methods from what’s called ‘causal inference’, but I’m not entirely convinced by the causal diagrams that they put in a supplement to their research paper.  They also use what’s called a ‘negative control’ – that is, they analysed data on hospitalisations from car and motorcycle accidents in relation to UV exposure, on the grounds that it seems very unlikely that these accidents would be related to UV exposure.  If an association between these accidents and UV exposure had emerged, that could indicate that there are confounding factors that hadn’t been taken account of.  No such associations were found. T hat does go some way to increase confidence that UV exposure might have some cause-and-effect association with CVD and cancer death rates – but in my view it still doesn’t mean that that cause-and-effect association has been proved to exist.

“It’s right in my view that they end up with a conclusion that is not completely definite about cause and effect in what they found.  This doesn’t at all mean that UV exposure definitely does not cause better health – just that there are other possible explanations.  The researchers are also right to point out that future studies are needed to clarify what’s going on, and that the position is likely to be very different to what they found in the UK in high-sunlight countries such as Australia.

Further information 

“The press release gives figures for how much the risk of death from CVD or cancer is lowered in people who live in sunny places like Cornwall, compared to people who live in less sunny places like Glasgow, and for people who said they used sunbeds at least once a year compared to those who said they used sunbeds less than that, or not at all.  The research paper makes it clear that they are talking about a relative measure of the risk of death in a certain period of time.  But, for people like those in the UK Biobank study, which provided the participants for this research, how big are the risks in absolute terms?  I’ll give an example.

“The press release says that the risk of dying from CVD in sunny Cornwall was 19% less than the risk in gloomier Glasgow.  The research paper reports that about 17 in a group of 1,000 people, like those in this study who lived in the parts of the UK in the gloomiest third (in terms of UV levels from the sun), died of CVD during the study period.  If we compare them to another 1,000 people, who had the same characteristics but lived in a place that was as sunny compared to where the first 1,000 lived as Cornwall is compared to Glasgow, then about 14 in the sunnier 1,000 would die of CVD according to the statistical results of this research.  That’s a reduction in deaths, but perhaps not a very large one, I’d say.  And we don’t really know to what extent that reduction is caused by the difference in UV exposure, as I’ve explained.

“The number of people in this study who died of CVD during the study is not very large, as these things go – about 6,700 – despite the fact there were almost 400,000 participants in all.  (That’s because the participants were middle-aged, and because participants in the UK Biobank study are healthier than the national average.)  This means that there is in fact quite a lot of statistical uncertainty in those numbers.  The number who would die of CVD out of the 1,000 living in sunnier places could plausibly be anywhere between 12 and 15 – yes, less than the 17 CVD deaths in the 1,000 living in the less sunny places, but not much less.

“On the other hand, living in a sunnier place, or otherwise getting more UV radiation, for a long time (say, one’s whole adult life), could mean that these benefits add up over a lifetime – if indeed it is the extra UV exposure that’s the cause of the health benefits.  But to say anything meaningful about how much it might add to life expectancy, one would need to make statistical assumptions that aren’t really justified.”

‘Higher ultraviolet light exposure is associated with lower mortality: An analysis of data from the UK biobank cohort study’ by Andrew C. Stevenson et al. was published in Health & Place.

DOI: 10.1016/j.healthplace.2024.103328

Declared interests

Prof Brian Diffey: “No conflicts of interest to declare.”

Rachel Richardson: “I have no conflicts of interest to declare.”

Prof Kevin McConway: “I am a Trustee of the SMC and a member of its Advisory Committee.  My quote above is in my capacity as an independent professional statistician.”