As with so many other cancer trends we have been watching over the last couple of decades, skin cancer trends have also been increasing at rates that don’t seem to make sense given the public wide campaign and uptake of sunscreen use. After several decades of using products with “UV filters” that are supposed to block the harmful DNA damaging effects of sunlight, we should by now be seeing a reduction in skin cancer cases…right?
When I came across this new UK Biobank study 1 I was intrigued…might this offer some explanation of what might be going on?
The study was large. It set out to study the important question of how inherited differences in DNA repair genes might interact with environmental and behavioral factors to influence skin cancer risk. It included more than 24,200 people with skin cancer (17,000 basal cell carcinoma, ~2,300 cutaneous squamous cell carcinoma, > 1,100 with in situ melanoma, ~ 3,800 with invasive melanoma). It compared this to >448,000 controls. It was an extensive genetic analysis of nearly 9,000 SNPs across 190 DNA repair genes and asked how these variants interacted with 11 demographic, environmental, and behavioral factors. One of the factors was the frequency and use of sunscreen.
Some of the results were expected. Darker skin and darker hair were associated with lower risk. Easier tanning was associated with lower risk. Sunlamp use and childhood sunburns emerged as important factors. These findings are consistent with what we already know: UV damages DNA, and inherited differences in DNA repair capacity modifies the skin cancer risk. This is obvious.
But one result was very uncomfortable.
Frequent use of sun or UV protection was associated with increased risk of all four skin cancer types.
They showed that the risk for basal cell carcinoma is 2.4 times higher, 2.3 times higher for cutaneous squamous cell carcinoma, roughly 2 to 3.5 times higher for localized melanoma, and for invasive melanoma, the risk ranged from about 1.4 to nearly 4 times higher!
These are not modest changes in risk. If we see a 2- to 4-fold increased association between an environmental factor and breast cancer, we would treat this as a major signal, not as something to casually explain away.
And therefore, it was reassuring to see that the authors acknowledged this directly in the discussion: “Surprisingly at first, frequent use of sunscreen was greatly associated with all cancers.”
And then they offered the standard explanations… Perhaps people who use sunscreen are also the people who spend more time in the sun…. Perhaps they do not reapply it…. Perhaps they started using sunscreen only after a skin cancer diagnosis…..
Of course all of these are possible. But they are not the only possibilities.
The study treats “use of sun/UV protection” as a behavioral variable. But sunscreen is not only a behavior. Sunscreen is a chemical exposure.
That distinction matters.
A hat is sun protection. Shade is sun protection. Staying indoors is sun protection. Mineral zinc oxide sunscreen is sun protection. And while a chemical sunscreen containing organic UV filters, stabilizers, preservatives, fragrances, and other compounds applied directly to skin is also considered sun protection, these are not equivalent exposures.
And yet, in this study, they are collapsed into a single questionnaire variable. That is a problem.
As I (and many others) have written about, some sunscreen chemicals are systemically absorbed. And some are endocrine disrupting chemicals. A randomized clinical trial published in JAMA2 found that several common sunscreen active ingredients — including avobenzone, oxybenzone, octocrylene, homosalate, octisalate, and octinoxate — reached plasma concentrations above the FDA threshold that triggers additional safety testing under maximal-use conditions3.
There is also evidence that some sunscreen ingredients are chemically unstable or can generate degradation products. Octocrylene, for example, has been reported to degrade into benzophenone over time in commercial sunscreen products. Benzophenone is not a trivial contaminant; it is a compound with toxicologic concern4.
So when a gene–environment study finds that sunscreen or sun-protection use is associated with skin cancer, why is the chemical composition of sunscreen not even discussed?
This is especially odd because the paper’s central premise is that DNA repair genes matter. The authors identified numerous SNP–environment interactions, including variants in DNA repair genes such as FANCA, ATM, MGMT, and RAD51B. If inherited variation in DNA repair modifies susceptibility to UV exposure, then it is not unreasonable to ask whether inherited variation in DNA repair might also modify susceptibility to chemical exposures applied to UV-exposed skin. Especially if that chemical has the ability to induce DNA damage?
The study cannot answer the question of whether sunscreen chemicals are causing the association but it certainly does not provide reassurance that is does not. And it also clearly suggests that we should be rethinking the accepted narrative of one-size-fits-all of sunscreen efficacy.
My point is not to discourage sun protection. UV radiation is a known carcinogen and for some people, especially those who cannot repair the DNA damage induced by this exposure efficiently, sun protection is an important option for preventing sunburns and reducing UV exposure. But “sun protection” should not be treated as synonymous with “chemical sunscreen use”.
And blanket chemical sunscreen use should not be treated as biologically inert safe alternative because sunscreen formulations contain biologically active and sometimes photoreactive chemicals. Formulation chemistry should be considered as a potential contributor or effect modifier for skin cancer risk.
This is not a radical statement. It is simply good environmental health science. And if we are going to call something a gene–environment interaction, we should be honest about what the environment includes.