Should I wear sunscreen?
In case you want to hear me read the article instead of reading it, I tried this out below. It’s 20 minutes. You miss the figures, links, and pictures, but save the energy of reading. To each their own.
My interest in this topic started when I saw the below article.
You can already guess what the answer they give is. This hits me with a familiar medical absolutism that from…prior experience… I have an instinctive repulsion to. Imagine putting on sunscreen to prevent the cold blue light of your laptop from burning your skin. I also saw this, in reference to Australian health authorities recommending some sun exposure.
Ok well, it sounds like there’s some debate bubbling here.
Sunscreen is sold as: must wear all the time or you’ll get skin cancer and look gross and die and look really stupid because why wouldn’t you just listen to the experts? In reviewing I wanted to look at:
· How much evidence is there for the sun/sunscreen skin cancer relationship?
· How much evidence is there for the sun/aging relationship?
· Are there any downsides to sunscreen?
Skin cancer:
First thing to know is that there are 3 main types of skin cancer. Basal cell carcinoma (BCC), squamous cell carcinoma (SCC), and melanoma. They originate from different skin cell types, melanoma is the worst one. All of them are very treatable with survival up to 100% if diagnosed early enough. Melanoma is bad if you don’t catch it early enough; when it’s metastasized it is hard to treat and survival is 10-15% at stage 4. But, hey, if you catch it early than no problem. According to The American Academy of Dermatology cases of skin cancer are increasing.
It is estimated that the overall incidence of BCC increased by 145% between 1976-1984 and 2000-2010, and the overall incidence of SCC increased 263% over that same period.
Melanoma rates in the United States have been rising rapidly over the past 30 years — doubling from 1982 to 2011 — but trends within the past decade vary by age. There has been a 31.5% increase between 2011 and 2019.
An obvious question to ask though: Why are skin cancer rates increasing?
The US, and world in general are using more sunscreen over the past 14 years (Euromonitor). Given the trend to virtual life, I would assume we are spending more time inside as well. So according to the sun hypothesis, we should be getting less skin cancer.
One potential explanation would be increased diagnosis. Maybe the sunscreen mania is rubbing off on the dermatologists, who now want to cut every mole off your body. It does seem like the diagnostic standards are constantly updating:
“Multiple studies show surprisingly high inter- and intra- observer variability between pathologists in the diagnosis of melanocytic neoplasms, particularly those with ambiguous histologic features.14,15,19,21-23 Much of the variability in diagnosis and prognosis for melanoma is due to the somewhat subjective visual observations used for melanoma diagnosis and prognosis, therefore, research into new, (****)) more objective methods to improve the accuracy and reproducibility in the field of melanoma diagnosis is ongoing.33
Ok, so, skin cancer rates could be increasing. Or they could be decreasing, no way to know without data I don’t have. Though I’ll come back to this because it matters a lot.
Next is a ubiquitous claim: “The majority of melanoma cases are attributable to UV exposure.”
In an article commenting on Australia’s updated guidelines: “We know that 90% of skin cancers come from daily unprotected UV exposure,” Mona A. Gohara, MD, a board-certified dermatologist and associate clinical professor of dermatology at the Yale School of Medicine.
Ok, but who’s “we” and how do “we know”?
(Funnily enough this professor engages in the exact kind of paternalism which made me interested in this topic: “It is a slippery slope,” Dr. Gohara said when we spoke about Australia's new public health guidance. “It is like saying it’s OK to smoke half a cigarette, but not the whole thing. UV light is a carcinogen any way you slice it.”)
Lady! It is ok to smoke half a cigarette.
I read a few of the studies cited by the American Academy of Dermatology to back up these claims . I expected them to be bad, but god are they bad.
First one: To come up with the 90% number what they did was look at a few years in Connecticut where there was low skin cancer, and compare that to rates in the US about 70 years later. To calculate the rate of melanoma attributable to UV, they assume that low melanoma in CT in that time period was because of low UV.
In their words: “The low melanoma burden in Connecticut during this period likely reflected UV exposure accumulated in the 1930s or earlier, when exposure was likely minimized by clothing style and limited recreational exposure.” Uhhh…
Just to make the problems with this more accessible let’s do a metaphor:
Imagine I love scratch off lottery tickets. One week in 2021 I won big on 5 scratchers in a row. Since then, things haven’t been going so hot. I’m down bad. I remember during that week I also might have eaten sushi 5 times. That must be it!
To quantify the effects of my decreased sushi intake on my financial loss, I guessed real hard at when I ate sushi over the last 3 years, because I don’t actually have a record. I count how much money I lost, and come up with a number for how much I could have won if only I had continued eating sushi.
“In this study we show undeniable evidence that Olek’s reduced sushi consumption habits are responsible for $453,230 in unrealized value.”
They provide no evidence that people in CT got less UV in 1930, and people certainly weren’t wearing sunscreen in 1930. It wasn’t until post WWII that common brands of sunscreen were invented.
So, this paper boils down to, melanoma has gone up. We assume that all the increase is because of increased UV. That means this much of melanoma is attributable to UV.
If there’s another explanation for why melanoma rates have “gone up” that would evaporate the claims in this paper.
There’s a much simpler explanation. Overdiagnosis. What if instead of the increase being attributable to UV, there was no increase at all?
This paper studied that question by looking at differences in mortality with no difference in treatment for melanoma. For overall melanomas (invasive and in situ) it found that 50% of melanomas in white men, and 65% in white women were overdiagnosed, meaning that you could have left them alone with no problem. For the in situ (early stage, about half of diagnoses), about 90% of melanomas were overdiagnosed! If melanoma was underdiagnosed in 1930 CT, that would completely account for the data in the paper.
It seems likely that melanoma has not gone up by nearly as much as is suggested by the AAD. A far more accurate conclusion than 90% of cases of melanoma are attributable to UV is 90% of cases of melanoma are attributable to overeager doctors.
If this sounds crazy to you, you don’t know the half of it. Did you know prostate cancer screening doesn’t save lives, breast cancer screening doesn’t save lives? This conversation with Professor Vinay Prasad on a podcast I like breaks it down. If you’re interested in the topic this conversation changed the way I think about cancer screening, and I can’t recommend it highly enough. I stan Vinay.
To drive home the point: here’s another study cited to make the claim, same problem: “Population attributable fractions and numbers of new melanoma cases in adults due to ambient UVR were calculated by age and sex for 153 countries by comparing the current melanoma burden with historical data, i.e., the melanoma burden observed in a population with minimal exposure to UVR”
I repeat unless you present compelling evidence that people got less UV in the past than they do now (and less melanoma isn’t evidence of that), you can’t make this argument.
This study includes another common line of evidence. They compare melanoma rates in dark skinned Africans to rates in other countries.
An aside: Melanin is basically sunscreen. Dark skin can provide about SPF 13.
Another aside: SPF is a division factor. This means that SPF 13 blocks 12/13 of the UV. So if you’re outside for an hour, you get the equivalent of 4 ½ minutes of sun. If you take anything away from this post, take this away. SPF 4 means an hour in the sun is 15 minutes. That’s still a lot of protection.
Last aside: I have a suspicion that secretly this is where the whole sun / skin cancer thing came from. Black people have diagnosed melanoma far less often than white people (1:30 ratio), so scientists guess this must be explained by the sunscreen effect of melanin. But an important fact to keep in mind is melanoma tends to be more fatal in black people, emblematic of later diagnosis.
There’s a few obvious problems here. One is that it is harder to diagnose melanoma on brown skin, given that it’s characterized as basically the appearance of a mole. Another is sub-saharan Africans are less likely to seek melanoma screening for a series of complex sociopolitical reasons.
So, part of this effect is differences in diagnostic rates, which is borne out by the different survival rates for melanoma; as the cancers are caught later in progression. Many of the diagnosed melanomas in white people likely wouldn’t need to be treated, but they’re still counted and chopped off.
Of course, there are a lot of other reasons you can’t compare sub-saharan Africans to white people in the US to determine the effect of UV on melanoma. There are a million things different about daily life in sub-saharan Africa vs. the US. Two of them being that I bet in sub-saharan Africa people spend a lot more time in the sun, and use less sunscreen than people in the US. It’s too facially ridiculous to even get into.
The last line of evidence people point to is that tanning beds are associated with skin cancer. This appears to be a bit more legitimate to me, but the evidence doesn’t point to anything near 90%. This meta-review (compilation of data from a bunch of studies) found:
Significant association between skin cancer and indoor tanning (melanoma, RR= 1.27, 95% CI 1.16–1.39; NMSC, RR = 1.40, 95% CI 1.18–1.65; squamous cell carcinoma (SCC), RR = 1.58, 95% CI 1.38–1.81; basal cell carcinoma (BCC), RR = 1.24, 95% CI 1.00–1.55)
The first thing I’d point out is none of the data going into their review is randomized. This means that a bunch of the studies going in to the review probably face similar issues to the ones above. Garbage in, garbage out, or so they say.
The second thing I’d point out is that the risk ratios across studies are all over the place, with some even finding fewer skin cancer cases in their sunbed group.
The third thing I’d point out is that even if their average risk across these studies comes out to correct, the ratios just aren’t that big. A relative risk of 1.27 for melanoma would mean that you’re 27% more likely to get melanoma. This would mean the fraction of melanoma cases you can attribute to tanning beds is 21%, and 79% are unrelated. Keep in mind these studies are for people who use tanning beds >10x a year, and many find no difference.
So overall: the evidence base for the link between UV and skin cancers is surprisingly weak.
One more place I looked was at the location of melanomas on the body. If sun exposure is responsible for the cancers, you would expect that the body parts that get the most sun would get the most cancer. You would expect the lowest rates in the areas where you get the least sun.
Is that the case? For women, it seems to be, for men the most common areas are chest/back. The articles about it also note that you can get melanoma in places which never see the sun, like nailbeds, soles of the feet, etc.
Head and Neck:
Female: 13.83%
Male: 27.19%
Trunk:
Female: 26.54%
Male: 39.15%
Upper Extremity/Shoulder:
Female: 29.44%
Male: 24.16%
Lower Extremity/Hip:
Female: 30.19%
Male: 9.50%
It’s not clear that there’s bias towards the areas that get the most sun. It’s also weird that women get it much more commonly on their shins, and men more commonly on their backs. Why are the distributions so different? I don’t think it’s explainable by dress style.
Ok well…
Someone has definitely tried to do a randomized trial, right? Took two groups, told one to put on sunscreen every day, another to do nothing differently, and tracked cancer over time? Yes. And the result was headlines like this:
Again, I’m frustrated by the scientist moralizing in the commentary of this article.
“The scientific facts are inescapable—regular use of sun protection reduces skin cancer and cancer precursors,” Leffell said by email.
“The benefits of sun protection and the incontrovertible evidence of sunburn and chronic sun exposure as a cause of about 60 percent of melanomas should inspire people to continue reasonable sun protection if they fit into the moderate to high risk groups, and even if they don’t,” Leffell added.”
There is no evidence provided about the source of this 60% claim, of course. I’m assuming he would point to some of those earlier studies which generated the 90% claim. A layperson reading this would think, oh he’s so confident, it must be true. But this sort of thing happens all the time, as does the emotional defense, I wrote about this here. Frankly, most doctors are just not very good at understanding medical evidence. Only 1 in 5 could correctly answer a basic statistical question about a diagnostic test.
Incontrovertible, Leffell? Au contraire, controvertible.
The Cochrane Review is a journal I’ve referred to elsewhere on my site, but it’s considered the gold standard in medicine. In commenting on the one randomized trial about sunscreen:
“They found that sun protection factor (SPF) 16 sunscreen applied on exposed sites of the body, such as the head, neck, and upper limbs, every morning and repeated after heavy sweating or bathing made no difference in terms of reducing the occurrence of new cases of BCC and cSCC confirmed or not by histopathology at 4.5 years, in comparison with discretionary use”
The number of cases was the same in both groups.
No additional evidence exists to support claims on melanoma due to rates too low to study. So this is where I’ll leave the skin cancer question. My conclusion would be that sunscreen could make a small difference in skin cancer likelihood, maxed out at about a 20% risk reduction on top of a low baseline risk. Reducing your chance of actual skin cancer from 1 in 40 to 1 in 50. Not nothing, but not much. It could also do nothing. No need to stress about reapplying rigorously, or using SPF 100. Don’t get too upset if you get a sunburn, it doesn’t seem to matter much.
Skin aging:
This section will be short. I think this is probably more intuitive but less important than skin cancer. Many people will have seen the below picture, ostensibly it’s a trucker who never wore sunscreen, so he had sun on one size of this face for decades.
Did sun cause this? Maybe. It’s worth noting though that you can’t find many more pictures of this condition. There are 3.54 million truck drivers in the US right now, and it has been a big industry for decades. If effects this extreme were common this is not something that would require leaning on such a rare anecdote.
The next best picture I found was this. I looked and couldn’t find a single other example. Feel free to google yourself.
I would poke holes in this line of evidence by asking why it’s such a rare condition that we have 2 pictures out of a profession that has employed 10’s of millions of people worldwide for decades. Why aren’t there 10’s to 100’s of thousands of people that look like this? Shouldn’t you be able to see someone with a half wrinkled face and say: “oh, truck driver” just like I can see someone with cauliflower ear and know “oh, fighter.”
That would beg the question of whether these patients have some other unique factors which are necessary for being extra sensitive to sun. Do they have some rare genetic variant, are they on some medications, are they heavy smokers, etc.
As with cancer, there are very limited data directly measuring aging in a randomized trial. Trials conducted in the same way, half of the group assigned to wear sunscreen rigorously, half told to do whatever they’ve been doing, measuring skin aging before and after 4 years, don’t show a dramatic effect. For the men in the trial there was no measurable difference. For women it seemed to help a bit, but the confidence intervals overlap.
Overall, for aging I think sunscreen helps, but it’s not as big of a difference as the “truck driver face” would imply. People obviously look hotter tan than pasty, so you’re trading off being hotter now for being a bit less wrinkly later. Different strokes.
Harms of sunscreen
And the grand finale; are there adverse health effects to sunscreen?
Frankly, I can’t make any strong claims about this. There is no randomized evidence to aid the question. Dermatologists will be commonly quoted as recommending a maximalist sunscreen regimen. That means rubbing chemicals on your skin multiple times per day.
FDA scientists have investigated whether some of the active ingredients of sunscreen are absorbed into the blood through the skin. In this study the participants put on sunscreen, and then had blood drawn to measure whether the chemicals got into their bodies. After just one application all of the chemicals investigated reached a blood concentration higher than what the FDA considers safe. Some of them reach 14x higher than the FDA’s safe threshold with just one application. Other studies have shown that some of the chemicals would also be found in breastmilk of nursing mothers. That is a bit concerning.
One more concern is that these chemicals are slow to be eliminated from the body, that means if you use them every day, the concentrations will keep increasing.
The problem is that no one really knows if the FDA limits make sense or what the chemicals do when inside the body. Some studies report that some chemicals are hormone disruptors but inject fish as the model, some studies try to look at correlations and find less testosterone in boys with the chemicals in their blood.
There are long-term concerning societal hormone related health trends, women are going through puberty earlier, testosterone and sperm counts are going down in men, but to link these to sunscreen would be taking a big leap. In general, my take is that ~80% of negative health trends can be explained with rising obesity, and decreased physical activity and most of the other stuff, like microplastics, sunscreen, etc. is marginal.
So, as with the flip side, you need randomized evidence to know whether this matters. For what it’s worth, Europe is more careful and doesn’t allow for the same chemical concentrations in sunscreen as the US, because of the hormone disruption concerns. In the interest of being careful with what chemicals you get put into your blood, this website has a decent but still fear-mongering breakdown of the chemicals of potential concern. Frankly, I would just wear sunscreen less often rather than obsessing about chemicals.
Conclusions:
I will end with a poorly evidenced claim. Getting sunshine is good for you.
Why it’s good for you, we don’t know; giving people a vitamin D pill doesn’t seem to have the same effects as sunshine on human health. There’s a lot of correlational evidence between vitamin D levels in the blood and good health outcomes, but it’s impossible to disentangle whether healthier people tend to spend more time in the sun, or if the sun itself is contributing to health.
If I were to guess, I would guess sunshine is good for you because of some regulation of circadian rhythm and the immune system/inflammation. I’ll trust my intuition, that I feel better, more energetic and healthy when I’ve gotten enough sun to get a good tan. Because I don’t want to absolutely fry on a beach day, I’ll put on SPF 15 when I go, or SPF 4 if I’m just on the beach for a couple hours. If I’m just out and about, inside and outside, I don’t put on sunscreen.
The public health wave has been to sell a maximalist vision of hiding from the sun. It’s time to stop listening.