Search This Blog

Wednesday 16 January 2019

Don't Drink (oil) and Fry (in the sun) - the link between polyunsaturated vegetable oil and skin cancer

It was good to see this well-researched burst of sunscreen scepticism doing the rounds this (southern) summer - Is Sunscreen the New Margarine? *
I was particularly struck by the unintentionally ironic title, because margarine's role in this story goes well beyond that of a handy analogy for a misguided public health initiative.

When I first read the Nanji and French research - countless rat experiments, consistent with what human epidemiology there was - showing that high PUFA oils potentiated the progression of alcoholic liver disease, were indeed essential fats if one wanted to develop this condition, I replaced the rice bran oil I had been conned into using with beef fat. One of the first things I noticed, being a clumsy cook, was that my skin stopped blistering when I burned myself. At the worst I might get a short-lived patch of dead skin, but I've had about 2 blisters in the 7 years since then and neither became painful. Then gradually I noticed I wasn't getting burnt in the sun as easily. I wasn't big on the sun in those days, but as I got more exposure to sun and less exposure to linoleic acid my resistance grew. I try to limit exposure, using clothes and shelter, to what seems reasonable, having no desire to turn nut-brown and wrinkled in my dotage, but occasionally I've been caught out for far longer than I intended, and no harm has ensued. In part this has been due to a policy of tanning early and often so that the protective pigment is in full effect once the summer reaches its peak, but dietary choice is a critical factor, as we shall see.

The first experiment I came across showing the photosentising effect of linoleic acid was this one (in hairless mice - obviously you can't use a hairy animal) [1];

However, dietary ALA inhibited the increase in erythema score after UVB irradiation compared with LA. The peroxidizability index of the skin total lipids was significantly higher, but UVB-induced prostaglandin E2 (PGE2) production was significantly lower in the group fed an ALA-rich diet compared with the group fed an LA-rich diet. 

There's a lot of this stuff recently, showing that omega-3 fats, and especially the long-chain fish oil ones, are protective - but also that their irradiation puts more strain on the antioxidant defense system, depleting vitamin C and glutathione. The Linus Pauling Institute website has an excellent summary of this research, including human trials, but if you think like me you want to know, is saturated fat protective too? I mean there are people out there exposed to very high UV levels eating not-especially-oily fish and coconut or ghee for generations - are they OK?

The Linus Pauling Institute doesn't tell us, nor does it really warn us about omega-6. The modern idea is just to increase our intake of flaxseed oil and salmon and antioxidant vitamins and minerals. Sounds expensive, for one thing, and I'm not sure our obsession with fish will prove to be sustainable if we're not allowed to eat anything else.

To answer this question we need to go back to the 1990's, when researchers were mostly looking at omega 6 fats and using saturated fats as controls.

We find:

When polyunsaturated fat intake had been increased, tumorigenesis had been exacerbated and a remarkable persistence of immunosuppression remained measurable. There was no difference apparent in the CHS responsiveness in mice fed 20% sunflower oil or saturated fat in the absence of UV irradiation, indicating that the persistent immunosuppression was likely to have been induced by the carcinogenic irradiation regime.[2]

Diet 1 is 20% hydrogenated cottonseed oil, diet 5 is 20% sunflower oil

Now 20% sunflower oil is exactly the kind of exposure you'd get if you replaced other fats with oil in a low fat diet, as the experts want you to. It's an amount of LA that the epidemiologists at Harvard Chan have no problem with, using as they do data collection methods unreliable enough to produce false negatives as well as false positives, and not controlling for UV exposure anyway. Note that the control here - hydrogenated cottonseed oil - is 69.25% trans fat, the rest all SFA.

Here's another, where there are two controls - 12% menhaden oil (basically fish oil) and 0.75% corn oil.[3]
The most interesting finding is that while menhaden oil is protective, crossing over to 12% menhaden oil from 12% corn oil is not; it's about the worst thing you can do; crossing over from 12% corn oil to 0.75% corn oil seems safer.

In summary, we have shown that (1) high dietary level of an omega-6 FA source (corn oil) enhances photocarcinogenic expression, both with respect to tumor latent period and multiplicity; (2) that this lipid-induced exacerbation of cancer expression occurs at the post-UV initiation, or promotion, stage of carcinogenesis; (3) modification of diet to low lipid level (corn oil), after UV-initiation, negates the enhancement of cancer expression exhibited by high level of corn oil; (4) high level of dietary menhaden oil containing omega-3 FA, when compared to an equivalent level of corn oil, inhibits photocarcinogenic expression; and (5) menhaden oil mediated inhibition of UV-carcinogenesis appears to occur during the initiation stage and by a mechanism dissimilar to that exerted by low levels of corn oil.

These data suggest a complexity of dietary lipid effects upon cancer expression that perhaps has not been widely appreciated.

But here's my favourite - finally someone uses butter and ghee as controls.[4]

They're measuring ear swelling in response to a hapten test, which I assume is some sort of irritant. Low swelling means the immune system is suppressed, and this predicts that the risk of cancers is increased, just as it did in the other experiments. Note this took 4 weeks of a 20% fat diet - the protective effect of butter wasn't obvious after only 2 weeks, but the harmful effect of sunflower oil was.

Now, how do we know that this applies to human populations or has anything to do with the high rates of melanoma (the most fatal skin cancer) in New Zealand or Australia?
There is a natural experiment we can look at. NZ, unlike Australia, is not a land of extreme natural sunlight - it's the "land of the long white cloud", so that travel to the tropics greatly increases UV exposure. (However, we do have greater UV variability at our lower latitude, described by NIWA's Richard Mackenzie here.) Kiwis didn't travel much in the past, especially to the tropics, with one important exception; around 140,000 men (mostly) served overseas in WW2, around 10% of the population, and most of them served in places like Greece, Crete, Egypt, Libya, Italy and the Solomon Islands. Wearing shorts and short-sleeved shirts - US naval officials in the Solomons were appalled by the lack of protection the NZ sailor's uniform gave against flash burns - with only a Tommy helmet for shade (too hot and heavy to wear all the time). Sunburn on arrival in the combat zone is mentioned in memoirs - in the Armed Forces you stay where you're told and work where you're told, shelter or not.

Soldiers of the 2nd NZEF20th Battalion, C Company marching in Baggush, Egypt, September 1941.
The NZ diet, and British Army rations, at this time were very low in omega 6. In NZ, margarine wasn't even legalised until 1972. We cooked with butter, hydrogenated coconut oil, and beef and lamb dripping - even chicken and pork were unusual foods until the 70's. Some time during that decade the use of oil (mostly soy and corn oil at first) took off and was entrenched by the 1980's.
Now if exposure to the excess UV rays of the tropics, rather than the combination of UV and oil, caused skin cancer we'd expect to see an age-specific curve in skin cancer mortality in NZ that rose after the war (starting in the mid-1950's would match the usual cancer latency) then dropped as that generation began to die off from all causes, and a generation raised on "slip slop slap" UV protection took over.
Instead we see this:

And if you check different age groups in this part of the Mortality Trends website, you find the same pattern (most obvious where there is highest mortality), despite the different rates of overseas service in each age group, and the dying off of each generation - mortality from skin cancer climbs rapidly from the 1970's, when the NZ nutritional transition to high-PUFA oils began, and has not declined at all as the Greatest Generation dies off and the lesser slip, slop, slap-happy generations take its place (melanoma mortality in NZ has remained stable from 2001 to the present, I can't find earlier stats specific for melanoma but of course it is the most lethal skin cancer). This data is a better fit for diet than it is for UV exposure. (Of course there's the ozone hole, dating from the 70's, but "The ozone hole does not have a large effect on the concentration of ozone over New Zealand. However, when the ozone hole breaks up in spring, it can send ‘plumes’ of ozone-depleted air over New Zealand." The relative unimportance of this is mentioned on page 6 of Mackenzie's paper.)
Which is not to say that UV exposure has unlimited safety, of course this is not the case, but that heart-healthy vegetable oil and margarine advice has made even limited exposure more dangerous than it needs to be. Much the same phenomenon we see with alcohol and liver disease (and, who knows, alcohol and the risk of other cancers).

Anyway, if you plan to go out in the sun for long enough to raise some vitamin D and nitric oxide, or drink enough alcohol to get drunk, or especially both, skip anything cooked in or made with a vegetable oil other than coconut or, within reason, olive oil (this includes mayonnaise), [Edit: olive oil is rich in the terpenoid antioxidant squalene, perfectly configured to quench singlet oxygen radicals released by UV exposure, likely to offset risk from its moderate 10-12% linoleic acid content], eat some butter or cheese or full-fat yoghurt, some fatty fish, chocolate, tomatoes and all the other antioxidant-rich foods with some evidence as UV-protective (caffeine seems to help too). Also keep insulin low - that means no refined carbs, intermittent fasting, low carb diet if necessary. Insulin and IGF-1 are important in skin repair and insulin resistance due to excessive insulin response to diet creates all sorts of skin problems, as well as underwriting most of those non-communicable diseases we hear so much about. (it's not easy to get insulin tested here but the TG/HDL ratio after a 12-14 hr fast is a decent proxy for the 2-hour insulin response).

[EDIT 11/02/19 - analysis of NHS/HPFS data has confirmed an association between omega 6 PUFA and skin cancers, up to 1.23 (1.08, 1.41) for squamous cell carcinoma. Of course this is an unreliable FFQ data set, but the finding runs contrary to the direction we'd expect conscientiousness bias to take in this population. It's not obvious what the range of omega 6 was, this information may be in supplementary tables I can't access.
Park MK, Li WQ, Qureshi AA, Cho, E. Fat intake and risk of skin cancer in u.s. adults. Cancer Epidemiol Biomarkers Prev. 2018 Jul;27(7):776-782. doi: 10.1158/1055-9965.EPI-17-0762. Epub 2018 Apr 10.]

* IMO the vitamin D supplement scepticism in this article is premature - vit D3 works if you use it for the right reason and take enough of it. It's just as good as sunlight for fixing my minor but annoying psoriasis, but I need 10,000 iu a day till it goes away. This only takes 6 days every month or two now - I used to need to take that amount all winter.


1] Takemura N, Takahashi K, Tanaka H, Ihara Y, Ikemoto A, Fujii Y, Okuyama H. Dietary, but not topical, alpha-linolenic acid suppresses UVB-induced skin injury in hairless mice when compared with linoleic acids. Photochem Photobiol. 2002 Dec;76(6):657-63. link

2] Reeve VE, Bosnic M, Boehm-Wilcox C. Dependence of photocarcinogenesis and photoimmunosuppression in the hairless mouse on dietary polyunsaturated fat. Cancer Lett. 1996 Nov 29;108(2):271-9. Full-text link

3] Homer S. Black  John I. Thornby  Janette Gerguis  Wanda Lenger. INFLUENCE OF DIETARY OMEGA‐6, ‐3 FATTY ACID SOURCES ON THE INITIATION AND PROMOTION STAGES OF PHOTOCARCINOGENESIS. Photochemistry and Photobiology Vol. 56, No. 2, pp. 195-199, 1992.

4] Cope RB, Bosnic M, Boehm-Wilcox C, Mohr D, Reeve VE. Dietary butter protects against ultraviolet radiation-induced suppression of contact hypersensitivity in Skh:HR-1 hairless mice. J Nutr. 1996 Mar;126(3):681-92. free full-text