The dietary skincare evidence quality ranking: dairy, sugar, fish oil, vitamin A

A reader asked me last summer to fact-check the dietary advice she had gathered from four different sources: her dermatologist, her nutritionist, a popular skincare account, and a podcast about hormones. The four sources contradicted each other on at least seven specifics, and she wanted to know which one to follow. The honest answer was none of them, because none of them were ranking their claims by evidence quality, and the question of “what should I eat for my skin” cannot be answered without that ranking.

I want to do that ranking here. There are about a dozen common dietary skincare claims, and they sit on completely different evidence tiers. Treating them as a single category called “diet matters for skin” produces the bad advice the reader had collected. Let me walk through the four most cited claims, with the actual papers and what they do and do not support.

Tier 1: dairy and acne (moderate evidence, plausible mechanism)

The dairy-acne hypothesis has more data behind it than most people realise, and it also has more uncertainty than the loudest accounts admit. Juhl et al. 2018 (PMID: 30096883) is the best place to start. This is a systematic review and meta-analysis covering 14 observational studies and around 78,500 participants, looking at dairy intake and acne prevalence. The pooled odds ratio for any dairy intake versus none was 1.25 (95% CI 1.15-1.36), with stronger effects for skim milk (OR 1.32) than for whole milk (OR 1.22) and weaker effects for cheese and yoghurt.

Several things to note about this. First, the effect size is real but not enormous. A 25-32% increase in acne odds is meaningful at the population level but it is not “stop drinking milk and your acne disappears.” Second, the effect is stronger for skim than for whole milk, which is the opposite of what most people guess. The leading hypothesis is that the fat in whole milk slows the absorption of milk-borne IGF-1 precursors, blunting the spike that drives sebocyte activity. Third, fermented dairy (yoghurt, kefir) showed weaker or null associations across studies. The fermentation may degrade some of the bioactive components.

The mechanistic story (IGF-1 signalling driving sebocyte differentiation and androgen receptor activity) is plausible and supported by Burris et al.’s 2013 review (PMID: 23438493). It is not airtight. We do not have a randomised controlled trial of dairy elimination on acne outcomes at scale. The Adebamowo studies (2005 in nurses, 2006 in adolescent girls) are the source of most of the observational signal and they are reasonable but they are not interventional.

What I tell people who ask: if you have moderate to severe acne and you drink more than two glasses of milk a day, particularly skim, it is worth a three-month elimination test. Cheese and yoghurt are lower priority. If your acne is mild and your dairy intake is moderate, the expected benefit from elimination is probably less than the benefit from other interventions. The data supports trying, not promising.

Tier 1: high glycaemic load and acne (moderate evidence, one good RCT)

This is the strongest interventional dietary signal in the acne literature. Smith et al. 2007 (PMID: 17448569) ran a 12-week parallel-arm RCT comparing a low-glycaemic-load diet to a conventional diet in 43 young men with acne. The low-GL arm showed a mean reduction of 21.9 total lesions versus 13.8 in the control arm, with corresponding improvements in inflammatory lesions. The same group published several follow-up studies with consistent results. The Australian dermatology group around Mann has produced most of the strong evidence here.

Subsequent studies have replicated this pattern with smaller effect sizes. The Korean and Turkish trials in the early 2010s showed similar but weaker effects. Burris et al. 2013 summarises this body of work and concludes that dietary glycaemic load probably matters for acne in genetically susceptible individuals.

The mechanism is again IGF-1 driven. High glycaemic load produces postprandial insulin spikes, which suppress IGFBP-3 (the IGF-1 binding protein), which increases free IGF-1, which acts on sebocytes. This pathway is the same one that explains the dairy effect, which is why the two interventions often appear together in clinical advice.

What I tell people who ask: a low-glycaemic-load diet for 12 weeks is one of the few dietary interventions with actual RCT support for skin outcomes. The effect size is moderate. It is most likely to help people with hormonally driven adult acne, particularly in the chin and jawline distribution. It is unlikely to help with non-inflammatory closed comedones or with rosacea or with eczema. Do not generalise.

Tier 2: fish oil and inflammatory skin (limited evidence, small RCTs)

The omega-3 story is older and weaker than the marketing suggests. Pilkington et al. 2011 (PMID: 21569104) reviews the photoprotection side of the argument, and there is genuine evidence that EPA and DHA supplementation reduces UV-induced erythema and may reduce some markers of photodamage. The effect sizes are real but small.

For atopic dermatitis, Kim et al. 2014 (PMID: 22455881) systematically reviewed omega-3 supplementation trials and found mixed results, with about half of the included studies showing modest improvement and the rest null. The pooled effect, if you trust the meta-analysis, is favourable but not large.

For acne specifically, the evidence is weaker still. A handful of small trials of EPA-rich supplements have shown modest reductions in inflammatory lesions over 10 to 12 weeks. These trials are usually underpowered, often unblinded, and have not been replicated at scale.

For rosacea, the evidence is essentially preliminary. There are mechanistic reasons to expect benefit (omega-3 derived resolvins are anti-inflammatory) but the clinical trial base is small.

What I tell people who ask: if you eat fish twice a week, you are probably getting the relevant dose without supplementation. If you do not, 1 to 2 grams of EPA+DHA combined per day is a reasonable supplemental dose for general anti-inflammatory effects, and it might marginally help inflammatory skin conditions. The skin benefit alone is not enough reason to take fish oil. The combined cardiovascular and skin argument is stronger. Buy a product that has been tested for oxidation (peroxide value matters more than brand). Refrigerate after opening.

Tier 3: dietary vitamin A and skin (almost no evidence outside deficiency)

This is the claim where the marketing massively overshoots the evidence. The argument goes: topical retinoids are vitamin A derivatives, vitamin A is essential for skin, therefore eating more vitamin A is good for skin. Each step of this argument is partially true and the conclusion does not follow.

In frank vitamin A deficiency (serum retinol below 0.7 micromol/L), skin shows real changes: follicular hyperkeratosis on the upper arms and thighs (phrynoderma), xerosis, and impaired wound healing. Correcting the deficiency reverses these changes. This is well established. It is also irrelevant to most readers in well-fed countries, where overt vitamin A deficiency is rare outside of malabsorption syndromes.

Above the deficiency threshold, the relationship between dietary vitamin A and skin appearance is essentially flat. Sorg and Saurat’s 2014 review (PMID: 24821234) covers this. Increased dietary vitamin A does not produce the anti-aging effects of topical tretinoin. The reason is pharmacokinetic. Topical retinoids deliver micromolar concentrations of all-trans retinoic acid directly to dermal fibroblasts. Dietary vitamin A is absorbed, esterified, stored in the liver, and released into circulation at tightly regulated levels. The skin receives a small fraction of what circulates, and the local conversion to retinoic acid is also regulated. You cannot bypass these controls by eating more sweet potato.

The supplementation literature is worse. High-dose vitamin A supplementation produces hepatotoxicity, hyperostosis, and teratogenicity. The thresholds for harm are not far above recommended intake (the upper limit for adults is 3000 micrograms retinol activity equivalents per day). The skin benefit at sub-toxic doses is nil. Most of the “vitamin A for skin” supplement market is selling beta-carotene, which has a regulated conversion to retinol and a much wider safety margin, but also a much weaker skin signal.

What I tell people who ask: eat a reasonable variety of food. Vitamin A deficiency is rare. Vitamin A toxicity is rarer but worse when it happens. The relationship between dietary vitamin A and skin appearance, above the deficiency threshold, is essentially absent. If you want the skin effects of retinoids, use a topical retinoid.

The tier I keep getting asked about: chocolate and acne (almost no evidence)

I want to address this briefly because it comes up constantly. The 1960s “Pennsylvania State Prison” study that purportedly cleared chocolate of acne involvement is methodologically poor. The studies since have been mixed. A small 2014 trial in adolescent males showed an increase in acne lesions after pure chocolate consumption, but it was tiny (n=14), uncontrolled for glycaemic load, and never replicated convincingly.

The honest answer is that chocolate is mostly sugar and dairy, and the dairy and high-glycaemic-load arguments above probably account for any chocolate effect. Pure cocoa with no sugar or dairy is probably neutral for acne. A milk chocolate bar is delivering both confounders simultaneously.

What I changed about my own diet

I dropped from two glasses of milk a day to none, six years ago, and the chin acne I had been treating for years with topicals improved over about four months. I cannot prove the milk was the cause. The before-and-after timeline was suggestive but not definitive. I added back cheese and yoghurt without obvious effects on skin and they have stayed in my diet since.

I do not eat low-glycaemic by design, but I do not eat sugary breakfast either, and my breakfast has been roughly the same eggs-plus-vegetables-plus-tea for years. I cannot tell you if this is the right framework or just the framework I am used to.

I eat fish about twice a week and do not supplement omega-3. My skin is fine. This is not strong evidence of anything except that the marginal effect of fish oil supplementation in someone who already eats fish is small.

I do not take any vitamin A or beta-carotene supplements and I would push back on anyone recommending them for skin benefit.

FAQ

Is dairy a problem for everyone with acne or just some people?

Probably just some. The Juhl meta-analysis showed an average effect but not a universal one. About 30 to 40 percent of acne patients show clinical improvement on three-month dairy elimination in clinic-based observation, with the rest showing no change. There is no good test for predicting who will respond. The only way to find out is to try.

Does collagen supplementation work?

I am leaving this question for a separate piece because it deserves more space than it would get here. The short answer is that there is some evidence for hydrolysed collagen peptides improving skin elasticity in older women, the effect size is modest, the mechanism is debated (probably not direct deposition; probably signalling through di- and tripeptide receptors), and the supplement market is wildly unregulated.

Are seed oils bad for skin?

The seed oil panic does not survive contact with the dermatological literature. There is no good evidence that linoleic acid or other omega-6 fats in seed oils worsen skin outcomes at normal dietary intake. The acne and rosacea data does not single out seed oils. The atopic dermatitis literature actually shows linoleic-rich oils being topically beneficial.

Should I cut sugar to clear my acne?

If by “cut sugar” you mean lower the overall glycaemic load of your diet, that has real RCT support. If you mean cut all sweetness from your diet for 30 days, the literature does not support that specifically. The mechanism is the postprandial insulin curve, not the sweetness on the tongue.

Does drinking water help skin?

If you are dehydrated, rehydration helps everything including skin. If you are already well hydrated, drinking more water produces no further skin benefit in any controlled study I have read. The eight-glasses-a-day folk wisdom for skin appearance is layered on a real but narrow physiological truth and does not generalise the way the wellness aisle pretends it does.

What this means for the routine

The dietary skincare claim space is uneven. Dairy and high-glycaemic load have real moderate evidence and are worth a structured trial if you have hormonally driven acne. Fish oil has weak-to-moderate evidence and is reasonable if you do not eat fish. Dietary vitamin A has essentially no evidence above deficiency. The single biggest error people make is treating all of these as one category called “diet for skin” and trying to do all of them at once, which makes the responsible interventions impossible to evaluate against the irrelevant ones. Pick one. Test it for 12 weeks. Document the result with photos. Then decide what is worth keeping.

Related Elelaf tools

• Cystic acne severity
• Stress acne quiz
• Skin type test

Sources

1. Juhl CR, Bergholdt HKM, Miller IM, Jemec GBE, Kanters JK, Ellervik C. Dairy Intake and Acne Vulgaris: A Systematic Review and Meta-Analysis of 78,529 Children, Adolescents, and Young Adults. Nutrients. 2018;10(8):1049. PMID: 30096883
2. Smith RN, Mann NJ, Braue A, Mäkeläinen H, Varigos GA. The effect of a high-protein, low glycemic-load diet versus a conventional, high glycemic-load diet on biochemical parameters associated with acne vulgaris: a randomized, investigator-masked, controlled trial. J Am Acad Dermatol. 2007;57(2):247-256. PMID: 17448569
3. Burris J, Rietkerk W, Woolf K. Acne: the role of medical nutrition therapy. J Acad Nutr Diet. 2013;113(3):416-430. PMID: 23438493
4. Pilkington SM, Watson RE, Nicolaou A, Rhodes LE. Omega-3 polyunsaturated fatty acids: photoprotective macronutrients. Exp Dermatol. 2011;20(7):537-543. PMID: 21569104
5. Kim H, Kim H, Lee E, Park HJ. Effects of dietary omega-3 fatty acid supplementation in patients with atopic dermatitis: a systematic review. J Dermatolog Treat. 2014;25(2):137-145. PMID: 22455881
6. Sorg O, Saurat JH. Topical retinoids in skincare: where do we stand? Dermatology. 2014;228(4):314-325. PMID: 24821234
7. Adebamowo CA, Spiegelman D, Berkey CS, et al. Milk consumption and acne in adolescent girls. Dermatol Online J. 2006;12(4):1. PMID: 17083856