Retinol, tretinoin and your oil glands: the skincare question we ask at the Dry Eye Evaluation

One of the first questions we ask at our Hamilton dry eye clinic can surprise people: what do you put on your face at night? Retinol, retinaldehyde, tretinoin and other vitamin A skincare are everywhere, usually applied close to the eyes. Because the meibomian (oil) glands in your eyelids are sebaceous glands, and vitamin A derivatives are used in dermatology precisely because they quieten sebaceous glands, it is fair to ask whether these products could be feeding a dry eye problem. A 2026 systematic review and meta-analysis in Investigative Ophthalmology & Visual Science has pulled the evidence together. The honest answer: there is a signal, but a blurrier one than the headlines suggest.

What the reviewers did

Abbott and colleagues searched four databases up to 25 August 2025 for any human study in which a vitamin A product was applied to the eye, eyelids or periorbital area. They registered the review in advance (PROSPERO CRD42024523189), used two reviewers for screening and extraction, rated risk of bias, and graded certainty of evidence using GRADE. They ended up with 89 studies and 10,794 participants; 55 studies provided harms data that could be pooled.

Three quarters (67 of 89) were randomised controlled trials, but most were dermatology trials of acne or photoageing where eye comfort was a side note. Nearly half (43, 48%) were industry funded and another 35% did not say who paid. Two thirds lasted under three months; only one went beyond a year. Crucially for a dry eye reader, 72 of the 89 studies (81%) reported no ocular surface outcomes at all.

What the pooled numbers show

Compared with non-vitamin A controls (usually the same cream without the active ingredient), vitamin A derivatives were associated with higher odds of several local harms, the three largest being:

• Pain: odds ratio 6.37 (95% CI 4.47–9.08; 10 studies, 2,914 participants; very low certainty)
• Dryness: odds ratio 2.69 (95% CI 2.11–3.42; 12 studies, 3,281 participants; low certainty)
• Burning: odds ratio 2.44 (95% CI 1.68–3.56; 8 studies, 1,600 participants; low certainty)

The odds of "any harm" were 1.71 times higher (95% CI 1.45–2.02; 15 studies, 3,579 participants; very low certainty). Serious adverse events were rare and not clearly increased: OR 1.21 (95% CI 0.53–2.76; 6 studies, 3,099 participants).

A quick word on those brackets. An odds ratio above 1 means the event was more common in the vitamin A group. The 95% confidence interval (CI) is the range of values the data are reasonably consistent with; if it includes 1.0, as it does for serious events, the study cannot rule out no difference. GRADE "low" and "very low" certainty mean further research could easily change the answer.

Prescription versus over-the-counter

The reviewers split the "any harm" analysis by strength. High-strength prescription retinoids (tretinoin, tazarotene, retinoic acid, palovarotene) were associated with 1.74 times the odds of any harm (95% CI 1.46–2.06; 11 studies, 3,220 participants). Low- and moderate-strength non-prescription products (retinol, retinyl esters, retinaldehyde and unspecified vitamin A) showed no clear association: odds ratio 1.33 (95% CI 0.68–2.59; 4 studies, 359 participants).

Before concluding that retinol is safe and tretinoin is not, notice two things. The non-prescription estimate rests on only four studies and 359 people, so it is imprecise rather than reassuring. And the formal test for a difference between the two groups was not significant (P = 0.45), so the data cannot confirm that prescription products are riskier than shop-bought ones. The authors describe over-the-counter retinyl esters as "less potent but not without risk, particularly near the eyes".

What about the oil glands themselves?

Here the evidence is thinner than the biology. Oral isotretinoin is known to cause meibomian gland atrophy, which is why the question matters. But among 89 studies, only two measured meibomian gland parameters, and only one gave usable data. One trial of vitamin A palmitate gel combined with cyclosporine in people with meibomian gland dysfunction reported improved gland secretion after 12 weeks; another, in cataract surgery patients, found no significant difference in meibography scores between a vitamin A eye gel and two other lubricants. Dry eye symptoms were reported in 13 studies, but the measures were too varied to pool; the review describes "small or negligible changes in symptoms, with no clear evidence of a consistent benefit or harm".

So the review does not show that topical vitamin A damages meibomian glands. It shows that almost nobody has looked properly, while these products are associated with more pain, dryness and burning.

How this shapes what we do

At Rose Optometry the optometrist team, including Jagrut Lallu and Jacqueline Rowe, ask about skincare as part of the history at the Dry Eye Evaluation, a paid consultation that includes meibography, tear film testing and a proper look at the lids. If you use retinoids near the eyes and have symptoms, we do not automatically tell you to stop; a good acne or anti-ageing routine matters. We check whether your glands and tear film show signs that fit, discuss keeping the product away from the lash line, and treat the underlying gland dysfunction with tools that have evidence behind them. Our overview of dry eye treatments and our earlier piece on what a dry eye treatment plan looks like explain how. If a dermatologist has prescribed tretinoin or an oral retinoid, we would rather coordinate with them than have you stop unilaterally.

What the studies can't tell us

• Whether harms were ocular or skin-related. Fifty-five percent of studies did not separate the two, so "dryness" often means dry skin, not dry eye.
• Whether application site matters. Reporting was too poor to compare direct-to-eye, periorbital and general facial use.
• Whether dose and duration matter. Most studies were shorter than three months; real-world retinol use runs for years.
• Whether over-the-counter products are genuinely safer. Four studies and 359 people cannot answer that.
• Whether meibomian gland structure changes. Two studies, one usable.
• The headline numbers are low or very low certainty, largely because of high risk of bias in the underlying trials and likely under-reporting of mild harms. Screening was also stopped early using an AI prediction model, a risk the authors judged minimal.

We could not identify the review's own funding source from the full text available to us; the included trials were largely industry funded.

If your eyes sting, burn or feel dry and you use vitamin A products on your face, mention it when you book. The Rose Optometry dry eye clinic in Hamilton treats the cause rather than the symptom, and skincare is now one of the causes we check for. Readers elsewhere can find a member practice nearby.

Where to get this looked at

If you live in Hamilton, the two places to start are Rose Optometry in the city, where the optometrist team (Jagrut Lallu, Jacqueline Rowe and colleagues) runs the dry eye clinic, and Visique Rototuna Optometrists on the north side of town. Both are member practices of the Dry Eye Specialist Group and work the same way: a paid Dry Eye Evaluation first, to find out what is actually driving your symptoms, and only then a treatment plan.

If you don't live in Hamilton, see your local Dry Eye Specialist Group member. The group is a network of independent practices across New Zealand that share this treat-the-cause approach: Illume Eye Care (Ponsonby, Auckland), Bay Eye Care (Tauranga), Feilding Visique Optometry, Naylor Palmer Optometrists (Palmerston North), Ashburton Eyecare and Milburn & Neill Optometrists (Dunedin), alongside the two Hamilton practices. Find your nearest one on the member practices page.

Reference

• Abbott KS et al. Potential Harms of Ocular and Periorbital Topical Vitamin A Application: A Systematic Review and Meta-Analysis. Invest Ophthalmol Vis Sci. 2026;67(11):11. PMID 42708560. https://doi.org/10.1167/iovs.67.11.11