TL;DR: Retinol is sold as something everyone over twenty-five needs forever. The honest answer is that some people should stop, some should switch, and a small group genuinely benefits from staying on it for years. The deciding factors are skin behavior, not age, and the breakup is usually less dramatic than the marketing suggests.
A reader in Brisbane wrote to me in March. She had been on a 0.5% retinol serum for almost four years, applied four nights a week, exactly as the brand told her. Her skin was thinner at the cheeks, flushing easily after coffee, stinging when she put on sunscreen. Her dermatologist told her to stop. The brand’s customer service told her she was probably “in a purge.” She wanted to know which one to believe.
I get a version of this email every six weeks. The conversation about retinoids has become so saturated with the message that everyone needs them, forever, that the off-ramp has gone missing. Some people genuinely should stop using retinol. Some should switch to a different molecule. A smaller group will get years of benefit from staying on it. The marketing treats all three the same way, which is the problem.
What the studies actually show about long-term use
The Mukherjee 2006 review in Clinical Interventions in Aging is the paper most often cited to defend long-term topical retinoid use (PMID: 18046911). It found benefits for photoaging, fine lines, and pigmentation across multiple trials. What gets quoted less often is the section on irritation. The review notes that retinoid dermatitis is “the most common adverse effect” and is dose-dependent. It does not say everyone should be on retinol forever. It says retinoids work when they are tolerated.
Kang 1995 in the Journal of Investigative Dermatology compared 0.025% and 0.1% tretinoin over forty-eight weeks (PMID: 7561157). Both improved photoaging. The higher concentration produced more irritation without proportionally more benefit at the end of the study. The graph that matters in that paper is the irritation curve, which stays elevated for the whole forty-eight weeks. It does not “fully normalize” the way brands like to claim. The skin tolerates it. That is not the same as adapting.
Bouloc 2015 looked at a low-dose retinol formulation over twelve weeks on photoaged skin (PMID: 25627659). The improvement was real but modest, around 15 to 20 percent on most measured parameters. The control group, using only the vehicle plus sunscreen, also improved. Sunscreen alone gets you a meaningful chunk of what retinol gets you, which is the awkward sentence the industry never puts in a campaign.
The thing the studies do not tell you, because they end at twelve or forty-eight weeks, is what happens at year three or year four. The answer is that we do not actually have good data. Long-term safety data on cosmetic-strength retinol over a decade does not exist in any rigorous form. We have prescription tretinoin data going back further, but those studies are on patients who are monitored. They are not on someone applying a 1% retinol with a 10% niacinamide and a fragrance every night for five years.
The four people who should stop
The first group is people whose barrier is visibly compromised. The signs are stinging on water, a low tolerance for any acid, flushing in normal indoor temperatures, and tightness that does not resolve with a basic moisturizer. The Bouloc paper and the Draelos work on barrier function (PMID: 18045358) both note that continued retinoid use on a damaged stratum corneum is the dermatology equivalent of running on a sprained ankle. The damage is real and it accumulates. Stopping for eight to twelve weeks is not a failure of discipline. It is the only thing that will let the lipid matrix rebuild.
The second group is people who developed perioral dermatitis, periorificial flushing, or rosacea-like symptoms after starting retinol. I have stopped being surprised at how often a multi-year retinol user shows up with what they think is acne around the mouth and nose. Retinoids can trigger or worsen perioral dermatitis in susceptible people. Stopping is the first step. Topical or oral treatment is the second. Going back on retinol after recovery is usually a mistake, even at a lower strength.
The third group is people on prescription tretinoin for a condition that has resolved. If you went on 0.025% tretinoin for inflammatory acne at twenty-two and you are now thirty-five with clear skin, you are not getting acne benefit from the cream anymore. You are getting photoaging prevention, which is a smaller and slower benefit than the marketing implies, and which sunscreen does most of. The decision to stay on it should be deliberate, not inertia.
The fourth group is pregnant people, people trying to conceive, and people breastfeeding. Topical retinoids are contraindicated in pregnancy by every major dermatology body. The data on cosmetic retinol is weaker than for tretinoin, but the precautionary default is to stop. The American Academy of Dermatology, the AAD, has been clear on this for years.
Who should switch instead of quitting
A different group does not need to stop using a retinoid. They need to stop using their specific retinoid.
If you are on a high-strength retinol (0.5% or 1%) and your skin tolerates it but does not seem to improve anymore, you have probably stopped getting incremental benefit. The dose-response curve flattens. Moving to a lower concentration, or moving to a different molecule like adapalene or a low-strength tretinoin, often resets the response. The Babcock 2015 paper on retinaldehyde noted that lower-irritation derivatives produce comparable outcomes at lower concentrations in some skin types. Switching is not downgrading. It is matching the tool to the job.
If you are using a retinol that is also formulated with fragrance, alcohol denat, or a strong essential oil blend, the irritation you are blaming on retinol is probably the formulation. Switching to a plain retinol in a basic vehicle, or to an over-the-counter adapalene (0.1%), removes that variable. I have seen people quit retinoids entirely when their problem was the perfume in the bottle.
If you started on retinol because a brand told you everyone over twenty-five needs it, and you do not have acne, hyperpigmentation, or visible photoaging that bothers you, you might consider whether you ever needed to be on it in the first place. The decision to start a retinoid should come from a defined goal. “Everyone uses it” is not a defined goal.
The small group who should keep going
There are people for whom long-term retinoid use is genuinely useful and well-tolerated. They tend to share a profile. Their skin is oilier than average or normal. They have a history of acne or active comedonal acne. They have visible photodamage they want to keep working on. They tolerate the molecule without needing to over-buffer it. They use sunscreen daily and consistently.
For this group, the dermatology evidence is decent. Kligman’s earlier work on tretinoin and photoaging, and the followup studies, show real changes in epidermal architecture over years (PMID: 2646000). The skin holds up better. Pigmentation is more even. Acne stays suppressed. This is the group the marketing is technically aimed at, but the marketing has expanded the perceived audience to roughly everyone with a face.
If you are in this group, the question is not whether to stop. It is whether your current strength and frequency still suits you. I would push back on the assumption that you need to keep increasing. Maintenance at a lower dose is a reasonable long-term position.
The contrarian section
I am not anti-retinoid. I have written extensively about tretinoin, and I think it is one of the few dermatology actives with real, repeatable evidence behind it. What I am against is the framing that retinol is a permanent membership.
The cosmetic industry has a structural reason to keep you on retinol forever. The molecule degrades, the bottle empties, you buy another one. Stopping is bad for revenue. The clinical reality is that skin benefits from cycles of use and recovery, and from being matched to changing life conditions. Pregnancy is one. Perimenopause is another. A bad stretch of stress, a new climate, a barrier event from a different product. Any of these can be a reasonable reason to take a break.
The phrase “you have to push through the purge” gets applied to situations that are not purging. Purging is an acceleration of existing comedones to the surface. It is mostly a tretinoin and adapalene phenomenon, and it resolves in six to eight weeks. What people describe as purging at month three or four of a retinol is usually irritant contact dermatitis or perioral dermatitis. Pushing through it makes it worse.
What I would tell my past self
You do not need to be loyal to a molecule. You started using retinol because you read it was important. You can stop using it because it is no longer serving you. There is no point system. There is no skincare permanent record. The skin you have at fifty will be most influenced by sun protection, by not smoking, by sleep, and by genetics. Topical retinoids contribute, but they are not the load-bearing wall the industry pretends they are.
If you stop, give it eight to twelve weeks before you decide whether to restart. The skin needs that long to recalibrate. Most people who quit retinol report that their baseline is calmer than they remembered. Some restart at a lower dose. Some never go back. Both are defensible.
Frequently asked
Will I lose my “results” if I stop retinol for a few months?
The photoaging benefit of retinoids is partially reversible if you stop, but reversal happens over months and years, not days. A three-month break does not undo what you built. A three-year break, with no sunscreen, would.
How do I tell purging from irritation?
Purging happens in areas you already break out, peaks at four to six weeks, and resolves by eight. Irritation appears anywhere, including areas you never break out, comes with redness or stinging, and gets worse the longer you keep using the product. If it is past two months and getting worse, it is not purging.
Is adapalene safer than retinol long-term?
Adapalene is more photostable, less irritating in head-to-head trials, and has better acne data. It is not “safer” in a pregnancy sense, but for daily use over years in non-pregnant adults, it has a more favorable risk profile than high-strength retinol for many people.
Can I just lower my retinol frequency instead of stopping?
Often yes. Dropping from four nights to two nights a week is a reasonable first step before quitting entirely. Use the skin cycling calculator to map it out.
What replaces retinol if I stop?
Sunscreen does most of the heavy lifting. Niacinamide, azelaic acid, and bakuchiol have evidence for parts of what retinol does. None of them are a one-to-one replacement, and that is fine.
If you are unsure whether your skin is reacting to your retinol or to something else, the barrier damage test is a starting point. For dosing and strength decisions, the retinol strength tool lays out the options.
References
- Mukherjee S, et al. Retinoids in the treatment of skin aging. Clin Interv Aging 2006;1(4):327-348. PMID: 18046911.
- Kang S, et al. Long-term efficacy and safety of tretinoin emollient cream 0.05% in the treatment of photodamaged facial skin. J Invest Dermatol 1995;105(4):549-556. PMID: 7561157.
- Kligman LH. The treatment of photoaged skin with topical tretinoin. Clin Geriatr Med 1989;5(1):169-186. PMID: 2646000.
- Bouloc A, et al. A skincare combined with a low concentration of retinol has a complementary effect to face wrinkles. J Eur Acad Dermatol Venereol 2015;29(8):1496-1500. PMID: 25627659.
- Draelos ZD. The effect of ceramide-containing skin care products on eczema resolution duration. Dermatol Ther 2007;20(5):330-336. PMID: 18045358.
- Babcock M, et al. Retinaldehyde: a versatile vitamin A derivative. Cosmetics 2015;2(1):71-87.
Related Elelaf tools
Sources
- Mukherjee S, et al. Clin Interv Aging 2006;1(4):327-348. PMID: 18046911.
- Kang S, et al. J Invest Dermatol 1995;105(4):549-556. PMID: 7561157.
- Kligman LH. Clin Geriatr Med 1989;5(1):169-186. PMID: 2646000.
- Bouloc A, et al. J Eur Acad Dermatol Venereol 2015;29(8):1496-1500. PMID: 25627659.
- Draelos ZD. Dermatol Ther 2007;20(5):330-336. PMID: 18045358.
- Babcock M, et al. Cosmetics 2015;2(1):71-87.