Your 30s, evidence first.

The trial every 30-something should know: Hughes 2013, Annals of Internal Medicine — 903 Queensland adults under 55, randomized to daily SPF 15+ versus discretionary use for 4.5 years, with skin aging measured by blinded microtopography casts. Daily users showed no detectable increase in skin aging; aging was 24% lower than in the discretionary group (relative odds 0.76). Publicly funded. The beta-carotene arm of the same trial did nothing.

Caveats: hand skin, Australian UV, a discretionary (not zero) comparator. Even so, it is the only long-term RCT of anything topical against photoaging in your age band, and the twin data point the same way (sunscreen users looked younger). Broad-spectrum SPF 30–50, every morning, reapplied outdoors; tinted iron-oxide formulas if you're melasma-prone.

A 2025 meta-analysis pooled eight vehicle-controlled RCTs (1,361 people, 16 weeks to 2 years): fine wrinkles improved by a mean 0.41 on the trial scales, coarse wrinkles 0.25, with epidermal thickening and increased procollagen on biopsy — and a threefold higher rate of dryness and peeling. The youngest trials enrolled people aged 29–50 with existing photodamage, so the 30s are inside the evidence; nobody has ever tested tretinoin as pure prevention on unlined skin.

How: 0.025% two or three nights a week over moisturizer, building to nightly over two months; expect 2–8 weeks of peeling. The 0.02–0.05% strengths give most of the benefit with less irritation.

Pregnancy: meta-analyses of accidental first-trimester exposure found no substantial excess of malformations — reassurance for accidents, but stop when trying to conceive; oral isotretinoin is a major teratogen.

If a prescription is a hurdle: a randomized split-face study found retinol 0.25–1.0% matched tretinoin 0.025–0.1% on every efficacy measure at 12 weeks (short, manufacturer-run, no placebo arm). Adapalene — OTC across much of Europe — was non-inferior to tretinoin 0.05% for photoaging in a randomized trial and improves lentigines, and it doubles as acne treatment — the obvious first retinoid for a 30-something with breakouts.

A 2023 systematic review found every RCT measuring skin topography favoured vitamin C over vehicle — including a 6-month double-blind trial of 5% ascorbic acid that reduced deep wrinkles — but the entire evidence base is about 139 volunteers, in photodamaged adults presumably older than you. The mechanism (antioxidant, collagen cofactor, UV-damage buffer under sunscreen) is sound; the clinical base is tiny. Worth it as a morning adjunct if the formulation is stable (opaque, airless, replaced every few months).

Bissett's 12-week double-blind split-face RCT (50 women, 5% niacinamide versus the same moisturizer without it) found improvements in fine lines, hyperpigmented spots, red blotchiness and sallowness. Consistent, well-run, and essentially all from Procter & Gamble. For the 30s it's the barrier-and-pigment support that pairs with a retinoid without irritation — and a ceramide booster (see our ceramides guide).

For the decade's signature pigment problem, azelaic acid is the topical you can use through pregnancy and breastfeeding. A meta-analysis of six RCTs (673 patients) found it comparable to hydroquinone for melasma; the classic double-blind trial in contraceptive users reached the same "not better, but valid" verdict. It also treats acne and post-acne marks. Slow, tingly at first, and genuinely useful.

Nothing in an eye cream is unavailable in a smaller tube from your face routine, and no trial shows the category outperforming a low-strength retinoid, vitamin C and sunscreen applied near the orbit. Under-eye changes in the 30s are thin skin, vascular show, pigment and the first orbital-rim change — and in a 60-woman study, poor sleepers scored worse on intrinsic aging measures including dark circles. Sleep, SPF and your existing actives are the evidence-adjacent moves.

Smoking cannot be randomized, so the identical-twin design is the best causal evidence available — and it is stark: in 186 pairs, the longer a twin smoked the older she looked, with each decade adding ~2.5 years and five years enough to register; a follow-up of smoking-discordant twins found the differences concentrated in the lower two-thirds of the face. Combined with sun, tobacco is the largest modifiable driver of how old you'll look at 45. Quitting in the 30s is the highest-yield "treatment" on this page.

The 30s are the last decade to bank bone and the first to defend it. A meta-analysis of controlled trials found resistance training raises lumbar-spine bone density in women, with premenopausal women represented in 13 studies; high-impact loading optimizes peak bone mass and the benefit is lost if training stops for six months. The skin bonus: in a 16-week trial of sedentary middle-aged women, both aerobic and resistance training improved elasticity, but only resistance training increased dermal thickness. Two or three sessions a week; the muscle you build now is the muscle you keep at 60.

Europe-wide, about 13% of people are severely deficient and ~40% below 50 nmol/L, more in winter and more in women. Correcting a measured deficiency matters for the bone you are banking this decade; supplementing sufficient people showed no benefit in large trials, and nothing links vitamin D to skin aging. Test in late winter, correct with 800–2,000 IU/day, skip the megadoses (our supplements guide has the full picture).

Twins who avoided alcohol looked significantly younger, though dose wasn't captured; a 3,267-woman survey spanning the 30s linked heavy drinking to upper-face lines, under-eye puffiness and midface volume loss — self-reported, industry-run, observational. Consistent direction, no interventional proof. Cutting evening drinking also removes a known sleep disruptor, which the next row cares about.

Oyetakin-White 2015 compared 30 poor sleepers (≤5 h, poor quality) with 30 good sleepers: the good sleepers had lower intrinsic-aging scores — fine lines, uneven pigment, laxity — lower water loss and 30% faster barrier recovery after tape-stripping. Small, cross-sectional, industry-collaborative. Nobody has shown that fixing sleep changes wrinkles; the barrier data make it plausible, and the downside of trying is nil.

Crane 2015 put sedentary older adults through 12 weeks of endurance training and found their skin biopsies shifted toward younger structure, with a muscle-derived signal (IL-15) implicated; the 2023 Japanese trial above found aerobic work improved elasticity but only resistance training thickened the dermis. Two small studies in older populations, extrapolated to yours — a bonus reason to move, not the reason.

The Rotterdam Study photographed 2,753 people over 50 and found adherence to a healthy-diet index associated with fewer wrinkles in women, red-meat-and-snack patterns with more, fruit-dominant patterns with fewer — cross-sectional, older, and confounded by everything that travels with healthy eating. Glycation (sugar cross-linking collagen) is real biochemistry with observational skin-autofluorescence data behind it, and no dietary RCT with skin endpoints. Eat well because of the rest of your body; the skin case is plausible, not proven.

A 2025 meta-analysis of 23 RCTs (1,474 participants) found collagen peptides improved hydration, elasticity and wrinkles overall — but no effect in trials without manufacturer funding, and none in high-quality trials. Participants were typically 40–60; there is essentially no 30s data. The full accounting is in our collagen guide; for this decade, the money buys more skin in a gym membership and a retinoid.