Your smile lines are a seam, not a wrinkle.

Sun and smoke thin the dermis and rob it of recoil, which is the skin's share of a deep fold. Daily sunscreen users in the Nambour trial showed no detectable increase in skin aging over 4.5 years — 24% less than discretionary users (Hughes 2013) — and among identical twins discordant for smoking, the smoker had the worse nasolabial folds (twin study). Graded moderate rather than strong because no trial has measured the fold itself, and because the fold's main drivers — fat descent and bone — are not skin. The habit that prevents the crease from printing; not the habit that lifts a cheek.

Tretinoin thickens the epidermis and rebuilds upper-dermal collagen; the meta-analysis of eight randomised trials in 1,361 patients found significant improvement in fine and coarse wrinkles over vehicle (meta-analysis). No trial has measured the nasolabial fold, and the mechanism can only reach the crease printed into thin skin — which is why the row sits at emerging for this problem despite strong evidence for wrinkles in general. For the skin type it is the base; for a heavy cheek it does nothing visible. The wrinkles guide covers the retinoid ladder.

Nobody should keep weight for a fold, but the trade should be known. Volumetric imaging shows deep and superficial facial fat both lost with weight loss, and deep fat is what props the cheek forward of the seam; people who lost more than 10% of body weight after 45 showed marked accentuation of the folds (ASPS), and a systematic review of massive weight loss found faces looking older — about five years older than their age to 40 blinded raters in one small study — with the most fat lost from the mid-cheek (systematic review). Slow loss gives the skin time to retract; weight cycling stretches it repeatedly; and the practical answer, for anyone on a weight-loss medication, is to budget for midface volume at the end of it rather than to stop.

A cheek pressed into a pillow experiences compression, shear and tension for hours every night, and the review that named "sleep wrinkles" traces a set of lines — including those beside the nose and mouth — that follow that distortion rather than any expression, and worsen as skin loses elasticity (Anson 2016; 2012 study). No trial has randomised sleep position, most people cannot hold one asleep, and the fold's deep drivers are untouched; a contoured pillow and a silk pillowcase are cheap experiments, not treatments.

This is the one concern where the facial-exercise evidence is relevant: the 20-week programme pilot in 16 middle-aged women found blinded dermatologists rating upper- and lower-cheek fullness improved and estimating faces about three years younger, the authors proposing muscle hypertrophy beneath the cheek (Alam 2018). It had no control group, measured no fold, and asked for half an hour a day; the same movements etch expression lines elsewhere. The one controlled study that graded the nasolabial folds — 18 people, seven weeks of daily exercises — found no difference there (only the upper lip was judged younger-looking), and its authors concluded "It cannot be concluded that facial exercises are effective" (De Vos 2013). Limited, then: the pilot's outcome is the tissue that props the fold, but the controlled data on the fold itself are negative.

The consumer "lifting" shelf is drainage and sensation. Gua sha and rollers leave the cheek flushed — a roller raised cheek blood flow for at least ten minutes (Miyaji 2018) — and when 34 women were randomised to one or the other for eight weeks, facial contour distances were 2.2 to 3.3 mm shorter in both groups, with no untreated group and no fold among the measurements (Ahn 2025). Microcurrent and muscle-stimulation devices claim to tone the muscles under the cheek: a randomised trial in 108 women found the zygomatic major 18.6% thicker on blinded ultrasound after 12 weeks without measuring the fold (Kavanagh 2012), and the only controlled trial we found that measured the fold, an industry-supported split-face study of a high-frequency stimulation device in 24 women, reported a shallower maximum fold depth on the treated side after eight weeks (Omatsu 2024; 2026 correction). Sheet masks hydrate the crease for an evening. None reaches the malar fat pad, the deep cheek fat or the pyriform bone, which are the fold's drivers. Fine as ritual; not a line on this page's plan.

Every modern filler earned its licence in the nasolabial fold, so the fold has the deepest randomised evidence in aesthetic medicine. The founding trial randomised 138 patients to hyaluronic acid or bovine collagen, one per side: at six months HA was judged superior in 56.9% of patients by wrinkle score and 62% by global improvement, against 9.5% and 8% for collagen, with less volume needed (Narins 2003). The meta-analysis of 51 randomised trials in 4,097 participants across 13 countries puts numbers on the class: mean wrinkle severity fell from 3.23 to 1.79 at one month, 2.02 at six and 2.46 at twelve, a "sustainable, up to one year" correction (meta-analysis). Newer gels are trialled against older ones for a year or more — a 64-week randomised within-subject study of a resilient HA (RHA trial), an 18-month open-label extension of a cohesive gel (Belotero extension), a registered 18-month controlled trial of a mid-viscosity gel (Volift trial) — and a 2024 network meta-analysis ranks them with little daylight between the major brands (network meta-analysis).

Placement decides the look: a medium gel in the deep dermis or just beneath it along the crease, threaded with a cannula from below, never enough to flatten the fold, and always after the support rows — a fold filled before the cheek is supported needs twice the gel and reads as a heavy lip. Hyaluronidase reverses it. The filler guide covers products, MRI persistence and the reversal drug.

Calcium hydroxylapatite microspheres in a gel carrier fill immediately and then provoke collagen around each sphere as the carrier is absorbed. Its pivotal trial randomised 117 patients with symmetric folds to calcium hydroxylapatite on one side and human collagen on the other: at six months it was rated superior in 79% of folds with about half the volume (decade review), and the 12-month multicentre split-face trial against non-animal stabilised hyaluronic acid found it more effective at every time point, with 79% of folds still improved at a year against 43% and 30% less volume used (split-face RCT); long-term follow-up confirmed safety (long-term study). Biopsies show it drives more active collagen remodelling than hyaluronic acid (histology study).

The trade is reversibility: there is no enzyme for it, it is too firm for thin skin and for the lip, and a vascular occlusion with it is harder to treat. Deep placement under a thick-skinned fold, by someone who uses it weekly.

"Treat the cheek first" is the injector's mantra, and it is half right. In the pivotal Voluma programme, 235 patients had midface volume restored with a control group, and patient satisfaction improved in untreated regions including the nasolabial folds at six months and two years, plausibly because the medial cheek compartment was re-supported (pivotal RCT; two-year outcomes). But when 77 patients were imaged in three dimensions before and after cheek filler, 3 ml in the cheek produced expansion over the injection site and no traction on the skin between it and the fold — no photographic change in the fold at all — and only those filled at the nose-to-cheek transition showed the medial face improve (PRS, 2018).

So the evidence supports support, not lift: gel in the deep medial cheek fat and along the nasojugal transition props the tissue above the seam; gel on the lateral cheekbone makes a cheekbone and leaves the fold alone. Ask exactly where the syringe is going. The filler guide and the jowls guide grade the midface tools.

Poly-L-lactic acid is a collagen stimulator rather than a filler: injected as a suspension, it provokes fibroblasts over months and the volume arrives slowly. The pivotal randomised, evaluator-blinded trial gave 116 patients PLLA and 117 human collagen for nasolabial folds: PLLA scored significantly higher from month 3 through 13, and correction in the PLLA arm persisted to 25 months (Narins 2010, JAAD; investigator ratings). A newer microsphere formulation was tested in a 252-person multicentre double-blind randomised trial against hyaluronic acid: 92.4% of folds still corrected at 48 weeks against 59.3% (2026 RCT), and a split-face non-inferiority trial compares brands (split-face RCT). Graded moderate rather than strong because the 25-month figure comes from an uncontrolled extension, the effect is slow and operator-dependent, and there is no reversal. The regenerative guide covers the biostimulators.

Polymethylmethacrylate microspheres in bovine collagen stay for life, and the fold is where its safety was measured: a 23-site open-label study followed 1,008 patients for five years after nasolabial-fold correction, with 871 completing; 83% were satisfied or very satisfied, treatment-related adverse events occurred in 11.7% (mostly lumps and redness, mostly in the first year), and biopsy-confirmed granulomas in 1.7%, almost all responding to treatment (five-year study; 2026 review). Unavailable in much of Europe, requiring a skin test for the bovine collagen, and permanent in a face that will keep changing around it — a granuloma at year eight is surgery, not an enzyme. Moderate for the data's length; a last choice for most faces.

Fat harvested from the abdomen or thigh and placed in the medial cheek and along the fold replaces the deep fat that deflated, with tissue that behaves like the tissue around it. The one randomised comparison enrolled 62 patients (57 completing) to fat or hyaluronic acid in both folds: blinded wrinkle scores showed no difference within nine months and a significant advantage for fat at twelve, with early side effects differing and later ones similar (Hu 2017). The trade is that it is an operation with a donor site, that a variable fraction of the graft survives so a second session is common, that it cannot be dissolved, and that it follows body weight afterwards. Best combined with a facelift or done for a whole midface rather than a fold alone.

Polycaprolactone microspheres in a carrier gel fill immediately and stimulate collagen as they degrade over one to four years depending on the version. The fold is where it was tested: in a multicentre randomised trial of 160 patients, 88.8% of folds treated with polycaprolactone were still improved at twelve months against 23.8% with a hyaluronic gel, whose effect faded from the third month (randomised trial); a 2025 retrospective comparison of 136 patients found it ahead of poly-L-lactic acid at three, six and twelve months with similar safety (comparative study); and it is among the products pooled in the 51-trial meta-analysis (meta-analysis). One randomised trial, in Chinese patients, keeps it at moderate. Longer than hyaluronic acid, not reversible, and with the same nodule caveats as the other stimulators; the regenerative guide grades the class.

Where the maxilla has receded, the fat-lined deep pyriform space beside the nostril enlarges, and injectors place a small bolus of high-G′ gel on the bone there to rebuild the shelf and lift the top of the fold. The evidence is technique-level: a retrospective clinical series of deep pyriform-space augmentation with a firm hyaluronic gel (retrospective study), cadaveric work mapping an adjacent "perialar space" for more stable placement because gel migrates out of the pyriform space (cadaveric study), and a combined-treatment proposal for the empty space (2025 proposal). The angular branch of the facial artery runs through this territory, which makes it a cannula-and-aspiration site for experienced hands only.

Fibroblasts grown from a punch biopsy of the patient's own skin and injected back along the fold aim to rebuild dermis rather than fill it. A pilot study compared cultured autologous fibroblast injections with hyaluronic acid filler for nasolabial folds and reported improvement that arrived more slowly than filler (pilot study). Biologically interesting, expensive, weeks of cell culture per treatment, and limited to a handful of research settings; the regenerative guide covers where cell therapies stand.

Weakening the lip elevators to soften a fold flattens the smile, so toxin is not a nasolabial-fold treatment. The one legitimate use is the gummy smile: a small dose into the levator labii superioris alaeque nasi beside each nostril lowers the lip for 12–24 weeks with high satisfaction, and the deep medial folds these patients carry soften with it (narrative review; systematic review of dose and site). A prospective comparison of 32 patients found toxin given an hour after thread lifting prolonged the threads' effect on the fold to six months (comparative study) — an adjunct, not a treatment. Limited for the fold itself.