Do polynucleotides work for under eyes?

The published record is small and mostly uncontrolled. Eight non-retracted studies have injected polynucleotides around the eyes; five compare people only with their own baseline, and the three randomized ones compared against another filler or topical saline, never a placebo injection. In the 27-person trial against hyaluronic-acid filler, the overall improvement ratings did not differ between the two sides.

How do polynucleotides compare with filler for under eyes?

One randomized, double-blind split-face trial (Lee 2022, 27 people) injected a polynucleotide on one side of the eye area and a non-crosslinked hyaluronic-acid filler on the other, three times two weeks apart. Visual-analogue and global improvement ratings were not significantly different; skin roughness and pore volume improved more on the polynucleotide side.

Do polynucleotides help dark circles?

One study looked at dark circles: an open-label observational study of 61 people in four countries, three injection sessions, with photographs analysed against each person's baseline (Amado 2026). It reported a shift toward less severe pigmentation at one and three months. There was no untreated or placebo group, and some authors report funding ties to filler makers.

How many sessions do under-eye polynucleotide studies use?

The studies vary. The 61-person dark-circle study used three injections two to three weeks apart; the 42-person study used two sessions of 2 mL; the 27-person comparison trial used three sessions two weeks apart. These describe what the studies did, not a schedule anyone should follow — that is a decision for a treating clinician.

Are polynucleotides FDA-approved for under eyes?

No. FDA's device-approval database returned no record for polynucleotide on 4 October 2026, and the studies above were run in South Korea, Thailand, Lebanon and Europe, where the products are sold. Where a US clinic offers them, the product is not an FDA-approved device.

Is a polynucleotide eye cream the same as an injection?

No. Every study on this page injected the product into the skin or applied it after microneedling. A leave-on cream puts DNA fragments on the skin's surface, which large molecules cross poorly. What eye creams declare is covered on our PDRN eye cream page.

Polynucleotides for Under Eyes: Every Periorbital Study, Counted

Eight published studies have injected polynucleotides around the eyes. Three are randomized; none compares against a placebo injection, the largest controlled one is 30 people, and a 218-person trial was retracted in 2026.

Polynucleotides — purified DNA fragments, usually from salmon or trout — are sold by clinics as an under-eye treatment for crepey skin, fine lines and dark circles. This page counts every published study that has injected them around the eyes, with its size, its design and what it was compared against. It reports the literature and does not recommend any treatment; whether a procedure suits a particular person is a question for a qualified clinician.

For what polynucleotides are and the wider evidence, see our polynucleotides for skin page; this page narrows to the eye area only.

How the studies were found

On 4 October 2026 we searched PubMed for polynucleotide, polydeoxyribonucleotide or PDRN in titles and abstracts, combined with periorbital, under-eye, infraorbital, dark circles, tear trough, crow's feet, eyelid or lower lid. The search returned 37 records. Many were irrelevant matches (the crow's-feet term also catches "crowding" in molecular-biology papers). We read every record and kept those reporting people whose eye area was treated with a polynucleotide product.

Eight remain, plus one retracted trial. Two reviews (a 2026 comparison of filler and polynucleotides for the under-eye area, and a multispecialty under-eye systematic review) and one manufacturer-authored laboratory paper with preliminary clinical observations (PMID 40076130) are not counted as studies.

The randomized studies — three, none against a placebo injection

Study	People	Design and comparator	What it found
Lee 2022 (PMID 32248707), Korea	27	Randomized, double-blind, split-face; polynucleotide filler one side, non-crosslinked hyaluronic-acid filler the other; 3 sessions, 2 weeks apart	Visual-analogue and global improvement ratings not significantly different; roughness and pore volume improved more on the polynucleotide side; dermal density no different
Jeong 2020 (PMID 31680395), Korea	30	Randomized, patient- and evaluator-blinded, split-face crow's feet; Rejuran polynucleotide filler vs a polycaprolactone filler	No significant difference at 12 weeks; crow's-feet grade improved in 12 of 29 on the polynucleotide side and 14 of 29 on the other
Yogya 2022 (PMID 35501660; TCTR20201105007), Thailand	30 (29 finished)	Whole face treated with radiofrequency microneedling three times; one side then given polynucleotides, the other saline, applied topically	3D-imaging showed faster improvement on the polynucleotide side at 2 months

Each of the three has an active comparator or a combined procedure. None compares a polynucleotide injection against a placebo injection, so none isolates what the polynucleotide itself does beyond what needle passes, a comparison filler or microneedling do. In the Thai study, the polynucleotide went onto skin just punctured by microneedles, which is a different delivery route from injection.

The single-group studies — five, measured against baseline

Study	People	What was done	What was measured
Amado 2026 (PMID 42164086), UK, Italy, Malaysia, Turkey	61	Open-label observational; trout-derived polynucleotide gel, 7.5 mg/mL, usually 1 mL per side, three sessions 2–3 weeks apart	Dark circles: digital photo analysis at 1 and 3 months against baseline; reported a highly significant shift toward less severe pigmentation
Ziade 2026 (PMID 41689167), Lebanon	42	Prospective observational; Pluryal Silk 7.5 mg/mL, 2 mL per session	FACE-Q patient questionnaires for the lower eyelid and crow's feet at 1, 3 and 6 months; improvements against baseline, greatest at 3 months after two sessions
Kim 2022 (PMID 35357558), Korea	30 (28 finished)	Single-arm; polynucleotide filler for crow's feet	Crow's-feet grade improved at 18 weeks; wrinkle, texture and pore values improved on 3D imaging; melanin did not change
Jeon 2026 (PMID 41361144)	15	Split-face, both sides polynucleotide: a conventional and an "ionized" version	Injection pain; wrinkle grading was secondary
Adel 2025 (PMID 41229581), Egypt, Canada, Serbia, Denmark	5	Case series; VAMP (Prollenium), 1 mL per side by cannula to the tear trough	Mild to moderate improvement at 3 months; authors say the size "precludes definitive conclusions"

A single-group study shows that people looked or felt better than at the start. It cannot separate the product from time, the act of injection, swelling in the first weeks, or the expectation of improvement, because no one received anything else. In the 61-person dark-circle study, the conflict-of-interest statement records that some authors, including the corresponding author, have received research grants and consultancy fees from producers or distributors of fillers, Mastelli among them, and that one author received support from Mastelli to attend a congress.

The retracted trial

The largest periorbital trial — 218 people, randomized and evaluator-blinded, comparing a polycaprolactone filler with Rejuran for crow's feet (Choi 2025, PMID 39313949) — was retracted by the Journal of Cosmetic Dermatology in March 2026 (retraction notice PMID 41834780). It is not counted here. Our polynucleotides for skin page lists the retracted polynucleotide papers across all face areas.

Why the eye area is its own question

Under-eye skin is thinner than the rest of the face. In a 2025 ultrasound study of 31 patients across the face (Kim 2025, PMID 41000432), the lower eyelid had the thinnest dermis, 0.53 mm on average, against 1.18 mm at the temple and posterior cheek. That study measured thickness to set an injection depth for an automatic injector; it did not test outcomes. The 5-patient tear-trough series above opens by naming thin skin and superficial blood vessels as what makes the area a clinical challenge.

Regulation

FDA's device-approval (PMA) database returned no record for polynucleotide on 4 October 2026. The studies above were run in South Korea, Thailand, Lebanon, Egypt and Europe. In the US, a polynucleotide injection is not an FDA-approved device; our skin boosters guide covers how the category is treated in different markets.

What the record does not show

Whether a polynucleotide injection improves under-eye skin more than a placebo injection. No published trial has tested it.
Results beyond six months: the longest follow-up here is the 42-person questionnaire study.
Any effect of a polynucleotide cream around the eyes — none of these studies used one. What eye creams declare, and the two small topical eye studies behind them, are on our PDRN eye cream page.

Side effects as reported

The studies report mild, short-lived reactions — injection pain, swelling, redness, bruising — and no serious adverse events, within follow-ups of three to six months and samples of 5 to 61 people. Studies this size cannot detect uncommon events. A 2026 study (Boonpethkaew, PMID 42457753) examined how polynucleotide injectables behave if injected into an artery (PMID 42457753), a risk shared by any injection near the eye's blood supply. Anyone considering an eye-area injection should discuss it with a qualified clinician.

PDRN explains how polynucleotides differ from PDRN by fragment length. For the injectable hyaluronic-acid booster with a US approval, see our SkinVive analysis. For eye-area products applied to the surface, see peptide eye cream.

Sources

PubMed E-utilities search, 4 October 2026: (polynucleotide*[tiab] OR polydeoxyribonucleotide[tiab] OR PDRN[tiab]) AND (periorbital OR "under-eye" OR "under eye" OR infraorbital OR "dark circles" OR "tear trough" OR crow* OR eyelid* OR "lower lid") [tiab] — 37 records, each read.
Lee YJ et al. Journal of Dermatological Treatment 2022 (PMID 32248707).
Jeong GJ et al. Journal of Cosmetic Dermatology 2020 (PMID 31680395).
Yogya Y et al. Dermatology and Therapy 2022 (PMID 35501660; Thai Clinical Trials Registry TCTR20201105007).
Amado R et al. Clinical, Cosmetic and Investigational Dermatology 2026 (PMID 42164086), including its conflict-of-interest statement.
Ziade G et al. Journal of Cosmetic Dermatology 2026 (PMID 41689167).
Kim JH et al. Aesthetic Plastic Surgery 2022 (PMID 35357558).
Jeon W et al. Efficacy of ionization-adjusted PN filler in minimizing injection pain. Aesthetic Plastic Surgery 2026 (PMID 41361144).
Adel N et al. Plastic and Reconstructive Surgery Global Open 2025 (PMID 41229581).
Choi SY et al. Journal of Cosmetic Dermatology 2025 (PMID 39313949), retracted; retraction notice PMID 41834780.
Kim JS et al. Dermal thickness across facial regions after polynucleotide injection with an automatic injector. Plastic and Reconstructive Surgery Global Open 2025 (PMID 41000432).
Boonpethkaew S et al. Intravascular behavior of cutaneous polynucleotide injectables following intra-arterial exposure. Scientific Reports 2026 (PMID 42457753).
openFDA device PMA endpoint, generic/trade name "polynucleotide" (no match), 4 October 2026.