Ingredient Beauty BenefitsRoute · Device-based
Radiofrequency: there are randomised trials, and every one of them compared it to another device
We read four randomised trials of monopolar radiofrequency. All of them — including a multicentre study analysing 212 people — used **another device as the control**, and all of them concluded 'not inferior'. We found no trial comparing it against doing nothing.
Last checked Sep 17, 2026
Radiofrequency (RF) devices pass current through skin to heat the dermis. The stated mechanism is collagen contraction followed by a repair response. In Korea the best-known brand is Thermage.
This is not a topical cosmetic. It is a device procedure performed in a clinic, and this layer is kept separate from the ingredient layer.
On regulatory class we will be plain: we have not opened the Korean medical-device approval record. The registry is indexed by item name, product name and model — not by brand — so a brand search returns nothing. We do not write down a classification we have not verified, so this page says 'being verified'.
What is established
In all four, the control arm was another device
The largest is a 2026 multicentre trial: 230 enrolled, 212 analysed, six months, assessor-blind. But the design is a non-inferiority trial — a new RF device against an existing RF device (Thermage CPT). At day 90 the improvement rates were 100% versus 98.10%, with no significant difference between groups.
The 2025 trial has the same shape. Forty people split between the new and the existing device; nasolabial fold depression volume fell significantly from baseline at 1, 3 and 6 months, but there was no difference between the two groups.
The 2019 and 2017 trials put RF against microfocused ultrasound: 20 people split face, 20 split neck. Both sides improved, and neither beat the other.
Read together, these four say something clear. 'RF is no worse than another device' is well established; 'RF beats doing nothing' is not something these trials can tell you. That is simply what a non-inferiority design means.
- Randomised controlled trial in people · 212 participants · 26 weeks · Compared with another active ingredient · Funding not declared · Endpoint scores from blinded assessors PMID 41758257
- Randomised controlled trial in people · 40 participants · 26 weeks · Compared with another active ingredient · Funding not declared · Endpoint scores from blinded assessors PMID 39957006
- Randomised controlled trial in people · 20 participants · 26 weeks · Compared with another active ingredient · Funding not declared · Endpoint scores from blinded assessors PMID 30531187
- Randomised controlled trial in people · 20 participants · 26 weeks · Compared with another active ingredient · Funding not declared · Endpoint scores from blinded assessors PMID 28902019
We found no trial with a sham control
Among the randomised trials we read, none used an untreated side or a sham procedure as the control. Every one is device against device.
It is worth spelling out why that matters. An improvement measured against baseline cannot be separated from swelling settling over time, from lighting and posture changing between photographs, or from knowing you were treated. Blinding the assessor does not fix this — what is blinded is the assessor, not the skin.
There is a real difficulty here: sham-controlling a device that produces heat and pain is hard, because the person on the table knows. But difficult and demonstrated are different things, and the difficulty does not fill in this blank.
Nasolabial fold volume was measured by instrument
The 2025 trial used Antera 3D imaging to measure the depression volume of the nasolabial folds. That is an instrument reading rather than a grader's eye, which makes it firmer than the scale scores above.
Significant reductions from baseline appeared at 1, 3 and 6 months. Still: forty people, the control is another RF device, and the change is against baseline. The limits above apply unchanged.
- Randomised controlled trial in people · 40 participants · 26 weeks · Compared with another active ingredient · Funding not declared · Endpoint instrument measurement PMID 39957006
Known risks
- Pain during treatment is reported; the trials above measured it on formal pain scales.
- Erythema and swelling are common and usually transient.
- Blistering, burns, pigment change and localised fat atrophy are reported rarely.
- The 2026 multicentre trial reported no serious adverse events. An absence of reports is not proof of absence — 212 people over six months cannot see rare events.
- Any adverse reaction is something to have seen by the clinic that performed the procedure.
What is not established
- We have not verified the Korean device approval class. When we do, it goes on this page as it reads.
- How long any effect lasts — follow-up in these trials tops out at six months.
- How it compares against doing nothing at all.
- How many sessions, at what interval — every trial above used a single treatment.
- Data in men and in darker skin are thin.
Whether to have it done is a decision to make with a doctor.
What this article relies on
- PubMed — 미국 국립의학도서관 문헌 데이터베이스National Library of Medicine · Checked on 2026-09-16