Silicone scar products: one of the most-used treatments there is, and Cochrane rated all of the evidence 'low' or 'very low'

The 2021 Cochrane review. Searched to April 2021 across the Cochrane Wounds Specialised Register, CENTRAL, Ovid MEDLINE, Ovid Embase and EBSCO CINAHL Plus, plus trial registries for ongoing and unpublished studies. No restriction by language, year of publication or setting.

Inclusion was randomised controlled trials enrolling people with hypertrophic scars and assessing silicone gel sheeting. 13 studies met it, with sample sizes from 10 to 60.

This entry is the 'sheeting versus no sheeting' comparison. Seven studies with 177 participants went into it.

But only two of them, 31 participants (32 scars), reported scar severity assessed by health professionals.

Result: it is uncertain whether severity of scarring differs between the groups — mean difference −1.83 (95% CI −3.77 to 0.12). Cochrane graded this very low certainty and said why: downgraded once for risk of bias and twice for serious imprecision.

Look at the interval. −3.77 to 0.12. Silicone could have been considerably better, or slightly worse. It crosses zero.

How to read that.

First, 31 people. Seven studies with 177 participants entered this comparison, and only two reported the endpoint that matters most. Having done a study and having reported the result are different things.

Second, we carry Cochrane's own account of reporting quality across. Many trials reported their methodology poorly, so the risk of bias was unclear. Which is why Cochrane's conclusion says it is 'not confident that most trials are free from risk of bias'.

Third, the 13 studies are clinically heterogeneous — different durations of follow-up, different scar sites. Six had a split-site design, and three had an unclear design, producing a mix of paired and clustered data.

So it sits at 'not established'. Not that silicone does nothing — that this material does not distinguish it from doing nothing.

The same Cochrane review. In the same comparison (sheeting versus none), one study with 34 participants reported pain.

Result: silicone gel sheeting may result in a slight reduction in pain compared with no treatment — mean difference −1.26 (95% CI −2.26 to −0.26). Low certainty, downgraded once for risk of bias and once for imprecision.

There is also a comparison against pressure garments. One study with 54 participants: sheeting may reduce pain compared with pressure garments — mean difference −1.90 (95% CI −2.99 to −0.81), also low certainty.

Why we grade this differently from the entry above.

The interval does not cross zero. −2.26 to −0.26, both ends on silicone's side. Set it beside the −3.77 to 0.12 above. Same review, same comparison, different interval depending on the endpoint.

But pain is what the person reports. And whether participants were masked in these trials is not stated in the abstract. You know whether a sheet is stuck to you. So we mark this endpoint on the badge as 'self-reported improvement'.

And it is 34 people.

So it sits at 'early evidence'. Consistent in direction, small in size.

One place where this entry reads practically. Raised scars itch and hurt. Reducing that is a different kind of benefit from changing the shape of a scar, and in this review the interval on symptoms is cleaner than the one on shape.

The same Cochrane review, which compared sheeting against ten things — no sheeting, pressure garments, silicone gel, topical onion extract, polyurethane, propylene glycol and hydroxyethyl cellulose sheeting, Kenalog injection, pulsed-dye laser, intense pulsed light, and a silicone gel bandage.

This entry is the comparison against onion extract. One trial, 32 participants.

Result: sheeting may slightly reduce severity of scarring compared with topical onion extract — mean difference −1.30 (95% CI −2.58 to −0.02), low certainty.

Why we give that number its own entry.

Onion extract is one of the signature ingredients of scar products. And this comparison is the only head-to-head we found. The result: silicone won, the margin is small, and one end of the interval is −0.02 — all but touching zero.

The other comparisons, for completeness.

• against silicone gel — one trial, 32 participants, unclear whether it differs (MD 0.40, 95% CI −0.88 to 1.68, very low certainty). Sheet against gel was not separated.
• against polyurethane — one trial, 60 participants, unclear (MD 0.50, 95% CI −2.96 to 3.96)
• against propylene glycol and hydroxyethyl cellulose sheeting — one trial, 38 participants, uncertain on pain
• against a silicone gel bandage — one trial, 60 participants, unclear on pain
• against Kenalog injection, pulsed-dye laser and intense pulsed light — there was a lack of reportable data

Read that last line again. The comparison against intralesional corticosteroid was in the review, and there was not enough reported data to analyse it.

So it sits at 'early evidence'. The interval does not cross zero, but it is 32 people at low certainty.

A 2026 systematic review and meta-analysis in Eur J Dermatol. This time about fluid silicone gel rather than sheeting.

Ovid MEDLINE, Embase, CINAHL and the Cochrane Library were searched under PRISMA, with 35 papers in the systematic review and 20 in the meta-analysis. Only randomised controlled trials were included, and silicone gel was compared against placebo, no treatment, silicone gel sheets and other conventionally used treatments.

That is a larger base than the Cochrane above: 35 papers, randomised trials only.

But the abstract gives no result figures.

What it says is that the review provides the efficacy of silicone gel fluids against different controls, presents the reported adverse effects, identifies gaps in current knowledge and provides a foundation for future studies. And that while silicone gel fluid is widely recognised as a leading non-invasive treatment, there is inconsistent evidence of superiority.

So the only thing we can carry across from this entry is that last sentence.

This site holds to not citing papers from memory and writing only what we verified. No numbers in the abstract means no numbers here. Which is why the participant field on the badge above is empty and the comparator field reads 'not stated' — several controls are named as a group, and we cannot pin down which.

Two reasons we include it anyway.

First, the existence of this review is itself information. A 35-paper review of randomised trials only appeared in 2026, and it summarised the position as 'inconsistent evidence of superiority'. That runs the same way as Cochrane's conclusion.

Second, to record that sheet and gel are different objects. The Cochrane above is about sheeting; this is about gel. And Cochrane's own 'sheet versus gel' comparison did not separate them in 32 people. Which is to say there is not yet an answer about which to use.

So it sits at 'not established'.