Cubierta de gel de silicona para prevenir y tratar las cicatrices hipertróficas y queloides
Resumen
Antecedentes
Las cicatrices queloides e hipertróficas son comunes y son causadas por una proliferación de tejido dérmico después de una lesión en la piel. Causan problemas funcionales y psicológicos a los pacientes y su tratamiento puede ser difícil. El uso de cubiertas de gel de silicona para prevenir y tratar las cicatrices hipertróficas todavía es relativamente nuevo y comenzó en 1981 con el tratamiento de las cicatrices por quemaduras.
Objetivos
Determinar la efectividad de la cubierta de gel de silicona para:
1) la prevención de cicatrices hipertróficas o queloides en pacientes con heridas recién cicatrizadas (p.ej., después de una cirugía);
2) el tratamiento de cicatrices establecidas en pacientes con cicatrices queloides o hipertróficas existentes.
Métodos de búsqueda
En mayo de 2013 se hicieron búsquedas en el Registro especializado del Grupo Cochrane de Heridas (Cochrane Wounds Group); en el Registro Cochrane Central de Ensayos Controlados (Cochrane Central Register of Controlled Trials, CENTRAL); en Ovid MEDLINE; en Ovid MEDLINE (In‐Process & Other Non‐Indexed Citations); en Ovid EMBASE; y en EBSCO CINAHL para esta segunda actualización.
Criterios de selección
Cualquier ensayo controlado aleatorizado o cuasialeatorizado, o ensayos clínicos controlados, que compararon la cubierta de gel de silicona para la prevención o el tratamiento de cicatrices hipertróficas o queloides con cualquier otro tratamiento no quirúrgico, ningún tratamiento o placebo.
Obtención y análisis de los datos
Se evaluó la calidad metodológica de todos los ensayos pertinentes. Tres autores de la revisión extrajeron los datos de forma independiente mediante un formulario estandarizado y verificaron los resultados de forma cruzada. Se evaluaron todos los ensayos que cumplieron con los criterios de selección de calidad metodológica.
Resultados principales
Se incluyeron 20 ensayos con 873 pacientes, con edades comprendidas entre 1,5 y 81 años. Los ensayos compararon la cubierta de gel de silicona adhesiva con ningún tratamiento; el apósito sin silicona; otros productos de silicona; la terapia láser; la inyección de acetónido de triamcinolona; el extracto de cebolla tópico y la terapia de presión. En los estudios de prevención, cuando se comparó con una opción de ningún tratamiento, aunque la cubierta de gel de silicona redujo la incidencia de cicatrices hipertróficas en los pacientes propensos a cicatrices (riesgo relativo [RR] 0,46; intervalo de confianza [IC] del 95%: 0,21 a 0,98), estos estudios fueron muy susceptibles a sesgo. En los estudios de tratamiento, la cubierta de gel de silicona produjo una reducción estadísticamente significativa del grosor de la cicatriz (diferencia de medias [DM] ‐2,00; IC del 95%: ‐2,14 a ‐1,85) y una mejoría del color (RR 3,49; IC del 95%: 1,97 a 6,15) pero, una vez más, estos estudios fueron muy susceptibles a sesgo.
Conclusiones de los autores
Hay escasa evidencia de que la cubierta de gel de silicona sea beneficiosa para prevenir las cicatrices anormales en pacientes con alto riesgo, pero la escasa calidad de los estudios de investigación hace que prevalezca una gran incertidumbre. Los ensayos que evaluaron la cubierta de gel de silicona como tratamiento para las cicatrices hipertróficas y queloides mostraron mejorías en el grosor y el color de las cicatrices, pero son de calidad baja y muy susceptibles a sesgo.
PICO
Resumen en términos sencillos
Cubierta de gel de silicona para prevenir el desarrollo de cicatrices hipertróficas y queloides y para tratar las cicatrices hipertróficas y queloides existentes
A medida que la herida se cura, se puede desarrollar una cicatriz. A veces las cicatrices se pueden desarrollar de manera anormal, formando cicatrices hipertróficas o queloides que se elevan, son antiestéticas y pueden causar problemas emocionales y de movimiento a las personas que las presentan. Estos tipos de cicatrices son difíciles de tratar.
Las cicatrices queloides son más comunes en la piel más oscura y se producen después de lesiones leves como picaduras de insectos, perforaciones en las orejas y vacunaciones. Las cicatrices queloides también se pueden extender a la piel que rodea la zona lesionada. La cicatrización hipertrófica es más común en pieles claras y se suele limitar a la zona lesionada. La cicatrización hipertrófica suele seguir a las cirugías o a las quemaduras. Las cicatrices hipertróficas y queloides tienen más probabilidades de desarrollarse si la lesión se produce en determinados lugares del cuerpo, por ejemplo, en la parte inferior de la cara, el cuello y la parte superior de los brazos.
La cubierta de gel de silicona es una lámina suave y autoadhesiva que se aplica a la piel intacta. Se cree que previene el desarrollo de nuevas cicatrices anormales y que también trata las cicatrices existentes. El objetivo de esta revisión fue evaluar la evidencia sobre si la cubierta de gel de silicona previene el desarrollo de cicatrices anormales en los pacientes con heridas recién cicatrizadas o si es una forma efectiva de tratar las cicatrices anormales existentes. La mayoría de los estudios identificados fueron de calidad baja y no está claro si la cubierta de gel de silicona ayuda a prevenir las cicatrices anormales o es eficaz para tratar las cicatrices anormales existentes.
Authors' conclusions
Background
Wounds, such as burns, surgical incisions and ulcers, are repaired through the deposition of components that form new skin. These components include blood vessels, nerves, elastin fibres (which give the skin some elasticity) and collagen fibres (for tensile strength), as well as glycosaminoglycans (GAGS) which form the gel‐like ground substance (or matrix) in which the structural fibres, nerves and blood vessels are embedded. In the early stages of healing, a cicatrix is formed. The cicatrix consists of a thin layer of skin (the pellicle) that covers the wound and subsequently contracts and becomes paler in colour, forming the scar.
Some scars develop abnormally, giving rise to keloid and hypertrophic scars. The scars arise from an excessive proliferation of dermal tissue following skin injury, with keloid scars developing in 5% to 15% of wounds (Wittenberg 1999). This proliferation of dermal tissue is due to both the production of fibrous tissue (fibroplasias) and the accumulation of abundant and randomly organised new collagen bundles.
O'Sullivan 1996 observed that although the terms 'keloid' and 'hypertrophic' have often been used synonymously, the two sorts of scarring are, in fact, significantly different. The principle clinical feature that distinguishes them is that in keloid scars the scar tissue progressively encroaches upon the normal skin surrounding it, producing a scar that appears irregular and pendulous in areas. Conversely, the hypertrophic scar is confined to the tissue damaged by the original injury. This type of scar increases in dimension by pushing out its margins, rather than invading surrounding tissue. Clinicians usually base diagnosis of keloid scarring on the overgrown boundaries and delayed onset of the scar (hypertrophic scars develop soon after injury) (Shaffer 2002).
Keloid scarring is reported to be more common in darker skin (Beers 1999; Niessen 1998), while hypertrophic scarring is more common in fair skin (Beers 1999). Examination of scars with an electron microscope shows keloid collagen to be thin and irregular with cross‐striations, suggesting immaturity, while keloid scars are deficient in lymphatics and their associated elastic fibres, and have a higher content of both water and soluble collagen than normal skin. Although hypertrophic scars have similar qualities in the early stages, after seven months the two become distinct as the water and collagen content of hypertrophic scars normalises (Raney 1993).
Hypertrophic scars tend to follow surgery and thermal injuries such as severe burns (Carney 1993; Eisenbeiss 1998; Shakespeare 1993), whereas keloid scars often originate after trivial injury such as ear piercing, insect bites and vaccination. The amount of scar tissue in a keloid scar exhibits little relation to the extent of the injury that caused it (O'Sullivan 1996).
Both types of scarring can cause functional and psychological problems for people, and their management can be difficult. Treatment options have included surgery, radiation therapy, steroid injections, pressure therapy, cryotherapy (treatment with liquid nitrogen) and laser therapy (Shaffer 2002). Many surgical techniques have been applied to remove keloids, either alone, or in combination with other treatments. Surgery alone has shown a high recurrence rate (Raney 1993).
Scars in specific sites of the body, including the lower face, presternum, pectoral area of the chest, upper back, ears, neck and outer (deltoid) area of the upper arms are more likely to develop abnormally (O'Sullivan 1996). People with scars in these high‐risk anatomical areas, or with a history of forming keloid scars, aim to prevent further scarring by observing certain principles that include: avoiding non essential cosmetic surgery, closing all wounds with minimal tension, and using pressure garments for four to six months after injury or surgery (O'Sullivan 1996).
The use of topical silicone for prevention and treatment of hypertrophic scarring is still relatively new. Silicone was first used, in gel form, for the treatment of burn scars at Australia's Adelaide Children's Hospital in 1981 (Perkins 1982). Silicone has since been produced in various forms, including: silicone cream compounds (Sawada 1992); silicone oil or gel with additives such as vitamin E (Palmieri 1995); in combination with other dressing media (Davey 1991); and as custom‐made silicone applications. This particular review is solely concerned with commercially produced adhesive silicone gel sheeting.
Silicone gel sheeting is a soft, self adhesive and semi‐occlusive sheet used for the treatment and prevention of both old and new hypertrophic and keloid scars. It is made from medical‐grade silicone (cross‐linked polydimethylsiloxane polymer) and reinforced with a silicone membrane backing (Katz 1992; Thomas 1997) thought to give it increased durability and make handling easier (Williams 1996).
Silicone gel sheeting is designed to be used on intact skin. It should not be used on open wounds and, according to the product information sheet supplied by the manufacturers (Smith & Nephew 2000), is contraindicated in people with dermatological conditions that disrupt the integrity of the skin (for example, severe acne or psoriasis).
The mode of action of silicone‐based products on scar tissue is unknown. Some researchers suggested that silicone may penetrate the skin, but studies by Ahn 1989 and Swanson 1974 found no evidence of silicone in the scar or stratum corneum. Quinn 1985 found that there was no significant difference in pressures obtained at the scar surface beneath the gel, and also concluded that there was no difference in scar surface temperature and oxygen tension, or water vapour transmissivity of the gel.
The cost of silicone gel sheeting (AUD 139 (Australian Dollars) recommended retail price for a 12 x 15 cm sheet, AUD 74 for a 12 x 6 cm sheet) may be moderated by the fact that, after rinsing, it can be reused by the patient or their carer. However, the fact remains that clinicians and funders of care will require clear evidence of its clinical effectiveness before recommending its use.
Objectives
The aim of this systematic review was to determine the effects of silicone gel sheeting in the:
-
prevention of hypertrophic or keloid scarring in people with newly healed wounds (e.g. post surgery); and
-
treatment of established scarring in people with keloid or hypertrophic scars after any type of wound.
Methods
Criteria for considering studies for this review
Types of studies
Any randomised controlled trials (RCTs) or quasi‐randomised controlled trials (QRCTs) (method for allocating participants to a treatment that is not strictly random, e.g. by date of birth, hospital record number, alternation) or controlled clinical trials (CCTs) (where an intervention group is compared to a comparison or control group) of interventions.
Types of participants
People with healed full‐thickness wounding (from any cause) where the skin was intact, with or without scarring at baseline.
Types of interventions
All comparisons of silicone gel sheeting with other conservative techniques (e.g. hydrocolloid dressings, non silicone gel sheeting, laser therapy or no intervention) were eligible.
We excluded comparisons of silicone gel sheeting with surgery. We excluded trials that reported only the absorption of silicone by the skin, but did not measure the effect on scar appearance.
Types of outcome measures
Primary outcomes
Prevention studies
The primary outcome measure was the number of people who developed keloid or hypertrophic scarring as determined by blood flow, hyperpigmentation, erythema (redness), scar thickness and regularity of scar.
Treatment studies
The primary measure was change in scar size (measured by area, length, volume, height, or width ‐ usually by ruler, taking an impression, or ultrasound).
Secondary outcomes
Prevention studies
Other measures of clinical outcome:
-
scar size (measured by area, length, volume, height, or width ‐ usually by ruler, taking an impression, or ultrasound);
-
scar colour (measured against standard colour charts), blood flow (measured using laser‐Doppler flowmetry) and scar appearance (measured on a three or five‐point scale with appropriate definitions);
-
skin elasticity (measured serially with the use of an elastometer);
-
development of complications (e.g. rashes, skin breakdown, measured on a numbered scale);
-
cosmetic appearance (cosmesis) as defined by patient opinion (using assessment scales) and physician observations;
-
patient tolerance, measured by reported side effects and adverse reactions;
-
preference for different modes of treatment, measured by patient choice after receiving at least two different types of treatment;
-
compliance, measured by physician and patient report.
Treatment studies
Other measures of clinical outcome:
-
scar colour (measured against standard colour charts), blood flow (measured using laser‐Doppler flowmetry) and scar appearance (measured on a three or five‐point scale with appropriate definitions);
-
skin elasticity (measured serially with the use of an elastometer);
-
development of complications (e.g. rashes, skin breakdown, measured on a numbered scale);
-
cosmesis as defined by patient opinion (using assessment scales) and physician observations;
-
patient tolerance, measured by reported side effects and adverse reactions;
-
preference for different modes of treatment measured by patient choice after receiving at least two different types of treatment;
-
compliance, measured by physician and patient report.
Search methods for identification of studies
For the search strategy for the first update of this review see Appendix 1.
Electronic searches
For this second update we searched the following databases in May 2013:
-
The Cochrane Wounds Group Specialised Register (searched 8 May 2013);
-
The Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2013, Issue 4);
-
Ovid MEDLINE (2007 to April Week 4 2013);
-
Ovid MEDLINE (In‐Process & Other Non‐Indexed Citations, May 07 2013);
-
Ovid EMBASE (2007 to 2013 Week 18);
-
EBSCO CINAHL (2007 to 3 May 2013)
We used the following search strategy to search CENTRAL:
#1 MeSH descriptor: [Keloid] explode all trees 61
#2 MeSH descriptor: [Cicatrix, Hypertrophic] explode all trees 84
#3 MeSH descriptor: [Hypertrophy] explode all trees 1125
#4 keloid* or hypertrophic or cicatrix 1030
#5 scar or scars or scarred or scarring 2163
#6 #1 or #2 or #3 or #4 or #5 3712
#7 MeSH descriptor: [Silicone Gels] explode all trees 34
#8 silicone next gel* 89
#9 silicone next sheet* 17
#10 silicone next dressing* 18
#11 #7 or #8 or #9 or #10 114
#12 #6 and #11 59
The search strategies for Ovid MEDLINE, Ovid EMBASE and EBSCO CINAHL can be found in Appendix 2. We combined the Ovid MEDLINE search with the Cochrane Highly Sensitive Search Strategy for identifying randomized trials in MEDLINE: sensitivity‐ and precision‐maximizing version (2008 revision) (Lefebvre 2011). The Ovid EMBASE and EBSCO CINAHL searches were combined with the trial filters developed by the Scottish Intercollegiate Guidelines Network (SIGN) (SIGN 2012). There were no restrictions with respect to language, date of publication or study setting.
Searching other resources
We examined the reference lists of relevant review articles and all included studies to identify further studies. We approached the major supplier of silicone gel sheeting (Smith and Nephew) for details of unpublished, ongoing and recently published trials. The search was not limited by language or publication status.
Data collection and analysis
Selection of studies
Two review authors (LOB, DJ) assessed the title and abstracts of potentially eligible trials independently. The review authors obtained papers that were potentially relevant and, using eligibility criteria, assessed their full text for inclusion independently. We resolved disagreements by discussion.
Data extraction and management
One review author extracted data and a second review author checked for accuracy. We used a standard data form to capture the following information:
-
characteristics of the study (design, method of randomisation, withdrawals/dropouts, funding source);
-
study participants (age, wound location, wound characteristics, scar type);
-
intervention (silicone gel, non silicone gel);
-
comparison intervention (e.g. laser therapy, compression, occlusive dressing);
-
duration of treatment;
-
outcome measures (type of scoring, timing of assessment, complications);
-
duration of follow‐up; and
-
results.
We requested additional unpublished data from primary authors and included these when available.
Assessment of risk of bias in included studies
Only RCTs, QRCTs or CCTs were included in this review because of the increased risk of bias with other types of study. Two review authors independently assessed the methodological quality of the included studies using the Cochrane Collaboration tool for assessing risk of bias (Higgins 2011) and any disagreement was discussed amongst all review authors to achieve a consensus. The 'Risk of bias' tool addresses six specific domains, namely sequence generation, allocation concealment, blinding, incomplete outcome data, selective outcome reporting and other issues (e.g. extreme baseline imbalance) (see Appendix 3 for details of criteria on which the judgement was based). We assessed blinding and completeness of outcome data for each outcome separately, and completed a 'Risk of bias' table for each eligible study.
We have presented our assessment of risk of bias using a 'Risk of bias' summary figure (Figure 1), which presents all of the judgements in a cross‐tabulation of study by entry. This display of internal validity indicates the weight the reader may give the results of each study.
'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.
Assessment of heterogeneity
We explored clinical heterogeneity by examining potentially influential factors such as age of people, cause of scar (e.g. if from recent surgery) and age of scar before treatment commenced. When statistical pooling was done, we tested for statistical heterogeneity using the Chi² test. If clinical heterogeneity was suspected, we combined the studies by narrative summary only. In the presence of statistical heterogeneity (i.e. when the Chi² was greater than degrees of freedom) but where other factors suggest pooling was appropriate, we used a random‐effects model. Otherwise we used a fixed‐effect model.
Data synthesis
The comparisons are as follows.
-
silicone gel sheeting compared with no treatment.
-
silicone gel sheeting compared with non silicone dressing.
-
silicone gel sheeting compared with other silicone products.
-
silicone gel sheeting compared with laser therapy.
-
silicone gel sheeting compared with triamcinolone acetonide injection treatment.
-
silicone gel sheeting compared with topical onion extract.
-
silicone gel sheeting compared with pressure therapy.
Data for prevention (i.e. for newly healed scars) and treatment (i.e. for existing keloid or hypertrophic scars) have been dealt with separately.
The analysis tables contain quantitative data from individual trial reports for prespecified outcomes and subgroups (e.g. those with a high risk of abnormal scarring versus normal population) for both dichotomous and continuous outcomes.
A narrative summary of results is presented. Results of dichotomous variables are presented as risk ratio (RR) with 95% confidence intervals (CI). We have used risk ratio rather than odds ratio, as event rates are high in these trials and odds ratios would give an inflated impression of the magnitude of effect. In addition, we have carried out statistical pooling on groups of studies which we considered to be sufficiently similar.
Results
Description of studies
Searches for this second update identified 13 potentially relevant articles. Independent scrutiny of the titles and abstracts by both review authors identified five new studies that met the inclusion criteria, bringing the total number of included studies to 20. Reasons for excluding the other studies can be found in the Characteristics of excluded studies.
We contacted 11 authors for additional trial data and five (de Oliveira 2001; Niessen 1998; Li‐Tsang 2006; Li‐Tsang 2010; Niessen 1998; Wigger‐Alberti 2009) kindly supplied these. We contacted the manufacturers of silicone gel sheeting (Smith & Nephew) and they supplied a categorised table of clinical trials conducted for key scar therapies. We checked this against the studies already sourced through the search strategy and ordered any papers not already considered, then subjected them to the same eligibility criteria as the other trials to determine whether they should be included. No further trials were identified from this source.
All 20 included trials compared silicone gel sheeting with either a control or another treatment. The studies were mainly single‐centre studies, although one included data from four hospitals (Niessen 1998). The studies were conducted in eight countries, with most being conducted in either North America (seven studies) or in Europe (eight studies). Where there were multiple trials for the same first author, we inspected for independence of study populations and found that that all were separate groups of participants. Prospective trial registration with unique trial numbering would help avoid duplication in systematic reviews.
The 20 included studies involved a total of 849 people aged between 1.5 to 81 years. The 'Characteristics of included studies' table provides details of individual studies. No age limits were explicitly applied, however where information was provided, most participants were adult.
In three studies (de Oliveira 2001; Gold 1994; Niessen 1998) a distinction was made between keloid and hypertrophic scarring and the results were discussed separately.
The trials made the following comparisons based on the objectives (i.e. to determine the effectiveness of silicone gel sheeting in preventing and treating hypertrophic and keloid scars):
(1) Silicone gel sheeting compared with no treatment
There were three prevention studies (Cruz‐Korchin 1996; Gold 2001; Niessen 1998) involving 245 people.
There were eight treatment studies (Ahn 1989; Carney 1994; Colom Majan 2006; de Oliveira 2001; Li‐Tsang 2006; Li‐Tsang 2010; Tan 1999; Wittenberg 1999) involving 219 people.
There were two studies that evaluated both prevention and treatment (Ahn 1991; Gold 1994) involving 82 people.
(2) Silicone gel sheeting compared with non silicone dressing
There were three treatment studies (de Oliveira 2001; Momeni 2009; Wigger‐Alberti 2009) involving 124 people.
(3) Silicone gel sheeting compared with other silicone products
There was one prevention study (Niessen 1998) involving 129 people.
There were three treatment studies (Carney 1994; Karagoz 2009; Palmieri 1995) involving 152 people.
(4) Silicone gel sheeting compared with laser therapy
There were two treatment studies (Pacquet 2001; Wittenberg 1999) involving 40 people.
(5) Silicone gel sheeting compared with triamcinolone acetonide injection treatment
There were three treatment studies (Kelemen 2007; Sproat 1992; Tan 1999) involving 58 people.
(6) Silicone gel sheeting compared with topical onion extract
There was one treatment study (Karagoz 2009) involving 30 people.
(7) Silicone gel sheeting compared with pressure therapy
There was one treatment study (Li‐Tsang 2010) involving 54 people.
There were many different measurement techniques and tools used, which made pooling of results difficult. The comparability of people at baseline was generally good, although one study (Pacquet 2001) provided no information on the control group, making it impossible to judge whether those groups were comparable. Most studies were also explicit about their inclusion and exclusion criteria, which allowed a clearer definition of the study population.
In most trials silicone gel sheeting was applied for at least 12 hours per day, with five studies (de Oliveira 2001; Li‐Tsang 2006; Li‐Tsang 2010; Momeni 2009; Niessen 1998) specifying 24 hours per day, another (Carney 1994) stating "as many hours per day as possible", and a third (Palmieri 1995) specifying 10 hours per day. However, two studies (Pacquet 2001; Wigger‐Alberti 2009) did not indicate the number of hours that the silicone gel sheeting was worn by participants. One study (Niessen 1998) changed the type of silicone gel sheeting used (from Sil‐K to Epiderm which is more adhesive) when the initial results from the first group of people (n = 80) were described by the authors as "disappointing". Another study (Carney 1994) also used two different types of gel (Silastic Gel Sheeting and Cica‐Care) and analysed the treatment subgroups separately.
Descriptions and definitions of the type of scar (hypertrophic versus keloid) were adequate in 13 out of the 20 studies. Despite not giving a full description of the distinction between hypertrophic and keloid scars, de Oliveira 2001 classified their participants' scars as either one or the other, and separated the scar types in their analysis. Gold 2001 compared high‐risk (i.e. those with a history of abnormal scarring) and low‐risk participant groups in their results. Most other studies combined hypertrophic and keloid scars in their analyses, raising questions about the appropriateness of the study design (Shaffer 2002).
Given the long‐term process of remodeling and scarring, it is recommended that follow‐up continues for at least one year (Shaffer 2002). Only three studies (Carney 1994; Colom Majan 2006; Niessen 1998) had follow up of 12 months. Six studies (Ahn 1989; Gold 1994; Palmieri 1995; Sproat 1992; Tan 1999; Wigger‐Alberti 2009) followed people for three months or less, which is clearly inadequate.
Risk of bias in included studies
The quality of trial methodology varied widely. The results for individual trials are presented in the 'Characteristics of included studies' table. Overall, the quality of the trial methodology in the included studies was poor. We judged only two studies (Momeni 2009; Wigger‐Alberti 2009) to be at overall low risk of bias.
Allocation
Random sequence generation
Five studies (Colom Majan 2006; Li‐Tsang 2010; Momeni 2009; Wigger‐Alberti 2009; Wittenberg 1999) explicitly reported their method of generating the randomisation sequence and we judged them to be at low risk of selection bias. Colom Majan 2006, , Wigger‐Alberti 2009 and Wittenberg 1999 used a computer‐generated randomisation list; Sproat 1992 used a prescribed randomised sequence; Li‐Tsang 2010 used the drawing of lots for randomisation; Momeni 2009 used a random number table. One study (Cruz‐Korchin 1996) reported an inadequate method of randomisation and we judged this to be at high risk of selection bias. Cruz‐Korchin 1996 allocated treatment based on the patient's dominant hand. The remainder did not describe their methods and we judged them at unclear risk of selection bias.
Allocation concealment
One study (Cruz‐Korchin 1996) did not conceal allocation (they used the dominant or non dominant hand of the participant to decide where the material was placed) and we judged this to be at high risk of bias. The remaining studies did not report allocation concealment and were judged to be at unclear risk of bias.
Blinding
Blinding of participants and personnel
We judged two studies (Momeni 2009; Wigger‐Alberti 2009) to have unclear risk of performance bias. One study (Momeni 2009) attempted to blind participants only by using placebo gel sheets on one‐ half of the scar but did not attempt to blind research personnel. The other study (Wigger‐Alberti 2009) blinded research personnel by using an independent nurse to apply and remove all dressings in the investigators' absence. We judged the remaining studies to be at high risk of performance bias as they were unable to blind either participants or personnel due to the nature of the intervention.
Blinding of outcome assessor
We judged four studies (Li‐Tsang 2006; Li‐Tsang 2010; Momeni 2009; Sproat 1992) to have adequately blinded outcome assessors. Two studies (Li‐Tsang 2006; Li‐Tsang 2010) used an independent research assistant to judge the outcome; Momeni 2009 used an independent plastic surgeon; Sproat 1992 used photographs showed to five blinded observers and trained a 'blindfolded observer' to undertake measurements. We judged these four studies to be at low risk of detection bias. One study (Karagoz 2009) used the same outcome assessor at the beginning and end of the treatment and was not blinded. We judged this study was judged to be at high risk of detection bias. The remaining studies did not comment on blinding of outcome assessment and were judged to be at unclear risk of detection bias.
Incomplete outcome data
We judged fifteen15 studies (Ahn 1989; Colom Majan 2006; Cruz‐Korchin 1996; Gold 1994; Karagoz 2009; Kelemen 2007; Li‐Tsang 2006; Momeni 2009; Niessen 1998; Pacquet 2001; Palmieri 1995; Sproat 1992; Tan 1999; Wigger‐Alberti 2009; Wittenberg 1999) to be at low risk of attrition bias. In each of these studies the numbers lost to follow up were low and adequate reasons were given for these losses. We judged the remaining studies to be at high risk of attrition bias. Three studies (Ahn 1991; Carney 1994; Gold 2001) had a loss to follow up greater than 20%. One study (Li‐Tsang 2010) reported moderately high losses but was judged to be at high risk of bias because two thirds of drop outs came from the control group. One study (de Oliveira 2001) did not comment on attrition bias nor can it be ascertained from the data. We judged this study to be at unclear risk of attrition bias.
Clear statements of evidence of intention‐to‐treat (ITT) analysis were rarely presented in trial reports, and only two studies (Wigger‐Alberti 2009; Wittenberg 1999) performed an ITT analysis.
Selective reporting
We judged one study (Kelemen 2007) to be at high risk of reporting bias, as only one "interesting" case from each group was reported, and patient ratings were not reported at all. We assessed all other studies in the review were assessed as having low risk of reporting bias. Although study protocols were not sought, all outcomes mentioned in the methods were reported in the results and clinically meaningful outcomes presented.
Other potential sources of bias
We judged seven studies (Ahn 1991; Colom Majan 2006; Gold 1994; Gold 2001; Li‐Tsang 2006; Li‐Tsang 2010; Wittenberg 1999) to be at high risk of bias due to the influence of companies supplying the silicone gel sheeting. Two studies (Ahn 1991; Gold 1994) reported receiving grant money from the companies supplying the silicone gel sheeting. Two studies (Colom Majan 2006; Gold 2001) stated that the research was supported by the company supplying the silicone gel but gave no further information. In the remaining three studies the company donated the silicone gel sheeting (Li‐Tsang 2006; Li‐Tsang 2010; Wittenberg 1999). The remaining studies appear free from other sources of bias.
Effects of interventions
Where available quantitative data are presented in the analysis tables.
How the results are presented and what the terms mean
Results of dichotomous variables are presented as risk ratio (RR) with 95% confidence intervals (CI). Risk ratio has been used rather than odds ratio as event rates are high in these trials and odd ratios would give an inflated impression of the magnitude of effect. Where statistically significant heterogeneity existed (i.e. the Chi² was greater than degrees of freedom) we used a random‐effects model.
The types of outcomes measured in the studies are listed in the 'Characteristics of included studies' table. The primary outcome measure for prevention studies was the proportion of people who developed abnormal scarring in postoperative cases (measured in terms of blood flow, hyperpigmentation, erythema, thickness and regularity of scar). There were many different measurement techniques and tools used, making pooling of results difficult.
Comparison: silicone gel compared with no treatment
There were 13 studies (Ahn 1989; Ahn 1991; Carney 1994; Colom Majan 2006; Cruz‐Korchin 1996; de Oliveira 2001; Gold 1994; Gold 2001; Li‐Tsang 2006; Li‐Tsang 2010; Niessen 1998; Tan 1999; Wittenberg 1999) in this category. Three of the studies (Cruz‐Korchin 1996; Gold 2001; Niessen 1998) studied the prevention of scars for people undergoing surgery, eight studied the effect of silicone gel sheeting on existing hypertrophic or keloid scars (Ahn 1989; Carney 1994; Colom Majan 2006; de Oliveira 2001; Li‐Tsang 2006; Li‐Tsang 2010; Tan 1999; Wittenberg 1999) and two studies (Ahn 1991; Gold 1994) included both prevention and treatment.
I: Prevention studies
Of the five trials that compared silicone gel sheet with no treatment for prevention of scarring, four (Ahn 1991; Cruz‐Korchin 1996; Gold 2001; Niessen 1998) included people with healed surgical wounds, and one (Gold 1994) included people who had had keloid scars removed with CO2 laser. Two of the trials described people according to their risk of developing abnormal scarring ‐ Gold 1994 only recruited 'high‐risk' people, while Gold 2001 recruited 'low' and 'high' risk people and presented the results of these two groups separately.
Primary outcome: development of keloid or hypertrophic scarring
Cruz‐Korchin 1996 reported that fewer incisions treated with silicone gel sheeting became hypertrophic, though this difference was not significant(risk ratio (RR) 0.45, 95% confidence interval (CI) 0.19 to 1.07). Individually, two small trials (Gold 1994; Gold 2001) found no significant difference between the silicone gel sheeting and the control groups in terms of abnormal scarring in high‐risk individuals only (people who were prone to scarring), but when pooled (random‐effects) we found that silicone gel sheeting was associated with significantly fewer abnormal scars(RR 0.46, 95% CI 0.21 to 0.98). Ahn 1991 found significantly fewer abnormal scars in people treated with silicone gel sheeting (RR 0.05, 95% CI 0 to 0.76), whilst Niessen 1998 found a significant difference in favour of the control group(RR 2.71, 95% CI 1.19 to 6.22). When all five trials were pooled (random‐effects, I² = 69%) there was no significant difference in the number of people developing abnormal scars (RR 0.55, 95% CI 0.21 to 1.45) (Analysis 1.1). All these trials are susceptible to bias as they did not describe allocation concealment, blinding of outcome assessors or an intention‐to‐treat (ITT) analysis.
Secondary outcomes
Cruz‐Korchin 1996 reported transient rash and minor skin maceration as complications, but there was no statistically significant difference between the groups. Niessen 1998 reported transient rash, which resolved on removal of the silicone gel sheeting. Pooling these studies (fixed‐effect, I² = 0%) demonstrated a statistically significant difference in favour of the control groups. This means that more complications developed in the groups treated with silicone gel (RR 8.00, 95% CI 1.02 to 62.83) (Analysis 1.2).
II: Treatment studies
Ten trials compared silicone gel sheeting with control for treating abnormal scarring (Ahn 1989; Ahn 1991; Carney 1994; Colom Majan 2006; de Oliveira 2001; Gold 1994; Li‐Tsang 2006; Li‐Tsang 2010; Tan 1999; Wittenberg 1999). The majority of control groups were untreated. In two studies the control group received lanolin and massage. Six trials included people with hypertrophic scars resulting from thermal burns (Carney 1994; Gold 1994; Li‐Tsang 2006; Li‐Tsang 2010) or surgery (Colom Majan 2006; Wittenberg 1999). Three (Ahn 1991; Ahn 1989; de Oliveira 2001) included people with hypertrophic and keloid scarring, and one (Tan 1999) only included people with keloid scarring.
Primary outcome
As the studies used different outcome measures it was impossible to pool results. We examined outcomes of reduction of scar length and width (de Oliveira 2001), scar thickness (Li‐Tsang 2006; Li‐Tsang 2010) and reduction in scar size by 50% (Tan 1999). The studies found no significant difference between silicone gel sheeting and control for reduction in scar length, width and reduction of size by 50% (Analysis 1.3; Analysis 1.4; Analysis 1.7) but significant results for scar thickness favouring silicone gel (RR ‐2.00, 95% CI ‐2.14 to ‐1.85) (Analysis 1.5) although this was only two studies (Li‐Tsang 2006; Li‐Tsang 2010) with relatively small numbers (N = 77).
Secondary outcomes
All studies except Wittenberg reported secondary outcomes. There were no statistically significant differences between the treatment or control groups for improvements in scar appearance, scar colour and the relief of itching and pain.
Five studies (Colom Majan 2006; de Oliveira 2001; Li‐Tsang 2006; Tan 1999de Oliveira 2001Li‐Tsang 2010; Tan 1999) showed a statistically significant amelioration of scar colour (defined as a significant improvement in erythema) with silicone gel (pooled RR 3.49, 95% CI 1.97 to 6.15, fixed‐effect, I² = 53%) (Analysis 1.8). When a random‐effects model is applied this result is still statistically significant. It should be noted, however, that with the exception of Li‐Tsang 2010 who used spectrocolorimetry, this is a subjective outcome and only two studies (de Oliveira 2001; Li‐Tsang 2010) masked the outcome assessor. Also, only Li‐Tsang 2010 reported the method of randomisation and no studies had adequate allocation concealment.
Four studies (Ahn 1989; Ahn 1991; Carney 1994; Li‐Tsang 2006) reported a statistically significant improvement in scar elasticity in those people treated with silicone gel sheeting. Data were presented graphically (mean percentage of stretch and standard error of mean in Ahn 1989 and Ahn 1991; percentage of extensibility of scar in Carney 1994; mean only in Li‐Tsang 2006) with P values, but actual measurement data were not reported. We requested further information from trial authors, with two replies (Li‐Tsang 2006; Li‐Tsang 2010) resulting in new data. We treated reported data as dichotomous (i.e. improvement in elasticity compared with no improvement) and due to the high heterogeneity likely caused by the different measurement methods (I² = 55%), pooled using a random‐effects model resulting in a statistically significant improvement in scar elasticity (RR 3.03, 95% CI 1.02 to 8.99) (Analysis 1.9).
Results for relief of pain and itch (Li‐Tsang 2006; Li‐Tsang 2010; Tan 1999) showed no statistically significant difference between the groups (Analysis 1.10).
Three studies (Ahn 1989; Carney 1994; Colom Majan 2006) reported complications such as transient skin rashes, pruritis, itching or superficial maceration. Authors reported that these resolved promptly when the silicone gel sheeting was withdrawn, or when correct hygiene was practised. Combining results from Ahn 1989 and Colom Majan 2006 we found statistically significantly more complications reported for silicone gel sheeting than in the control group (RR 9.52, 95% CI 1.35 to 67.10, fixed‐effect, I² = 0%) (Analysis 1.11). No raw data were reported by Carney and email communication with the author did not produce further data.
Comparison: silicone gel compared with non silicone dressing
I: Prevention studies
No prevention studies were identified.
II: Treatment studies
Three studies (de Oliveira 2001; Momeni 2009; Wigger‐Alberti 2009) including 124 people compared silicone gel sheeting with non silicone gel sheeting. One study (de Oliveira 2001) classified scars as either hypertrophic or keloid, and the other two only included hypertrophic scars.
Primary outcome
There was no statistically significant difference between the two groups for reduction of scar width or scar length (Analysis 2.1; Analysis 2.2).
Secondary outcomes
There was no statistically significant difference between the two groups for amelioration of scar colour (RR 1.01, 95% CI 0.87 to 1.17) (Analysis 2.3).
Two studies reported complications. de Oliveira 2001 reported irritative contact dermatitis which was resolved by washing the skin and removing the silicone gel sheeting for five hours. Wigger‐Alberti 2009 reported two adverse events described as local dermatitis which occurred in the silicone gel group. The final study (Momeni 2009) reported no adverse events in either group.
Comparison: silicone gel sheeting compared with silicone gel sheeting with different contact layers
I: Prevention studies
One study (Niessen 1998) involved 155 women undergoing bilateral breast reduction. This trial had three arms and scars were either treated with adhesive silicone gel sheeting (Epiderm adhesive), non adhesive silicone gel sheeting (Sil‐K) or covered with Micropore alone.
Primary outcome
The authors reported that 12 months after surgery no difference in hypertrophic scar development was found between the adhesive and non adhesive silicone gel sheets. However, they did not present separate data for the two intervention groups, but reported combined data for the silicone gel sheeting groups (adhesive plus non adhesive) compared with the control group. This trial was poorly reported and the method of allocating treatment to scar site was unclear. There was no blinded outcome assessment and no ITT analysis.
Secondary outcomes
Results obtained from the trial author (Niessen) by email on 238 scars (114 adhesive silicone gel group, 124 non adhesive silicone gel group) showed no statistically significant difference in results for scar width, height, colour and perfusion at 12 months post surgery (Analysis 6.1; Analysis 6.2; Analysis 6.3; Analysis 6.4).
II: Treatment studies
Three studies were included in this category. One (Palmieri 1995) compared silicone gel sheets to silicone gel sheets with added vitamin E. This study involved 80 people with established hypertrophic and keloid scars resulting from either surgery or thermal burns. One (Karagoz 2009) compared silicone gel sheets to a paint‐on silicone gel and included a total of 30 people with hypertrophic scarring post‐thermal burns in this part of the study. The third study (Carney 1994) compared adhesive (Cica‐Care) with non adhesive (Silastic) silicone gel sheeting, and included 42 people with 47 hypertrophic scars.
Primary outcome
Palmieri 1995 reports that photographs of scar size, colour and cosmesis were objectively scored on a scale of zero to five, however these results appear to have been combined with patient self ratings of itching and pain on a Scott‐Huskisson scale. We contacted the authors for clarification, but they did not reply, therefore these data could not be used and it was impossible to draw conclusions about the effectiveness of either treatment for change in scar size, or determine whether the assessors were blinded to treatment allocation. Size of scar was not measured by Carney 1994 or Karagoz 2009.
Secondary outcomes
Carney 1994 did not provide a statistical analysis of the comparison between the two silicone gel sheets, but stated that after six months of treatment, 88.9% of scars in the non adhesive gel group, and 100% of the scars in the adhesive group were improved for colour, and 100% of both groups were improved for scar softness. We contacted the lead author and asked them for data, however the author replied that the actual data had not been retained.
Palmieri 1995 reported a combined subjective and objective score, which showed that 75% of people treated with silicone gel sheeting had improvements in cosmesis, pain and itching of at least 50%, compared with 90% of those treated with silicone gel plates with added vitamin E. There was a statistically significant improvement in favour of silicone gel sheet with added vitamin E (RR 0.79, 95% CI 0.65 to 0.96) (Analysis 3.5). No complications were reported.
Karagoz 2009 used the Vancouver Scar Scale, reporting mean total scores and sub‐scale scores before and after treatment. They concluded that there was no significant difference between the silicone gel sheets and the paint‐on silicone group. Patient ratings of improvement on a four‐point scale were presented, but no statistical analysis was undertaken.
Given the presentation of results in these studies, it is impossible to draw a conclusion regarding the effectiveness of either treatment for the primary outcome (change in scar size).
Comparison: silicone gel sheeting compared with laser therapy
I: Prevention studies
No prevention studies were identified.
II: Treatment studies
Two studies (Pacquet 2001; Wittenberg 1999) involving 40 people compared the use of silicone gel sheeting with 585 nm pulsed dye laser therapy.
Primary outcome
Although scar size was measured in both studies, Pacquet 2001 did not report results at all and Wittenberg 1999 presented results for volume in graphical form only. We contacted both, but no responses were received.
Secondary outcomes
Pacquet 2001 found no statistically significant difference in scar erythema between people receiving silicone gel sheeting compared with those receiving 585 nm pulsed dye laser therapy. Similarly, Wittenberg 1999 also found no statistically significant difference in pain or burning, scar elasticity or fibrosis in people receiving these treatments. Since the published results in both papers were presented graphically and specific numerical data were not provided, no analyses tables or graphs are available in this review. No complications were reported by Pacquet 2001, although Wittenberg 1999 reported that one patient withdrew because of pain on laser treatment, and one patient was unable to use the silicone gel sheeting because of skin irritation.
Comparison: silicone gel sheeting compared with triamcinolone acetonide injection treatment
I: Prevention studies
No prevention studies were identified.
II: Treatment studies
Three studies involving 58 people were identified (Kelemen 2007; Sproat 1992; Tan 1999). Triamcinolone acetonide injections are an existing treatment for hypertrophic scars but can be painful.
Primary outcome
Tan 1999 reported that two out of the 17 people (12%) treated with silicone gel sheeting had a statistically significant reduction (defined as at least 50%) in the size of keloid scars, in contrast to the 16 out of 17 people (94%) who had a significant reduction when treated with intralesional injections of triamcinolone acetonide (40 mg/ml). The RR was 0.13 (95% CI 0.03 to 0.46) (Analysis 4.1). Sproat 1992 reported changes in scar height and width graphically. We contacted the researcher, but they had not kept specific numerical data, so no analyses tables or graphs are available in this review for these measures. Sproat 1992 reported that scar height decreased for both treatment groups, but that scar width increased in both (more so with triamcinolone acetonide injection). This trial report was not supported by any data analysis and therefore must be viewed with caution. Kelemen 2007 concluded that the intralesional steroid injection group had greater improvement of Vancouver scale scores after eight weeks than the silicone group, however they also did not provide any data analysis, so again these conclusions must be viewed with caution.
Secondary outcomes
Tan 1999 reported that people treated with the injections showed a statistically significant improvement in erythema compared to those treated with silicone gel sheeting; the RR was 0.10 (95% CI 0.01 to 0.70) (Analysis 4.2). There was no statistically significant difference for symptomatic relief of itching and pain (Analysis 4.3). Sproat 1992 reported a statistically significant difference in mean time to symptomatic improvement (mean difference ‐2.90 days, 95% CI ‐3.93 to ‐1.87) (Analysis 4.4) and patient preference in favour of the silicone gel sheeting (RR 5.50, 95% CI 1.48 to 20.42) (Analysis 4.5). Kelemen 2007 reported a faster reduction in patient‐reported symptoms (measured on a Likert scale), but provided no data to support this claim.
Sproat 1992 reported that statistically significantly more participants in the triamcinolone injection group experienced complications (including severe pain (71% of people), skin atrophy, pigmentary changes and white bead‐like skin deposits (64% of people)) compared with one instance of superficial rash in the silicone gel sheeting group (which resolved on discontinuation of the sheeting for two days) (RR 0.10, 95% CI 0.01 to 0.68) (Analysis 4.6). Tan 1999 and Kelemen 2007 both reported that no adverse reactions occurred with either treatment.
Comparison: silicone gel sheeting compared with topical onion extract
I: Prevention studies
No prevention studies were identified.
II: Treatment studies
One study involving 30 people was identified (Karagoz 2009).
Primary outcome
This study used spectrocolorimeter and Tissue Ultrasound Palpation System (TUPS) to measure scar colour and thickness respectively. The Vancouver Scar Scale was used to measure scar pliability. All scores were reported as mean/standard deviation (SD) of total scores for each group. After six months of intervention, there was a significant difference between the total score for the silicone gel sheeting and topical onion extract groups in favour of the silicone gel sheeting group (MD 1.90, 95%CI 0.62 to 3.18 Analysis 5.1). Patient ratings of improvement on a four‐point scale were presented, but no statistical analysis was undertaken.
Two participants in the silicone gel sheeting group developed skin maceration and pruritis, but this resolved after interrupting treatment for a week.
Comparison: silicone gel sheeting compared with pressure therapy
I: Prevention studies
No prevention studies were identified.
II: Treatment studies
One study involving 54 people was identified (Li‐Tsang 2010).
Primary outcome
This study used a spectrocolorimeter to measure colour (lightness, redness and yellowness), Tissue Ultrasound Palpation System (TUPS) to measure thickness and the Vancouver Scar Scale to measure pliability. Pain and itch were measured using a 10‐point visual analogue scale. Means and SDs were reported for all measures up to six months post commencement of treatment. No significant differences were found between the pressure therapy and the silicone gel sheeting groups on measures of scar thickness, scar colour lightness and yellowness, or pliability (Analysis 6.1; Analysis 6.2). Pain and itch were both significantly lower in the silicone gel sheeting group (Analysis 6.3; Analysis 6.4).
No complications were reported in either group.
Discussion
The introduction of silicone gel sheeting as both a prevention and treatment intervention in the early 1980s has led to a number of research trials of varied quality.
Whilst there is some weak evidence from two small trials (total 51 participants ‐ see comparison 1.1.1 in Analysis 1.1) that silicone gel‐treated incisions are less likely to become hypertrophic in high‐risk people, these trials had a high potential for selection and detection bias (method of randomisation unclear; no blinding of outcome assessors) and therefore must be viewed with a great deal of caution.
Similarly the findings that silicone gel sheeting improves the elasticity and colour of keloid scars came from low‐quality studies. Whilst triamcinolone appeared more effective at improving keloid scarring than silicone gel sheeting this finding too was from a single study with high susceptibility to bias (unclear randomisation; lack of blinding of outcome assessors).
Complications of silicone gel sheeting use including itch and skin rashes were reported in three trials and were more common than in control groups.
In this review, only randomised controlled trials (RCTs), quasi‐RCTs (QRCTs) and controlled clinical trials (CCTs) were considered, leading to a relatively small number of studies (20) and people (873) for evaluation. Few studies compared similar interventions or measured similar criteria. Several trials had methodological problems, and reported inadequately on their randomisation protocols and/or allocation concealment, or failed to undertake an intention‐to‐treat analysis. Blinding of outcome assessors, which would not have been difficult to achieve, was poorly reported. None of the included trials addressed health‐related quality of life, the minimum meaningful difference in scar characteristics such as size or colour, or the cost of treatments. In addition there is the potential for unit of analysis errors, some trials used the person as the unit of analysis, others the scar and in some cases the person was used as their own control, multiple scars on one person treated as being independent in the analysis would inflate precision of any pooled estimates, in general the reporting in this area was poor.
There was also some inconsistency in instruments of measurement, for example, different patient rating scales for pain/irritation/itch were used, making it difficult to compare results.
All but three of the studies had short duration of follow‐up (i.e. less than 12 months), which is inadequate given that scar remodeling and collagen synthesis continues for over a year.
It is interesting to note the difference between the results from the Niessen 1998 and Cruz‐Korchin 1996 trials when their clinical experiments were so similar (both treated women who had recently undergone breast reductions). Both researchers defined the difference in their trials, via letters published in Annals of Plastic Surgery (1997), in an attempt to explain their results. Niessen stated that the most important difference was the application of Micropore (3M) which provided support around the control (untreated) scars, thus demonstrating that "...it is not the silicone material itself that prevents the development of hypertrophic scar tissue". In her response, Cruz‐Korchin 1997 agreed that support would tend to reduce scar width, but silicone sheets would reduce width and flatten the hypertrophic scar. She also observed that her study population was composed mainly of Hispanics who are more prone to forming hypertrophic scars, and compared this to Niessen's study population of "fair‐skinned Caucasians, in whom hypertrophic scarring seldom occurs". At present this trial is the only one to have compared Micropore against silicone gel sheeting and, therefore, more research is needed to investigate whether the physical support of the scar is as effective as silicone gel sheeting.
In summary the effects of silicone gel sheeting on hypertrophic and keloid scarring are unclear and warrant rigorous evaluation.
'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.
Comparison 1 Silicone gel versus no treatment (control), Outcome 1 Development of abnormal scarring ‐ prevention.
Comparison 1 Silicone gel versus no treatment (control), Outcome 2 Development of complications ‐ prevention.
Comparison 1 Silicone gel versus no treatment (control), Outcome 3 Reduction of scar length ‐ treatment.
Comparison 1 Silicone gel versus no treatment (control), Outcome 4 Reduction in scar width ‐ treatment.
Comparison 1 Silicone gel versus no treatment (control), Outcome 5 Scar thickness ‐ treatment.
Comparison 1 Silicone gel versus no treatment (control), Outcome 6 Scar pliability ‐ treatment.
Comparison 1 Silicone gel versus no treatment (control), Outcome 7 Reduction of keloid scar size by 50% ‐ treatment.
Comparison 1 Silicone gel versus no treatment (control), Outcome 8 Scar colour amelioration ‐ treatment.
Comparison 1 Silicone gel versus no treatment (control), Outcome 9 Improvement in scar elasticity ‐ treatment.
Comparison 1 Silicone gel versus no treatment (control), Outcome 10 Symptomatic relief of itching and pain ‐ treatment.
Comparison 1 Silicone gel versus no treatment (control), Outcome 11 Development of complications ‐ treatment.
Comparison 2 Silicone gel versus non silicone dressing, Outcome 1 Reduction of scar width.
Comparison 2 Silicone gel versus non silicone dressing, Outcome 2 Reduction of scar length.
Comparison 2 Silicone gel versus non silicone dressing, Outcome 3 Scar colour improvement.
Comparison 3 Silicone gel versus silicone gel with different contact layers, Outcome 1 Scar width ‐ prevention.
Comparison 3 Silicone gel versus silicone gel with different contact layers, Outcome 2 Scar height ‐ prevention.
Comparison 3 Silicone gel versus silicone gel with different contact layers, Outcome 3 Scar colour ‐ prevention.
Comparison 3 Silicone gel versus silicone gel with different contact layers, Outcome 4 Scar perfusion ‐ prevention.
Comparison 3 Silicone gel versus silicone gel with different contact layers, Outcome 5 Improvement >50% in cosmesis, itching and pain ‐ treatment.
Comparison 4 Silicone gel versus triamcinolone acetonide injection treatment, Outcome 1 Reduction of keloid scar size by 50%.
Comparison 4 Silicone gel versus triamcinolone acetonide injection treatment, Outcome 2 Improvement in erythema.
Comparison 4 Silicone gel versus triamcinolone acetonide injection treatment, Outcome 3 Symptomatic relief of itching and pain.
Comparison 4 Silicone gel versus triamcinolone acetonide injection treatment, Outcome 4 Average time (in days) to improvement.
Comparison 4 Silicone gel versus triamcinolone acetonide injection treatment, Outcome 5 Patient preference.
Comparison 4 Silicone gel versus triamcinolone acetonide injection treatment, Outcome 6 Development of complications.
Comparison 5 Silicone gel versus topical onion extract, Outcome 1 Improvement in Vancouver Scar Scale.
Comparison 6 Silicone gel versus pressure therapy, Outcome 1 Scar thickness ‐ treatment.
Comparison 6 Silicone gel versus pressure therapy, Outcome 2 Scar pliability ‐ treatment.
Comparison 6 Silicone gel versus pressure therapy, Outcome 3 Pain.
Comparison 6 Silicone gel versus pressure therapy, Outcome 4 Itching.
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Development of abnormal scarring ‐ prevention Show forest plot | 5 | 402 | Risk Ratio (M‐H, Random, 95% CI) | 0.55 [0.21, 1.45] |
| 1.1 High risk of scarring | 2 | 51 | Risk Ratio (M‐H, Random, 95% CI) | 0.46 [0.21, 0.98] |
| 1.2 Low risk of scarring | 1 | 31 | Risk Ratio (M‐H, Random, 95% CI) | 0.35 [0.02, 8.08] |
| 1.3 Risk not stated | 3 | 320 | Risk Ratio (M‐H, Random, 95% CI) | 0.57 [0.10, 3.40] |
| 2 Development of complications ‐ prevention Show forest plot | 2 | 350 | Risk Ratio (M‐H, Fixed, 95% CI) | 8.0 [1.02, 62.83] |
| 2.1 Prevention | 2 | 350 | Risk Ratio (M‐H, Fixed, 95% CI) | 8.0 [1.02, 62.83] |
| 3 Reduction of scar length ‐ treatment Show forest plot | 1 | 27 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 4 Reduction in scar width ‐ treatment Show forest plot | 1 | 27 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 5 Scar thickness ‐ treatment Show forest plot | 2 | 77 | Mean Difference (IV, Fixed, 95% CI) | 0.00 [‐2.14, ‐1.85] |
| 6 Scar pliability ‐ treatment Show forest plot | 2 | 68 | Mean Difference (IV, Fixed, 95% CI) | ‐0.74 [‐0.83, ‐0.64] |
| 7 Reduction of keloid scar size by 50% ‐ treatment Show forest plot | 1 | 34 | Risk Ratio (M‐H, Fixed, 95% CI) | 5.0 [0.26, 97.00] |
| 8 Scar colour amelioration ‐ treatment Show forest plot | 5 | 151 | Risk Ratio (M‐H, Fixed, 95% CI) | 3.49 [1.97, 6.15] |
| 9 Improvement in scar elasticity ‐ treatment Show forest plot | 5 | 154 | Risk Ratio (M‐H, Random, 95% CI) | 3.03 [1.02, 8.99] |
| 10 Symptomatic relief of itching and pain ‐ treatment Show forest plot | 3 | 97 | Risk Ratio (M‐H, Random, 95% CI) | 1.23 [0.78, 1.96] |
| 11 Development of complications ‐ treatment Show forest plot | 2 | 39 | Risk Ratio (M‐H, Fixed, 95% CI) | 9.52 [1.35, 67.10] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Reduction of scar width Show forest plot | 1 | 30 | Mean Difference (IV, Fixed, 95% CI) | ‐0.03 [‐0.15, 0.09] |
| 2 Reduction of scar length Show forest plot | 1 | 30 | Mean Difference (IV, Fixed, 95% CI) | ‐0.02 [‐0.07, 0.03] |
| 3 Scar colour improvement Show forest plot | 2 | 136 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.01 [0.87, 1.17] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Scar width ‐ prevention Show forest plot | 1 | 238 | Mean Difference (IV, Fixed, 95% CI) | 0.80 [‐0.01, 1.61] |
| 2 Scar height ‐ prevention Show forest plot | 1 | 238 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [‐0.10, 0.10] |
| 3 Scar colour ‐ prevention Show forest plot | 1 | 233 | Mean Difference (IV, Fixed, 95% CI) | ‐0.20 [‐0.78, 0.38] |
| 4 Scar perfusion ‐ prevention Show forest plot | 1 | 235 | Mean Difference (IV, Fixed, 95% CI) | ‐1.40 [‐4.25, 1.45] |
| 5 Improvement >50% in cosmesis, itching and pain ‐ treatment Show forest plot | 1 | 80 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.79 [0.65, 0.96] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Reduction of keloid scar size by 50% Show forest plot | 1 | 34 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.13 [0.03, 0.46] |
| 2 Improvement in erythema Show forest plot | 1 | 34 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.1 [0.01, 0.70] |
| 3 Symptomatic relief of itching and pain Show forest plot | 1 | 18 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.6 [0.20, 1.79] |
| 4 Average time (in days) to improvement Show forest plot | 1 | 28 | Mean Difference (IV, Fixed, 95% CI) | ‐2.9 [‐3.93, ‐1.87] |
| 5 Patient preference Show forest plot | 1 | 28 | Risk Ratio (M‐H, Fixed, 95% CI) | 5.5 [1.48, 20.42] |
| 6 Development of complications Show forest plot | 1 | 28 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.1 [0.01, 0.68] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Improvement in Vancouver Scar Scale Show forest plot | 1 | 30 | Mean Difference (IV, Fixed, 95% CI) | 1.90 [0.62, 3.18] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Scar thickness ‐ treatment Show forest plot | 1 | 48 | Mean Difference (IV, Fixed, 95% CI) | ‐0.24 [‐0.97, 0.49] |
| 2 Scar pliability ‐ treatment Show forest plot | 1 | 48 | Mean Difference (IV, Fixed, 95% CI) | ‐0.35 [‐0.85, 0.15] |
| 3 Pain Show forest plot | 1 | 48 | Mean Difference (IV, Fixed, 95% CI) | ‐1.90 [‐2.99, ‐0.81] |
| 4 Itching Show forest plot | 1 | 44 | Mean Difference (IV, Fixed, 95% CI) | ‐2.04 [‐3.16, ‐0.92] |
