Topical benzoyl peroxide for acne

  • Protocol
  • Intervention


  • Zhirong Yang,

    Corresponding author
    1. School of Public Health, Peking University, Centre for Evidence Based Medicine and Clinical Research, Department of Epidemiology and Biostatistics, Beijing, China
    2. Shantou University Medical College, Shantou-Oxford Clinical Research Unit, Shantou, Guangdong, China
    • Zhirong Yang, Centre for Evidence Based Medicine and Clinical Research, Department of Epidemiology and Biostatistics, School of Public Health, Peking University, 38 Xueyuan Road, Haidian District, Beijing, 100191, China.

    Search for more papers by this author
  • Yuan Zhang,

    1. McMaster University, Department of Clinical Epidemiology and Biostatistics, Hamilton, ON, Canada
    Search for more papers by this author
  • Elvira Lazic Mosler,

    1. General Hospital "Dr. Ivo Pedišić", Department for Dermatology and Venereology, Sisak, Croatia
    Search for more papers by this author
  • Hang Li,

    1. Peking University First Hospital, Dermatologic Department, Beijing, Xicheng District, China
    Search for more papers by this author
  • Jing Hu,

    1. School of Public Health, Peking University, Centre for Evidence Based Medicine and Clinical Research, Department of Epidemiology and Biostatistics, Beijing, China
    Search for more papers by this author
  • Yanchang Zhang,

    1. Yale School of Public Health, Department of Chronic Disease Epidemiology, New Haven, Connecticut, USA
    Search for more papers by this author
  • Jia Liu,

    1. Xuanwu Hospital, Capital Medical University, Department of Neurology, Beijing, China
    Search for more papers by this author
  • Qian Zhang

    1. The University of Nottingham, c/o Cochrane Skin Group, Nottingham, UK
    Search for more papers by this author


This is the protocol for a review and there is no abstract. The objectives are as follows:

To assess the effects of benzoyl peroxide for acne.


Description of the condition

Acne vulgaris is a common, chronic inflammatory disease of pilosebaceous units. It is characterised by increased sebum production and the formation of comedones, erythematous papules, pustules and nodules, which may lead to scarring (Archer 2012).

For an explanation of the terminology used throughout the text, please refer to the glossary in Table 1.

Table 1. Glossary of unfamiliar terms
Acne inversaA chronic disease of the apocrine glands occurring mainly in the axillae and the groin regions. It is caused by poral occlusion with secondary bacterial infection, evolving into abscesses which eventually rupture. The chronic phase is characterised by ulcers, sinus tracts, fistulas, fibrosis and scarring
Acne vulgarisChronic acne involving mainly the face, chest and shoulders, which is common in adolescents, and characterised by the intermittent formation of discrete papular and/or pustular lesions, sometimes resulting in scarring
Alpha-hydroxy acidsOrganic acids, such as glycolic, lactic, citric and mandelic acid, containing a hydroxyl group bonded to the carbon atom adjacent to the carboxylic acid group. They are used in skin care preparations for their exfoliating properties
AndrogensA steroid hormone, such as testosterone or androsterone, which controls the development and maintenance of masculine characteristics. They stimulate sebaceous glands to grow and produce sebum, and therefore cause acne
Azelaic acidAzelaic acid is a natural material that kills bacteria in the skin and decreases the production of keratin. It is used to treat and prevent mild and moderate acne that is caused by bacteria
Bacterial resistanceThe ability of bacteria to resist the effects of an antibiotic
Benzoyl peroxideAn organic compound in the peroxide family used for acne treatment. It works as a peeling agent. It increases skin turnover, clearing pores and reducing the bacterial count (specifically P. acnes) as well as acting directly as an antimicrobial agent
ClindamycinA lincosamide antibiotic, commonly used for topical treatment of acne
ColonisationThe presence of bacteria on a body surface (like on the skin, mouth, intestines or airway) without causing disease in the person
ComedoneA blocked pore in the form of a yellow or black bump or plug on the skin
ComedolyticThe term used to describe a product or medication that inhibits the formation of comedones. Comedolytic products work by helping the skin to shed more effectively, keeping the pores from becoming plugged
CorticosteroidsAny of a class of steroid hormones formed in the cortex of the adrenal gland or chemically similar synthesised hormones that have anti-inflammatory properties
CystA closed sac, having a distinct membrane compared to the nearby tissue, which may contain air, fluids or semi-solid material
CytokinesA diverse group of soluble molecules important for cell signalling in the generation of an immune response, where they act as intercellular mediators or signalling molecules
Drug-induced acneAcne caused or exacerbated by several types of drugs, such as antiepileptics, halogens and steroids
DifferentiationThe process by which a less specialised cell becomes a more specialised cell
EczemaAn acute or chronic non-contagious inflammation of the skin, often caused by allergy and characterised by itching, scaling and blistering
ErythemaBlanching reddening of the skin due to local vasodilatation
ErythromycinA macrolide antibiotic, commonly used for topical treatment of acne
ChloracneAn acneiform eruption due to exposure to chlorine compounds
HypercolonisationAbnormal increase in the number of bacteria otherwise normally present on a body surface without causing disease in the person
Infantile acneAcne which presents at the age of two to six months and persists until the age of three to four years
KeratinisationThe process of keratin production in order to form an epidermal barrier in stratified squamous epithelial tissue
MicrocomedonesMicroscopic comedones, not visible to the naked eye
NoduleA deep skin-seated dome-shaped solid lump
Occupational acneAcne causes by exposure to extraneous agents or adverse conditions in a working environment. The agents and conditions that most commonly cause problems are: oils, tars or excessive humidity
PapuleSmall, solid, raised lesion, usually dome-shaped
Pilosebaceous unitThe hair follicle and sebaceous gland
Polycystic ovarian syndromeA condition caused by the imbalance of female sex hormones. It is associated with absence of ovulation resulting in irregular menstrual cycles and infertility, insulin resistance causing obesity, as well as high levels of masculinising hormones causing excessive hair growth and acne
PustuleA visible collection of pus
Reactive oxygen speciesChemically reactive molecules containing oxygen. Increased levels of reactive oxygen species may result in significant damage to cell structures, which is called an oxidative stress
ResorcinolA dihydroxy benzene compound used in many acne treatment products. It helps prevent comedones by removing buildup of dead skin cells
RetinoidsA class of chemical compound related chemically to vitamin A, topically used for acne treatment due to the way they regulate the epithelial cell growth
RosaceaA chronic dermatitis of the face, especially of the nose and cheeks, characterised by a red or rosy coloration, caused by dilation of capillaries, and the appearance of acne-like pimples
Sebaceous glandsGlands that produce sebum and deliver it to the surface of the skin. They are larger and greater in number on the face and upper parts of the trunk, which makes these the acne-prone areas
SebumAn oily substance produced by the sebaceous glands of the skin. Its main function is to protect and waterproof the hair and skin. Oily skin and acne are the result of excessive sebum production
ScarThe fibrous tissue replacing normal tissues destroyed by injury, disease or surgery
Sodium sulphacetamideA sulphonamid antibiotic used topically for fighting bacteria on the skin in the treatment of acne, dandruff and seborrhoeic dermatitis
TetracyclineA broad-spectrum antibiotic synthesised from chlortetracycline or derived from certain micro-organisms of the genus Streptomyces
Topical therapyA medication in the form of a cream, foam, gel, lotion or ointment, which is applied to body surfaces in order to treat ailments


Acne vulgaris affects nearly all adolescents and adults at some time in their lives (Webster 2002). It is estimated that up to 40 to 50 million individuals in the USA have acne, with an 85% prevalence in those aged 12 to 24 years (Bhate 2013; White 1998). Moderate to severe acne constitutes 15% to 20% of all cases (Bhate 2013; Dréno 2010; Law 2010; Wei 2010). Girls are likely to suffer from acne earlier than boys (Archer 2012), but boys appear to be more susceptible to the disease (Halvorsen 2011). Acne may decrease with age, but 64% of people aged 20 to 29 years and 43% of people aged 30 to 39 years may still have visible acne (Bhate 2013; Schäfer 2001). Globally, acne is the second most disabling skin disease after eczema (Murray 2012).


Multiple factors are involved in the development of acne. An increased level of androgens at puberty, greater sebum production and abnormal hyperproliferation of keratinocytes leads to the development of small microscopic lesions called microcomedones. In this lipid-rich and anaerobic environment Propionibacterium acnes (P. acnes), which is present in normal follicles, proliferates abnormally. Conventionally, it is believed that abnormal colonisation of P. acnes initiates the production of inflammatory and chemotactic mediators, which drives the inflammatory processes (Brown 1998; Burkhart 1999; Cunliffe 2000; Gollnick 2003). There is also evidence suggesting the involvement of inflammation at all the stages of acne development (Jeremy 2003; Tanghetti 2013), and the exact sequence of events and the interaction between these events and other possible factors (genes, diet, smoking, sunlight, etc.) remains unclear (Williams 2012).

Diagnosis and outcome measures

Clinical diagnosis of acne is usually straightforward. The condition tends to affect the face (99%), the back (60%) and the chest (15%) (Archer 2012), where the lesions are comedones (whiteheads and blackheads), which are non-inflamed lesions (Simpson 2008). Inflammatory lesions such as papules, pustules, nodules and cysts may develop after the non-inflamed lesions (Layton 2010). Papules and pustules are superficial lesions 5 mm or less in diameter, but they may evolve into deep pustules or nodules in more severe forms of the disease. In conglobate acne, suppurative nodules can extend deeply and over larger areas, forming exudative sinus tracts and tissue destruction, resulting in extensive and disfiguring scarring.

Classification of acne severity at the time of diagnosis is important because guidelines for subsequent treatment are based on the severity of disease (Nast 2012; Strauss 2007; Thiboutot 2009). Acne can be assessed and subsequently classified from two perspectives: as objective disease activity based on measurement of the visible signs of acne by an investigator, or as a patient assessment of the impact on their quality of life (Nast 2012). More than 25 acne assessment scales have been described and they are inconsistently used across different trials (Lehmann 2002). This does not allow a direct comparison of the results of separate trials (Nast 2012; Zarchi 2012). Additionally, grading is a subjective measure that may vary from one dermatologist to another (Ramli 2012). In clinical trials, assessment of the severity of acne before and after the intervention is essential to determine the therapeutic effect (Zarchi 2012). Grading and lesion counting appear to be most frequently used for this purpose (Zarchi 2012), as is described in the revised Leeds acne grading system, which includes numerical grading systems for the back and chest as well as for the face (Lehmann 2002).

Description of the intervention

Treatments for acne target the pathophysiological processes and a wide range of topical and systemic treatments are currently available (Katsambas 2004). Topical therapies, including benzoyl peroxide, tretinoin, antibiotics and salicylic acid, can be used for non-inflammatory comedones or mild to moderate inflammatory acne (Strauss 2007; Thiboutot 2009). The underlying mechanism can be action primarily against comedones (retinoids and salicylic acid) or against inflammatory lesions (antibacterials and antibiotics).

Benzoyl peroxide is an oxidising agent that is bactericidal for P. acnes. Besides its primary bactericidal effect on P. acnes, it also has mild anti-inflammatory, as well as comedolytic activity (Patel 2010; Strauss 2007). Treatment of acne vulgaris with benzoyl peroxide alone or in combination with other topical treatments (antibiotics, retinoid, salicylic acid or zinc) at concentrations of 2% to 5% is the standard of care for mild to moderate acne (Bojar 1994; Dutil 2010; Gollnick 2003; Lookingbill 1997; Strauss 2007). The most common fixed-combination products containing benzoyl peroxide are clindamycin with benzoyl peroxide, erythromycin with benzoyl peroxide and adapalene with benzoyl peroxide (Layton 2009; Taylor 2004). Besides benzoyl peroxide, other potentially efficient over-the-counter agents for acne treatment include azelaic acid, alpha-hydroxy acids, resorcinol, sulphur and zinc, but evidence of their effectiveness from randomised controlled clinical trials and studies comparing their efficacy with other topical treatments is still lacking.

There is also an increasingly wide range of non-drug-based approaches that have been developed for treating acne, among which low-concentration chemical peels with glycolic, salicylic or trichloroacetic acid are beneficial for the reduction of comedones (Kempiak 2008; Rendon 2010). In addition, comedo extractions, light electrocautery, electrofulguration and cryotherapy present other therapeutic options for comedonal acne. In addition, acne can be treated by photodynamic therapy, utilising topical 5-aminolevulinic acid together with various light sources (e.g. blue, red, intense pulsed) or lasers (e.g. pulsed dye, 635 nm red diode), as well as methyl aminolevulinate plus red light. Blue or intense pulsed light alone and lasers such as the pulsed dye, the 1320 nm neodymium:YAG and especially the 1450 nm diode may be of therapeutic benefit for inflammatory acne (Rai 2013). For deep, inflamed nodules and cysts, intralesional injections of corticosteroids, such as triamcinolone acetate, are beneficial (Levine 1983; Strauss 2007).

Commercially available over-the-counter preparations of benzoyl peroxide include gels, creams, lotions, soaps and washes, ranging from 2.5% to 10% in concentration (Strauss 2007; Zaenglein 2006). The choice of vehicle depends largely on skin type and the person's preference (Brown 1998). Irritant dermatitis (erythema, scaling, burning and itching) is the primary limitation of benzoyl peroxide for some people; this primarily occurs within the first few days of treatment but generally subsides with continued use (Gollnick 2003; Sagransky 2009). However, when in contact with hair, clothing and other fabrics benzoyl peroxide can cause bleaching (Bojar 1995; Sagransky 2009).

How the intervention might work

Benzoyl peroxide acts through three fundamental mechanisms: it is bactericidal to P. acnes, it has mild comedolytic and anti-inflammatory properties (Dutil 2010; Patel 2010; Strauss 2007), and it is lipophilic, concentrating inside the sebaceous follicles to produce benzoic acid and reactive oxygen species. By oxidising bacterial proteins, benzoyl peroxide can inhibit protein and nucleotide synthesis, and mitochondrial activity (Dutil 2010; Fakhouri 2009; Krakowski 2008). The response to benzoyl peroxide appears to be rapid; it has been shown that significant reductions in surface and follicular micro-organisms may be obtained after 48 hours treatment with 5% benzoyl peroxide in aqueous gel (Bojar 1995), and clinical improvement has been noted as early as five days after beginning treatment (James 2005). A novel finding is the report of a significant reduction in P. acnes within 20 hours of a single application of 5% benzoyl peroxide in solution, which implies that the vehicle in topical therapy is important (Ramirez 2006).

Nowadays, the development of fixed-concentration combinations of agents is the basis of topical treatment of acne. Combinations of benzoyl peroxide with topical antibiotics or retinoids exert a synergistic effect, allowing several pathogenic factors to be targeted by a single product, which makes them an ideal choice from the point of efficacy, tolerability, compliance and decreased bacterial resistance (Gamble 2012).

Resistance to P. acnes, which commonly develops during monotherapy with topical antibiotics, has not been reported with benzoyl peroxide because of its direct toxicity to P. acnes, which is due to its ability to inhibit bacterial protein and nucleotide synthesis, metabolic pathways and mitochondrial activity (Dutil 2010). This mechanism allows benzoyl peroxide to be used as a long-term therapy for acne, either as monotherapy or in combination with topical antibiotics, without the hazard of the development of bacterial resistance. However, any relationship between skin colonisation with antibiotic-resistant P. acnes and treatment outcomes remains unclear.

Besides topical antibiotics, topical retinoids (adapalene and tazarotene) are frequently used as combination therapy with benzoyl peroxide. Retinoids regulate the differentiation and proliferation of keratinocytes, and have an anti-inflammatory effect (Chivot 2005; Williams 2012). However, because benzoyl peroxide oxidises retinoids if applied simultaneously, it has been suggested that it should be used in the morning and the retinoid at night to minimise any possible interaction (Gollnick 2003; Kraft 2011). However, modern formulations allow the stable combination of topical retinoids and benzoyl peroxide (Tan 2009).

Why it is important to do this review

Effective treatment of this condition is important. Acne is a common skin disease in adolescence, which can cause psychological harm to an individual and the possibility of long-term scarring. It is also an economic burden.

More than half of people with acne may experience shame, embarrassment, anxiety, lack of confidence and impaired social contact (Bach 1993; Cunliffe 1986; Jowett 1985). Severe acne may increase anger and anxiety (Layton 2010). Acne itself induces stress, which may also exacerbate the condition (Archer 2012). Furthermore, acne episodes impose a financial burden on healthcare providers as well as the person themselves: an acne episode costs a total of USD 690 on average, ranging from USD 360 to USD 870 (Gamble 2012; Yentzer 2010). The average cost of a 30-day supply of a topical treatment depends on the drugs, but ranges from USD 21 to more than USD 100, while generic benzoyl peroxide costs USD 21 to USD 60 per 30-day course of treatment (Gamble 2012; Krakowski 2008). The annual direct cost of acne management is over USD 2.5 billion and among skin diseases it ranks second only to the cost of treating skin ulcers and wounds (Bickers 2006).

Multiple treatments for acne have been developed, among which topical benzoyl peroxide has been recommended for the first-line treatment of mild or moderate acne.

Evidence from two recently published systematic reviews assessing the use of benzoyl peroxide for treating acne was insufficient to inform clinical practice (Mohd Nor 2012; Seidler 2010). One compared the efficacy of 5% benzoyl peroxide, clindamycin with a range of concentrations between 1% and 1.2%, 5% benzoyl peroxide with salicylic acid and a combination of benzoyl peroxide with clindamycin (Seidler 2010). The other focused only on benzoyl peroxide-containing products but restricted the length of follow-up to at least six weeks (Mohd Nor 2012). Neither review gave details about the search strategies they used, nor was it clear whether their outcomes of interest were defined a priori. Further, there was no third author in either review to act as an arbiter if there was a dispute when extracting data from the studies or evaluating study quality.

A comprehensive and transparent assessment of the efficacy and safety of topical benzoyl peroxide treatment for acne is important and this is what we plan to do in our Cochrane review. We will not apply any restriction on the concentration of benzoyl peroxide or the length of follow-up. We hope to provide sufficient evidence to inform physicians when treating people with this skin condition.


To assess the effects of benzoyl peroxide for acne.


Criteria for considering studies for this review

Types of studies

We will consider randomised controlled trials (RCTs) with a parallel or cross-over design eligible for our review. We will only include the first-phase data from cross-over trials, because of the possibility of the effect of the acne interventions carrying over to subsequent periods.

Types of participants

We will consider participants with a clinical diagnosis by a healthcare practitioner of acne vulgaris on the face or trunk, regardless of age, gender, severity, setting and previous treatment. We will exclude studies where individuals were diagnosed with rosacea, chloracne, acne inversa, infantile acne, occupational acne, drug-induced acne and acne associated with polycystic ovary syndrome.

Types of interventions

We will include any treatment regimen of benzoyl peroxide (including wash-off and leave-on), dose, duration and vehicle (including gel, cream, lotion, etc.) used alone or as any fixed-combination product containing benzoyl peroxide, compared with placebo, no treatment or other active topical treatments (such as retinoids, antibiotics, azelaic acid, salicylic acid, etc.). We will consider comparisons between different formulations and concentrations of benzoyl peroxide, if available. We will allow other co-interventions if they were offered equivalently to both groups of the trial.

Types of outcome measures

We will consider short-term (less than four weeks), medium-term (four or more weeks and less than eight weeks) and long-term (eight weeks or more) follow-up periods for each outcome.

Primary outcomes
  1. Participant global self assessment of acne improvement.

  2. Withdrawal due to adverse effects.

Secondary outcomes
  1. Change in lesion counts (total, inflamed and non-inflamed, separately). We will define a clinically significant total lesion count as more than 10 lesions. According to the number of the lesions, acne severity is graded as follows: mild acne (fewer than 20 comedones, fewer than 15 inflammatory lesions or total lesion count fewer than 30); moderate acne (20 to 100 comedones, 15 to 50 inflammatory lesions or total lesion count 30 to 125); severe acne(more than five cysts, total comedone count more than 100, total inflammatory count more than 50 or total lesion count more than 125) (Lehmann 2002).

  2. The percentage of participants rated 'clear' or 'almost clear' on the Investigator's Global Assessment (IGA) scale of acne severity.

  3. Change in quality of life (assessed with a validated instrument such as Skindex-16, Skindex-29 or Cardiff Acne Disability Index).

  4. Reduction in P. acnes strains (total and resistant).

  5. Total number of adverse events.

Search methods for identification of studies

We aim to identify all relevant randomised controlled trials (RCTs) regardless of language or publication status (published, unpublished, in press or in progress).

Electronic searches

We will search the following databases for relevant trials:

  • the Cochrane Skin Group Specialised Register;

  • the Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library;

  • MEDLINE via OVID (from 1946);

  • EMBASE via OVID (from 1974); and

  • LILACS (Latin American and Caribbean Health Science Information database, from 1982).

We have devised a draft search strategy for randomised controlled trials (RCTs) for MEDLINE (OVID), which is displayed in Appendix 1. We will use this as the basis for search strategies for the other databases listed.

Trials registers

We will search the following trials registers:

Searching other resources

References from published studies

We will check the bibliographies of included primary studies and related (systematic) reviews to identify further references to relevant trials.

Unpublished literature

We will try to identify unpublished studies and grey literature via correspondence with authors and pharmaceutical companies.

Adverse effects

We will not perform a separate search for adverse effects of the target intervention. However, from the included studies we identify, we will critically examine data on adverse effects and the methods used to collect these.

Data collection and analysis

We plan to include at least one 'Summary of findings' table in our review. In this we will summarise the primary outcomes for the most important comparison. If we feel there are several major comparisons or that our findings need to be summarised for different populations we will include further 'Summary of findings' tables

Selection of studies

Two authors (YZ, ELM) will examine the titles and abstracts of each reference retrieved in the search to determine those potentially related to our review. Based on this first assessment, we will then obtain the full text of these articles to ascertain further whether they will be included in our review. Throughout this process we will select studies independently and resolve any disagreement by referring to a third author (ZY). We will record the reasons for exclusion of each study and draw a PRISMA flow-chart of study selection (Liberati 2009; Moher 2009).

Data extraction and management

Using standard data extraction templates (Appendix 2), two authors (ZY, JH) will independently extract the data from the full text of eligible studies, with any disagreements being resolved by a third author (HL). We will compile the following information from the included studies:

  1. publication details (e.g. year, country, authors);

  2. study design;

  3. setting, inclusion and exclusion criteria;

  4. population characteristics (e.g. age, severity of the acne);

  5. details of the intervention (e.g. regimen, concentration, duration, co-interventions);

  6. outcome measures and corresponding data;

  7. dropouts;

  8. length of follow-up; and

  9. type of data analyses (e.g. intention-to-treat (ITT), modified ITT).

We will use this information to populate the 'Characteristics of Included Studies' table for each included study (Appendix 2).

Assessment of risk of bias in included studies

Independently, two review authors (ZY, YCZ) will perform assessment of risk of bias using the Cochrane Collaboration tool (Higgins 2011). If necessary, we will resolve any disagreement through discussion with a third review author (YZ).

We will assign 'low', 'high' or 'unclear' risk of bias for each of the following 'Risk of bias' domains based on the judgement criteria (Higgins 2011) (Appendix 2):

  1. random sequence generation;

  2. allocation concealment;

  3. blinding of participants and personnel;

  4. blinding of outcome assessment;

  5. incomplete outcome data;

  6. selective outcome reporting;

  7. other sources of bias.

Measures of treatment effect

We will express dichotomous outcomes as risk ratios (RR), with 95% confidence intervals (CIs). We will express continuous outcomes as mean differences (MD), with 95% CIs. If continuous outcomes are measured with different methodology across studies, we will use standardised mean differences (SMD) with 95% CIs. We will treat metric scales (ordinal data outcomes) as continuous outcomes.

Unit of analysis issues

We will include cross-over trials; however, only the first phase of intervention will be included, as we expect that interventions for acne may have a lasting effect and there is minimal data on the effectiveness of wash-out periods for the treatment. In the case of within-participant studies, where different treatments have been allocated to different body parts, we will consider the body part as a unit of analysis.

Dealing with missing data

We will contact the authors of the included trials to acquire missing information (Appendix 3). We will conduct ITT analysis where data are available, in which case we will analyse all participants in the group to which they were allocated, regardless of whether or not they received the allocated intervention. When we fail to obtain missing data for ITT analysis, we will use the method of last observation carried forward for continuous data if available. For dichotomous data we will conduct sensitivity analysis based on consideration of 'best case' and 'worst case' scenarios (Gamble 2005).

Assessment of heterogeneity

We will assess statistical heterogeneity using the I² statistic. When the I² statistic is greater than 80%, which indicates substantial heterogeneity, we will only make a narrative description rather than undertake any meta-analysis. When the I² statistic is greater than 50%, we will explore the potential sources of heterogeneity through pre-specified subgroup analyses, which we have suggested in Subgroup analysis and investigation of heterogeneity.

Assessment of reporting biases

We will generate funnel plots using the log of the effect measure for the individual studies by its standard error to assess reporting bias for each primary outcome, if more than 10 studies are included in the meta-analysis. Additionally, we will perform Egger's test for the assessment publication bias (Egger 1997).

Data synthesis

We expect that clinical heterogeneity will exist across included studies, mainly because of the various tools for grading acne and various concentrations and vehicles of benzoyl peroxide, which may introduce differences in treatment effects across studies. Therefore we will use a random-effects model a priori for data synthesis of all the outcomes.

For dichotomous outcomes, we will pool risk ratios (RR) and corresponding 95% confidence intervals (CI) from individual studies. For continuous outcomes, we will pool the mean difference (MD) and corresponding 95% CIs from individual studies. In case of continuous outcomes measured with different outcome scales, we will calculate and pool standardised mean differences (SMD) with 95% CIs from individual studies.

We will only make direct comparisons between topical benzoyl peroxide (alone or in combination) and placebo, no treatment and other active topical treatments (alone or in combination); indirect comparisons are not suitable because we expect that an assumption of consistency is invalid in the current therapeutic field of acne. We will not include outcomes that are not considered in our review, but which a study has evaluated, in the analyses but will describe them in a narrative way.

Subgroup analysis and investigation of heterogeneity

We will perform exploratory subgroup meta-analyses (if each subgroup includes at least three studies) to assess whether the treatment effects are different between subgroups. If there are fewer than three studies within each subgroup, we will only conduct a narrative analysis for the treatment effects. Possible subgroups could be as follows:

  • Disease factors at baseline

    • Site (face versus trunk)

    • Severity (mild versus moderate versus severe)

  • Treatment factors

    • Concentration

    • Vehicle

    • Co-intervention (with versus without co-interventions)

Sensitivity analysis

If possible (at least three studies are included in meta-analysis), we will perform exploratory sensitivity analyses to examine the treatment effects by excluding studies of low methodological quality (more than three items rated 'high risk' or 'unclear risk' using the Cochrane Collaboration tool for assessing risk of bias).


Many thanks to Laura Prescott, Finola Delamere and Hywel Williams for providing us with useful suggestions on the title registration and the development of the protocol. Many thanks to Trials Search Co-ordinator, Liz Doney, for helping us revise and establish the search strategy. Many thanks also to other members of the Cochrane Skin Group and all the peer reviewers who contributed to our protocol.

The Cochrane Skin Group editorial base wishes to thank Robert Dellavalle who was the Cochrane Dermatology Editor for this protocol; Matthew Grainge and Esther van Zuuren who were the Statistical and Methods Editors, respectively; the clinical referee, Anne Eady; and the consumer referee, who wishes to remain anonymous.


Appendix 1. MEDLINE (OVID) search strategy

1. exp Acne Vulgaris/
3. 1 or 2
4. exp Benzoyl Peroxide/
5. benzoyl peroxide$.ti,ab.
6. peroxide dibenzoyl.ti,ab.
7. dibenzoyl peroxide.ti,ab.
8. benzoyl superoxide.ti,ab.
9. panoxyl.ti,ab.
10. diphenylglyoxal superoxide.ti,ab.
11. benzoperoxide.ti,ab.
12. or/4-11
13. randomized controlled
14. controlled clinical
15. randomized.ab.
16. placebo.ab.
17. clinical trials as
18. randomly.ab.
19. trial.ti.
20. 13 or 14 or 15 or 16 or 17 or 18 or 19
21. exp animals/ not
22. 20 not 21
23. 3 and 12 and 22

Appendix 2. Characteristics of included studies table


Study design: parallel/cross-over design

Duration of follow-up:


Inclusion criteria:

Exclusion criteria:

Diagnostic criteria:

Treatment before study:

Sex: [male %]

Age: [mean (SD)/range years, or as reported]

Sites of acne:

Duration of acne: [mean/range years (SD), or as reported]

Severity of acne and corresponding criteria of judgement:


Topical treatment:

Combination: yes (specify)/no

Regimen: wash-off/leave-on


Vehicle: gel/cream/lotion/other (specify)





Topical treatment: no treatment/placebo/active treatments (specify)

Regimen: wash-off/leave-on


Vehicle: gel/cream/lotion/other (specify)






Time point of measurement:

Number of event (percentage) in each group:

Effect size:

Study details

Study period:



Number of study centres:

Study terminated before regular end (for benefit/because of adverse events): yes/no

Publication details

Language of publication:

Funding: commercial/non-commercial/other funding

Publication status: full article/journal supplement/conference paper/other (specify)

Stated aim for study

Quote from publication: "..."





Risk of bias table

BiasCriteria for judgementAuthors' judgementSupport for judgement
Random sequence generation (selection bias)

Low risk: any one of the following random sequence generations: a random number table, a computer random number generator, coin tossing, shuffling cards or envelopes, throwing dice, drawing of lots, or minimisation

High risk: any one of non-random sequence generations (other methods except those suggested as low risk)

Unclear risk: insufficient information to permit judgement of 'low risk' or 'high risk'


Quote from publication: "..."
Comment: ...


Allocation concealment (selection bias)

Low risk: allocation based on any one of the following methods: central allocation, sequentially numbered drug containers of identical appearance, sequentially numbered, opaque, sealed envelopes

High risk: allocation based on any other methods (other methods except those suggested as low risk)

Unclear risk: insufficient information to permit judgement of 'low risk' or 'high risk'


Quote from publication: "..."
Comment: ...


Blinding of participants and personnel (performance bias)

Low risk: blinding of participants and key study personnel, and unlikely that the blinding could be broken

High risk: no blinding or incomplete blinding, or blinding of participants and key study personnel could be broken

Unclear risk: insufficient information to permit judgement of 'low risk' or 'high risk'


Quote from publication: "..."
Comment: ...


Blinding of outcome assessment (detection bias)

Low risk: blinding of outcome assessment and unlikely that the blinding could be broken

High risk: no blinding or incomplete blinding, or blinding of outcome assessment could be broken

Unclear risk: insufficient information to permit judgement of 'low risk' or 'high risk'


Quote from publication: "..."
Comment: ...


Incomplete outcome data (attrition bias)

Low risk: any one of the following situations: no missing outcome data, missing outcome data (proportion of loss of follow-up ≤ 10%) balanced in numbers across intervention groups with similar reasons, or missing data have been imputed using data complement or ITT

High risk: any other situation except those suggested above

Unclear risk: insufficient information to permit judgement of 'low risk' or 'high risk'


Quote from publication: "..."
Comment: ...


Selective reporting (reporting bias)

Low risk: the study protocol is available and all of the study's pre-specified (primary and secondary) outcomes that are of interest in the protocol have been reported in the pre-specified way

High risk: any other situation except those suggested above

Unclear risk: insufficient information to permit judgement of 'low risk' or 'high risk'


Comment: ...


Other sources of bias

We will assess the baseline characteristics (e.g. age, gender, duration, location, severity, previous treatment) imbalance between groups

Low risk: no statistical significance (P ≥ 0.05) of baseline imbalance for each characteristic

High risk: statistical significance (P < 0.05) of baseline imbalance for each characteristic

Unclear risk: insufficient information to permit judgement of 'low risk' or 'high risk'

 Comment: ...

Appendix 3. Survey of authors providing information on trials


Study ID

Study author contactedStudy author repliedStudy author asked for
additional information
Study author provided data
Study 1    
Study 2    
Study 3    


n: no; y: yes

What's new

17 March 2015AmendedAuthor information (affiliation) updated

Contributions of authors

ZY was the contact person with the editorial base.
ZY co-ordinated the contributions from the co-authors and wrote the final draft of the protocol.
ZY, YZ, ELM and JH worked on the methods sections.
ZY, HL, ELM and JL drafted the clinical sections of the background and responded to the clinical comments of the referees.
ZY, YZ and YCZ responded to the methodological and statistical comments of the referees.
All the authors contributed to writing the protocol.
QZ was the consumer co-author and checked the protocol for readability and clarity. She also ensured that the outcomes are relevant to consumers.
ZY is the guarantor of the final review.


The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the NIHR, NHS or the Department of Health, UK.

Declarations of interest

Zhirong Yang: Nothing to declare.
Yuan Zhang: Nothing to declare.
Elvira Lazic Mosler: Nothing to declare.
Hang Li: Nothing to declare.
Jing Hu: Nothing to declare.
Yanchang Zhang: Nothing to declare.
Jia Liu: Nothing to declare.
Qian Zhang: Nothing to declare.

Sources of support

Internal sources

  • Peking University Health Science Center, China.

  • Shantou-Oxford Clinical Research Unit, China.

External sources

  • Cochrane Skin Group, UK.

  • The National Institute for Health Research (NIHR), UK.

    The NIHR, UK, is the largest single funder of the Cochrane Skin Group.