Interventions for actinic keratoses

Aditya K Gupta; Maryse Paquet; Elmer Villanueva; William Brintnell

doi:10.1002/14651858.CD004415.pub2

. 2012 Dec 12;2012(12):CD004415. doi: 10.1002/14651858.CD004415.pub2

Interventions for actinic keratoses

Aditya K Gupta ^1,^2,^✉, Maryse Paquet ², Elmer Villanueva ³, William Brintnell ²

Editor: Cochrane Skin Group

PMCID: PMC6599879 PMID: 23235610

Abstract

Background

Actinic keratoses are a skin disease caused by long‐term sun exposure, and their lesions have the potential to develop into squamous cell carcinoma. Treatments for actinic keratoses are sought for cosmetic reasons, for the relief of associated symptoms, or for the prevention of skin cancer development. Detectable lesions are often associated with alteration of the surrounding skin (field) where subclinical lesions might be present. The interventions available for the treatment of actinic keratoses include individual lesion‐based (e.g. cryotherapy) or field‐directed (e.g. topical) treatments. These might vary in terms of efficacy, safety, and cosmetic outcomes.

Objectives

To assess the effects of topical, oral, mechanical, and chemical interventions for actinic keratosis.

Search methods

We searched the following databases up to March 2011: the Cochrane Skin Group Specialised Register, CENTRAL in The Cochrane Library, MEDLINE (from 2005), EMBASE (from 2010), and LILACS (from 1982). We also searched trials registers, conference proceedings, and grey literature sources.

Selection criteria

Randomised controlled trials (RCTs) comparing the treatment of actinic keratoses with either placebo, vehicle, or another active therapy.

Data collection and analysis

At least two authors independently abstracted data, which included adverse events, and assessed the quality of evidence. We performed meta‐analysis to calculate a weighted treatment effect across trials, and we expressed the results as risk ratios (RR) and 95% confidence intervals (CI) for dichotomous outcomes (e.g. participant complete clearance rates), and mean difference (MD) and 95% CI for continuous outcomes (e.g. mean reduction in lesion counts).

Main results

We included 83 RCTs in this review, with a total of 10,036 participants. The RCTs covered 18 topical treatments, 1 oral treatment, 2 mechanical interventions, and 3 chemical interventions, including photodynamic therapy (PDT). Most of the studies lacked descriptions of some methodological details, such as the generation of the randomisation sequence or allocation concealment, and half of the studies had a high risk of reporting bias. Study comparison was difficult because of the multiple parameters used to report efficacy and safety outcomes, as well as statistical limitations. We found no data on the possible reduction of squamous cell carcinoma.

The primary outcome 'participant complete clearance' significantly favoured four field‐directed treatments compared to vehicle or placebo: 3% diclofenac in 2.5% hyaluronic acid (RR 2.46, 95% CI 1.66 to 3.66; 3 studies with 420 participants), 0.5% 5‐fluorouracil (RR 8.86, 95% CI: 3.67 to 21.44; 3 studies with 522 participants), 5% imiquimod (RR 7.70, 95% CI 4.63 to 12.79; 9 studies with1871 participants), and 0.025% to 0.05% ingenol mebutate (RR 4.50, 95% CI 2.61 to 7.74; 2 studies with 456 participants).

It also significantly favoured the treatment of individual lesions with photodynamic therapy (PDT) compared to placebo‐PDT with the following photosensitisers: aminolevulinic acid (ALA) (blue light: RR 6.22, 95% CI 2.88 to 13.43; 1 study with 243 participants, aminolevulinic acid (ALA) (red light: RR 5.94, 95% CI 3.35 to 10.54; 3 studies with 422 participants), and methyl aminolevulinate (MAL) (red light: RR 4.46, 95% CI 3.17 to 6.28; 5 studies with 482 participants). ALA‐PDT was also significantly favoured compared to cryotherapy (RR 1.31, 95% CI 1.05 to 1.64).

The corresponding comparative risks in terms of number of participants completely cleared per 1000 were as follows: 313 with 3% diclofenac compared to 127 with 2.5% hyaluronic acid; 136 with 0.5% 5‐fluorouracil compared to 15 with placebo; 371 with 5% imiquimod compared to 48 with placebo; 331 with ingenol mebutate compared to 73 with vehicle; 527 to 656 with ALA/MAL‐PDT treatment compared to 89 to 147 for placebo‐PDT; and 580 with ALA‐PDT compared to 443 with cryotherapy.

5% 5‐fluorouracil efficacy was not compared to placebo, but it was comparable to 5% imiquimod (RR 1.85, 95% Cl 0.41 to 8.33).

A significant number of participants withdrew because of adverse events with 144 participants affected out of 1000 taking 3% diclofenac in 2.5% hyaluronic acid, compared to 40 participants affected out of 1000 taking 2.5% hyaluronic acid alone, and 56 participants affected out of 1000 taking 5% imiquimod compared to 21 participants affected out of 1000 taking placebo.

Based on investigator and participant evaluation, imiquimod treatment and photodynamic therapy resulted in better cosmetic outcomes than cryotherapy and 5‐fluorouracil.

Authors' conclusions

For individual lesions, photodynamic therapy appears more effective and has a better cosmetic outcome than cryotherapy. For field‐directed treatments, diclofenac, 5‐fluorouracil, imiquimod, and ingenol mebutate had similar efficacy, but their associated adverse events and cosmetic outcomes are different. More direct comparisons between these treatments are needed to determine the best therapeutic approach.

Plain language summary

Interventions for actinic keratoses

Actinic keratoses are a skin disease caused by long‐term sun exposure. Damaged skin shows small, red, rough, scaly, flat spots called actinic keratoses or lesions, which feel like patches of dry skin. Symptoms such as bleeding and pain can be associated with actinic keratoses. Moreover, actinic keratoses have the potential to develop into skin cancer if left untreated. The reasons for treatment may include cosmetic appearance, relief of symptoms, or prevention of skin cancer. Treatment can be directed either at individual lesions or to larger areas of the skin where several visible and less visible lesions occur (field‐directed treatment).

This systematic review included results from 83 randomised controlled clinical trials evaluating 24 treatments, with a total of 10,036 participants diagnosed with actinic keratosis. We included 18 topical creams or gels applied to a skin area by the participants: adapalene gel, aretinoid methyl sulfone (Ro 14‐9706), betulin‐based oleogel, calcipotriol (vitamin D), colchicine, diclofenac, 2‐(difluoromethyl)‐dl‐ornithine (DFMO), 5‐fluorouracil, ß‐1,3‐D‐glucan, imiquimod, ingenol mebutate (PEP005), isotretinoin, masoprocol, nicotinamide, resiquimod, sunscreen, DL‐α‐tocopherol (vitamin E), and tretinoin. One treatment, etretinate, was taken orally. Clinical staff administered two mechanical treatments (carbon dioxide and Er:YAG laser resurfacing) on a skin area, and they administered three chemical treatments: cryotherapy on individual lesions, photodynamic therapy on individual lesions or a skin area, and trichloroacetic acid peel on a skin area.

The clinical effects resulting from the treatment of actinic keratoses were reported differently from one study to another. In spite of this inconsistency, it can be concluded that several good treatment options exist for the treatment of actinic keratoses. Actinic keratoses were successfully treated with cryotherapy, diclofenac, 5‐fluorouracil, imiquimod, ingenol mebutate, photodynamic therapy, resurfacing, and trichloroacetic acid peel. These different treatments were generally comparably effective. Skin irritation was associated with some of these treatments, such as diclofenac and 5‐fluorouracil, but other side‐effects were uncommon. The final cosmetic appearance varies from one treatment to another. Imiquimod treatment and photodynamic therapy resulted in better cosmetic appearance than treatment with cryotherapy and 5‐fluorouracil.

Treatment with photodynamic therapy gives better therapeutic and cosmetic results than cryotherapy for individual lesions. For field‐directed treatments, diclofenac, 5‐fluorouracil, imiquimod, and ingenol mebutate are good options associated with different side‐effects and cosmetic results. Thus, the choice of treatment option for actinic keratosis depends on the number of lesions, the individual's desired results, and tolerance to the treatments.

Background

Description of the condition

Disease definition

Actinic keratoses are scaly lesions on the skin resulting from abnormal growth of atypical epidermal keratinocytes. They are localised at the surface of the skin on the sun‐exposed parts of the face or hands, particularly among older fair‐skinned individuals. Actinic keratoses are markers for increased rate of non‐melanoma skin cancer (Ramsay 2003) and shows the morphological and histological features of squamous cell carcinoma (Cockerell 2000; Feldman 2011). An actinic keratosis could be considered a precancerous lesion or carcinoma in situ based on the fact that the majority of invasive squamous cell carcinomas arise from actinic keratoses. Actinic keratoses are confined to the epidermis, whereas squamous cell carcinoma extends more deeply into the dermis. Thus, to limit the morbidity and mortality associated with squamous cell carcinoma, treatment of actinic keratoses is strongly recommended.

Actinic keratosis is also known as solar keratosis, senile keratosis, senile hyperkeratosis, keratoma senile, keratosis senilis, and actinic cheilitis (actinic keratosis on the lip) (Marks 1993; Rigel 2008; Schwartz 1997).

Clinical Features

The conventional clinical actinic keratosis lesion is a pink, red, or brown scaly patch on the skin, less than one centimetre in diameter (Roewert‐Huber 2007). Often, the scaliness of a lesion can be felt before it can be seen; this may progress into thickened or hypertrophic (increased bulk, due to an increase in lesion size) lesions. Actinic keratoses can be clinically graded with grade 1, slightly palpable; grade 2, moderately thick and visible; and grade 3, very thick and hyperkeratotic (Cockerell 2000; Olsen 1991). Accurate clinical diagnosis requires careful observation under adequate lighting conditions and palpation of the lesion texture (Marks 1993). Actinic keratoses are diagnosed histologically with a skin biopsy (Cockerell 2000; Marks 1993). Detectable actinic keratosis lesions are often associated with field change where the surrounding skin is also altered, and subclinical lesions may be present (Vatve 2007).

There are different classifications based on the clinical appearance of actinic keratoses: atrophic, hyperkeratotic, bowenoid, acantholytic, lichenoid, and pigmented (Rigel 2008; Roewert‐Huber 2007). Atrophic actinic keratoses are dry, scaly‐appearing lesions on a reddened base (due to dilated blood capillaries) without distinct margins. Hyperkeratotic actinic keratoses are papules and plaques with scale or scale‐crust that also possibly have cutaneous horns or conical masses. Bowenoid actinic keratoses are scaling red plaques with sharply‐established borders that simulate Bowen's Disease (a solitary red plaque with distinct borders) in that the abnormal cells are found throughout the depth of the epidermis. Acantholytic actinic keratoses have focal acantholysis (separation from other cells) occasionally accompanied by clefts. Lichenoid actinic keratoses show dense band‐like infiltration of lymphocytes in the papillary dermis and vacuolar alteration at the dermoepidermal junction. Pigmented actinic keratoses have a hyperpigmented or reticulated appearance. Differential diagnosis of actinic keratosis includes Bowen's disease, squamous cell carcinoma, keratoacanthoma, basal cell carcinoma, seborrhoeic keratosis, and lentigo maligna (Holmes 2007).

Symptoms of actinic keratosis include tenderness, itchiness, burning, and a sandpaper‐like texture. Over time, lesions may remain unchanged, proliferate, regress, reappear, or develop into squamous cell carcinoma. Microscopically, actinic keratosis lesions show abnormal tissue development (dysplasia) in the skin cells (keratinocytes). During early development of a lesion, the lower layers of the epidermis show the most dysplastic keratinocytes. As a lesion develops, the dysplastic cells permeate the epidermis and form conical‐shaped scales when the surface of the epidermis is reached. Acceleration of growth of the epidermal layer and abnormal cellular maturation leads to excessive production of immature adherent scales with a sandpaper or gritty feel (Marks 1993). The lower skin layer (dermis) undergoes patchy inflammation as seen by an increased number of white blood cells (lymphocytes) noted in the dermis (Marks 1993).

Pathogenesis and epidemiology

The anatomical distribution of actinic keratosis lesions correlates with areas of the body that receive the most long‐term, chronic, and intense exposure to ultraviolet radiation in sunlight (Marks 1993; Schwartz 1997). More than 80% of the lesions occur on the head, neck, back of the hands, and forearms (Salasche 2000). Chronic exposure to ultraviolet (UV) radiation, mainly UVB (290 to 320 nm), is the major agent leading to mutagenesis (disordered regulation of growth) in keratinocytes (Callen 1997). In fact, mutations in the p53 tumour suppressor gene have been found in 53% of those with actinic keratoses and 69% of squamous cell carcinoma biopsies (Nelson 1994). Ultraviolet radiation can also contribute to suppression of the immune system, resulting in a decreased ability to eliminate over‐proliferating cells (Holmes 2007). Moreover, UV light could directly activate human papillomavirus replication. The virus, in turn, degrades a proapoptotic protein BAk, also preventing elimination of tumour cells (Holmes 2007). Thus, sunlight initiates and promotes the formation of non‐melanoma skin cancer.

The cause of actinic keratosis involves an interaction between skin colour (melanin protects by absorbing UVB radiation); advancing age (cumulative sun exposure and decrease in the effectiveness of the immune system); gender (actinic keratosis is more prevalent in men); history of severe sunburn in childhood; and sun exposure, which is influenced by latitude and the integrity of the ozone layer (Holmes 2007; Lebwohl 2003; Salasche 2000). Other factors may include occupation (working outdoors), socioeconomic status, and diet (Lebwohl 2003; Marks 1993; Salasche 2000; Schwartz 1997). Immunosuppressive therapy, e.g. in organ transplant recipients, and exhibition of genetic diseases of skin hypopigmentation (low pigmentation), such as xeroderma pigmentosum or albinism (Holmes 2007; Moy 2000), are also risk factors.

The first National Health and Nutrition Examination Survey (NHANES I) found that in healthy white people in the US, the age‐adjusted prevalence rate for actinic keratoses was 6.5%. This increases significantly with advancing age: In 65‐ to 74 year‐old men with high sun exposure, the prevalence rate was 55.4% and 18.5% for low sun exposure (Engel 1988). In Australia, where prevalence of actinic keratosis is the highest, as many as 40% of white adults may have an actinic keratosis. For younger adults, aged 30 to 39 years, the rate was 22% for men and 8% for women. In older adults aged 60 to 69 years, 83% of men and 64% of women have an actinic keratosis. For this population of adults, 42% developed at least 1 new lesion within the year (Frost 2000). Although known to be precancerous, the probability of a lesion undergoing malignant transformation to a squamous cell carcinoma is not clear, but ranges from 0.025% to 16% per year (Glogau 2000; Jeffes 2000).

Description of the intervention

An actinic keratosis may potentially become cancerous; therefore, monitoring is advised. Because of the prevalence of actinic keratoses among an ageing population, treatment has been sought by an increasing number of people (Warino 2006). Reasons for treatment include prevention of cancer development; relief of symptoms, such as bleeding; and improvement of cosmetic appearance. Interventions for actinic keratoses could be divided into individual treatment of lesion and field‐directed treatment, i.e. applied to an area of sun‐damaged skin where there may be multiple lesions. Individual lesion treatment (spot) might relieve symptoms or cosmetic concerns, whereas field‐directed treatment might be more appropriate for prevention of transformation into squamous cell carcinoma. Most of the field‐directed treatments are topical treatments where efficacy depends on patient compliance.

Behaviour modifications, including limiting sun exposure between 10am and 4pm, the use of sunscreens with a SPF (sun protection factor) rating of at least 15, and the use of protective clothing, are the best methods for the prevention of actinic keratosis and will help reduce the need for treatment (Schwartz 1997; Wilkerson 1984).

Various strategies for the treatment of actinic keratoses have been developed; these include physician‐administered cryotherapy for a few lesions, and topical 5‐fluorouracil, topical imiquimod, topical masoprocol, topical diclofenac in 2.5% hyaluronic acid gel, and photodynamic therapy for large numbers of lesions. Salicylic acid may also be used for early lesions, while dermabrasion and laser resurfacing are beneficial when there is coexistent photodamage or multiple recalcitrant lesions. Excision (removal of the lesion, often using a scalpel blade) and chemical peels (use of a caustic agent that causes the lesion to slough off) are both appropriate for hyperkeratotic or recalcitrant lesions. Interferon and oral retinoids are uncommon treatments, and they are still under development. These treatments have varying efficacies and adverse effect profiles (Dinehart 2000; Ibrahim 2009; Marks 1993; Wilkerson 1984).

Thus, the factors to consider when making decisions about treatment include efficacy, tolerability, number of lesions to treat, spot or field‐directed treatment, compliance, history of skin cancer, immunosuppression, previous treatment history, and cosmetic appearance.

How the intervention might work

Topical Interventions

Diclofenac gel

One topical treatment for actinic keratoses is the non‐steroidal anti‐inflammatory drug (NSAID) diclofenac in 2.5% hyaluronic acid gel. The hyaluronic acid vehicle contributes to the success of this treatment by delivering and then retaining diclofenac at the epidermis, protecting against UV radiation and its cosmetic properties (Brown 2005). Although the precise mechanisms of action are not clear, diclofenac is thought to target several aspects of actinic keratosis pathophysiology. One mechanism that has been proposed is the inhibition of cyclooxygenase 2 (COX‐2) (Hemmi 2002), which leads to a reduction in prostaglandin synthesis (Rivers 2004). This COX‐2 inhibition or other mechanisms may be responsible for diclofenac’s inhibition of cell differentiation in vitro, induction of apoptosis in vitro and in vivo, alteration of cell proliferation, and inhibition of angiogenesis (Adamson 2002; Alam 1995; Lu 1995; Seed 1997). Diclofenac has also been shown to activate the nuclear hormone receptors, peroxisome proliferator‐activated receptors (PPARs), in vitro; these receptors are involved in many cellular functions including cell differentiation and apoptosis (Adamson 2002).

5‐Fluorouracil (5‐FU)

This topical agent causes a decrease in cell proliferation and an induction of cell death, particularly in cells with high mitotic (cell division) rates. This occurs through the inhibition of thymidylate synthetase, which blocks the methylation reaction of deoxyuridylic acid to thymidylic acid, thereby, interfering with DNA and RNA synthesis (Berman 2006; Chakrabarty 2004; Eaglstein 1970; Robins 2002b).

Imiquimod

This topical treatment for actinic keratoses is a synthetic compound belonging to the imidazoquinolone family of drugs (Hemmi 2002). It acts as an immune modulator by activating toll‐like receptors, ultimately resulting in the modulation of the mRNA expression of many immunomodulatory genes, which induces the production of cytokines by monocytes, macrophages, and epidermal keratinocytes (Correale 2002; Stanley 1999). This has the effect of enhancing innate and acquired immune responses, which leads to strong antiviral and antitumoural activity (Vidal 2006). Imiquimod also induces pro‐apoptotic pathways through a variety of mechanisms (Amini 2010).

Chemical Interventions

Cryotherapy

Cryotherapy is often the treatment of choice for individual actinic keratosis lesions (Goldberg 2010). It uses liquid nitrogen to freeze and destroy the epidermis containing actinic keratoses (Goldberg 2010), with efficacy increasing as a function of freezing duration (Thai 2004).

Photodynamic therapy (PDT)

Photodynamic therapy involves the selective accumulation of a photosensitising agent in premalignant or malignant cells (Gold 2008; Juarranz 2008). This is achieved by the application of 5‐aminolevulinic acid (5‐ALA) or MAL (ALA methyl ester), which are precursors to protoporphyrin IX (PpIX), a potent photosensitiser (Fink‐Puches 1997). This causes an excess of PpIX, which selectively accumulates in neoplastic cells. Subsequently, the photosensitiser is activated by visible light, causing the generation of reactive oxygen species in the presence of oxygen. These reactive oxygen species [mainly singlet oxygen (ˈO₂)] start a cascade of biochemical events that induce damage and the death of neoplastic cells through an apoptotic mechanism (Juzeniene 2007; Moan 1991).

Why it is important to do this review

The existing evidence for use of the various treatment agents for actinic keratoses is varied, and there are concerns regarding adverse events and cosmetic outcomes. It is vital to critically assess data in terms of the benefits as well as the risks associated with treatment.

Objectives

To assess the effects of interventions for actinic keratoses.

Methods

Criteria for considering studies for this review

Types of studies

This review included randomised controlled trials comparing the treatment of actinic keratoses to either placebo, vehicle, other current therapies, or variation in treatment conditions (e.g. different concentrations of the active ingredient or types of light sources for phototherapy). We included cross‐over trials and parallel and intraindividual (e.g. left‐ or right‐side comparison) studies.

Types of participants

We included participants with clinical signs of actinic keratoses as assessed by a medical practitioner or histological diagnosis. Diagnostic criteria, such as the Marks definition (Marks 1993) or the Salasche or Schwartz characterisation (Salasche 2000; Schwartz 1997), were acceptable, as was the diagnosis of actinic keratoses by a dermatologist using the terms 'actinic keratosis', 'solar keratosis', 'senile keratosis', 'senile hyperkeratosis', 'keratoma senile', or 'keratosis senilis'. We included studies with immunocompetent and immunosuppressed participants.

Types of interventions

We considered the following interventions:

prescription‐based topical treatments, e.g. diclofenac in hyaluronic gel, 5‐fluorouracil, or imiquimod;
prescription‐based oral drugs, e.g. oral retinoids;
mechanical interventions, e.g. curettage, dermabrasion, or resurfacing;
chemical interventions, e.g. chemical peels, cryotherapy, or photodynamic therapy; and
combinations of topical and oral treatments with mechanical or chemical interventions.

The comparators were vehicle, placebo, another active compound or intervention, or a variation of the treatment (duration, concentration, etc).

Types of outcome measures

For actinic keratoses, the outcomes can be expressed per lesion or per participant. Because the participants or body parts of the participants (intraindividual design), not the lesions, were generally randomised, only per‐participant outcomes could be included in meta‐analyses. Thus, the included outcomes in this review were outcomes reported per participant.

Efficacy outcomes for studies on actinic keratoses are generally based on the clearance of individual lesions. Lesions present at baseline are generally identified, graded (grade I: slightly palpable, better felt than seen; grade II: moderately thick, easily seen and felt; and grade 3: very thick, hyperkeratotic, or both), and mapped. Use of transparencies and photography might help with this process. Sometimes distinction is made between lesions present at the baseline and new lesions appearing during the study. At the end of the study, the assessors evaluate the clearance, or not, of the lesions.

Ideally, complete clearance of actinic keratosis lesions at follow‐up would be measured (i.e. number of participants with 100% clearance of target (present at baseline) or all actinic keratosis lesions).

A second outcome measurement, such as partial clearance, is also often used. The definition of partial clearance is subjective but frequently indicates the number of participants with 75% or more of actinic keratosis lesions being completely cleared, i.e. a reduction in the number of lesions by at least 75%.

Alternatively, the mean reduction of total number of lesions at baseline per participant is also used, i.e. the difference between the mean number of lesions at baseline and the mean number of lesions at assessment. The results are then presented as absolute mean or mean percentage of reduction in lesion counts compared to baseline.

We only included outcomes expressed as number of participants experiencing adverse events in this review.

Cosmetic outcomes are really varied from global assessment to individual characteristics, such as changes in pigmentation. We only included outcomes expressed as number of participants or mean per participant in this review.

Primary outcomes

Efficacy outcomes

Subjective assessment: global degree of improvement in symptoms or signs as rated by a medical practitioner or participant, or global improvement indices (GII) for completely improved or cleared.
Objective assessment: participant complete (100%) or partial (> 75%) clearance.
Objective assessment: mean reduction in lesion counts (absolute number or percentage).

Secondary outcomes

Safety and cosmetic outcomes

Withdrawal due to adverse events.
Skin irritation.
Minor adverse events excluding skin irritation.
Cosmetic outcomes: cosmetic changes, including pigmentation and scarring.

Search methods for identification of studies

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

Electronic searches

We searched the following databases up to 23 March 2011:

the Cochrane Skin Group Specialised Register using the terms: ((actinic or solar or senile) and keratos*) or hyperkeratos*;
the Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library using the search strategy in Appendix 1:
PUBMED/MEDLINE via OVID (from 2005) using the strategy in Appendix 2;
EMBASE via OVID (from 2010) using the strategy in Appendix 3; and
LILACS (Latin American and Caribbean Health Science Information database, from 1982) using the strategy in Appendix 4.

The UK and US Cochrane Centres have an ongoing project to systematically search MEDLINE and EMBASE for reports of trials, which are then included in the Cochrane Central Register of Controlled Trials. Searching has currently been completed in MEDLINE from inception to 2004 and in EMBASE from inception to 2009. Further searches of these two databases were undertaken for this review by the Cochrane Skin Group to cover the years not searched by the UK and US Cochrane Centres for CENTRAL.

A final prepublication search for this review was undertaken on 4 April 2012. Although it has not been possible to incorporate RCTs identified through this search within this review, relevant references are listed under 'Studies awaiting classification'. They will be incorporated into the next update of the review.

Trials Registers

We searched the following trials registers on 10 March 2011 using the search terms ((actinic, senile, or solar) and keratos) or hyperkeratos.

The metaRegister of Controlled Trials (www.controlled‐trials.com).
The US National Institutes of Health Ongoing Trials Register (www.clinicaltrials.gov).
The Australian New Zealand Clinical Trials Registry (www.anzctr.org.au).
The World Health Organisation International Clinical Trials Registry platform (www.who.int/trialsearch).
The Ongoing Skin Trials Register (www.nottingham.ac.uk/ongoingskintrials).

Searching other resources

Unpublished literature

We conducted online searches (via pharmaceutical company websites, the U.S. Food and Drug Administration (FDA) website, or both) for the following products and drug companies:

3M/Graceway Pharmaceuticals (imiquimod, Aldara, or Zyclara);
Actavis Mid‐Atlantic LLC (imiquimod);
Allergan (5‐fluorouracil, Fluoroplex);
Apotex (imiquimod);
Dermik/Sanofi Aventis (5‐fluorouracil, Carac);
DUSA Pharmaceuticals (aminolevulinic acid, Levulan Kerastick);
Galderma (adapelene, Differin);
ICN (5‐fluorouracil, Efudex);
Leo Pharmaceuticals (calcipotriol, Dovonex, or Daivonex);
Mochida Pharmaceuticals (imiquimod, Beselna);
PharmaDerm/NycoMed US (diclofenac, Solaraze);
Pharmacia & Upjohn (5‐fluorouracil);
Photocure ASA/Galderma (methyl aminolevulinate, Metvix, or Metvixia);
Roche (etretinate, Tegison);
Stiefel/GlaxoSmithKline (isotretinoin, Isotrex, or Isotrexin); and
URL Pharma (colchicine, Colcrys).

Conference proceedings

We scanned the conference proceedings of the British Association of Dermatologists and the European Academy of Dermatology from 2007 to 2011 for further references to relevant trials. We examined the conference proceedings for 2009 and 2010 of the Annual Meeting of the American Academy of Dermatology, the Annual Meeting of the European Society for Dermatological Research, the Congress of the European Association of Dermatol‐Oncology, the Annual Meeting of the British Association of Dermatologists, and the Annual Meeting of the Australasian College of Dermatologists. We scanned the conference proceedings for the 2012 Annual meeting of the American Academy of Dermatology.

Language restrictions

We imposed no language restrictions when we searched for publications. We electronically translated articles published in languages other than English.

Adverse effects

We did not perform a separate search for adverse effects of interventions for actinic keratoses. We looked at reports of adverse events or side‐effects in the RCTs identified as a result of our searches, as part of our secondary outcomes.

Data collection and analysis

Selection of studies

At least two authors (WB and MP) independently checked titles and abstracts identified from the searches. We obtained the full text of all studies of possible relevance for independent assessment by two authors (MP and WB). The authors decided which trials fit the inclusion criteria and recorded their methodological quality (MP and WB). They resolved any disagreement by discussion between the authors and a third party arbitrator (AG). Previous contributors also participated in this process in earlier versions of the review.

Data extraction and management

At least two authors extracted and summarised, using data collection forms, the details of eligible trials. One author (MP) double‐checked and entered data. The authors were not blinded to the names of the trial authors, journals, or institutions.

Assessment of risk of bias in included studies

Assessment of risk of bias included the Review Manager 5.1 'Risk of bias' assessment tool shown in the 'Risk of bias' tables. In addition, GradePro "quality of evidence" was also used for selected outcomes, and the results are shown in the overview tables for five selected interventions.

Measures of treatment effect

We performed a meta‐analysis for each treatment comparison to calculate a weighted treatment effect across trials. We expressed the results as a risk ratio (RR) with 95% confidence intervals (CI) for dichotomous outcomes, and a mean difference (MD) with 95% CI for continuous outcomes. We calculated the number needed to treat (NNT) for significantly different dichotomous outcomes using the following formula: NNT = Ι 1/ ACR * (1‐RR)Ι where the risk ratios (RR) from the meta‐analysis and the moderate assumed control risk (ACR) calculated in GRADEpro was used. For ACR, a mean baseline risk from the study was used for analysis with only one study; and low, median, or high control‐group risk were used based on the variation in the included studies in meta‐analysis. This previous method would not be applicable to outcomes with an ACR of 0%, i.e. no event in the control group, because of the numerical problems that would ensue.

Unit of analysis issues

The unit of analysis was the participant. We analysed cross‐over trials using data from the first phase only and pooled, where possible, with parallel‐design studies. We divided results from withinparticipant trials (intraindividual, e.g. split face) into 2 categories: 1) outcomes expressed as number of participants (e.g. participant complete clearance), which could not be included in meta‐analyses, were only reported in the text; and 2) outcomes expressed as mean with standard deviation (e.g. mean reduction in lesion counts = mean of reductions observed in each participant), which could be included in meta‐analyses using the inverse‐variance method. We combined together data from studies with multiple treatments when appropriate (e.g. "all treatment groups" versus "placebo"), or we split the data from the shared group. If studies were using more than one outcome included in this review, we included all outcomes in the analyses.

Assessment of heterogeneity

We assessed heterogeneity using an I² statistic value expressed as a percentage. We excluded results from meta‐analyses with an I² statistic value of 80% or higher. We explored reasons for heterogeneity in studies, and if necessary, sensitivity analyses examined the effects of excluding a study, e.g. those studies with lower methodological quality.

Many studies do not distinguish between the physical location of actinic keratosis lesions on the body. This can introduce heterogeneity, as actinic keratoses of the face and scalp are often more effectively treated by certain topical formulations than lesions located elsewhere. In some studies, pretreatment of lesions to remove hyperkeratosis essentially negated the differences encountered by lesion location, as lower response has been associated with greater hyperkeratosis. Most of the studies included limited their investigation to grade one or two lesions, i.e. minimally to moderate thick lesions. However, when comparing efficacy results from two separate studies using the same treatment, studies incorporating pretreatment of any kind may have accounted for different efficacy rates.

Data synthesis

A random‐effects model was prespecified for all meta‐analyses. The Mantel‐Haenszel method was used for dichotomous outcomes (e.g. cure rates), and an inverse variance model was used for continuous outcomes (e.g. mean reduction in lesion counts).

Subgroup analysis and investigation of heterogeneity

Where appropriate, we undertook subgroup analysis (subgroups of participants) in an attempt to decrease heterogeneity between studies (for example, when different dosing regimens were used or to keep information separated, i.e. when blue or red light was used for photodynamic therapy). In addition, if data were presented for several assessment time points or anatomical locations, we performed subgroup analyses.

Results

Description of studies

See the 'Characteristics of included studies', 'Characteristics of excluded studies', and 'Characteristics of ongoing studies'.

Results of the search

We identified 1001 references from searching bibliographic and trials databases, as well as 28 references through other sources. After removing duplicate references and ongoing studies without results, we had 469 records to screen. We excluded 318 records based on titles and abstracts as they did not meet our eligibility criteria (non‐randomised studies, reviews, not interventions to cure). We assessed the full texts of the remaining 151 records. We then excluded a further 55 records, leaving 96 studies. We included 83 of these in our qualitative analysis; 12 are listed under studies awaiting classification, and 1 is an ongoing study. We included 75 studies in our meta‐analysis.

The PRISMA study flow chart in Figure 1 summarises the results of the search for studies.

(Please note that in all tables in the results section, the "X" means that the associated outcome was reported, and when there was no participant withdrawal it is specified between parentheses.)

Included studies

We included 83 randomised studies in the review, encompassing 10,036 participants in total.

Design

We only included the randomised (participants or right/left side in intraindividual studies) clinical trials if the interventions were covered by this review and if they reported numerical results for at least one of the review outcomes. This criterion excluded the outcome 'withdrawal due to adverse events', which is generally reported in all studies.

Some studies had more than one design. The design of the studies is summarised in the following table.

	Placebo/vehicle‐controlled	Active‐ controlled¹
Parallel groups	46 studies (including part I of 1 cross‐over study)	17 studies
Intraindividual²	12 studies	10 studies

Number of studies	Outcomes	Equivalence or transformed into outcome
Primary outcomes
12	Global improvement indices expressed per participant (Investigator, participant, or both)	Physician global assessment improvement, global therapeutic response or treated area, investigator assessment scale, investigator global assessment, overall response
53	Participant complete clearance (number of participants, rate, proportion, percentages)	Complete responders, total clearance, response to treatment, proportion of participants achieving total clearance, field complete clearance, complete remission, complete response of lesional area, participant's complete resolution, complete clearing, number of participants with 100% clearance, complete participant response, target lesion number score = 0, complete healing, cumulative lesion number score = 0, 100% lesions cleared, percentage of participants who experienced 100% clearance of all target lesions, number of participants with all cleared lesions
20	Participant partial (> 75%) clearance (number of participants, rate, proportion, percentages)	At least 75% reduction in the number of lesions, at least 75% of lesions cleared, percentage of participants who experienced 75% or greater clearance of all target lesions, therapy responders with at least 75% of clearing of the lesions, participant partial (> 80%) clearance rates
50	Mean reduction in lesion counts (absolute values or percentages)	Mean reduction in the number of actinic keratoses, mean changes of lesion counts, mean numbers of lesions at baseline and assessment time point, mean percentage reduction in the number of actinic keratoses, average changes in lesion counts, mean per cent changes from baseline for all actinic keratoses, mean per cent lesions cleared
Other outcomes
3	Global improvement indices expressed as scores	Physician global assessment, global improvement score
29	Lesion complete response (per lesions)	Reduction rate in number of actinic keratoses, clearance of individual lesions, rate of totally healed lesions, number of lesions with 0% of remaining area, complete clinical clearance rate on lesion basis, complete clearance rate of lesions, individual lesion clearance, lesion counts at baseline and assessment, percentage lesion reduction, proportion of baseline lesions cleared at the end of treatment, lesions remitted, total lesion counts
9	Median per cent reduction of baseline lesions	Median per cent changes from baseline for all actinic keratoses
6	Participant histological clearance	Histological clearance, histological confirmation
5	Recurrence	‐
3	Participant partial (> 50%) clearance	Participant with 50% or greater reduction, clearance = resolution of > 50% of the lesions
5	Reduction in lesion size	Overall reduction in lesion area, partial remission (50% size reduction of 75% of lesions), mean diameter of target lesion at baseline and assessment
2	Median number of lesions at baseline and assessment time point	‐
1	Participant partial (> 66%) clearance	‐
1	Participant partial (% not specified) clearance	‐
1	Total lesion number score (0 = 0 lesions, 1 = 1 to 3 lesions, 2 = 2 to 4 lesions, 3 = > 6 lesions)	‐
1	Negative predictive value, i.e. ratio between histological and clinical clearance	‐
1	Participant's perception of efficacy	‐
1	Efficacy on a visual analogue scale for field‐directed treatment	‐
1	Relapse	‐

Number of studies	Outcomes
10	Participant's satisfaction
8	Rest periods or temporary interruption during treatment
7	Compliance
6	Participant's preference
2	Biological and immunological outcomes
1	Skin concentrations of drug and products due to its mechanism of action
1	Investigator's preference
1	Lesion severity index
1	Quality of life on a visual analogue scale
1	Number of spray cooling for photodynamic therapy

Author	Topic	Clarification
Kurt Gebauer	Type of analysis used in the study Gebauer 2003	Intention‐to‐treat
Joseph Jorizzo	Type of analysis used in the studies Jorizzo 2002 and Jorizzo 2006	The type of analysis could not be confirmed
Iraji Fariba	Outcome presented was 'lesions complete response' or 'participant complete clearance' in the study Fariba 2006	No response received
Emil Tanghetti	Type of analysis used in the study Tanghetti 2007	Intention‐to‐treat

Data not separated for actinic keratoses	Did not meet review criteria for outcomes	Unacceptable or unclear randomisation	No numerical values, graphical data, or not enough information	Prevention of actinic keratosis lesions	Follow‐up reports on included studies
Alberts 2004 Green 1998 Humphreys 1996	Apalla 2010b Babilas 2007 Babilas 2008 Bartels 2009 Biecha‐Thalharnmer 2003 Braathen 2009 Dirschka 2010 Edwards 1986 Epstein 2006 Ericson 2004 Jury 2005 Kurwa 1999 Morales 2010 Puizina‐Ivic 2008a Radakovic‐Fijan 2005 Shuttleworth 1989 Smith 2006 Sotiriou 2011 Wulf 2006	Alexiades‐Armenakas 2003 Babilas 2006 Berlin 2008 Gold 2006 Goldman 2003 Griffin 1991 Grimaître 2000 Marrero 1998 Tsoukas 2010 Valeant 2004	Apalla 2010a Breza 1976 de Sévaux 2003 Dermik 2003 Gupta 2004 Robins 2002a Rosen 2010 Simmonds 1973 Spencer 2010 Touma 2004 Weinstock 2010 Yamauchi 2002	Apalla 2010c Elmets 2010 Naylor 1995 Wennberg 2008	Fowler 2002 Hanke 2011 Stockfleth 2004 Szeimies 2010a

Diclofenac in 2.5% hyaluronic acid compared to interventions for actinic keratoses in immunocompetent participants
Intervention/Comparison intervention	*Illustrative comparative risks (95% CI)**		Relative effect  (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	Comments
Intervention/Comparison intervention	Assumed risk With comparator	Corresponding risk With intervention	Relative effect  (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	Comments
Participant complete clearance
3% diclofenac in 2.5% hyaluronic acid/2.5% hyaluronic acid	Study population		RR 2.46   (1.66 to 3.66)	420  (3 studies)	⊕⊕⊕⊝  moderate	For all lesions, data from 30, 60, and 90 day treatments pooled together, assessment at 30 days after the end of treatment (Analysis 6.5)
	127 per 1000	313 per 1000  (211 to 466)
	Moderate
	132 per 1000	325 per 1000  (219 to 483)
3% diclofenac in 2.5% hyaluronic acid/5% imiquimod	‐	‐	‐	‐	‐	Not reported
3% diclofenac in 2.5% hyaluronic acid + ALA‐red light PDT/2.5% hyaluronic acid + ALA‐red light PDT	‐	‐	‐	‐	‐	Not reported
Mean reduction in lesion counts
3% diclofenac in 2.5% hyaluronic acid/2.5% hyaluronic acid	The mean reduction in lesion counts in the control groups was  2.5 lesions	The mean reduction of lesion counts in the intervention groups was  2.55 higher  (1.56 to 3.53 higher)	‐	345  (2 studies)	⊕⊕⊕⊕  high	Data from 30, 60, and 90 day treatments pooled together, assessment 30 days after the end of treatment (Analysis 6.12)
3% diclofenac in 2.5% hyaluronic acid/ 5% imiquimod	‐	‐	‐	‐	‐	Not reported
3% diclofenac in 2.5% hyaluronic acid + ALA‐red light PDT/2.5% hyaluronic acid + ALA‐red light PDT	See comment	See comment	Not estimable	10  (1 study)	⊕⊕⊝⊝  low	Intraindividual study: at 6 weeks; diclofenac/hyaluronic acid (HA) + ALA‐PDT = 10.13, HA + ALA‐PDT= 9.9, at 6 months; diclofenac/HA + ALA‐PDT = 11.56, HA + ALA‐PDT = 10.56, at 12 months; diclofenac/HA + ALA‐PDT = 12.5, HA + ALA‐PDT = 8.8
Mean percentage of reduction in lesion counts
All comparisons	‐	‐	‐	‐	‐	Not reported
Withdrawal due to adverse events
3% diclofenac in 2.5% hyaluronic acid/2.5% hyaluronic acid	Study population		RR 3.59   (1.92 to 6.7)	592  (4 studies)	⊕⊕⊕⊕  high	(Analysis 6.13) Additional data from intraindividual study: no participant withdrew because of adverse events (N = 20). GRADE = low.
	40 per 1000	144 per 1000  (77 to 269)
	Moderate
	43 per 1000	154 per 1000  (83 to 288)
3% diclofenac in 2.5% hyaluronic acid/5% imiquimod	0 per 1000	0 per 1000	Not estimable	49  (1 study)	⊕⊕⊕⊝  moderate	There were no participant withdrawals due to adverse events.
3% diclofenac in 2.5% hyaluronic acid + ALA‐red light PDT/2.5% hyaluronic acid + ALA‐red light PDT	0 per 1000	0 per 1000	Not estimable	10  (1 study)	⊕⊕⊕⊝  moderate	There were no participant withdrawals due to adverse events.
Skin irritation
3% diclofenac in 2.5% hyaluronic acid/2.5% hyaluronic acid	See comment	See comment	Not estimable	20  (1 study)	⊕⊕⊝⊝  low	Intraindividual study reported irritation only on the diclofenac treated side of 8 out of 20 participants
3% diclofenac in 2.5% hyaluronic acid/5% imiquimod	‐	‐	‐	‐	‐	Not reported
3% diclofenac in 2.5% hyaluronic acid + ALA‐red light PDT/2.5% hyaluronic acid + ALA‐red light PDT	‐	‐	‐	‐	‐	Not reported

5‐fluorouracil (5‐FU) compared to interventions for actinic keratoses in immunocompetent participants
Intervention/Comparison intervention	*Illustrative comparative risks (95% CI)**		Relative effect  (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	Comments
Intervention/Comparison intervention	Assumed risk With comparator	Corresponding risk With intervention	Relative effect  (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	Comments
Participant complete clearance
0.5% 5‐FU/Vehicle	Study population		RR 8.86   (3.67 to 21.40)	522  (3 studies)	⊕⊕⊕⊕  high	Data from 1, 2, and 4 week treatments were pooled together (Analysis 9.1)
	15 per 1000	136 per 1000  (56 to 328)
	Moderate
	0 per 1000	0 per 1000  (0 to 0)
0.5% 5‐FU with cryotherapy/Vehicle with cryotherapy	71 per 1000	291 per 1000  (116 to 731)	RR 4.08   (1.63 to 10.23)	142  (1 study)	⊕⊕⊝⊝  low	1 cycle (Analysis 62.1)
0.5% 5‐FU/ALA‐PDT	292 per 1000	499 per 1000  (239 to 1000)	RR 1.71   (0.74 to 3.98)	48  (1 study)	⊕⊝⊝⊝  very low	Data from blue light and pulsed dye laser were pooled (Analysis 11.1)
0.5% 5‐FU/5.0% 5‐FU	See comment	See comment	Not estimable	21  (1 study)	⊕⊝⊝⊝  very low	Intraindividual study: 0.5% and 5.0% 5‐FU = 9/21
5% 5‐FU with 0.05% tretinoin /5% 5‐FU with placebo	‐	‐	‐	‐	‐	Not reported
5% 5‐FU /10% masoprocol	‐	‐	‐	‐	‐	Not reported
5% 5‐FU/5% Imiquimod	Study population		RR 1.85   (0.41 to 8.33)	89  (2 studies)	⊕⊝⊝⊝  very low	(Analysis 13.1)
	600 per 1000	1000 per 1000  (246 to 1000)
	Moderate
	555 per 1000	1000 per 1000  (230 to 1000)
5% 5‐FU/Carbon dioxide laser resurfacing	‐	‐	‐	‐	‐	Not reported
5% 5‐FU/Er:YAG laser resurfacing	‐	‐	‐	‐	‐	Not reported
5% 5‐FU/Cryotherapy	680 per 1000	959 per 1000  (721 to 1000)	RR 1.41   (1.06 to 1.87)	49  (1 study)	⊕⊕⊝⊝  low	Only data after the treatment (Analysis 14.1)
5% 5‐FU/Trichloroacetic acid peel	‐	‐	‐	‐	‐	Not reported
Mean reduction in lesion counts
0.5% 5‐FU/Vehicle	The mean reduction in lesion counts in the control groups was  4 lesions	The mean reduction in lesion counts in the intervention groups was  5.40 higher  (2.94 to 7.86 higher)	‐	142  (1 study)	⊕⊕⊕⊝  moderate	Data from 1, 2, and 4 week treatment were pooled. (Analysis 9.2) Results from another study (N = 177) with no SD: placebo: 2.7 lesions, 5‐FU = 8.8 lesions, GRADE = moderate
0.5% 5‐FU with cryotherapy/Vehicle with cryotherapy	The mean reduction in lesion counts in the control groups was  6.6 lesions	The mean reduction in lesion counts in the intervention groups was  2 higher  (0.49 lower to 4.49 higher)		142  (1 study)	⊕⊕⊕⊝  moderate	1 cycle (Analysis 62.2)
0.5% 5‐FU/ALA‐PDT	‐	‐	‐	‐	‐	Not reported
0.5% 5‐FU/5.0% 5‐FU	See comment	See comment	Not estimable	21  (1 study)	⊕⊕⊝⊝  low	Intraindividual study: results with no SD: 0.5% 5‐FU = 8.8 lesions, 5.0% 5‐FU = 6.1 lesions
5% 5‐FU with 0.05% tretinoin /5% 5‐FU with placebo	The mean reduction in lesion counts in the control groups was  11.1 lesions	The mean reduction in lesion counts in the intervention groups was  1.2 higher  (3.24 lower to 5.64 higher)	‐	19  (1 study)	⊕⊕⊝⊝  low	(Analysis 12.1)
5% 5‐FU /10% masoprocol	The mean reduction in lesion counts in the control groups was  11.3 lesions	The mean reduction in lesion counts in the intervention groups was  1.5 higher  (2.36 lower to 5.36 higher)	‐	49  (1 study)	⊕⊕⊝⊝  low	(Analysis 15.2)
5% 5‐FU/5% Imiquimod	‐	‐	‐	‐	‐	Not reported
5% 5‐FU/Carbon dioxide laser resurfacing	‐	‐	‐	‐	‐	Not reported
5% 5‐FU/Er:YAG laser resurfacing	See comment	See comment	Not estimable	55  (1 study)	⊕⊕⊝⊝  low	Results with no SD: number of lesions at 3 months:5‐FU = 13.2, resurfacing = 13.8, at 6 months:5‐FU = 12.5, resurfacing = 13.9, at 12 months: 5‐FU = 12.4, resurfacing = 14.2
5% 5‐FU/Cryotherapy	‐	‐	‐	‐	‐	Not reported
5% 5‐FU/Trichloroacetic acid peel	‐	‐	‐	‐	‐	Not reported
Mean percentage of reduction in lesion counts
0.5% 5‐FU/Vehicle	The mean percentage of reduction in lesion counts ranged across control groups from  28.8 per cent	The mean percentage of reduction in lesion counts in the intervention groups was  33.60 higher  (22.88 to 44.32 higher)	‐	142  (1 study)	⊕⊕⊕⊝  moderate	Data from 1 week treatment.(Analysis 9.3) Results from two other studies with no SD 1) (N = 207) placebo = 21.6%, 5‐FU = 69.5%, GRADE = low, 2)(N = 177) placebo = 34.4%, 5‐FU = 78.5%, GRADE = moderate
0.5% 5‐FU with cryotherapy/Vehicle with cryotherapy	The mean percentage of reduction in lesion counts in the control groups was  45.6 per cent	The mean percentage of reduction in lesion counts in the intervention groups was  21.4 higher  (5.1 to 37.7 higher)	‐	142  (1 study)	⊕⊕⊕⊝  moderate	(Analysis 62.3)
0.5% 5‐FU/ALA‐PDT	‐	‐	‐	‐	‐	Not reported
0.5% 5‐FU/5.0% 5‐FU	See comment	See comment	Not estimable	21  (1 study)	⊕⊕⊝⊝  low	Intraindividual study: results with no SD: 0.5% 5‐FU = 67% and 5.0% 5‐FU = 47%
5% 5‐FU with 0.05% tretinoin /5% 5‐FU with placebo	‐	‐	‐	‐	‐	Not reported
5% 5‐FU /10% masoprocol	The mean percentage of reduction in lesion counts in the control groups was  77.6 percent	The mean percentage of reduction in lesion counts in the intervention groups was  20 higher  (11.82 to 28.18 higher)	‐	49  (1 study)	⊕⊕⊕⊝  moderate	(Analysis 15.3)
5% 5‐FU/5% Imiquimod	See comment	See comment	Not estimable	39  (1 study)	⊕⊕⊝⊝  low	Results with no SD: 5% 5‐FU = 94%, 5% imiquimod = 66%
5% 5‐FU/Carbon dioxide laser resurfacing	The mean percentage of reduction in lesion counts in the control groups was  92 percent	The mean percentage of reduction in lesion counts in the intervention groups was  8.80 lower  (20.76 lower to 3.16 higher)	‐	14  (1 study)	⊕⊝⊝⊝  very low	(Analysis 16.1 )
5% 5‐FU/Er:YAG laser resurfacing	See comment	See comment	Not estimable	55  (1 study)	⊕⊕⊝⊝  low	Results with no SD: at 6 months: 5‐FU = 79.2%, resurfacing 94.5%, at 12 months: 5‐FU = 76.6%, resurfacing = 91.1%
5% 5‐FU/Cryotherapy	‐	‐	‐	‐	‐	Not reported
5% 5‐FU/Trichloroacetic acid peel	The mean percentage of reduction in lesion counts in the control groups was  89 per cent	The mean percentage of reduction in lesion counts in the intervention groups was  5.8 lower  (15.38 lower to 3.78 higher)	‐	18  (1 study)	⊕⊝⊝⊝  very low	(Analysis 18.1)
Withdrawal due to adverse events
0.5% 5‐FU/Vehicle	0 per 1000	N/A (5/119 = 42/1000)	RR 5.41   (0.3 to 96.18)	177  (1 study)	⊕⊝⊝⊝  very low	Data from 1, 2, and 4 week treatments were pooled.(Analysis 9.4) Another study reported 24/207 participants withdrew because of adverse events and 12 of them were in 4 week 5‐FU group. GRADE = low
0.5% 5‐FU with cryotherapy/Vehicle with cryotherapy	See comment	See comment	Not estimable	142  (1 study)	⊕⊕⊝⊝  low	There were no participant withdrawals in the first part of this three part study (incomplete data were given for the whole study).
0.5% 5‐FU/ALA‐PDT	0 per 1000	N/A (1/12 = 83/1000)	RR 5.77   (0.25 to 131.92)	36  (1 study)	⊕⊕⊝⊝  low	Data from blue light and pulsed dye laser were pooled (Analysis 11.2)
0.5% 5‐FU/5.0% 5‐FU	See comment	See comment	Not estimable	21  (1 study)	⊕⊕⊝⊝  low	Intraindividual study: 16/21 discontinued treatment but did not withdraw: 4 because of 0.5%, 8 because of 5.0% , 4 because of both creams.
5% 5‐FU with 0.05% tretinoin /5% 5‐FU with placebo	See comment	See comment	Not estimable	19  (1 study)	⊕⊕⊝⊝  low	Intraindividual study: 1 participant withdrew because of irritation but associated treatment was not specified.
5% 5‐FU /10% masoprocol	0 per 1000	N/A (1/30 = 33/1000)	RR 2.71   (0.12 to 63.84)	57  (1 study)	⊕⊕⊝⊝  low	(Analysis 15.4)
5% 5‐FU/5% Imiquimod	0 per 1000	0 per 1000	Not estimable	89  (2 studies)	⊕⊕⊝⊝  low	There were no participant withdrawals due to adverse events.
5% 5‐FU/Carbon dioxide laser resurfacing	250 per 1000	45 per 1000  (2 to 817)	RR 0.18   (0.01 to 3.27)	17  (1 study)	⊕⊕⊝⊝  low	(Analysis 16.2)
5% 5‐FU/Er:YAG laser resurfacing	0 per 1000	N/A (1/27 = 37/1000)	RR 3.11   (0.13 to 73.11)	55  (1 study)	⊕⊕⊝⊝  low	(Analysis 17.1)
5% 5‐FU/Cryotherapy	0 per 1000	0 per 1000	Not estimable	49  (1 study)	⊕⊕⊕⊝  moderate	There were no participant withdrawals due to adverse events.
5% 5‐FU/Trichloroacetic acid peel	0 per 1000	0 per 1000	Not estimable	18  (1 study)	⊕⊕⊝⊝  low	There were no participant withdrawals due to adverse events.
Skin irritation
0.5% 5‐FU/Vehicle	654 per 1000	948 per 1000  (830 to 1000)	RR 1.45   (1.27 to 1.65)	384  (2 studies)	⊕⊕⊕⊝  moderate	Data from 1, 2, and 4 week treatments were pooled (Analysis 9.5)
0.5% 5‐FU with cryotherapy/Vehicle with cryotherapy	‐	‐	‐	‐	‐	Not reported
0.5% 5‐FU/ALA‐PDT	‐	‐	‐	‐	‐	Not reported
0.5% 5‐FU/5.0% 5‐FU	1000 per 1000	1000 per 1000	‐	21  (1 study)	⊕⊕⊕⊝  moderate	Intraindividual study: All participants reported facial irritation in association with both creams
5% 5‐FU with 0.05% tretinoin /5% 5‐FU with placebo	See comment	See comment	Not estimable	19  (1 study)	⊕⊕⊕⊝  moderate	Intraindividual study: 12 had more irritation with tretinoin, 4 had more with placebo, and 3 had equal irritation.
5% 5‐FU /10% masoprocol	‐	‐	‐	‐	‐	Not reported
5% 5‐FU/5% Imiquimod	‐	‐	‐	‐	‐	Not reported
5% 5‐FU/Carbon dioxide laser resurfacing	‐	‐	‐	‐	‐	Not reported
5% 5‐FU/Er:YAG laser resurfacing	429 per 1000	703 per 1000  (429 to 1000)	RR 1.64   (1 to 2.69)	55  (1 study)	⊕⊕⊝⊝  low	At the end of treatment (Analysis 17.2)
5% 5‐FU/Cryotherapy	‐	‐	‐	‐	‐	Not reported
5% 5‐FU/Trichloroacetic acid peel	‐	‐	‐	‐	‐	Not reported

Cryotherapy compared to interventions for actinic keratoses in immunocompetent participants
Intervention/ Comparison intervention	*Illustrative comparative risks (95% CI)**		Relative effect  (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	Comments
Intervention/ Comparison intervention	Assumed risk	Corresponding risk	Relative effect  (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	Comments
‐	With comparator	With intervention	‐	‐	‐	‐
Participant complete clearance
Cryotherapy /Betulin‐based oleogel	643 per 1000	784 per 1000  (489 to 1000)	RR 1.22   (0.76 to 1.97)	28  (1 study)	⊕⊝⊝⊝  very low	(Analysis 42.1)
Cryotherapy/cryotherapy with betulin‐based oleogel	714 per 1000	786 per 1000  (514 to 1000)	RR 1.1   (0.72 to 1.69)	28  (1 study)	⊕⊝⊝⊝  very low	(Analysis 61.1)
Cryotherapy/5% 5‐FU	958 per 1000	680 per 1000  (518 to 901)	RR 0.71   (0.54 to 0.94)	49  (1 study)	⊕⊕⊝⊝  low	Assessment after treatment (Analysis 43.1)
Vehicle with cryotherapy/0.5% 5‐FU with cryotherapy	292 per 1000	70 per 1000  (29 to 178)	RR 0.24   (0.1 to 0.61)	142  (1)	⊕⊕⊝⊝  low	1 cycle (Analysis 63.1)
Cryotherapy /Imiquimod	846 per 1000	677 per 1000  (499 to 931)	RR 0.8   (0.59 to 1.10)	51  (1 study)	⊕⊝⊝⊝  very low	5% imiquimod (Analysis 44.1)
Cryotherapy with vehicle /Cryotherapy with imiquimod	Study population		RR 0.2   (0.05 to 0.73)	311  (2 studies)	⊕⊕⊕⊕  high	Pooled data (5% and 3.75% imiquimod)(Analysis 64.1) Results from an additional intraindividual study: cryotherapy + vehicle = 5/27, cryotherapy+imiquimod = 8/27 GRADE = moderate
	287 per 1000	57 per 1000  (14 to 209)
	Moderate
	264 per 1000	53 per 1000  (13 to 193)
Cryotherapy /ALA‐red light PDT	581 per 1000	442 per 1000  (354 to 558)	RR 0.76   (0.61 to 0.96)	297  (1 study)	⊕⊕⊝⊝  low	(Analysis 46.1)
Cryotherapy/MAL‐red light PDT	‐	‐	‐	‐	‐	Not reported
Mean reduction in lesion counts
Cryotherapy /Betulin‐based oleogel	‐	‐	‐	‐	‐	Not reported
Cryotherapy/cryotherapy with betulin‐based oleogel	‐	‐	‐	‐	‐	Not reported
Cryotherapy/5% 5‐FU	‐	‐	‐	‐	‐	Not reported
Vehicle + cryotherapy/0.5% 5‐FU + cryotherapy	The mean reduction in lesion counts in the control groups was  8.6 lesions	The mean reduction in lesion counts in the intervention groups was  2 lower  (4.49 lower to 0.49 higher)	‐	142  (1)	⊕⊕⊕⊝  moderate	1 cycle (Analysis 63.2)
Cryotherapy /Imiquimod	‐	‐	‐	‐	‐	Not reported
Cryotherapy with vehicle /Cryotherapy with imiquimod	‐	‐	‐	‐	‐	Not reported
Cryotherapy /ALA‐red light PDT	‐	‐	‐	‐	‐	Not reported
Cryotherapy/MAL‐red light PDT	‐	‐	‐	‐	‐	Not reported
Mean percentage of reduction in lesion counts
Cryotherapy /Betulin‐based oleogel	‐	‐	‐	‐	‐	Not reported
Cryotherapy/cryotherapy with betulin‐based oleogel	‐	‐	‐	‐	‐	Not reported
Cryotherapy/5% 5‐FU	‐	‐	‐	‐	‐	Not reported
Vehicle with cryotherapy/0.5% 5‐FU with cryotherapy	The mean percentage of reduction in lesion counts in the control groups was  67 percent	The mean percentage of reduction in lesion counts in the intervention groups was  21.4 lower  (37.7 to 5.1 lower)	‐	142  (1)	⊕⊕⊕⊝  moderate	(Analysis 63.3)
Cryotherapy /Imiquimod	‐	‐	‐	‐	‐	Not reported
Cryotherapy with vehicle /Cryotherapy with imiquimod	See comment	See comment	‐	301  (2 studies)	⊕⊝⊝⊝  very low	High heterogeneity (I²=86%) between 3.75% (parallel group, MD ‐34.10, 95% CI ‐41.38 to ‐26.82)) and 5.0% (intraindividual, MD ‐11.20, 95% CI ‐26.53 to 4.13) imiquimod studies. (Analysis 64.4)
Cryotherapy /ALA‐red light PDT	‐	‐	‐	‐	‐	Not reported
Cryotherapy/MAL‐red light PDT	See comment	See comment	Not estimable	240  (2 studies)	⊕⊝⊝⊝  very low	Intraindividual studies with no SD: at 12 weeks: cryotherapy = 74.5%, MAL‐PDT= 84.4%, at 24 weeks: cryotherapy = 83.9‐87%, MAL‐PDT = 75‐86.7%
Withdrawal due to adverse events
Cryotherapy /Betulin‐based oleogel	0 per 1000	0 per 1000	Not estimable	28  (1 study)	⊕⊕⊕⊝  moderate	There were no participant withdrawals due to adverse events.
Cryotherapy/cryotherapy with betulin‐based oleogel	0 per 1000	0 per 1000	Not estimable	28  (1 study)	⊕⊕⊕⊝  moderate	There were no participant withdrawals due to adverse events.
Cryotherapy/5% 5‐FU	0 per 1000	0 per 1000	Not estimable	49  (1 study)	⊕⊕⊕⊝  moderate	There were no participant withdrawals due to adverse events.
Vehicle with cryotherapy/0.5% 5‐FU with cryotherapy	See comment	See comment	Not estimable	142  (1 study)	⊕⊕⊝⊝  low	There were no participant withdrawals due to adverse events in the first part of this three part study (incomplete data were given for the whole study).
Cryotherapy /Imiquimod	0 per 1000	0 per 1000	Not estimable	51  (1 study)	⊕⊕⊕⊝  moderate	There were no participant withdrawals due to adverse events.
Cryotherapy with vehicle /Cryotherapy with imiquimod	Study population		RR 0.93   (0.28 to 3.07)	312  (2 studies)	⊕⊕⊕⊝  moderate	Pooled data (5% and 3.75% imiquimod) (Analysis 64.6)
	33 per 1000	30 per 1000  (9 to 100)
	Moderate
	21 per 1000	20 per 1000  (6 to 64)
Cryotherapy /ALA‐ red light PDT	0 per 1000	0 per 1000	Not estimable	297  (1 study)	⊕⊕⊕⊝  moderate	There were no participant withdrawals due to adverse events.
Cryotherapy/MAL‐ red light PDT	11 per 1000	11 per 1000  (2 to 75)	RR 1.06   (0.16 to 7.16)	379  (2 studies)	⊕⊕⊕⊝  moderate	(Analysis 45.2) Two additional intraindividual studies: 4 of 119 and 2 of 121 participants withdrew because of adverse events and one of them was related to MAL‐PDT. GRADE = low
Skin irritation
Cryotherapy /Betulin‐based oleogel	‐	‐	‐	‐	‐	Not reported
Cryotherapy/cryotherapy with betulin‐based oleogel	‐	‐	‐	‐	‐	Not reported
Cryotherapy/5% 5‐FU	‐	‐	‐	‐	‐	Not reported
Vehicle with cryotherapy/0.5% 5‐FU with cryotherapy	‐	‐	‐	‐	‐	Not reported
Cryotherapy /Imiquimod	‐	‐	‐	‐	‐	Not reported
Cryotherapy with vehicle /Cryotherapy with imiquimod	Study population		RR 0.39   (0.1 to 1.54)	311  (2 studies)	⊕⊕⊕⊝  moderate	Pooled data (5% and 3.75% imiquimod) (Analysis 64.7)
	83 per 1000	32 per 1000  (8 to 128)
	Moderate
	125 per 1000	49 per 1000  (13 to 192)
Cryotherapy /ALA‐red light PDT	372 per 1000	100 per 1000  (59 to 171)	RR 0.27   (0.16 to 0.46)	297  (1 study)	⊕⊕⊝⊝  low	Assessment one day after the treatment (Analysis 46.2)
Cryotherapy/MAL‐red light PDT	‐	‐	‐	‐	‐	Not reported

Imiquimod compared to interventions for actinic keratoses in immunocompetent participants
Intervention/Comparison intervention	*Illustrative comparative risks (95% CI)**		Relative effect  (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	Comments
Intervention/Comparison intervention	Assumed risk	Corresponding risk	Relative effect  (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	Comments
‐	With comparator	With intervention	‐	‐	‐	‐
Participant complete clearance
2.5% imiquimod/placebo	62 per 1000	277 per 1000  (148 to 518)	RR 4.49   (2.4 to 8.39)	486  (2 studies)	⊕⊕⊕⊕  high	(Analysis 20.1)
3.75% imiquimod/placebo	Study population		RR 6.45   (3.87 to 10.73)	730  (3 studies)	⊕⊕⊕⊕  high	(Analysis 20.1)
	53 per 1000	343 per 1000  (206 to 571)
	Moderate
	50 per 1000	322 per 1000  (193 to 536)
Cryotherapy + 3.75% imiquimod/Cryotherapy + vehicle	33 per 1000	301 per 1000  (111 to 820)	RR 9.12   (3.36 to 24.79)	247  (1 study)	⊕⊕⊕⊝  moderate	For all lesions (Analysis 65.1)
5% imiquimod/placebo	Study population		RR 7.70   (4.63 to 12.79)	1871  (9 studies)	⊕⊕⊕⊕  high	(Analysis 20.1)
	48 per 1000	371 per 1000  (223 to 617)
	Moderate
	32 per 1000	246 per 1000  (148 to 409)
5% imiquimod/3% diclofenac in 2.5% hyaluronic acid	‐	‐	‐	‐	‐	Not reported
5% imiquimod /5% 5‐FU	See comment	See comment	‐	89  (2 studies)	⊕⊝⊝⊝  very low	The two studies was associated with high heterogeneity (I²= 93%) and the results could not be pooled together. One study favoured 5‐FU (RR 0.31, 95% CI 0.14 to 0.67] whereas the other did not (RR 0.88, 95% CI 0.73 to 1.06] (Analysis 22.1)
5% imiquimod/Cryotherapy	680 per 1000	843 per 1000  (619 to 1000)	RR 1.24   (0.91 to 1.7)	51  (1 study)	⊕⊝⊝⊝  very low	(Analysis 23.1)
Cryotherapy + 5% imiquimod/Cryotherapy + vehicle	91 per 1000	225 per 1000  (64 to 796)	RR 2.48  (0.70 to 8.76)	64  (1 study)	⊕⊕⊝⊝  low	For all lesions.(Analysis 65.1) Results from an additional intraindividual study: cryotherapy + imiquimod side (8/27 = 30%), cryotherapy alone side (5/27 = 19%), GRADE = low
5% imiquimod/ALA‐PDT				‐		Not reported
ALA‐PDT + 5% imiquimod/ALA‐PDT + placebo	See comment	See comment	Not estimable	25  (1 study)	⊕⊕⊝⊝  low	Intraindividual study: ALA‐PDT + 5% imiquimod = 2/25; ALA‐PDT + placebo = 2/25
Mean reduction in lesion counts
2.5% imiquimod/placebo	‐	‐	‐	‐	‐	Not reported
3.75% imiquimod/placebo	‐	‐	‐	‐	‐	Not reported
Cryotherapy + 3.75% imiquimod/Cryotherapy + vehicle	‐	‐	‐	‐	‐	Not reported
5% imiquimod/placebo	The mean reduction in lesion counts in the control groups was  0.6 lesions	The mean reduction in lesion counts in the intervention groups was 2.20 higher  (1.05 lower to 5.45 higher)	‐	12  (1 study)	⊕⊕⊝⊝  low	(Analysis 19.5) Results from an additional intraindividual study with no SD (N = 21): 5% imiquimod: 3.9 lesions, placebo = 0.5 lesions, GRADE = very low
5% imiquimod/3% diclofenac in 2.5% hyaluronic acid	‐	‐	‐	‐	‐	Not reported
5% imiquimod /5% 5‐FU	‐	‐	‐	‐	‐	Not reported
5% imiquimod/Cryotherapy	‐	‐	‐	‐	‐	Not reported
Cryotherapy + 5% imiquimod/Cryotherapy + vehicle	‐	‐	‐	‐	‐	Not reported
5% imiquimod/ALA‐PDT	‐	‐	‐	‐	‐	Not reported
ALA‐PDT + 5% imiquimod/ALA‐PDT + placebo	See comment	See comment	Not estimable	25  (1 study)	⊕⊕⊝⊝  low	Results from intraindividual study without SD: ALA‐PDT + 5% imiquimod= 19.9 lesions; ALA‐PDT + placebo= 16.0 lesions
Mean percentage of reduction in lesion counts
2.5% imiquimod/placebo	‐	‐	‐	‐	‐	Not reported
3.75% imiquimod/placebo	The mean percentage of reduction in lesion counts in the control groups was  21.1 per cent	The mean percentage of reduction in lesion counts in the intervention groups was 46.90 higher  (36.68 to 57.12 higher)	‐	247  (1 study)	⊕⊕⊕⊝  moderate	(Analysis 20.3)
Cryotherapy + 3.75% imiquimod/Cryotherapy + vehicle	The mean percentage of reduction in lesion counts in the control groups was  43.3 per cent	The mean percentage of reduction in lesion counts in the intervention groups was 34.1 higher  (26.82 to 41.38 higher)	‐	247  (1 study)	⊕⊕⊕⊝  moderate	For all lesions (Analysis 65.2)
5% imiquimod/placebo	‐	‐	‐	‐	‐	Not reported
5% imiquimod/3% diclofenac in 2.5% hyaluronic acid	‐	‐	‐	‐	‐	Not reported
5% imiquimod /5% 5‐FU	See comment	See comment	Not estimable	39  (1 study)	⊕⊕⊝⊝  low	Results with no SD: 5% imiquimod = 66%, 5% 5‐FU = 94%
5% imiquimod/Cryotherapy	‐	‐	‐	‐	‐	Not reported
Cryotherapy + 5% imiquimod/Cryotherapy + vehicle	The mean percentage of reduction in lesion counts in the control groups was  62 per cent	The mean percentage of reduction in lesion counts in the intervention groups was 11.2 higher  (4.13 lower to 26.53 higher)	‐	27  (1 study)	⊕⊕⊝⊝  low	For all lesions.(Analysis 65.2) Results from an additional intraindividual study: cryotherapy‐5% imiquimod = 73.2+27.1%, cryotherapy + vehicle = 62.0+30.3%. GRADE = moderate
5% imiquimod/ALA‐PDT	‐	‐	‐	‐	‐	Not reported
ALA‐PDT + 5% imiquimod/ALA‐PDT + placebo	See comment	See comment	Not estimable	25  (1 study)	⊕⊕⊝⊝  low	Results from intraindividual study without SD: ALA‐PDT + 5% imiquimod = 86.7% ; ALA‐PDT + placebo = 73.1 %
Withdrawal due to adverse events
2.5% imiquimod/placebo	19 per 1000	9 per 1000  (2 to 50)	RR 0.5   (0.09 to 2.7)	486  (2 studies)	⊕⊕⊕⊝  moderate	(Analysis 20.5)
3.75% imiquimod/placebo	19 per 1000	17 per 1000  (4 to 73)	RR 0.92   (0.22 to 3.93)	483  (2 studies)	⊕⊕⊕⊝  moderate	(Analysis 20.5)
Cryotherapy + 3.75% imiquimod/Cryotherapy + vehicle	32 per 1000	41 per 1000  (11 to 150)	RR 1.3   (0.36 to 4.73)	247  (1 study)	⊕⊕⊝⊝  low	(Analysis 65.3)
5% imiquimod/placebo	Study population		RR 2.59   (1.59 to 4.23)	2290  (8 studies)	⊕⊕⊕⊝  moderate	(Analysis 20.5) Four small sample size studies with no participant withdrawal are not included in meta‐analysis: pooled data, imiquimod 0/79 and placebo 0/31. Additional two intraindividual studies: no participant withdrew because of adverse events (0/42) GRADE = very low (both studies had more than 20% participant lost).
	21 per 1000	56 per 1000  (34 to 91)
	Moderate
	5 per 1000	13 per 1000  (8 to 22)
	High
	0 per 1000	0 per 1000  (0 to 0)
5% imiquimod/3% diclofenac in 2.5% hyaluronic acid	0 per 1000	0 per 1000	Not estimable	49  (1 study)	⊕⊕⊕⊝  moderate	There were no participant withdrawals due to adverse events.
5% imiquimod /5% 5‐FU	0 per 1000	0 per 1000	Not estimable	50  (1 study)	⊕⊕⊕⊝  moderate	There were no participant withdrawals due to adverse events.
5% imiquimod/Cryotherapy	0 per 1000	0 per 1000	Not estimable	51  (1 study)	⊕⊕⊕⊝  moderate	There were no participant withdrawals due to adverse events.
Cryotherapy + 5% imiquimod/Cryotherapy + vehicle	30 per 1000	10 per 1000  (0 to 246)	RR 0.34   (0.01 to 8.13)	65  (1 study)	⊕⊕⊝⊝  low	(Analysis 65.3)
5% imiquimod/ALA‐PDT	0 per 1000	0 per 1000	Not estimable	30  (1 study)	⊕⊕⊝⊝  low	There were no participant withdrawals due to adverse events.
ALA‐PDT + 5% imiquimod/ALA‐PDT + placebo	0 per 1000	0 per 1000	Not estimable	25  (1 study)	⊕⊕⊝⊝  low	There were no participant withdrawals due to adverse events.
Skin irritation
2.5% imiquimod/placebo	6 per 1000	21 per 1000  (4 to 117)	RR 3.45   (0.63 to 18.97)	486  (2 studies)	⊕⊕⊕⊝  moderate	(Analysis 20.6)
3.75% imiquimod/placebo	6 per 1000	30 per 1000  (6 to 159)	RR 4.86   (0.92 to 25.83)	484  (2 studies)	⊕⊕⊕⊝  moderate	(Analysis 20.6)
Cryotherapy + 3.75% imiquimod/Cryotherapy + vehicle	8 per 1000	56 per 1000  (7 to 445)	RR 6.72   (0.84 to 53.83)	247  (1 study)	⊕⊕⊝⊝  low	(Analysis 65.4)
5% imiquimod/placebo	5 per 1000	18 per 1000  (4 to 79)	RR 3.68   (0.86 to 15.74)	708  (3 studies)	⊕⊕⊕⊝  moderate	(Analysis 20.6) Additional intraindividual study: similar mild irritation between the two treatment sides (N = 20) GRADE = very low
5% imiquimod/3% diclofenac in 2.5% hyaluronic acid	‐	‐	‐	‐	‐	Not reported
5% imiquimod/5% 5‐FU	‐	‐	‐	‐	‐	Not reported
5% imiquimod/Cryotherapy	‐	‐	‐	‐	‐	Not reported
Cryotherapy + 5% imiquimod/Cryotherapy + vehicle	121 per 1000	194 per 1000  (61 to 622)	RR 1.6   (0.5 to 5.13)	64  (1 study)	⊕⊕⊝⊝  low	(Analysis 65.4)
5% imiquimod/ALA‐PDT	‐	‐	‐	‐	‐	Not reported
ALA‐PDT + 5% imiquimod/ALA‐PDT + placebo	‐	‐	‐	‐	‐	Not reported

Study	Global Improvement indices for completely improved or cleared (N = 90 participants)	Participant complete clearance	Participant partial (> 75%) clearance	Mean reduction (changes) in lesion counts (N = 90 participants)
Kang 2003	Investigator	‐	‐	Absolute values

Study	Withdrawal due to adverse events (N = 90 participants)	Skin irritation	Minor adverse events excluding skin irritation (N = 90 participants)	Cosmetic outcome (N = 42 participants)
Kang 2003	x	‐	x	Graph

Number of participants with the following improvements	Ro14‐9706	Tretinoin
Complete response	0/25	2/25
Partial response	12/25	10/25
No response	13/25	11/25
Worsening	0/25	2/25

Characteristic	McEwan 1997	Wolf 2001	Rivers 2002	Solaraze study 2	Gebauer 2003	Fariba 2006	Ulrich 2010
Study design	Parallel‐group	Parallel‐group	Parallel‐group	Parallel‐group	Parallel‐group	Intraindividual	Parallel‐group
Anatomical locations	Face, scalp, ear, neck, lower arm/elbow, hand, lower leg/knee	Forehead, central face, scalp, arms, hands	Forehead, central face, scalp, dorsum of hands	Face, scalp, forehead, arm, forearm, back of hands	Head/neck, hands, or arms	Face or scalp	Face, scalp, hands
Diclofenac concentration (%)	3	3	3	3	3	3	3
Frequency of treatment	Twice daily	Twice daily	Twice daily	Twice daily	Twice daily	Twice daily	Twice daily
Duration given (days)	56 to 168	90	30, 60	90	90	90	112
Assessment	At the end of 24‐week treatment	4 weeks after the end of treatment	4 weeks after the end of treatment	4 weeks after the end of treatment	At the end of 12‐week treatment and 4 weeks after the end of treatment	At the end of 12‐week treatment	At the end of 16‐week treatment

Mean percentage of reduction in lesion counts	Placebo	5‐fluorouracil (1 week)	5‐fluorouracil (2 weeks)	5‐fluorouracl (4 weeks)	5‐fluorouracl (pooled)
Jorizzo 2002 (N = 207 participants)	21.6%	69.5%	86.1%	91.7%	82.4%
Jorizzo 2004 (N = 144 participants)	28.8% + 32.6% (SD)	62.4% + 32.6% (SD)	N/A	N/A	62.4%
Weiss 2002 (N = 177 participants)	34.4%	78.5%	83.6%	88.7%	83.6%

Study	Global improvement indices for completely improved or cleared	Participant complete clearance (N = 16 participants)	Participant partial (> 75%) clearance	Mean reduction (changes) in lesion counts (N = 16 participants)
Akar 2001	‐	x	‐	Absolute values

Study	Withdrawal due to adverse events (N = 644 participants)	Skin irritation (N = 20 participants)	Minor adverse events excluding skin irritation (N = 462 participants)	Cosmetic outcome (N = 32 participants)
McEwan 1997	x	‐	‐	‐
Wolf 2001	x	‐	In general (excluding dermatology because it could include skin irritation) and specific adverse events based on body system	‐
Rivers 2002	x	‐	x	‐
Solaraze study 2	‐	‐	‐	‐
Gebauer 2003	x	‐	x	‐
Fariba 2006	x (intraindividual)	x	‐	‐
Ulrich 2010	x	‐	‐	x

Study	Withdrawal due to adverse events (N = 384 participants)	Skin irritation (N = 384 participants)	Minor adverse events excluding skin irritation (N = 384 participants)	Cosmetic outcome
Jorizzo 2002	x	x	In general (excluding dermatology because it could include skin irritation) and specific adverse events based on body system	‐
Weiss 2002	x	x	In general (excluding dermatology because it could include skin irritation) and specific adverse events based on body system	‐

Study	Withdrawal due to adverse events (N = 21 participants)	Skin irritation (N = 21 participants)	Minor adverse events excluding skin irritation (N = 21 participants)	Cosmetic outcome
Loven 2002	x	x	x	‐

Study	Anatomical locations	Imiquimod percentage	Number of doses/week	Number of weeks	Number of doses	Time of assessment
Persaud 2002	Face, arms, legs	5	3	8 or less	24 or less	8 weeks after the end of treatment
Stockfleth 2002	Face, scalp, forehead, dorsal forearm, neck, back of hands	5	3	12 or less	36 or less	At the end of the 12‐week treatment
Chen 2003	Face, forehead and temples, cheeks	5	3	3 or 6	9 or 18	4 weeks after the end of treatment
Lebwohl 2004	Face or scalp	5	2	16 or less	32 or less	8 weeks after the end of treatment
Szeimies 2004	Face or bald scalp	5	3	16 or less	48 or less	8 weeks after the end of treatment
Korman 2005	Face or bald scalp	5	3	16	48	8 weeks after the end of treatment
Ooi 2006	Scalp, extremities, or upper trunk	5	3	16 or less	48 or less	At the end of treatment
Alomar 2007	Face or bald scalp	5	3	4 or 8	12 or 24	4 weeks after the end of treatment
Jorizzo 2007	Head	5	3	4 or 8	12 or 24	4 weeks after the end of treatment
Ulrich 2007	Face, forehead, or bald scalp	5	3	16	48	8 weeks after the end of treatment
Gebauer 2009	Dorsal of 1 or both forearms and hands	5	2,3,5,7	8	16,24,40,56	8 weeks after the end of treatment
Zeichner 2009	Head	5	1	24	24	4 weeks after the end of treatment
Hanke 2010	Face or bald scalp	2.5, 3.75	7	6 (3 on, 3 off, 3 on)	42	8 weeks after the end of treatment
Jorizzo 2010	Face	3.75	7	4 (2 on, 2 off, 2 on)	28	20 weeks after the end of treatment
Ortonne 2010	Head (bald scalp or face)	5	3	8 (4 on, 4 off, 4 on)	24	At week 20 (6 weeks after the end of treatment)
Swanson 2010a	Face or bald scalp	2.5, 3.75	7	4 (2 on, 2 off, 2 on)	28	8 weeks after the end of treatment
NCT00828568 Aldara	Face or bald scalp	5	2	16	32	8 weeks after the end of treatment
NCT00828568 Taro	Face or bald scalp	5	2	16	32	8 weeks after the end of treatment

Study	Withdrawal due to adverse events (N = 3444 participants)	Skin irritation (N = 1677 participants)	Minor adverse events excluding skin irritation (N = 700 participants)	Cosmetic outcome (N = 1691 participants)
Persaud 2002	x (none lost ‐ intraindividual not included)	‐	‐	‐
Stockfleth 2002	x (none lost)	‐	‐	‐
Chen 2003	x (none lost)	‐	‐	‐
Lebwohl 2004	x	‐	‐	x
Szeimies 2004	x	x	‐	x
Korman 2005	x	‐	‐	x (imiquimod, not included)
Ooi 2006	x (none lost)	‐	x	‐
Alomar 2007	x	‐	‐	‐
Jorizzo 2007	x	‐	‐
Ulrich 2007	x	‐	x (imiquimod only, not included)	x (qualitative, not included)
Gebauer 2009	x	‐	x	‐
Zeichner 2009	x (none lost ‐ intraindividual not included)	Qualitative (not included)	‐	‐
Hanke 2010	x	x	x	x
Ortonne 2010	x (none lost)	‐	‐	‐
Swanson 2010a	x	x	‐	x
NCT00828568 Taro	x	x	‐	‐
NCT00828568 Taro	x	x	‐	‐

Body system	Minor adverse event	Placebo (N = 60)	0.025% ingenol mebutate for 3 days (N = 50)	0.05% ingenol mebutate for 3 days (N = 57)	0.05% ingenol mebutate for 2 days (N = 55)
Body as a whole	Chills	0	0	1	0
Body as a whole	Fever	0	0	0	1
Body as a whole	Flu or cold	0	0	0	1
Dermatologic	Contact dermatitis	0	0	1	0
Dermatologic	Impetigo	0	0	1	0
Hemic and lymphatic	Traumatic hematoma	0	0	1	0
Metabolic and nutritional disorders	Increase in creatine phosphokinase	0	0	0	1
Musculoskeletal and connective tissue	Muscle spasms	1	0	0	1
Nervous system	Headache	0	0	1	0
Renal and urogenital	Proteinuria	0	0	0	1
Respiratory	Nasal congestion	0	1	0	0
Incidences
Each treatment arm		1	1	5	5
3 days versus 2 days			6		5
0.025% versus 0.05%			1	10

Study	Withdrawal due to adverse events (N = 100 participants)	Skin irritation (N = 92 participants)	Minor adverse events excluding skin irritation	Cosmetic outcome
Alirezai 1994	x	x (in general and severe)	‐	‐

Higher versus lower resiquimod concentrations
	Participant complete clearance (after 1 cycle)	Participant complete clearance (after 1 or 2 cycles)	Participant partial clearance (after 1 or 2 cycles)
0.1% vs 0.01%	>	>	>
0.1% vs 0.03%	>	<	=
0.1% vs 0.06%	>	>	>
0.06% vs 0.01%	>	=	=
0.06% vs 0.03%	<	<	<
0.03% vs 0.01%	>	>	>

Minor adverse events excluding skin irritation
Higher versus lower resiquimod concentrations
	Musculoskeletal and connective tissue (in general)				Nervous system (in general)		Skin and subcutaneous tissue disorders (in general)
0.1% vs 0.01%	>				>		>
0.1% vs 0.03%	>				=		>
0.1% vs 0.06%	=				=		<
0.06% vs 0.01%	>				>		>
0.06% vs 0.03%	>				=		>
0.03% vs 0.01%	>				>		<
	Body as a whole		Musculoskeletal and connective tissue		Nervous system
	Fatigue	Rigors	Arthralgia	Myalgia	Headache	Lethargy	Psychiatric disorders
0.1% vs 0.01%	>	>	>	>	>	>	>
0.1% vs 0.03%	=	>	>	>	=	>	=
0.1% vs 0.06%	>	<	<	>	=	=	<
0.06% vs 0.01%	>	>	>	>	>	>	>
0.06% vs 0.03%	<	>	>	=	=	>	>
0.03% vs 0.01%	>	>	=	>	>	>	>

PERMALINK

Interventions for actinic keratoses

Aditya K Gupta

Maryse Paquet

Elmer Villanueva

William Brintnell

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Background

Description of the condition

Disease definition

Clinical Features

Pathogenesis and epidemiology

Description of the intervention

How the intervention might work

Topical Interventions

Diclofenac gel

5‐Fluorouracil (5‐FU)

Imiquimod

Chemical Interventions

Cryotherapy

Photodynamic therapy (PDT)

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Efficacy outcomes

Secondary outcomes

Safety and cosmetic outcomes

Search methods for identification of studies

Electronic searches

Trials Registers

Searching other resources

Unpublished literature

Conference proceedings

Language restrictions

Adverse effects

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Unit of analysis issues

Assessment of heterogeneity

Data synthesis

Subgroup analysis and investigation of heterogeneity

Results

Description of studies

Results of the search

1.

Included studies

Design

Sample sizes

Interventions

Topical treatments

Oral treatments

Mechanical interventions

Chemical interventions

Participants

Outcomes

Efficacy outcomes

Safety outcomes

Cosmetic outcomes

Other outcomes

Excluded studies

Risk of bias in included studies

2.

Allocation

Mean reduction in lesion counts
Study	Intervention	At 3 months	At 6 months	At 12 months
Ostertag 2006	5‐fluorouracil	13.2	12.5	12.4
Ostertag 2006	Er:YAG laser resurfacing	13.8	13.9	14.2

Mean percentage of reduction in lesion counts
Study	Assessment	Resurfacing	5‐FU
Ostertag 2006	At 6 months	94.4%	79.2%
Ostertag 2006	At 12 months	91.1%	76.6%

Study	Withdrawal due to adverse events (N = 55 participants)	Skin irritation (N = 55 participants)	Minor adverse events excluding skin irritation (N = 55 participants)	Cosmetic outcome (N = 55 participants)
Ostertag 2006	x	Overtime	Overtime	x

Study	Withdrawal due to adverse events (N = 297 participants)	Skin irritation (N = 297 participants)	Minor adverse events excluding skin irritation	Cosmetic outcome
Hauschild 2009b	x (none lost)	x	Qualitative	‐

Characteristic	Szeimies 2002	Freeman 2003	Morton 2006	Kaufmann 2008
Study design	Parallel	Parallel	Intraindividual	Intraindividual
Anatomical locations	Face, scalp, others (< 10%)	Face or scalp	Face and scalp	Upper and lower extremities (98%), trunk, neck
Prior preparation of lesions (scale and crust removal)	Cryotherapy: yes PDT: yes	Cryotherapy: no PDT: yes	PDT: yes	PDT: yes (except mild lesions = 12%)
Number of treatment cycle	1 (face and scalp) or 2 (other locations)	Cryotherapy : 1 PDT: 2	1 or 2	1 or 2
Number of weeks between treatments	1	1	12	12
Number of freeze‐thaw cycles per treatment	2	1	2	2
Total freezing time (sec)	24 + 18	12 to 26	16	20 + 14
Individual lesion or field‐directed treatment (MAL)	Individual lesions	Individual lesions	Individual lesions	Individual lesions
Occlusion time with 16% MAL (hour)	3	3	3	3
PDT intensity (mW/cm²)	70 to 200	50 to 250	N/A	N/A
PDT dose (J/cm²)	75	75	37	37
Type of light source	Non‐coherent light	(CureLIght)	LED (Aktilite CL 128 lamp)	LED (Aktilite CL 128 lamp)
Time of assessment	12 weeks after the end of treatment	12 weeks after the end of treatment	12 weeks after the end of treatment	12 weeks after the end of treatment

Mean percentage of reduction in lesion counts
Study	Assessment at	Cryotherapy	MAL‐PDT
Kaufmann 2008	12 weeks	N/A	N/A
Kaufmann 2008	24 weeks	87%	75%
Morton 2006	12 weeks	74.5%	84.4%
Morton 2006	24 weeks	83.9%	86.7%

Parameter	Szeimies 2002 (N = 122 participants)	Freeman 2003 ( N = ? participants)	Morton 2006	Kaufmann 2008
Evaluation by investigator	X	X	X	X
Evaluation by participant	X	X	N/A	N/A
Outcome	1) excellent or good 2) fair or poor	Excellent	1) excellent 2) good 3) fair 4) poor	1) excellent 2) good 3) fair 4) poor
Reported per participant	X (only for participants with > 75% reduction of total lesions)	X (only for participants with 100% reduction of total lesions)	N/A	N/A
Reported per lesion	N/A	X	X	X

	Blue light		Red light
Characteristic	Jeffes 2001	Piacquadio 2004	Hauschild 2009a and Hauschild 2009b	Szeimies 2010b
Study design	Assessor‐blinded intraindividual	Assessor‐blinded parallel	Double‐blinded parallel	Double‐blinded parallel
Anatomical locations	Face and scalp	Face or scalp	Head	Face, bald scalp, or both
Prior preparation of lesions (e.g. scale and crust removal)	N/A	N/A	No	Yes
Number of treatment cycle	1 or 2	1 or 2	1	1 or 2
Number of weeks between treatments	8	8	N/A	12
Individual lesion or field‐directed treatment	Individual lesions	Individual lesions	Individual lesions	Individual lesions
ALA formulation	20% cream	20% cream	Patch containing 8 mg	BF‐200 gel
Occlusion time (hour)	14 to 18	14 to 18	4	3
PDT intensity (mW/cm²)	3, 5, 10	10	N/A	Aktilite: 50‐70 PhotoDyn 750: 196
PDT dose (J/cm²)	2, 5,10	N/A	37	Aktilite: 37 PhotoDyn 750: 170
Illumination time (seconds)	N/A	1000	N/A	Aktilite: N/A PhotoDyn 750: 900
Type of light source	Non‐laser fluorescent (Dusa BLU‐417)	visible (Blu‐U)	LED (Aktilite CL 128 lamp or Omnilux)	LED (Aktilite CL 128 lamp) or incoherent (PhotoDyn 750)
Time of assessment	8 weeks after the end of each treatment	8 weeks after the end of each treatment	12 weeks after the end of each treatment	12 weeks after the end of each treatment

Study	Withdrawal due to adverse events (N = 701 participants)	Skin irritation (N = 300 participants)	Minor adverse events excluding skin irritation (N = 543 participants)	Cosmetic outcome (see table below)
Jeffes 2001	x (none lost)	‐	‐	x
Piacquadio 2004	x (none lost)	‐	x	x
Hauschild 2009a Hauschild 2009b	x (none lost)	x	x	x
Szeimies 2010b	x (none lost)	‐	Intraindividual study not included in meta‐analysis	x

Assessment at (months)	MAL‐PDT	5‐fluorouracil
1	4/4	0/4
3	4/4	1/4
6	4/4	1/4