Interventions for preventing keratinocyte cancer in high-risk groups not receiving immunosuppressive therapy

  • Protocol
  • Intervention


  • Martha Alejandra Morales-Sánchez,

    Corresponding author
    1. Dermatological Center, "Dr. Ladislao de la Pascua", Education and Research Unit, México City, Mexico
    • Martha Alejandra Morales-Sánchez, Education and Research Unit, Dermatological Center, "Dr. Ladislao de la Pascua", Dr. José María Vértiz No. 464 Col. Buenos Aires, México City, 06780, Mexico.

  • María Luisa Peralta-Pedrero,

    1. Dermatological Center, "Dr. Ladislao de la Pascua", Education and Research Unit, México City, Mexico
  • Fermín Jurado-Santa Cruz,

    1. Dermatological Center, "Dr. Ladislao de la Pascua", Education and Research Unit, México City, Mexico
  • Hyemin Pomerantz,

    1. Hofstra Northwell School of Medicine, Department of Dermatology, Hempstead, New York, USA
  • Leticia A Barajas-Nava

    1. Hospital Infantil de México Federico Gómez (HIMFG), Health National Institute, Evidence-Based Medicine Research Unit, México City, Mexico
    2. Iberoamerican Cochrane Network, Barcelona, Barcelona, Spain


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

To determine the efficacy and safety of interventions for preventing keratinocyte cancer in high-risk groups not receiving immunosuppressive therapy.


Description of the condition

Keratinocyte cancer (KC) is the most diagnosed malignancy worldwide in white populations (Perera 2015). The term 'keratinocyte cancer' refers to basal cell carcinoma (BCC) and cutaneous squamous cell carcinoma (cSCC), which account for 80% of skin cancer cases (Albert 2003). The term 'non melanoma skin cancer' (NMSC) includes BCC; cSCC; and other skin cancers, such as Merkel cell cancer.

Keratinocyte cancer arises from keratinocytes, the cells of the outer layer of the skin, which is called the epidermis (Albert 2003). The incidence of keratinocyte cancer (KC) varies around the world, but Australia has the highest incidence with more than 1000 cases per 100,000 person-years, while Africa has the lowest incidence, reporting less than one case per 100,000 person-years (Lomas 2012). The USA estimates that 3.5 million cases of KC were diagnosed in 2010 compared with 1.63 million cases of all other cancers (Perera 2013). However, the prevalence of skin cancer is underestimated because basal cell carcinoma (BCC) is not routinely captured in national registries of cancer.

Both BCC and cSCC arise from keratinocytes, but BCC originates specifically from basal keratinocytes (Chen 2013), and cSCC from squamous cells (Rahimi 2013). Basal cell carcinoma grows slowly and rarely metastasises, while cSCC is more invasive with higher metastatic potential (Rahimi 2013). In fact, deaths from invasive cSCC are equivalent to the number of deaths from renal cancer and melanoma (Karia 2013).

Basal cell carcinoma appears in sun-exposed areas, such as the head and neck, while cSCC also develops on the backs of the hands, arms, and trunk (Telfer 2008). Basal cell carcinoma has different clinical variants, such as superficial, nodular, cystic, and sclerosing or scar-like lesions, all of which can have pigment or ulceration (Telfer 2008). On the other hand, cSCCs are exophytic tumours with erythema (redness), scales, and a warty appearance (Warszawik-Hendzel 2015).

If keratinocyte cancer is diagnosed at an early stage when the tumour is small, it can be removed leaving a small scar with an acceptable cosmetic result for the individual. However, when the tumour grows, it destroys more skin, can cause disfigurement, and has the potential to spread to other organs, particularly in the case of cSCC. The larger the KC is the more complex its treatment will be because after surgical removal of the tumour, the wound has to be repaired using a skin graft or flap (NCCN 2014).

The incidence rates of KCs are increasing steeply, creating a huge burden to those affected and to healthcare providers (Guy 2015). Whilst mortality is low compared with other cancers, morbidity of surgical procedures on cosmetically sensitive body sites is very high (Gordon 2015). A recent systematic review found that the USA spends the most funding on skin cancer, followed by Australia, Germany, and the UK. However, in terms of population size, Australia and New Zealand have the highest cost burden for skin cancer, followed by Denmark and Sweden (Gordon 2015). The total annual cost of keratinocyte cancer has been estimated to be 500 million dollars in Australia (Fransen 2012).

It is a challenge for health systems worldwide to manage people at high risk of KC because of the rapidly increasing burden of disease, and for cSCC, there is also the increased chance of a poor outcome, such as loco-regional metastasis or disease-specific death. The cost burden increases exponentially due to the need for multiple surgical procedures and the morbidity associated with these procedures. So, in this group of people, as well as early diagnosis of KC and medical follow-up, ideally, it is preventive interventions that are the key to reducing the cost burden of medical care for the community and maintaining quality of life for the individual (Bath-Hextall 2007).

Causes and risk factors for KC

Ultraviolet (UV) light is the most important risk factor for KC because it causes the malignant transformation of keratinocytes (Bauer 2011). People are exposed to UV light during occupational or recreational outdoor activities. Ultraviolet B (UVB) (290 to 320 nanometers of wavelength) is responsible for most of the changes in DNA (deoxyribonucleic acid) that cause modifications in skin cells; ultraviolet A (UVA) (320 to 440 nanometers) can also augment the carcinogenic effects of UVB (Mancebo 2014). Ultraviolet light exposure induces the dimerisation of pyrimidines and breaks in the DNA strands (Table 1), the changes to which may promote carcinogenesis in keratinocytes (Pfeifer 2013). In fact, markers of photodamage to the skin, like actinic keratosis, solar elastosis (abnormal elastic tissue in the skin), solar lentigines, and telangiectasia (small dilated blood vessels in the skin, also known as spider veins), have been associated with high risk for KC, mainly for BCC (Khalesi 2013). Thinning of the ozone layer and geographical altitude and latitude are variables that modify the sun exposure of individuals (Lucas 2015). Epidemiological studies have shown that chronic cumulative sun exposure is the major risk factor for KC and even occupational exposure increases this risk to 77% in individuals (Schmitt 2011). Cumulative sun exposure increases with age, so the risk is higher in the elderly. Other individual characteristics that predispose people to develop BCC are red hair, fair skin colour, having skin that burns and never tans, freckling in childhood, and having melanocytic nevi on the arms (Khalesi 2013b).

Table 1. Glossary
5-aminolevulinate (ALA)A chemical compound produced during the synthesis of the haemoglobin used for the treatment of actinic keratosis
AzathioprineAn immunosuppressant, a drug that inhibits the function of the immune system (also called the defence system), used in organ transplant receptors to avoid rejection and in other autoimmune diseases
BasalThe term used to refer to the first layer of the epidermis, the basal cell layer
Chronic fistulaeAn abnormal passage or connection between an organ like the gut and another structure like the skin that lasts months
Chronic sinuses of osteomyelitisCavities that were formed after a bone infection due to the destruction of the bone tissue
CongenitalSomething (characteristic or disease) that exists from birth
Curettage & electrodesiccationA procedure in which a dermatologist scrapes a skin lesion with an instrument called a curette, and after that, with an electrosurgical device that destroys the remaining lesion, burning the tissue. This procedure is repeated until the dermatologist has completely removed the skin lesion
CysticRelating to a cyst, a tissue that resembles a sac whose content may be air, liquid, or a solid material like keratin
Dimerisation of pyrimidinesLinkages or double bonds formed in the nitrogen bases of the DNA due to exposure to ultraviolet light; these changes in DNA are considered molecular lesions that can evolve into a skin cancer. Thymine and cytosine are nitrogen bases of DNA called pyrimidines
Epidermodysplasia verruciformisAn inherited skin disease characterised by a chronic infection caused by the human papillomavirus
Epigenetic regulatorsEnzymes that modulate the methylation of DNA and may modify gene expression and have an impact on cancer prevention
Exophytic tumourA tumour that grows outward, i.e. warty in appearance
Follicular atrophodermaA skin condition characterised by follicular openings without hair on the hands, arms, and legs
Genus beta HPV seropositivityHaving blood serum antibodies against human papillomavirus of the genus beta
Histone acetylationThe process of gene regulation that increases gene expression by transferring acetyl groups to histones (proteins that package the DNA)
HyperkeratoticThickening of the skin by increasing the quantity of the stratum corneum, the upper layer of the skin
HypotrichosisReduction of body and scalp hair due to the change of vellus instead of terminal hair
IsoformVariant of a specific protein that is produced by the same gene
Melanocytic neviBenign tumour of the skin composed of melanocytes, the cells of the skin that produce pigment or colour. They are also called skin moles
MeningiomaBenign tumours arising from the membranes that cover the brain and spinal cord
MethylationAddition of a methyl group (1 carbon atom bonded to 3 hydrogen atoms) to a chemical compound
Methyl aminolevulinate (MAL)A drug that changes to protoporphyrin IX with exposure to red light, used in photodynamic therapy
NodularRelating to nodules or skin protuberances
Nucleotide excision repair (NER) pathwayMolecular mechanism of the DNA to repair the damage produced by physical or chemical agents, like ultraviolet radiation
OsteomyelitisInfection of the bone that may cause tissue death
OsteosarcomaA malignant tumour of the bone
PoikilodermaA skin condition characterised by having telangiectasia, areas of depigmentation, or hyperpigmentation and thinning of the skin
PolyphenolsAntioxidant derived from plants that avoids the damage caused by free radicals
ProstaglandinsChemical compounds, produced in almost all human cells, that activate platelets and endothelium and mast cells, and whose levels are increased in inflammation processes
Protoporphyrin IXOrganic compound produced by the action of the enzyme protoporphyrinogen oxidase during the synthesis of the heme group for the haemoglobin
Pyrimidine dimersMolecular damage to DNA caused by ultraviolet radiation
Pyrimidines1 of the 2 kinds of nitrogen bases that form DNA; the pyrimidines are uracil, thymine, and cytosine
SclerosingCharacterised by induration of the skin due to the increase of collagen in the skin
SeroconversionProduction of antibodies in response to a specific antigen (like in viral infections)
SeropositivityA positive result in a laboratory test for the presence of a specific antibody
Solar lentiginesPigmented spots in the skin due to chronic sun exposure
SuperficialRelating to the surface of the skin
TelangiectaticRelating to the dilation of the superficial and small blood vessels of the skin
TeratogenicAnything that disturbs the normal growth and development of the fetus

Human papillomavirus (HPV) infection may play a role in KC development (Farzan 2013). Recently, a study showed weak associations between seropositivity for HPV and keratinocyte cancer risk; HPV type 5 (HPV-5) seroconversion was strongly associated with risk of cSCC (odds ratio (OR) 3.2, 95% confidence interval (CI) 1.3 to 7.6) (Faust 2016). HPV type 5 was also associated with the development of cSCC in people with epidermodysplasia verruciformis (Khalid 2014). The role of HPV infection and immunosuppressive therapy in carcinogenesis is related to the immune system, which should be able to detect the early development of tumours and destroy them (Yu 2014). In fact, any circumstance that causes impairment of the immune system may facilitate the development of skin cancer (Yu 2014), although the nucleotide excision repair (NER) pathway generally repairs skin damage from sun exposure to prevent the development of cancer (Rass 2008). However, there are rare genetic disorders, like xeroderma pigmentosum, that predispose to skin cancer because they originate by mutations in the NER pathway (Rass 2008). Recently, researchers have found chromosomal abnormalities in BCC and cSCC that contribute to carcinogenesis (Carless 2014).

High-risk groups

Certain types of people are more likely to develop KC than the general population if they have a condition that induces the formation of a tumour or impairs the immune system (Bath-Hextall 2007). Below we list the known high-risk groups for skin cancer, but for this systematic review, we will focus on those not receiving immunosuppressive therapy.

Personal history of KC

People diagnosed with KC are at an increased risk of developing a further skin cancer (Flohil 2013). A recent meta-analysis found that for people with BCC, the risk of developing another BCC within the following five years is 36.2% (11.0% to 49.9%); for developing another cSCC after cSCC, it is 37% (30% to 50%); and for developing either BCC or cSCC after a previous KC, it is 36.2% (22.3% to 50%) (Flohil 2013). A previous study estimated that the risk of a subsequent KC is 38% for people with one or two KCs but increases to 93% if they were diagnosed with three to nine prior KCs (Marcil 2000). Individuals with 10 or more KCs will develop a new tumour within two years (Marcil 2000).

Diagnosis of precursor lesions

People with precursor lesions, like actinic keratosis (AK) and Bowen's disease (BD), are at a higher risk of developing a cSCC (Tsatsou 2012). Actinic keratosis is an intraepithelial keratinocytic dysplasia (skin with changes in the nuclei of the keratinocytes) with a progression rate to cSCC from 0% to 0.075% per lesion-year (Werner 2013). In those with a previous KC, the progression risk increases to 0.53% (Werner 2013). "The typical patient has 6 to 8 AK; therefore, a patient with multiple AK has a [sic] annual risk of developing invasive cSCC ranging from 0.15% to 80%" (Ratushny 2012).

Bowen's disease or cSCC in situ

People with Bowen's disease are at risk of their lesions evolving from cSCC in situ to the invasive form (Peterka 1961). The rate of progression of Bowen's disease to invasive cSCC is from 3% to 5% (Kao 1986). Bowen's disease has a good prognosis because it is a slow-growing lesion that responds to topical treatments, like photodynamic therapy, 5-fluorouracil, and imiquimod (Bath-Hextall 2013).

Long-standing ulcer or scar

People suffering from chronic wounds that cause chronically inflamed skin are at risk of KC (Menendez 2006). Cutaneous squamous cell carcinoma arises from previously traumatised and scarred tissue from burns, chronic sinuses of osteomyelitis, post-traumatic wounds, pressure sores, chronic fistulae, and as a complication of Fournier's gangrene (an infection of the perineum that causes tissue death) (Chalya 2012). The term Marjolin's ulcer is used to describe tumours arising in burn scars, of which 75% to 96% of cases lead to cSCC (Chalya 2012). The incidence of Marjolin's ulcers ranges from 1% to 2% in all burn scars (Ochenduszkiewicz 2006). The mean time between injury and the onset of malignant transformation is 11 years (Ochenduszkiewicz 2006).

Personal history of long-term psoralen and ultraviolet A (PUVA) treatment

Psoralen and ultraviolet A (PUVA) is a widely used treatment for psoriasis, vitiligo, atopic dermatitis, and mycosis fungoides. However, long-term exposure to PUVA increases the risk of cSCC because it has a mutagenic and immunosuppressive effect on the skin (Beissert 2002; Stern 1998). People exposed to high-dose PUVA (more than 200 treatments or 2000 J/cm2) have a 14-fold higher risk of developing cSCC during the following five-year period than those exposed to low-dose PUVA (less than 100 treatments or fewer than 1000 J/cm2) (Stern 1998). The risk of cSCC and BCC in people with psoriasis increases linearly with the number of sessions and persists after discontinuation of treatment (Archier 2012).

Chronic arsenic exposure

The World Health Organization (WHO) estimates that more than "200 million persons worldwide might be chronically exposed to arsenic in drinking water at concentrations above the safety standard of 10 micrograms/L" (Naujokas 2013). The International Agency for Research on Cancer (IARC) defines arsenic as a human carcinogen (IARC 2012). Data from epidemiological studies have revealed that arsenic in drinking water is a health problem in countries like Argentina, Bangladesh, Chile, China, Ghana, India, Mexico, Taiwan, Vietnam, and some regions in the United States. Arsenical keratoses are cutaneous markers of chronic arsenic exposure whereas darkening of the skin or hyperpigmentation is the initial manifestation (Hong 2014). Arsenical keratoses are wart-like skin lesions from 2 nm to 4 mm that appear mainly on the palms and soles, and a sudden increase in their size or bleeding suggests malignant transformation to cSCC (Naujokas 2013). Basal cell carcinoma is also associated with chronic arsenic exposure (IARC 2012).

History of radiation therapy

Individuals who received a dose of 35 Gy or more of radiation therapy have a 40 times higher risk of developing KC years or even decades after the treatment (Walt 2012). This evidence comes from the study of Japanese atomic bomb survivors and from the children treated with ionising radiation for tinea capitis (Boaventura 2012; Kodama 1996). The risk is mainly for BCC (Walt 2012).


Genodermatoses are inherited genetic skin diseases with systemic involvement. A minority of cases of KC occur in individuals with the following hereditary cancer syndromes (Bath-Hextall 2007).

Xeroderma pigmentosum (XP)

Xeroderma pigmentosum is a rare autosomal recessive disease that compromises the repair of DNA damage by ultraviolet radiation (DiGiovanna 2012). It is characterised by sun sensitivity and the development of skin cancers at a young age (Lehmann 2011). Affected individuals are unable to repair UV-induced DNA damage and develop BCCs and cSCCs at a median age of eight years (Naik 2013). The increase in the risk for KC is 1000-fold in people under the age of 20 years (Kraemer 2014).

Oculocutaneous albinism (OCA)

Oculocutaneous albinism is a rare group of congenital disorders characterised by decreased or absent pigmentation in the skin, hair, and eyes (Mabula 2012). The risk of KC, cSCC mainly, has been reported to be as high as up to 1000-fold when compared with the general population (Mabula 2012). People with albinism develop KC before the age of 40 years, and the most frequent site is the head and neck followed by the limbs (Mabula 2012).

Epidermolysis bullosa (EB)

Epidermolysis bullosa comprises a group of congenital disorders characterised by blister formation (Intong 2012). Only junctional and dystrophic EB are associated with KC (Fine 2009). The cumulative lifetime risk of cSCC has been estimated to be from 18% to 25% in those with junctional EB (Fine 2009; Yuen 2011). Data from the National EB Registry in the US have shown that by the age of 45, 85% of people with this condition will have at least one cSCC, and by the age of 20, their cumulative risk is 6% (Fine 2008).

Epidermodysplasia verruciformis (EDV)

Epidermodysplasia verruciformis is a rare heritable disease characterised by an extensive infection with HPV (Gül 2007). HPV type 5 and HPV type 8 (HPV-8) are major types associated with the development of cSCC on sun-exposed skin during the third or fourth decades of life in about 20% to 30% of people with this condition (Khalid 2014).

Fanconi anaemia (FA)

Fanconi anaemia is a genetic disease whose mutation affects the FA pathway, which is responsible for the repair of DNA damage known as interstrand cross-links (Auerbach 2009). Mutation in the FA pathway blocks DNA replication (Auerbach 2009). People with FA have a 500- to 700-fold increase in the incidence of cSCC (Scheckenbach 2012; van Monsjou 2013), mainly in the oral cavity. The risk of being diagnosed with head and neck cancer is around 14% for those who live to the age of 40 (Romick-Rosendale 2013).

Dyskeratosis congenita (DC)

Dyskeratosis congenita is characterised by a triad of abnormal nails, net-like skin pigmentation, and white oral lesions (Ballew 2013). Dyskeratosis congenita may be inherited as X-linked, autosomal dominant or recessive (Kirwan 2009). People with DC are at high risk of developing cancer, such as cSCC of the head and neck (Khincha 2013). Their cumulative risk of cancer is 40% by 50 years of age, with a 1100-fold increased risk of cSCC of the tongue (Alter 2010). Cutaneous cSCC has been reported in 1.5% of cases and tends to appear at younger ages (Alter 2010).

Rothmund-Thomson syndrome (RTS)

Rothmund-Thomson syndrome is an autosomal recessive disorder characterised by redness, swelling, and blistering of the face, ears, buttocks, and extensor surface of the extremities, which appears at six months of age and results in poikiloderma two years later (Larizza 2010; Wang 2001). Mutations in the RECQL4 gene are found in 40% to 66% of patients (Lindor 2000). RECQL4 encodes a DNA helicase, an enzyme that is involved in DNA repair (Mehollin-Ray 2008). The second most common cancer that those with this syndrome develop is cSCC, followed by BCC, Bowen's disease, and verrucous carcinoma (Castori 2012).

Bloom syndrome (BS)

Bloom syndrome, also known as congenital telangiectatic erythema, is a rare autosomal recessive disorder characterised by a triad of erythema, photosensitivity, and intrauterine growth restriction (Arora 2014). Mutations in the BLM gene encoding of a DNA helicase involved in DNA replication are responsible for the disorder (Ellis 1996). Affected individuals are 150 to 300 times more likely to develop malignancy involving the gastrointestinal tract, genitalia, urinary tract, and skin (Arora 2014). Multiple BCCs of the face and neck occur in people with BS at a mean age of 31.7 years (Sanz 2013).

Werner syndrome (WS)

Werner syndrome is an extremely rare premature-aging disease, and affected individuals have a shorter lifespan and an increased risk of malignancy (Lauper 2013; Sugimoto 2014). The most frequent neoplasms in people with the syndrome are thyroid carcinoma, malignant melanoma, meningioma, soft tissue sarcomas, leukaemia, and osteosarcoma (Goto 2013). However, cSCC and BCC account for 4.8% of all malignancies (Lauper 2013).

Nevoid basal cell carcinoma syndrome (NBCCS)

Nevoid basal cell carcinoma syndrome, also referred to as Gorlin-Goltz syndrome (GGS), is due to mutations in the PTCH (patched tumour suppressor) gene (Kiwilsza 2012; Scully 2010). People develop multiple BCCs, jaw cysts, and palmar and plantar pits (Lo Muzio 2008), and they may present with from one to more than 100 BCCs, having a median number of eight with a diameter of 1 mm to 10 mm, mainly on the face, neck, and trunk (Nikolaou 2012).

Rombo syndrome (RS)

Rombo syndrome is a X-linked dominant disorder characterised by hair twists (pili torti), facial milia, and multiple benign hair follicle tumours called trichoepithelioma (Michaelsson 1981). Patients develop atrophy around hair follicles (atrophoderma vermiculatum) and photosensitivity (Michaelsson 1981). They are at an increased risk of developing basal cell carcinoma (van Steensel 2001). People with RS develop BCCs around the age of 35 depending on the time of ultraviolet exposure (Michaelsson 1981; Parren 2011).

Basex-Dupré-Christol syndrome (BDCS)

Basex-Dupré-Christol syndrome is a X-linked dominant disorder characterised by the triad of hypotrichosis, follicular atrophoderma, and BCCs (Abuzahra 2012). Basal cell carcinomas develop after the first decade of life (Parren 2011).

Clinical features

Basal cell carcinoma is the most common type of skin cancer in Australia and appears as a pearly pink dome-shaped tumour with telangiectatic vessels in its surface (Cancer Australia & AIHW 2008). It has five major histological patterns: nodular, superficial, micronodular, infiltrative, and morpheaform. In 90% of cases, the tumour occurs on the face (Firnhaber 2012). Slow growth and low capacity for metastatic spread characterises basal cell carcinoma (NCCN 2014). Skin biopsy, in which the typical nests of basaloid cells are found, confirms the diagnosis (Dubas 2013).

Cutaneous squamous cell carcinoma is the second most frequently occurring skin cancer, and unlike BCC, actinic keratosis is its precursor, which is also called a precancerous lesion or keratinocyte dysplasia (Ratushny 2012). Cutaneous squamous cell carcinoma appears as a firm hyperkeratotic tumour, with or without ulceration, and in more than 50% of cases, occurs on the face followed by the upper extremities (Ratushny 2012). It spreads by local infiltration and has the capacity for lymphatic and haematogenous dissemination (Dubas 2013). Actinic keratoses are found predominantly in fair-skinned people and may progress to cSCC at a rate estimated to be between 0.025% and 16% per lesion per year (Werner 2013).


The diagnosis of KC is through physical examination of the skin, and medical history may reveal the risk factors for skin cancer. If a clinician finds a lesion suspected to be KC, a biopsy should be performed, and histology assessment is able to determine the type of KC, BCC, or cSCC (NCCN 2014). The whole physical examination is performed to determine if the KC has been disseminated and to locate the sites of metastasis (Grégoire 2010).

Description of the intervention

The review will focus on interventions for preventing keratinocyte cancer. Prevention in high-risk groups will help to reduce the need for surgical excision; minor surgical procedures, such as curettage and shave excision (Grégoire 2010; Mydlarz 2015; NCCN 2014); and other destructive therapies, such as liquid nitrogen application for low-risk tumours that allow the possibility of treating multiple lesions at a single patient visit (NCCN 2014), all of which are the current treatment options to treat keratinocyte cancer.

We will consider pharmacological, educational, dietary, and photodynamic treatments that aim to reduce the incidence of keratinocyte cancer in high-risk groups not receiving immunosuppressive therapy. In general, these interventions will be able to protect the skin from sun damage and will also reverse the skin changes induced by ultraviolet radiation (Bath-Hextall 2007). The promotion of the use of sunscreen by educational interventions are part of primary and secondary prevention procedures in high-risk groups (Hirst 2012). Topical and oral pharmacological treatments, such as retinoids, non-steroidal anti-inflammatory drugs, and antioxidants, may also have the capacity to inhibit the growth of keratinocyte cancer (Bath-Hextall 2007). In fact, photodynamic therapy is used to treat precursor lesions of skin cancer (Vergilis-Kalner 2013).

How the intervention might work

Although skin cancer is the most common cancer, it is also a preventable cancer because the main risk factors may be avoidable. Prevention strategies are classified as primary, secondary, and tertiary, and their objectives are as follows: to reduce the risk of KC; diagnose the disease in early stages; and avoid recurrences, respectively. All of the strategies are necessary to prevent KC in high-risk groups.

Topical treatments


These chemical or physical agents protect the skin from sun exposure (UVA and UVB) by absorbing or blocking ultraviolet radiation, and their measure of protection is provided by the sun protection factor (SPF) (Bens 2014). Dermatologists recommend applying a sunscreen with a SPF of 30 or higher every two hours when doing outdoor activities (Latha 2013). The daily use of a broad-spectrum sunscreen is a cost-effective intervention to reduce the development of skin cancers, especially cutaneous squamous cell carcinoma (Hirst 2012).

The Nambour Skin Cancer and Actinic Eye Disease Prevention Trial demonstrated that the application of a sunscreen with SPF 16 every morning decreases the incidence of cSCC after eight years of use (Green 1999). The incidences of cSCC and BCC in the group randomised to use the sunscreen daily decreased by 35% and 25%, respectively (van der Pols 2006). In people who had received organ transplants, the use of sunscreen at least 5.6 times per week decreased the incidence of invasive cSCC and the count of new AKs (Ulrich 2009).

DNA repair enzymes

A previous Cochrane systematic review found that topical T4N5 liposome lotion containing DNA repair enzymes reduced the incidence of AKs and BCCs in people with XP (Bath-Hextall 2007). Photolyase, a DNA repair enzyme, was tested in 10 volunteers to assess if it could prevent the formation of pyrimidine dimers (Berardesca 2012). After being exposed to ultraviolet radiation, the skin protected with photolyase and sunscreen produced less photo products than the skin protected only with sunscreen (Berardesca 2012). The TPF50 is a topical product that contains sunscreen SPF 50, DNA repair enzymes (photolyase, endonuclease, and 8-oxoguanine glycosylase), and antioxidants (carnosine, arazine, ergothionine) (Emanuele 2014). This chemical compound applied to the skin diminishes the DNA damage caused by ultraviolet radiation exposure (Emanuele 2014).


Retinoids are synthetic derivatives of vitamin A (Micali 2014). Topical retinoids include tretinoin, adapalene, and tazarotene, while oral retinoids include isotretinoin, alitretinoin, etretinate, acitretin, and bexarotene (Micali 2014). Oral retinoids have been used for chemoprevention of skin cancer in those at high risk (Bettoli 2013). Isotretinoin has been used for xeroderma pigmentosum and nevoid basal cell carcinoma syndrome; and acitretin, for transplant recipients (Bettoli 2013; Marquez 2010). The mechanism of action involved in the prevention of KC is not well understood, but it has been demonstrated that retinoids inhibit cell growth and induce apoptosis (scheduled cell death) of skin cells (Alizadeh 2014; Ianhez 2013). This effect may inhibit tumour formation (Alizadeh 2014). However, the results of the Veterans Affairs Topical Tretinoin Chemoprevention Trial showed that the application of 0.1% tretinoin for 1.5 to 5.5 years did not have an effect on the incidence of BCC or cSCC or on the AK count (Weinstock 2012). The adverse effects of topical retinoids, like skin irritation manifested by erythema and scaling, can worsen after 12 months of daily use (Weinstock 2012). Systemic administration of retinoids may cause dryness of the skin, nose, lips, and eyes; and may also increase liver enzymes, cholesterol, and triglycerides (Micali 2014). Exposure to retinoids during pregnancy may be teratogenic (Micali 2014).

Non-steroidal anti-inflammatory drugs (NSAIDs)

Non-steroidal anti-inflammatory drugs are agents with anti-inflammatory, analgesic, antipyretic (reduce or avoid fever), and platelet-inhibitory properties; their function is to block the synthesis of prostaglandins by inhibiting cyclooxygenase, an enzyme with two isoforms: COX-1 and COX-2 (Liebman 2013). Non-steroidal anti-inflammatory drugs may prevent KC because they inhibit the inflammation cascade, which plays a role in the development and progression of tumours (Liebman 2013). In the SKICAP-AK trial (part of the retinoid skin cancer prevention trials), NSAID use showed a protective effect against BCC and cSCC (hazard ratio (HR) 0.43, 95% confidence interval (CI) 0.25 to 0.73 and HR 0.49, 95% CI 0.28 to 0.87) (Clouser 2009). By contrast, a meta-analysis of several studies in 2014 found that NSAIDs had no protective effect against BCC or cSCC. Even when any NSAIDS, aspirin, non-aspirin NSAIDs, or celecoxib were compared with no NSAID, there was no difference in effect (Zhang 2014). However, in 2015 another meta-analysis found that the use of NSAIDs reduced the risk of cSCC from 6% to 22%, mainly in people with previous AK (Muranushi 2015). Topical diclofenac, a COX-1 and COX-2 NSAID, has been successfully used for the treatment of AK, so chemoprevention with COX inhibitors may prevent KC in high-risk people (Liebman 2013).


Reactive oxygen species (ROS) generate free radicals that cause permanent damage to DNA and act as a trigger for tumour development (Landry 2014). Reactive oxygen species modulate the activity of oncogenes, which participate in the initiation, promotion, and progression of cancer (Landry 2014). Antioxidants are compounds that are able to donate electrons and neutralise free radicals, avoiding cell damage (Pisoschi 2015). So, antioxidants can provide some benefits for the prevention of cancer, especially in the early stages of carcinogenesis (Saeidnia 2013). There are synthetic and natural antioxidants; the latter are obtained from the diet (Saeidnia 2013). The most important antioxidants are vitamin E, vitamin C, alpha lipoic acid, coenzyme Q10, flavonoids, carotenoids, and glutathione (Saeidnia 2013; Wölfle 2014). Topically applied vitamin E or dl-alpha tocopherol has no effect on AKs; however, it decreases the levels of the chemical compounds that promote tumour growth (like the polyamines) (Foote 2009). Topically applied vitamin C, alone or in combination with vitamin E, before ultraviolet radiation exposure can decrease the formation of pyrimidine dimers that are associated with skin cancer (Murray 2008; Oresajo 2008). Supplementation with beta-carotene in individuals with low levels of beta-carotene has no protective effect on the development of BCC and cSCC according to some studies (Dorgan 2004; Schaumberg 2004). In fact, the SU.VI.MAX (Supplementation en Vitamines et Minéraux Antioxidants) study measured the efficacy of daily supplementation with antioxidants in reducing cancer and cardiovascular diseases (Hercberg 1998). The participants were randomised to receive a daily supplementation of 120 mg of vitamin C, 30 mg of vitamin E, 6 mg of beta-carotene, 100 µg of selenium, and 20 mg of zinc, or a placebo, for a period of 7.5 years (Hercberg 1998). After this time, the incidence of all cancers decreased in men but not in women, and this effect disappeared at the five-year follow up (Hercberg 2004; Hercberg 2010). Contrary to what researchers expected, in women, the incidence of skin cancer increased and after the interruption of supplementation, the incidence decreased (Ezzedine 2010). Finally, a recent Cochrane Review found that selenium was associated with an increase in the risk of KC (risk ratio (RR) 1.44, 95% CI 0.95 to 1.17) (Vinceti 2014).

Photodynamic therapy (PDT)

Photodynamic therapy is a novel treatment that consists of applying a photosensitiser, like methyl aminolevulinate (MAL) or 5-aminolevulinate (ALA), to the skin, followed by an incubation period to let these agents be converted to protoporphyrin IX inside the skin cells (Vergilis-Kalner 2013). After this period, the skin is illuminated by blue light to activate the photosensitisers in order to destroy premalignant lesions of AK (Vergilis-Kalner 2013). Photodynamic therapy may prevent the development of KC by completely clearing 76% of AKs on the skin, but the recurrence rate is 24% after 12 months of treatment (Tschen 2006). Photodynamic therapy with MAL completely removes the AKs of organ transplant recipients after one or two cycles of treatment according to a case series of 16 patients (Hasson 2012). A recent trial assessed the efficacy of PDT, using a novel ALA, to prevent the development of new skin cancers in people with a previous KC; however, after three years, both groups had almost the same incidence of new skin cancers (Dixon 2014). In another study of renal transplant recipients who received PDT, this treatment achieved a decrease in the number of new AKs (Togsverd-Bo 2015).

Educational interventions

Behavioural recommendations for preventing KC include the following: avoiding the sun from 10 am to 4 pm (and avoiding potential sunburn), using sunscreen, wearing protective clothing and sunglasses when outside, and seeking shade when outdoors (Agbai 2014). The practice of these preventive behaviours may be achieved by educational interventions that include the participation of community; mass media campaigns; and environmental changes, such as providing sunscreen and shade for outdoor activities (Lin 2011). The U.S. Preventive Services Task Force has systematically reviewed educational interventions and concluded that the best strategies to promote sun-avoidance behaviours are those focused on primary care settings and schools (Lin 2011).

Dietary modifications

Diet can be a potential intervention for preventing cancer because its components are nontoxic and can be consumed without interruption (Saha 2013). Two of the most studied components of the diet are sulforaphane (SFN) from broccoli and epigallocatechin-3-gallate (EGCG) from green tea. These are known as epigenetic regulators of skin cancer cell function; they interfere with DNA methylation and histone acetylation (Saha 2013). Sulforaphane may suppress proliferation and increase the death of cancer cells, especially in the epidermis, and prevent tumour formation (Chinembiri 2014). Extracts of broccoli induce enzymes that protect cells from oxidative damage and inhibit inflammation in the skin (Dinkova-Kostova 2010). Ingestion or topical treatment with polyphenols of green tea can inhibit KC development (Saha 2013). High dietary fat intake has been linked to an increase of KC, but the mechanism of action is unclear (Meeran 2009). Nicotinamide, the amide of nicotinic acid (vitamin B3 or niacin), inhibits the immunosuppression and depletion of energy in keratinocytes induced by ultraviolet radiation (Park 2010; Surjana 2013). Nicotinamide also promotes DNA repair after sun damage (Thompson 2014). Previous studies found that 500 mg to 1500 mg daily of nicotinamide protects skin against sun damage; nicotinamide is found in meat, eggs, and nuts (Yiasemides 2009).

Why it is important to do this review

Keratinocyte cancer is one of the most common malignancies around the world and its incidence is rising (Cancer Australia & AIHW 2008; Leiter 2014). Although its mortality rate is low, KC represents an economic burden to health systems (Cakir 2012). Currently, physicians can identify high-risk groups for developing KC, but they need to know what intervention is the most effective to prevent KC, apart from avoiding sun exposure and periodic medical surveillance. Those people in high-risk groups may develop more than one KC, and it is better to invest in prevention than in multiple treatments. We consider that people receiving immunosuppressive therapy are different from other high-risk groups because theoretically the best prevention measure could be the suspension of immunosuppressive drugs. However, suspension of immunosuppressive drugs in people who need these drugs to control existing medical conditions would be unethical. Many treatments have been described for the prevention of KC, but it is necessary to conduct a systematic review that includes the interventions related to education and new ones like photodynamic therapy in high-risk groups not receiving immunosuppressive therapy.

In this review, we aim to include interventions for precursor lesions, such as actinic keratosis, that may prevent the progression to KC in addition to their effect in the clearance of the lesions. Finally, the interventions in people with genetic diseases, like albinism and xeroderma pigmentosum, should be analysed independently of other groups of people because they develop multiple tumours at younger ages (Mabula 2012; Naik 2013).


To determine the efficacy and safety of interventions for preventing keratinocyte cancer in high-risk groups not receiving immunosuppressive therapy.


Criteria for considering studies for this review

Types of studies

We will include only randomised controlled trials (RCTs) that aim to prevent keratinocyte cancer in high-risk groups not receiving immunosuppressive therapy. We will include only studies that compare any combination of interventions and any comparison with placebo or no treatment, without restriction in dose and duration. We will exclude quasi-experimental studies.

Types of participants

We will include studies whose participants have the following characteristics:

  • personal history of keratinocyte cancer confirmed by biopsy;

  • precursor lesions of keratinocyte cancer like actinic keratoses;

  • cutaneous squamous cell carcinoma in situ or Bowen's disease;

  • long-standing ulcer or scar;

  • personal history of long-term psoralen and ultraviolet A (PUVA) treatment (long-term PUVA is defined as more than 50 PUVA treatments) (Stern 2012);

  • chronic arsenic exposure;

  • history of radiation therapy; or

  • genodermatoses or hereditary cancer syndromes, like xeroderma pigmentosum, oculocutaneous albinism, epidermolysis bullosa, epidermodysplasia verruciformis, fanconi anaemia, dyskeratosis congenita, Rothmund-Thomson syndrome, Bloom syndrome, Werner syndrome, nevoid basal cell carcinoma syndrome, Rombo syndrome, or Basex-Dupré-Christol syndrome.

We will include studies with participants of any age, gender, ethnic background, or socioeconomic status. We will exclude studies where participants are receiving immunosuppressive therapy. We will consider that a participant is receiving immunosuppressive therapy if he or she is a transplant recipient or is receiving TNF-alpha antagonists or RAF inhibitors for the treatment of an autoimmune disease or other type of cancer - including leukaemia and lymphoma. We will also exclude studies where participants are receiving azathioprine monotherapy for inflammatory disorders, such as inflammatory bowel disease and inflammatory arthritis. Human immunodeficiency virus (HIV)+ individuals are at high risk of developing keratinocyte cancers, and the increased risk is due to immunodeficiency (as is the case with organ transplant recipient populations), so we have decided to exclude this population from the review, even though they are not receiving immunosuppressive therapy (Grulich 2007; Silverberg 2013).

Types of interventions

We will include all topical and oral agents used for preventing keratinocyte cancer compared with placebo; no treatment; other topical or oral agents; or a different formulation, concentration, dose, frequency, or duration of the same agent. We will include photodynamic therapy used for preventing keratinocyte cancer. We will also include educational interventions to promote sun protective behaviours in high-risk groups and dietary modifications.

  1. Topical treatments, except retinoids

  2. Topical retinoids, such as tretinoin, adapalene, and tazarotene

  3. Oral retinoids, such as acitretin, retinol, and isotretinoin

  4. Non-steroidal anti-inflammatory drugs, such as diclofenac or celecoxib

  5. Antioxidants, such as selenium and beta-carotene

  6. Photodynamic therapy

  7. Educational interventions

  8. Dietary modifications, such as low-fat diet.

Comparisons will include the following:

  • any topical agent versus placebo, vehicle, or no treatment;

  • any topical agent versus another topical agent;

  • any oral agent versus placebo, vehicle, or no treatment;

  • any oral agent versus another oral agent;

  • any topical agent versus any oral agent;

  • any different formulation, concentration, dose, frequency, or duration of the same topical or oral agent;

  • any topical agent versus photodynamic therapy;

  • any oral agent versus photodynamic therapy;

  • any photodynamic therapy versus no treatment;

  • any educational intervention versus another educational intervention; or

  • any dietary modification versus another dietary modification.

Types of outcome measures

Primary outcomes
  • Incidence of new keratinocyte cancers.

  • Adverse events serious enough to lead to withdrawal.

Secondary outcomes
  • Recurrence.

    • Time to recurrence, defined as the time between the start of the intervention and the return of the keratinocyte cancer (KC) at the same site of the body.

    • Proportion of participants with a recurrent KC.

  • Second or subsequent KC.

    • Time to second or subsequent KC.

    • Proportion of participants who develop a subsequent KC (either annually or at five or 10 years of follow up after starting the intervention).

  • Mortality.

    • Proportion of participants who died due to KC (disease-specific mortality) at the end of the trial (or at an annual or five or 10 years of follow up after the start of the intervention).

  • Adverse effects.

    • Proportion of participants with adverse effects during the trial and proportion of those with serious adverse events leading to withdrawal.

  • Quality of life.

    • Quality of life score assessed by participants using a validated scale for skin cancer patients, such as the Skin Cancer Quality of Life (SCQoL) questionnaire (Vinding 2013), the Skin Cancer Quality of Life Impact Tool (SCQOLIT) (Burdon-Jones 2013), the Skin Cancer Index (SCI) (Rhee 2006), the Dermatology Life Quality Index (DLQI) (Finlay 1994), or the Children's Dermatology Life Quality Index (CDLQI) (Lewis-Jones 1995).

  • Markers of photodamage.

    • Proportion of participants with complete clearance of precursor lesions of KC, such as actinic keratosis (AK), at the end of the intervention.

    • Proportion of participants with a decrease in the number of cutaneous markers of photodamage, such as solar elastosis, solar lentigines, and telangiectasia, at the end of the intervention.

Search methods for identification of studies

We aim to identify all relevant 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);

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

  • SciELO (Scientific Electronic Library Online) via Biblioteca Virtual en Salud (BVS).

We have devised a draft search strategy for RCTs for MEDLINE (Ovid), which is displayed in Appendix 1. This will be used as the basis for search strategies for the other databases listed.

Trials registries

We will search the following trials registries and portals. We will use a combination of the following terms: skin cancer and prevention.

Searching other resources

References from included studies

We will check the bibliographies of included studies and articles cited as references in order to identify further relevant trials.

Unpublished trials

We will search for unpublished trials by contacting the corresponding and first author of included studies. We will contact known specialists in the review topic area and relevant pharmaceutical companies.

Citation indexes

We will search in the Science Citation Index Expanded™ via Web of Science™ from 1900 (Thomson Reuters) and Scopus® database via Elsevier from 1960.

Dissertations and theses databases

We will search Open Access Theses and Dissertations ( for relevant trials.

Grey literature

We will search the System for information on Grey Literature via OpenSIGLE ( for relevant trials.

Adverse effects

We will not perform a separate search for adverse effects of the target interventions. However, we will examine data on adverse effects from the included studies that we identify.

Data collection and analysis

Some parts of the methods section of this protocol uses text that was originally published in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

Selection of studies

We will merge search results using Mendeley (a free reference manager and PDF organiser) and remove duplicate records of the same report, identifying them by title, authors, and article identifier or digital object identifier (DOI). Two authors will independently examine the titles and abstracts to decide which reports are potentially relevant and should be reviewed in full text. They will resolve any disagreement by discussion between themselves or by the judgement of a third author. We will retrieve the full text of the relevant reports.

Two different authors will independently compare the reports by author names, location, setting, details of the interventions, number of participants, baseline data, date, and duration of the study in order to identify duplicate publications or multiple reports from the same study. We will assess, using a data extraction form (DEF), the full text of relevant reports to identify which of them meet the inclusion criteria. If the full text contains insufficient information or it is unclear, we will contact the corresponding author to clarify study eligibility. If at this stage a report does not meet all of the inclusion criteria, we will exclude it. We will record the number of reports identified, included, and excluded in a flow diagram with the reasons for exclusion, according to the PRISMA statement.

Data extraction and management

We will collect data from included studies using an electronic DEF created in Microsoft® Access. Two authors - one an expert in skin cancer and the other an expert in methodology - will independently extract data from every report. If there is insufficient or unclear information about a report, we will contact the corresponding author or any of the study authors to obtain the missing data. The two authors will discuss their disagreements, and if they cannot resolve their discrepancies, they will consult a third author. If this process is unsuccessful, we will report the disagreement in the review. For multiple reports of the same study, we will extract data from each report separately and then combine the information contained in the DEFs. We will extract the following information from each article (Higgins 2011):

  • the citation and corresponding author's information;

  • reasons for including or excluding the article according to the inclusion criteria;

  • the study design, duration, and the country where the trial took place;

  • the inclusion criteria for participants to verify that they belong to a high-risk group;

  • the number of participants randomised and data about follow-up to identify withdrawals for each group of interventions;

  • a description of the interventions: number of groups, duration of the treatment, dosage, and route of administration;

  • the number of previous KCs (basal cell carcinoma (BCC) or cutaneous squamous cell carcinoma (cSCC)) per participant and their treatments;

  • skin type according to the Fitzpatrick scale;

  • the number of participants who developed a KC or a subsequent KC;

  • the time to recurrence and a second or subsequent KC;

  • the number of participants 1) who died due to KC; 2) with complete clearance of actinic keratoses; 3) with a decrease in the cutaneous markers of photodamage; 4) with adverse effects;

  • the scores of quality of life questionnaires;

  • limitations of the study; and

  • the funding source.

We will collect BCC and cSCC data separately if available.

Assessment of risk of bias in included studies

Two authors will independently assess the risk of bias of the included studies using the Cochrane tool. This tool addresses the following domains: random sequence generation; allocation concealment; blinding of participants and personnel; blinding of outcome assessment; incomplete outcome data; and selective outcome reporting.

We will categorise our judgements as 'low risk', 'high risk', or 'unclear risk' of bias, using the criteria available in the Cochrane Handbook for Systematic Reviews of Interventions (Table 8.5.d) (Higgins 2011). We will record the information in the 'Risk of bias' tables in Review Manager (RevMan) (Review Manager 2014) to create the 'Risk of bias' graph and summary. The graph will illustrate the proportion of studies categorised as low, high, and unclear risk, while the summary will present all of the judgements for each study. We will summarise the risk of bias for primary and secondary outcomes within and across studies.

We will consider a study low risk when all domains for each outcome are low risk, high risk when at least one domain is judged high risk, and unclear risk when at least one domain is classified as unclear.

Measures of treatment effect

We will calculate risk ratios (RR) and their corresponding 95% confidence intervals (95% CI) for dichotomous outcomes (incidence of KC, incidence of subsequent KC, mortality due to KC, clearance of AK, cutaneous markers of photodamage, and adverse effects). However, we will report the difference in the mean number of new KC lesions between the intervention group and the control group per year of follow-up. Additionally, we will calculate the rates by dividing the number of new KC lesions between the number of person-years of follow-up per group of intervention and the rate ratio of both groups.

We will also compute the number needed to treat for an additional beneficial outcome (NNTB) and the number needed to treat for an additional harmful outcome (NNTH) for each intervention.

For continuous outcomes, such as quality of life, we will calculate mean differences (MD) and standard deviations (SD) when studies use the same scale; otherwise, we will calculate the standardised mean difference (SMD).

We will calculate the hazard ratio and 95% CI for time-to-event data, such as time to recurrence.

Unit of analysis issues

The analysis will take into account the level of randomisation, so in a parallel group design, the unit of analysis will be the participant.

In cluster-randomised trials (for educational interventions), we will reduce the size of each trial to its "effective sample size" by dividing the original sample size between a quantity called the "design effect" (Rao 1992). For continuous data, we will only reduce the sample size, and for dichotomous data, we will divide both the number of participants and the number experiencing the event by the design effect. After obtaining the effective sample size, we will enter the data into RevMan for analysis.

In the case of cross-over trials, we will calculate the paired t-test. If the paired t-test cannot be computed due to missing data, we will exclude the study from the meta-analysis. In the case of trials with repeated observations of participants during follow-up, we will select the following time-points: one year, five years, and 10 years. In hereditary cancer syndromes, the KC events can happen to a person more than once, which is why we will consider them as count data, and we will compare the difference in the mean number of events. In other high-risk groups, we will consider these repeated events as counts of rare events, and we will compute the rate ratio.

We will analyse the studies where multiple parts of the body receive the same topical intervention as cluster-randomised trials. However, if different parts of the body are randomised to different topical interventions, we will analyse the data as a cross-over trial if the study included an adequate washout period (Higgins 2011). In case of studies with more than two intervention groups, we will include groups that are relevant to the systematic review or for the meta-analysis. To overcome a unit of analysis error in studies with multiple groups, we will combine all relevant experimental intervention groups of the study into a single group and all relevant control intervention groups into a single control group.

Dealing with missing data

We will contact the corresponding trial author to request missing data. If there are missing standard deviations for continuous data, we will calculate these from standard errors, confidence intervals, t values, and P values. We will conduct intention-to-treat (ITT) analysis including all participants who did not receive the assigned intervention according to the protocol as well as those who were lost to follow up. We will also keep participants in the intervention groups to which they were randomised, regardless of which intervention they received. For the ITT analysis, we will use imputation and we will assume that all missing participants experienced the poor outcome for dichotomous data and that no change took place from baseline outcomes for continuous data.

We will also perform sensitivity analyses to assess how results change with the assumptions that are made about missing data. In the Discussion section, we will address the impact of missing data on the findings of the review.

Assessment of heterogeneity

If where we wish to combine studies for a particular outcome and comparison we find that the I² statistic is > 50% or > 75%, we will consider heterogeneity as substantial or considerable, respectively. In such cases, we will conduct a meta-analysis using the random-effects model. If the requirements for conducting a meta-analysis are not fulfilled or the I² statistic is over 75%, we will describe the data qualitatively.

Assessment of reporting biases

To reduce publication and time lag bias, we will search for ongoing, recently completed, and unpublished trials. We will search for trials in databases of theses and dissertations. Searching in databases worldwide and including non-English language studies will reduce the location and language bias. However, we will construct a contour-enhanced funnel plot if there are at least 10 studies included and the sample sizes are not similar. If there is asymmetry, we will consider publication bias as one of the possible explanations. Other explanations could be the methodology design, inadequate analysis, heterogeneity, and chance (Higgins 2011).

Data synthesis

If the assessments of risk of bias and the reporting bias are low, we will perform a meta-analysis only if the studies are not clinically heterogeneous (i.e. variability in the participants, interventions, and outcomes across the studies) and methodologically heterogeneous (i.e. variability in the study design and risk of bias). In the event of marked clinical heterogeneity, we will limit our analysis to high-risk group participants if possible. We will enter data into RevMan to perform the analysis. We will assume that heterogeneity in intervention effects is due to the categories of interventions and will use the random-effects model for the analysis. In addition, we will perform fixed-effect model analysis as a sensitivity analysis. If the requirements for conducting a meta-analysis are not fulfilled or the I² value is over 75%, we will describe the data qualitatively.

Where results are estimated for individual studies with low numbers of outcomes (< 10 in total) or where the total sample size is less than 30 participants and a risk ratio is used, we will report the proportion of outcomes in each treatment group together with a P value from a Fisher's Exact test.

Subgroup analysis and investigation of heterogeneity

We will undertake subgroup analyses for the primary outcomes if there are at least two studies per subgroup. We will analyse hereditary cancer syndromes as a whole subgroup. We will compare the following subgroups.

  1. Intermitent-dose versus daily-dose of treatments: daily use of topical retinoids cause more adverse effects than their use on alternate days in long-term usage.

  2. Low-cumulative versus high-cumulative dose of treatments: adverse effects may increase with higher cumulative doses of systemic retinoids.

  3. High-risk versus low-risk groups for cardiovascular diseases: non-steroidal anti-inflammatory drugs (NSAIDs) may cause cardiovascular adverse effects in individuals with a high-risk profile.

  4. Short-term versus long-term use of treatments: the longer the duration of the treatment the longer the time without the development of KC.

  5. Treatments in participants with genodermatoses versus other high-risk groups: participants with hereditary cancer syndromes will develop multiple KCs that may require higher doses and long-term use of treatments.

If data are available, we will perform a subgroup analysis for BCC versus cSCC.

Sensitivity analysis

In order to determine the robustness of our results, we will perform the following sensitivity analyses:

  1. comparing the results of the fixed-effect method analysis with the random-effects method analysis;

  2. excluding studies with high or unclear risk of bias;

  3. excluding studies where we will need to perform intention-to-treat analysis;

  4. excluding studies with outliers; and

  5. excluding pharmaceutical industry-sponsored studies.

'Summary of findings' table

We plan to include at least one 'Summary of findings' table in our review to summarise the essential primary outcomes (Higgins 2011) and will assess the quality of the body of evidence using the five Grading of Recommendations Assessment, Development and Evaluation (GRADE) considerations (study limitations, consistency of effect, imprecision, indirectness, and publication bias). 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.


The authors wish to thank the Cochrane Skin Group and Mahmoud Tawfik, who provided consumer input, checking the protocol for readability and clarity and ensuring that the outcomes are relevant to consumers.

The Cochrane Skin Group editorial base wishes to thank Sue Jessop, who was the Dermatology Editor for this protocol; Ben Carter, who was the Statistical Editor; Esther van Zuuren, who was Methods Editor; the clinical referee, Charlotte Proby; and the consumer referee, Johanna Small.


Appendix 1. MEDLINE (Ovid) search strategy

  1. Keratosis, Actinic/

  2. actinic keratos$.ti,ab.

  3. Bowen's Disease/

  4. bowen$ disease.ti,ab.

  5. squamous cell carcinoma in situ.ti,ab.

  6. marjolin's ulcer$.ti,ab.

  7. Fournier Gangrene/

  8. fournier gangrene.ti,ab.

  9. PUVA Therapy/

  10. puva.ti,ab.

  11. Arsenic/

  12. arsenic.ti,ab.

  13. Radiotherapy/

  14. (radiation therapy or radiotherapy).ti,ab.

  15. Xeroderma Pigmentosum/

  16. xeroderma pigmentosum.ti,ab.

  17. Albinism, Oculocutaneous/

  18. albinism.ti,ab.

  19. Epidermolysis Bullosa/

  20. epidermolysis bullosa.ti,ab.

  21. Epidermodysplasia Verruciformis/

  22. epidermodysplasia verruciformis.ti,ab.

  23. Fanconi Anemia/

  24. fanconi anemia.ti,ab.

  25. Dyskeratosis Congenita/

  26. Dyskeratosis Congenita.ti,ab.

  27. Rothmund-Thomson Syndrome/

  28. Rothmund Thomson Syndrome.ti,ab.

  29. Bloom Syndrome/

  30. bloom syndrome.ti,ab.

  31. Werner Syndrome/

  32. werner syndrome.ti,ab.

  33. Basal Cell Nevus Syndrome/

  34. Nevoid basal cell carcinoma syndrome.ti,ab.

  35. gorlin goltz syndrome.ti,ab.

  36. rombo syndrome.ti,ab.

  37. Basex Dupre Christol syndrome.ti,ab.

  38. Skin Ulcer/

  39. skin ulcer$.ti,ab.

  40. Cicatrix/

  41. Phototherapy/

  42. Skin Diseases, Genetic/

  43. Arsenic Poisoning/

  44. genodermatos$.ti,ab.

  45. (scar or scars or scarring).ti,ab.

  46. or/1-45

  47. exp Skin Neoplasms/pc [Prevention & Control]

  48. keratinocyte cancer$.ti,ab.

  49. exp Carcinoma, Basal Cell/

  50. exp Carcinoma, Squamous Cell/

  51. nmsc.ti,ab.

  52. non melanoma skin cancer$.ti,ab.

  53. basal cell carcinoma$.ti,ab.

  54. squamous cell carcinoma$.ti,ab.

  55. basal keratinocyte$.ti,ab.

  56. exp Neoplasms, Basal Cell/

  57. bcc.ti,ab.

  58. basal cell cancer$.ti,ab.

  59. basalioma$.ti,ab.

  60. exp Neoplasms, Squamous Cell/

  61. squamous cell epithelioma$.ti,ab.

  62. scc.ti,ab.

  63. or/47-62

  64. randomized controlled

  65. controlled clinical

  66. randomized.ab.

  67. placebo.ab.

  68. clinical trials as

  69. randomly.ab.

  70. trial.ti.

  71. 64 or 65 or 66 or 67 or 68 or 69 or 70

  72. exp animals/ not

  73. 71 not 72

  74. 46 and 63 and 73

[Lines 64-73: Cochrane Highly Sensitive Search Strategy for identifying randomized trials in MEDLINE: sensitivity- and precision-maximizing version (2008 revision)]

Contributions of authors

MM was the contact person with the editorial base.
MM co-ordinated the contributions from the co-authors and wrote the final draft of the protocol.
MM, MP, and LB worked on the methods sections.
MM, LB, and FJ drafted the clinical sections of the background and responded to the clinical comments of the referees.
MM, MP, and LB responded to the methodology and statistics comments of the referees.
MM, MP, LB, HP, and FJ contributed to writing the protocol.
MM is the guarantor of the final review.


This project was supported by the National Institute for Health Research (NIHR), via Cochrane Infrastructure funding to the Cochrane Skin Group. The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the Systematic Reviews Programme, NIHR, NHS or the Department of Health.

Declarations of interest

Martha Alejandra Morales-Sánchez: nothing to declare.
María Luisa Peralta-Pedrero: nothing to declare.
Fermín Jurado-Santa Cruz: nothing to declare.
Leticia A Barajas-Nava: nothing to declare.
Hyemin Pomerantz: "I was involved in the publication of secondary outcomes of two trials: Pomerantz 2014, Pomerantz 2015, Pomerantz 2016, and Pomerantz H, Chren MM, Lew R, Weinstock MA. Validation and comparison of quality-of-life measures for topical 5-fluorouracil treatment: results from a randomized controlled trial. Clinical and Experimental Dermatology [Accepted]. The primary outcomes of the trials were 1) chemopreventive effects of topical tretinoin, and 2) chemopreventive effects of topical 5-fluorouracil."

Sources of support

Internal sources

  • No sources of support supplied

External sources

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

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