Lifestyle factors of smoking, BMI and alcohol consumption on the risk of Non-Melanoma Skin cancer in adults: Systematic review

Review Questions/Objectives The objective of the systematic review is to synthesize the best available research evidence on the epidemiological association between non-melanoma skin cancer NMSC (and its subtypes) and smoking, alcohol and BMI. Background Around 97% of skin cancers are epithelial in origin and are either basal cell carcinomas (BCCs) or squamous cell carcinomas (SCCs), collectively known as non-melanoma skin cancer (NMSC). The incidence of non-melanoma skin cancer (NMSC) is increasing worldwide 1-5. Globally an estimated 2-3 million new cases on NMSC are recorded each year 6. Basal cell carcinoma is defined as a slow-growing, locally invasive, malignant, epidermal skin tumour that mainly affects people with light coloured skins 7. They are the most malignant growth found in humans and originate from basal cells of the epidermis. Approximately 80% of BCCs occur on the head and neck but other common areas can include the trunk and lower limbs. There are several subtypes that can occur and these include: nodular/cystic; superficial and morphoeic (high risk of aggressive tissue invasion and destruction). Basal cell carcinomas present with non-healing lesions (i.e. recurrent scabbing or bleeding), they frequently have a pearly edge and telangiectasia. The pearly edge can be seen more easily when the area of skin is stretched, this is particularly useful for visualisation of nodular and superficial basal cell carcinoma. The percentage of tumours that metastasise (the spread of the tumour from its primary location to a secondary location via the blood, lymph or across body cavities) is very low (0.0028%-0.55%) 8. For most nodular and superficial basal cell carcinomas, treatment is simple and effective; the major risk factor is new primary tumours. However if left untreated, or if inadequately treated, the lesion can cause extensive local tissue destruction, particularly on the face. Neglected cases may even infiltrate bone and deeper structures such as the brain 9. Death from basal cell carcinoma is extremely rare, but may occur in neglected cases or those with underlying immunosuppression. Guidelines for the management of basal cell carcinoma suggest that topical therapy curettage and cautery, cryotherapy, simple excision, and photodynamic therapy may be used for low-risk disease; however treatment options for high-risk basal cell carcinomas are wide surgical excision, radiotherapy, and Mohs micrographic surgery 10. The choice of therapy depends on a number of prognostic factors such as tumour size, site, definition of clinical margins, histological subtype, failure of previous treatment, immunosupression and patients’ preference. Squamous cell carcinoma is generally more aggressive than basal cell carcinomas and originates in skin cells that produce keratin. Unlike basal cell carcinoma which has no reported precursor lesions, there are two principal precursors of squamous cell carcinoma: actinic (solar) keratoses (AKs) and Bowen’s disease (intra-epidermal carcinoma; IEC), both of which are described as carcinoma-in-situ. Squamous cell carcinoma is distinguished from carcinoma-in-situ by having an invasive component (i.e. involving connective tissue and blood vessels in the dermis), which can be determined histologically Approximately 50-60% of squamous cell carcinomas occur on the skin of the head and neck with other common sites being the hands, forearms, upper trunk and lower legs. The lips and ears are particularly high risk areas 11. Squamous cell carcinomas usually present as enlarging, sometimes hyperkeratotic, nodules on sun-exposed sites. They may be ulcerated if poorly differentiated and may occur in sites of previous scarring. The dull raised edge of a squamous cell carcinoma lacks the translucency of the pearly edge of a basal cell carcinoma. Treatment for cutaneous squamous cell carcinoma is complete removal of the primary tumour and any metastases. Guidelines suggest surgical excision for all resectable tumours with wide margins or histological margins for high-risk tumours. The treatment of choice for high risk tumours or locally recurrent tumours is Mohs micrographic surgery. Radiotherapy may be used for non-resectable tumours and curettage and cautery could be useful for small (<2cm), well differentiated tumours. For cutaneous squamous cell carcinomas, tumour location, size, depth, histological differentiation, host immunosuppression, previous treatment and type of treatment are all factors that can affect the metastatic potential. Small well differentiated tumours have 5-year survival rates of 96% whereas larger poorly differentiated tumours with perineural involvement may recur in half of cases. Early diagnosis substantially reduces the chance of metastasis and, hence, improves survival rates. Developing an SCC increases the risk of having another skin cancer in the future and the 3 year cumulative risk of developing another SCC averages at 18% (9%-23%) and the risk of developing a BCC, following the initial SCC, is 43% 12. The most important risk factors for NMSC are thought to be age, skin type, and exposure to UV radiation including sunlight and sun-beds. Human papillomavirus infection is also thought to play a part in skin cancer carcinogenesis 13. People with precursor lesions (actinic keratosis and Bowen’s disease), previous skin cancers, recipients of transplanted organs, immunosupression, exposure to arsenic, treatment with psoralen ultraviolet and rare inherited genetic disorder and albinism are also at greater risk of developing NMSC. It is unclear at present whether smoking modifies the risk of basal cell carcinoma since some research has shown no relationship with smoking 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, whereas other research has suggested smoking may increase risks of basal cell carcinoma in heavy smokers 24, 25, 26 or could be protective of basal cell carcinoma 27. There is evidence that smoking is a risk factor for oral squamous cell carcinoma however it is unclear if smoking modifies the risk of cutaneous squamous cell carcinoma since some research has shown no relationship 28and other research has suggested it may increase the risk 29, 22. Limited evidence has been published about the risk of basal cell carcinoma or squamous cell carcinoma with alcohol intake and Body Mass Index (BMI). The risk of basal cell may be reduced in those with an increased BMI 25 30. There are large numbers of observational studies looking at the relative risk of developing NMSC from a variety of exposures; however, to date no systematic review or meta-analysis has been performed to collate evidence of smoking, alcohol or BMI. With basal cell carcinoma and squamous cell carcinoma making up two of the most common cancers, as well as their relatively high incidence and economic cost, a review of all current evidence is necessary to help develop disease prevention strategies in public health. Inclusion Criteria Types of Participants We will include adults (18+ years old) of either gender from any ethnicity. Patients with genetic syndrome, for example Gorlin Syndrome, Xeroderma Pigmentosa, albinism; or any other condition that increases the risk of developing a BCC/SCC; patients with recurrent tumours and patients taking immunosuppressants (for example, transplant patients) will be excluded. The phenomena of interest The focus of interest for this systematic review is the epidemiological association between exposure to modifiable lifestyle factors such as smoking, alcohol consumption and body mass index (BMI) and non-melanoma skin cancer NMSC (and its subtypes). Types of Exposures Smoking measured either using self-reported status or consumption, or using biochemical markers (for example cotinine) will be eligible. We will include all forms of tobacco exposure, including cigarettes, cigars, chewing tobacco and roll-ups. Self-reported smoking status that was verified by either carbon monoxide measurements will be used in preference to sole self-reported status included. Body mass index (BMI) measured either using self-reported status or through measured BMI, derived from height and weight measurements, will be included. We will additionally include other anthropometric measurements including waist to hip ratio and the percentage body fat. Alcohol consumption measures either using self-reported status or consumption using a measureable quantity of alcohol, such as drinks per week or units of alcohol, will be included. Types of Outcomes We will assess the effect of the above exposures on: The risk of squamous cell carcinoma The risk of basal cell carcinoma The risk of non-melanoma skin cancer We will include any eligible lesion location anywhere on the skins surface, excluding the lip. Additionally, we will exclude internal head and neck cancers. Types of Study We will consider all comparative analytical observational epidemiological studies, such as case-control, cohort and analytical cross-sectional studies, in the systematic review. Randomised control trials of interventions, ecological studies and observational descriptive studies will be excluded. Search strategy Studies will be identified using a three step approach. Initially MEDLINE (1950-2011) and CINAHL (1980-2011), will be searched to identify key words used in the titles and abstracts. Second a more extensive search in other databases will be undertaken to further identify articles. Third, reference lists and bibliographies of retrieved articles will be reviewed to identify research not located through other search strategies. The databases to be searched include: CINAHAL: from 1980 to February 2011; Cochrane: to February 2011; EMBASE: from 1980 to February 2011; MEDLINE: from 1950 to February 2011; Details of the search strategy are presented in Appendix I. CAB Abstracts (from 1910 to present) will be searched using adapted terms from the MEDLINE Searches using the following (OR/4-11) AND ((OR/17-24) OR (OR/28-32) OR (OR/35-38)) AND ((OR/47-59) OR (OR/63-66) OR (OR/73-78)). The references and abstracts identified from the search will be assessed against the inclusion/exclusion criteria independently by two reviewers and full text will be obtained of relevant studies. If the title and abstracts are inconclusive the full text will be obtained for further review. The references will be entered into a bibliographic software package EndNote. If any data are missing from the trial report, attempts will be made to obtain that data by contacting the authors. Studies that have been published in duplicate will be included only once. Decision for study eligibility will be made by two reviewers and will be checked by the third. Assessment of methodological quality The Newcastle-Ottawa Scale (NOS) 31 will be used to assess the methodological quality of the studies included in the systematic review. The scale consists of three categories: Selection (4 stars), Comparability (2 stars) and Outcome (3 stars) which, when combined, give a maximum score of nine stars. The scale has only been developed for cohort and case-control studies, therefore when assessing cross- sectional studies adaptations to the new scale needs to be implemented, this will be achieved by combining sections from both the case-control and cohort criteria. The new criteria will be taken from existing categories relating to: i) Selection (2 stars) contained criterion 1 from the selection section of both case-control and cohort studies; ii) Comparability (2 stars) remaining the same as it is for case-control and cohort studies; iii) Outcome (3 stars) consists of criterion 1-3 (adapted) from the exposure/outcome section of case-control studies. Therefore a maximum score of seven stars can be achieved. The scores for each study will be recorded in table format and analysed in the results section. We have elected to use the Newcastle-Ottawa Scale to assess the quality of the included studies as this is a validated tool for critical appraisal of observational studies, and allows us to perform subgroup and sensitivity analyses based on the 3 components of the scale (selection, comparability and ascertainment) to assess their impact on the pooled results.(See Appendix II for the details of the critical appraisal tool) Data collection Data will be extracted from the eligible studies by two authors working independently (TE & JLB or TE & FBH). Disagreements will be resolved through consensus between authors, or through discussion with the third author. A data extraction form was designed and will be piloted to record key information from each eligible study (see Appendix III). It will consist of the following five sections; 1. Description of study - Type of study design, timing, setting and measurement of exposure. 2. Participants - Inclusion criteria, exclusion criteria and numbers of patients throughout the study. 3. Methodological quality of studies - NOS for case control, cohort and cross sectional studies. 4. Outcome measures - Definition of outcome and method of diagnosis 5. Results - To allow statistics to be extracted Data synthesis Measures of effect The results will be expressed as odds ratios (OR) with 95% confidence intervals (CI) for dichotomous outcomes, mean differences (MD) with 95% CI for continuous outcomes. We will use adjusted estimate in preference, were available. Assessment of Heterogeneity Heterogeneity will be quantified using I2 32. Data will be synthesised using meta-analysis techniques if I2<85%. Assessment of Reporting Biases Publication bias (small study bias) will be tested by the use of a funnel plot, where adequate data are available for similar types of exposures. Meta-analysis For studies with similar exposures, categorised as smoking, BMI, alcohol consumption, a meta-analysis will be performed, to estimate a weighted measure of effect across studies. We shall estimate pooled effects using a random effects (DerSimonian and Laird) model to allow for heterogeneity due to inherent biases within the studies. Where it is not possible to perform a meta-analysis, for example, where estimates of heterogeneity exceed 85%, the data will be summarised individually for each study. Subgroup analysis and investigations of heterogeneity If substantial heterogeneity is detected between the studies, reasons for heterogeneity will be explored, based on the date of publication, and the definition of exposure (for example, for smoking we will subgroup the data based on current smoking, former smoking, ever smoking). Sensitivity analysis Sensitivity analysis will be performed to observe the effect of excluding studies with lower methodological quality. Data will be analysed using Review Manager 5. This systematic review will be conducted in accordance with Joanna Briggs Institute (JBI) guidance on systematic reviews and meta-analysis of quantitative evidence and MOOSE guidelines20 of how to report a meta-analysis of observational studies33.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.