Long-term topical corticosteroid use and risk of skin cancer: a systematic review

Objective: The objective of this systematic review was to synthesize available research evidence to determine the risk of skin cancer in patients with long-term use of topical corticosteroids (TCS). Introduction: Topical corticosteroids are one of the most commonly prescribed medicines in dermatology and the mainstay of the treatment of atopic dermatitis and other skin conditions such as psoriasis. They are often required for months or years to control the disease and ultimately restore patients’ quality of life. In some patients, TCS may have a local immunosuppressive effect and theoretically increase the risk of skin cancer, whilst on the other hand TCS may decrease the risk of skin cancer in patients where TCS are used to treat inflammatory skin disease. To date, no systematic review has been performed to collate evidence on the effect of long-term TCS use on the risk of skin cancer. Inclusion criteria: This review considered studies that included people of all ages, genders and ethnicities, including HIV and transplant participants or participants with genetic diseases (for example, Gorlin-Goltz syndrome) This review considered studies that evaluated long-term use of topical corticosteroids. “Long-term” was defined as using TCS more than once a week for a month or longer. The review included cohort, cross-sectional and case-control observational studies exploring the association between the stated intervention and outcomes. The primary outcome measures of interest were: non-melanoma skin cancer (keratinocyte carcinoma), cutaneous squamous cell carcinoma (cSSC), basal cell carcinoma (BCC) or melanoma skin cancer. Genital and oral skin cancers are considered to be slightly different so we did not include them in this review. Methods: We performed a comprehensive search of MEDLINE, Embase and LILACS on November 9, 2017 to identify observational epidemiological studies assessing the association between long-term TCS use and skin cancer. We also searched EThOS at the British Library and three drug safety databases to identify unpublished work. The titles, abstracts and full text identified from the search were assessed independently by two authors against pre-specified inclusion/exclusion criteria. Methodological quality was not assessed as no articles were found which met the inclusion criteria. Data extraction was not possible as no articles were found which met the inclusion criteria. It was not possible to complete data synthesis as no articles were found which met the inclusion criteria. Results: A total of 1703 potentially relevant studies were identified following a comprehensive electronic search. After abstract and title screening, 51 full texts were assessed for eligibility criteria. Of these, no study met the inclusion criteria. No additional records were identified from searching unpublished literature. Conclusions: We did not find any studies that could help us establish if long-term TCS use is associated with skin cancer. Future research using primary care databases might give a better understanding regarding long-term use of TCS and skin cancer.


72
There are two types of skin cancers: melanoma and non-melanoma (keratinocyte). Around 97% of skin 73 cancers are non-melanoma (NMSC), comprising mainly of basal cell carcinomas (BCCs) or cutaneous 74 squamous cell carcinomas (cSCCs). The incidence of NMSC is increasing worldwide 9-13 with an 75 estimated 2-3 million new cases of NMSC recorded each year. 14 With respect to cutaneous malignant 76 melanoma (CMM), this is the most serious form of skin cancer and has been increasing steadily in 77 incidence over the past 30 years. 15 Mortality due to CMM is much higher than that of NMSC. 16

78
There are several observational studies that have looked at the relative risk of developing skin cancer 79 due to oral corticosteroid exposure. [17][18]

101
On the other hand, it is possible that treating skin inflammation with TCS may reduce the risk of skin 102 cancer. Several systematic reviews and meta-analyses report the benefits of anti-inflammatory drugs 103 in reducing the risk of cancer, including skin cancers. 27,275,26 The management of certain types of 104 inflammatory skin diseases includes the rationale that reducing inflammation reduces the risk of cSCC 105 development in vulval and penile lichen sclerosus as well as hypertrophic lichen planus. It is also known 106 that chronic inflammation is a risk for the development of cSCC, such as in chronic ulceration and the 107 development Marjolin's ulcer. 29,307,28 This mainly holds true for cSCC but less is known about BCC and 108 melanoma. .Therefore, overall TCS may decrease the risk of skin cancer in patients where TCS are 109 used to treat inflammatory skin disease.

110
To date, no published systematic review or meta-analysis have been performed to collate evidence on 111 long-term TCS use on the risk of skin cancer. The review group examined MEDLINE and EMBASE,

112
Prospero and JBI Database of Systematic Reviews and Implementation Reports and did not find any 113 current or planned reviews on the same topic. Immunosuppression induced by TCS, either local or 114 systemic, may allow these cancers to emerge from reduced immunosurveillance. However, TCS may 115 also reduce the risk of skin cancer in patients where TCS are used to treat inflammatory skin disease.

116
With TCS use being one of the most commonly prescribed drugs in the clinical field of dermatology and 117 the increasing incidence of skin cancer there is a need to review all current evidence about the possible 118 association. The protocol for this systematic review has recently been published. 3129

Review question 121
The objective of this systematic review was to synthesize the available research evidence to determine 122 the risk of skin cancer in patients on long-term use of topical corticosteroids.

123
In people using long-term (more than once a week regular use over for one1 month or longer) topical 124 corticosteroids what is the risk of developing skin cancer (clinically or histologically confirmed non-

135
This review considered studies that evaluate long-term use of topical corticosteroids. We define here 136 'long-term' as using TCS more than once a week for a month or longer.

202
The study by Landi et al. 320 was the only one study identified which had a primary objective of 203 determining whether steroid treatment was associated with skin cancer. However this study did not 204 meet our eligibility criteria as there was no information on frequency and duration of TCS exposure.

205
The authors conducted a case-control study in Italy, from 1994 to 1999, which included patients with

212
and they took into account ascertainment bias by adjusting for frequency of moles removed in 213 addition to other covariates. The authors concluded that people without melanoma were more likely to Created by XMLmind XSL-FO Converter. have been exposed to glucocorticoid-based therapy than those with melanoma; there was no effect 215 modification by reason for treatment or route of administration. Larger studies would be needed to 216 confirm these findings.

217
There were three studies which investigated the risk of skin cancer amongst people with 218 dermatological conditions. Ming et al. 620 conducted a case-control study in the US, between 1998 and 219 2001, with 1378 NMSC cases and 1533 controls with other dermatologic conditions to explore 220 whether people with NMSC are more likely to have had atopic dermatitis (AD) than those without 221 NMSC. The authors report TCS use was a confounder for the association between AD and NMSC.

222
After adjusting for age, sex, ethnicity and TCS use the odds of AD was 0.78 (96% CI 0.61, 0.98) for 223 those who had NMSC compared to those who did not. The authors also conducted a secondary 224 analysis and report the odds of being a TCS user was about 30% less in those with a NMSC as 225 compared with those without a NMSC. However, again, no information on frequency of TCS use was 226 available, a limitation cited by the authors. The inclusion criteria for this review focused on patients who were exposed to long-term TCS use 236 (more than once a week for a month or longer). There were no studies whose primary objective was 237 to investigate long-term TCS exposure per se, as opposed to any TCS use. It is therefore 238 unsurprising that no study included information on both duration and frequency of use. We believe 239 that using a specific definition of long-term use based on clinical experience was necessary to 240 maximize the external validity of this review.

242
We did not find any studies that might help us establish if long-term TCS use is associated with skin 243 cancer. Future research using primary care databases might give a better understanding regarding 244 long-term use of TCS and skin cancer.

Implications for practice
Created by XMLmind XSL-FO Converter.
There was an absence of evidence identified in the review to make clinical recommendations.

248
There is a significant gap in the evidence base in the area of long-term TCS use and the risk of skin 249 cancer, and therefore future research needs to be conducted to answer this important question. There 250 are several published papers in the area of oral and genital sites, therefore a systematic review in this 251 specific area could be of potential benefit to the dermatology community.

253
The review team has no conflicts of interest.