Aspirin and Nonsteroidal Anti-inflammatory Drug Use and Keratinocyte Cancers

A Large Population-based Cohort Study of Skin Cancer in Australia

N. Pandeya; C.M. Olsen; B.S. Thompson; J.C. Dusingize; R.E. Neale; A.C. Green; D.C. Whiteman; for the QSkin Study


The British Journal of Dermatology. 2019;181(4):749-760. 

In This Article

Abstract and Introduction


Background: Nonsteroidal anti-inflammatory drugs (NSAIDs) have been postulated as chemopreventive agents for basal cell carcinoma (BCC) and squamous cell carcinoma (SCC), but findings from observational studies have been inconsistent, and clinical trial data are scant.

Objectives: To examine the association between aspirin and NSAID (nonaspirin) use and the risk of BCC and SCC in a large cohort specifically designed for skin cancer outcomes.

Methods: We used data from the QSkin Study, a prospective cohort of 43 764 residents of Queensland, Australia (34 630 were included in this study and 23 581 were used in our primary analyses). We used Cox proportional hazards models to estimate the hazard ratios (HRs) between self-reported aspirin and NSAID use 1 year prior to study baseline and the first histologically confirmed BCC or SCC for high-risk (history of skin cancer excisions or more than five actinic lesions treated) and average-to-low-risk participants (no history of skin cancer excision and at most five actinic lesions treated).

Results: After a median of 3 years of follow-up, 3421 participants developed BCC and 1470 SCC (2288 BCC and 932 SCC with complete covariate information). Among the high-risk group (1826 BCC and 796 SCC), compared with never use, frequent (at least weekly) NSAID use was associated with reduced risk of BCC (HR 0·84, 95% confidence interval 0·71–0.99) but not SCC. Aspirin use was associated with reduced risk of SCC (HR 0·77, 95% confidence interval 0·64–0·93) only among infrequent (less than weekly) users and was not associated with BCC. We observed no association between NSAID or aspirin use and the risk of BCC or SCC among average-to-low-risk participants.

Conclusions: While some weakly inverse associations were observed between prior use of aspirin or NSAIDs and skin cancer, the inconsistent patterns of associations do not provide convincing evidence that these medications reduce subsequent skin cancer risk. Further data on doses, duration and long-term follow-up may help us to comprehend the cumulative dose effect.


Keratinocyte cancers (KCs) – namely basal cell carcinoma (BCC) and squamous cell carcinoma (SCC) of the skin – are the most common cancers in fair-skinned populations and their incidence is rising.[1] While fatalities are relatively rare, the costs of diagnosis and treatment are very high.[2,3] Exposure to ultraviolet radiation is the principal causal factor for KCs, and recent primary prevention campaigns have led to some sustained changes in behaviours in relation to sun exposure and sun protection.[4] However, sun protection is a long-term preventive strategy and alternative approaches to prevention need to be explored, particularly among those already at high risk due to cumulative sun damage.

Chemoprevention is one possible avenue for skin cancer prevention. Recently, a number of randomized controlled trials (RCTs) have investigated the effect of different agents on the incidence of KC.[5–7] For example, oral nicotinamide[5] and an oral retinoid[7] have proven beneficial as chemopreventive agents for SCC in people with a history of KC. Nonsteroidal anti-inflammatory drugs (NSAIDs) have also been postulated as agents with potential for preventing or retarding the growth of epithelial cancers, based on preclinical data and follow-up of long-term cardiovascular trials.[8–11] A clinical trial has observed a protective effect of oral celecoxib on KC among those with high numbers of actinic keratoses.[6] Systematic reviews of observational studies[12–15] have also suggested a potential benefit of NSAIDs in reducing BCC and SCC incidence. However, significant heterogeneity exists in the study findings, which might be explained by variations in exposure measurement, study design or residual confounding. As for nicotinamide, the putative benefit of NSAIDs appears to be greatest among people at high risk of skin cancer, as opposed to the general population.

While RCTs are optimal for establishing causality, compelling evidence of benefit from high-quality observational studies on aspirin or NSAIDs with KC as the primary outcome would be required to justify a trial, given the potential for adverse effects such as increased risks of major haemorrhage, other vascular events and gastrointestinal side-effects.[16–18] Population-based cohort studies with sufficient information on skin cancer risk factors and potential confounding factors can assist in elucidating the relationship between use of NSAIDs and KC. To date, prospective studies that have assessed this association were not designed primarily for skin cancer, and thus typically lack information on important skin cancer risk factors to control adequately for confounding.

Here we present the findings of an investigation into the association between use of aspirin and nonaspirin NSAID (hereafter NSAID) and risk of cutaneous BCC and SCC. We used data from a large cohort study that was designed to examine skin cancer outcomes. Our primary aim was to examine these associations among those at high risk for skin cancer including those with a history of skin cancer excision or more than five actinic lesions. In secondary analyses, we examined the associations among those with no history of such treatments.