A paper published in the Journal of Investigative Dermatology has examined the use of non-steroidal anti-inflammatory drugs (NSAIDs) and occurrence of one type of skin cancer, reporting that NSAIDs may have potential for preventing it. Roundup comments accompanied this Before the Headlines analysis.
Title, Date of Publication & Journal
‘Aspirin and non-steroidal anti-inflammatory drugs can prevent cutaneous squamous cell carcinoma: a systematic review and meta-analysis’, Chiho Muranushi et al.
Journal of Investigative Dermatology
Thursday 18 December 2014
Study’s main claims – and are they supported by the data
This paper does not prove that NSAIDs “can prevent cutaneous squamous cell carcinoma” (SCC, a type of skin cancer), as suggested in the paper’s title. (NSAIDs are ‘Non-Steroidal Anti-inflammatory Drugs’, and aspirin, referred to separately in the paper’s title, is one particular NSAID.) Although the paper adds to the body of evidence suggesting that NSAIDs may reduce the risk of SCC, particularly among those at greatest risk of developing this type of cancer, further research is needed before firm conclusions can be drawn.
This was a meta-analysis that combined the results from nine different studies. The results from these nine studies were combined to produce an overall average effect of NSAIDs on the risk of SCC. This was done separately for aspirin and for non-aspirin NSAIDs, as well as for all NSAIDs together.
This paper reports on a systematic review and meta-analysis that was done to appropriately high and generally accepted standards in statistical terms although, as pointed out above, the title of the paper is somewhat misleading.
However, there are limitations, largely stemming from the nature of the studies that were available to be reviewed. These limitations are pointed out in the paper by its authors (and, indeed, also in the press release).
The main limitation is that most of the studies were not able to accurately measure use of NSAIDs – all but one study, which was a randomised controlled trial in which the drug was administered to individuals, rely either on patients’ reports of their NSAID use or on pharmacy databases, which would only include prescribed NSAIDs (not those purchased over-the-counter). This could affect the accuracy of the results.
Further, there was a large amount of heterogeneity between the studies for all NSAIDs and for aspirin. In other words, the impact of NSAID use was inconsistent between studies. In particular, the authors found evidence that the impact of NSAID use may be greater among individuals at high risk of developing SCC. Larger studies would be needed to confirm this.
A linked issue is that all but one of the studies were observational – that is, they collected data on NSAID use and SCC, and investigated whether the two were associated. In such studies, it is always difficult to conclude that one thing (NSAID use, in this case) is actually causing the other (reduced SSC risk), because there may be confounding variables that account for some or all of the association. Statistical analysis can take account of some of the possible confounding variables but, as the authors describe in the paper, the studies they reviewed differed in the ways in which this was done. In particular, some of the studies could not allow for different levels of sun exposure, known to be a risk factor for skin cancer.
The authors clearly state that more research is needed in order to address these limitations.
Meta-analysis: an analysis in which results from a number of different studies are combined to obtain an overall effect size; this combined result is a weighted average of the estimated effect in the different studies, such that larger, more precise studies are given greater weight than smaller, less precise studies.
NSAIDs: non-steroidal anti-inflammatory drugs.
SCC: squamous cell carcinoma, a type of skin cancer.
Before The Headlines is a service provided to the SMC by volunteer statisticians: members of the Royal Statistical Society (RSS), Statisticians in the Pharmaceutical Industry (PSI) and experienced statisticians in academia and research. A list of contributors, including affiliations, is available here.