Figure 1 shows the study selection process. Database searches yielded a total of 558 unique publications whose title and abstract were screened. We discarded 536 because of no relevance (n=416), no unique data presented (for example, editorial article or review, n=110), or language other than English (n=10). We examined the remaining 22 articles in full text. Sixteen studies met the inclusion criteria and are included in table 1. We included 12 studies in the quantitative meta-analysis. Excluded from meta-analysis were one study that reported only mean numbers of exposures (numbers of indoor tanning visits),18
one study that grouped patients with basal cell carcinoma, squamous cell carcinoma, and melanoma,24
and two studies that grouped basal cell carcinoma and squamous cell carcinoma together.23
These last two studies also used patients admitted to hospital with solid organ cancer as controls, which are unlikely to be representative of the population from which the cases arose.23
All four studies showed non-significant associations, although one showed a statistically significant positive association when sunbed exposure was grouped with medical phototherapy.24
Fig 1PRISMA flow diagram of literature search and study selection for meta-analysis of indoor tanning and non-melanoma skin cancer. BCC=basal cell carcinoma; NMSC=non-melanoma skin cancer; SCC=squamous cell carcinoma
The 12 studies included in the meta-analysis were published between 1985 and 2012, used data collected between 1977 and 2010 in six different countries, and included 80
661 total participants and 9328 cases of NMSC. All 12 studies reported or provided raw data for effect estimates as odds ratios or hazard ratios (table 2). When available, we preferentially present effect estimates reported using multivariate models.
Table 2 Outcome effect sizes for studies included in primary meta-analysis
Effect estimates for ever exposure to indoor tanning compared with never exposure were available for 10 out of 12 studies. A meta-analysis of these studies yielded summary relative risks of 1.29 (95% confidence interval 1.08 to 1.53) for basal cell carcinoma and 1.67 (1.29 to 2.17) for squamous cell carcinoma (fig 2). A χ2
test for heterogeneity was non-significant for both basal cell carcinoma (P=0.14) and squamous cell carcinoma (P=0.09); I2
statistics were 36.8% (95% confidence interval 0% to 72%) for basal cell carcinoma and 47.1% (0% to 79%) for squamous cell carcinoma. To include all available studies, we did a sensitivity analysis using the effect statistics from all 12 studies, including the two studies that reported only higher dose exposure.17
Both studies considered only basal cell carcinoma; with these two studies included, the summary relative risk for basal cell carcinoma was 1.25 (1.01 to 1.55). Funnel plots assessing publication bias were symmetrical for both basal cell carcinoma and squamous cell carcinoma, and all Begg’s and Egger’s tests were not statistically significant, suggesting that publication bias was unlikely. Our results did not change appreciably in a sensitivity analysis excluding four retrospective studies that did not fully adjust for confounders.14
Fig 2Relative risk of basal cell carcinoma and squamous cell carcinoma in participants ever exposed to indoor tanning compared with participants never exposed to indoor tanning
To assess the presence of a dose-response effect, we did a sub-analysis on studies that included effect estimates for frequent or multiple (high dose) exposures to indoor tanning (table 3, top). High dose exposure was associated with a relative risk of 1.50 (0.81 to 2.77) for basal cell carcinoma. To assess the potential effect of exposure to indoor tanning at a young age, we did a sub-analysis on studies that included effect estimates for early life exposure (table 3, bottom). Indoor tanning exposure before age 25 was associated with a relative risk of 1.40 (1.29 to 1.52) for basal cell carcinoma and 2.02 (0.70 to 5.86) for squamous cell carcinoma.
Table 3 Outcome effect sizes for studies included in subset analyses of high dose exposure and young age at exposure
Population attributable risk
Applying our summary risk estimates to the prevalence of exposure to indoor tanning in the United States, we calculated the population attributable risk fraction at 3.7% for basal cell carcinoma and at 8.2% for squamous cell carcinoma. This corresponds to 98
408 cases of basal cell carcinoma and 72
244 cases of squamous cell carcinoma, making 170
652 cases of non-melanoma skin cancer each year attributable to indoor tanning (see supplementary figure).