No association between whole‐body ultraviolet A1 phototherapy and skin cancers in humans: a cancer registry linkage study

Dear Editor, Ultraviolet (UV)A1 phototherapy was first developed in the 1980s.1 Its use started in the UK in Dundee in 1996.2 It is not available in most centres despite being useful for treating some of the sclerosing/fibrosing conditions and granulomatous conditions. Where available, UVA1 can be very helpful for other indications such as atopic eczema.3 One concern is that sufficient exposures might increase the risk of skin cancer.4 UVA1 has been shown to produce cyclobutane pyrimidine dimers in humans,5 and so has the potential to be carcinogenic. There are case reports of people who have received UVA1 phototherapy who then developed cutaneous melanoma and Merkel cell carcinoma, but there is no clear evidence that this form of phototherapy increases the risk of any form of skin cancer in humans.6 7 As part of a Chief Scientist Office – Scottish Government study linking treatment with phototherapies and the Scottish cancer registry, we assessed UVA1 exposure in patients who had been exposed to UVA1 between 1996 and 2016 in Dundee. Most courses of whole‐body UVA1 are delivered with a high‐output metal halide ‘bed’ rather than our lower‐output UVA1 cabinet. Photosys data were linked to national Scottish cancer registry data for patients from Tayside and Fife (using deterministic data linkage by the University of Dundee Health Informatics Centre). We followed up using this registry linkage for 297 patients receiving 419 whole‐body UVA1 courses [mean age at first UVA1 treatment was 46·5 years (range 12·9–91·4), 199 women (67%); the majority of sun‐reactive skin phototypes were types I (52 patients, 18%), II (136 patients, 46%) and III (89 patients, 30%)]; (i) from 1 June 1979 until first UVA1 exposure and then (ii) from first UVA1 exposure until 31 December 2016 (or earlier ‘censoring’ owing to, for example, death). Of the 297 patients in this cohort who were followed up for up to 20 years after first UVA1 treatment, 232 patients had received less than 50 cumulative UVA1 treatments, 61 patients received between 51 and 350 treatments and four patients received 351 or more whole‐body UVA1 treatments. We did not collect data on doses (the number of treatments is likely more relevant especially as doses are individualized for each patient); a typical course involves about 200 J cm−2 UVA1. We identified six skin cancers in five separate patients. None of these patients had also received UVB phototherapy or psoralen plus UVA. Of the two patients (one who received 30 UVA1 exposures for morphoea and one who had a course of 17 exposures for widespread granuloma annulare) with basal cell carcinoma (BCC), one patient developed this 4 years after UVA1 but the other developed BCC at the same time as the first UVA1 exposure (so it is implausible that UVA1 caused the BCC). Of those who developed cutaneous squamous cell carcinoma (SCC), one patient (who received a single course of 10 treatments for morphoea) was diagnosed with two SCCs simultaneously 3 years after UVA1 treatment, one was diagnosed with an SCC 6 years after first exposure to UVA1 (of which she had a cumulative 152 treatments for eczema) and one (who had received 31 treatments for morphoea and also developed cancer before first phototherapy) had SCC 10 years after first UVA1 treatment. We estimated standardized incidence ratios (SIRs) and 95% confidence intervals for BCC (only the first BCC is recorded in the Scottish cancer registry; all other cancer types are recorded), cutaneous SCC, cutaneous melanoma, non‐Hodgkin lymphoma and prostate cancer (in men) compared with the general population of Scotland, taking into account age and sex, both before and after UVA1 exposure. The SIRs for each tumour type are provided in Table 1. It is reassuring that none of the SIRs after UVA1 treatment were statistically significant, that the absolute numbers of skin cancers were small, and that they were not more commonly found in those who had received many treatments. Cancers in people in Tayside and Fife treated with ultraviolet (UV)A1, adjusted for sex and age PY, person‐years; O, number of observed cases; E, number of expected cancer cases; SIR, standardized incidence rate; CI confidence interval; NA, not applicable. aComparison with the general Scottish population adjusted for age and sex. No Merkel cell carcinoma or melanomas were found after UVA1 treatment. Cancers in people in Tayside and Fife treated with ultraviolet (UV)A1, adjusted for sex and age PY, person‐years; O, number of observed cases; E, number of expected cancer cases; SIR, standardized incidence rate; CI confidence interval; NA, not applicable. aComparison with the general Scottish population adjusted for age and sex. No Merkel cell carcinoma or melanomas were found after UVA1 treatment. The numbers are still modest and follow‐up was ≤ 20 years. Therefore, we cannot discount any increased risk among patients treated with UVA1. It is of interest that in this cohort there was an increased risk of cutaneous SCC and of non‐Hodgkin lymphoma before UVA1 treatment, suggesting a risk in this cohort of patients (perhaps in part because many were on systemic immunosuppressive drugs for conditions such as morphoea). Furthermore, although these were not significant (they might have been due to chance), there were decreased risks of melanoma and prostate cancer and a slightly increased risk of BCC before the UVA1 exposure; the SIRs were lower after UVA1 treatment. Thus, UVA1 could be related to the lack of increased risks or, potentially, could even be associated with decreasing the risks of these cancer types. When considering any intervention it is necessary to balance risks against expected benefits, and for many of the indications for UVA1, such as widespread morphoea, alternative treatments carry greater risks. The authors thank the Health Informatics Centre at the University of Dundee for providing access to databases and computational resources. Photosys, the database of Photonet, the National Managed Clinical Network for Phototherapy, was essential. Funding sources: This research was supported by the Chief Scientist Office, under grant TCS/16/47 to the University of Dundee and by the Dundee Photobiology Trust. Conflicts of interest: The authors declare they have no conflicts of interest.