Poster Presentation Skin Cancer 2024

The counterintuitive association of sun exposure and melanoma-specific mortality in a cohort of Norwegian women (#187)

Ashley Ahimbisibwe 1 , Flavie Perrier 1 , Morten Valberg 2 3 , Adele C. Green 4 5 , Reza Ghiasvand 2 6 , Corina S. Rueegg 2 , Elisabete Weiderpass 7 , Tonje B. Braaten 8 , Trude E. Robsahm 6 , Marit B. Veierød 1
  1. Oslo Centre for Biostatistics and Epidemiology, University of Oslo, Oslo, Norway
  2. Oslo Centre for Biostatistics and Epidemiology, Oslo University Hospital, Oslo, Norway
  3. Department of Community Medicine and Global Health, Institute of Health and Society, University of Oslo, Oslo, Norway
  4. Department of Population Health, QIMR Berghofer Medical Research Institute, Brisbane, Brisbane, Australia
  5. Cancer Research UK Manchester Institute, University of Manchester, Manchester, United Kingdom
  6. Research Department, Cancer Registry of Norway, Oslo, Norway
  7. 7International Agency for Research on Cancer, Lyon, France
  8. Department of Community Medicine, Faculty of Health Sciences, The Arctic University of Norway, Tromsø, Norway

Introduction: Counterintuitively, some studies have found that sun exposure, a well-known melanoma risk factor, may be associated with improved survival among melanoma patients. Using the Norwegian Women and Cancer cohort (NOWAC), we investigated the association between pre-diagnostic sunburns and melanoma-specific death, and potential impact of unmeasured confounding and selection bias (common causes of counter-intuitive associations in observational studies).

Methods: We ascertained all first primary cutaneous melanoma in NOWAC via the Cancer Registry of Norway. The association between pre-diagnostic sunburns and melanoma-specific death was estimated with Cox regression. Observed covariate E-values were used to investigate the potential impact of unmeasured confounding based on the observed covariates (histopathological variables, age at diagnosis, hair color, body mass index, education, UVR exposure and sunbathing vacations). Frailty analysis quantified the potential magnitude of selection bias.

Results: Among 2234 melanoma patients, 168 died from melanoma after mean follow-up of 8.8 years. We found reduced risk of melanoma-specific death for ever- versus never-sunburn (hazard ratio: 0.46, 95% confidence interval: 0.27–0.77) and significant negative trend for cumulative number of sunburns (ptrend=0.005). Unmeasured covariates altering the association is unlikely as they needed to be more strongly associated with both sunburn and melanoma-specific death than all the variables in our analysis. Frailty analysis suggested the observed relative risk between sunburn and melanoma-specific death may reflect selection bias.

Conclusion: The observed reduced risk of melanoma-specific death in women with pre-diagnostic sunburn is probably not explained by unmeasured confounding but could be explained by selection bias, i.e., including only patients.