Sun exposure and ocular melanoma in Australia

Abstract

The objectives of this project were to measure the incidence of ocular melanoma in Australia and to identify aspects of sun exposure, sun-related behaviours, artificial ultraviolet exposures, and host and genetic factors that may influence its risk. A population-based ocular melanoma incidence survey was conducted; a retrospective survey of cancer registries (1990-1995) and a prospective survey of cancer registries and ophthalmologists (1996-1998). These surveys identified around 20% under-enumeration by the Australian cancer registries. In 1996 to 1998, the incidence of ocular melanoma in Australia was similar to that in other populations of predominantly European origin, and higher in men (1.07 per 100,000, 95% CI 0.95- 1.19) than women (0.78 per 100,000, 95% CI 0.68-0.88). In men, the incidence was also significantly higher in rural compared to urban Australia. Consistent with previous studies, melanomas developed at the most sun—exposed locations within each ocular tissue.

A population-based case-control study was conducted in 1997 and 1998. Cases (n=290; 92% participation) were Australian residents aged 18 to 79 years diagnosed with ocular melanoma between January 1996 and July 1998 and notified to the ocular melanoma incidence survey prior to 315‘ July 1998. Controls (n=916; 67% participation) were randomly selected from the electoral rolls, frequency matched to cases by age, sex and State.

Subjects completed a self—administered postal questionnaire and a half-hour telephone interview with experienced interviewers masked to their case-control status and the hypotheses under study. The postal questionnaire included questions about pigmentary characteristics and a lifetime personal residence and work calendar, information from which was used to prompt recall of outdoor hours during the telephone interview. The sun exposure questionnaire was developed for this study by abbreviating a “gold standard” face-to-face questionnaire. Other information collected during the telephone interview included the use of protective devices, exposure to emissions from welding instruments and sunlamps, and history of eye and skin disease. Three measures of ambient solar irradiance were obtained for each residence, UVB, UVA and global solar radiation. Two non—white and 21 Asian-bom controls were excluded from all pigmentary and exposure analyses.

Unconditional logistic regression, including covariates for the matching variables, was used to identify risk factors by sub-site within the eye.

Risk of choroid and ciliary body melanoma (n=246) was increased in people with grey (OR=2.9), hazel (OR=2.2) and blue eyes (OR=1.7) compared to brown eyes. Risk was also increased in those with 4 or more naevi on their back, those unable to tan, and those who squinted when outdoors as a child. Risk was reduced in people born outside Australia and New Zealand (OR=0.7). Non-brown eye colour was a risk factor for iris melanoma (n=25). No constitutional risk factors were identified for conjunctival melanoma (n=19). Eye colour is the strongest constitutional predictor of choroid and ciliary body melanoma, and may indicate a protective effect of melanin density at these sites.

Choroid and ciliary body melanoma was positively associated with outdoor hours, specifically occupational, but not ambient solar irradiance. The association was consistent for parallel sun exposure measures. Risk increased with increasing exposure; the OR was 1.8 (95% CI 1.1-2.8) for the highest quarter of sun exposure on working days up to 40 years of age for men and women. The strongest positive associations were for lifetime exposure (OR=3.3), lifetime working-day exposure (OR=2.6), and exposure around 20 years of age (OR=2.8), in men. Other measures positively associated with risk were residence by the water during early life, rural residence at diagnosis, history of cataract and history of cutaneous melanoma. No sun-related risk factors were identified for iris melanoma, while conjunctival melanomas were positively associated with history of cutaneous melanoma.

Choroid and ciliary body melanoma was positively associated with duration of welding and use of both arc and oxy-acetylene instruments, but not with exposure to welding done by others. Risk was also significantly increased with use of sunlamps, particularly in men, with the strongest associations for lifetime hours of use (OR=4.4), duration of use (OR=3.8), first use at a young age, and first use after 1980. There were no significant associations with iris or conjunctival melanoma.

This research presents consistently strong evidence that sun exposure is an independent risk factor for choroidal and ciliary body melanoma in Australia. Evidence of association of these melanomas with exposure to artificial sources of ultraviolet radiation — welding and sunlamps — adds support.