Epidemiological evidence supports the aetiological role of natural (sunlight) and artificial (sunlamps and sunbeds) sources of ultraviolet radiation (UV-A) on the occurrence of cutaneous malignant melanoma (CMM).¹ Occupational studies have suggested increased risks for CMM among chemists, telecommunications and electronics workers,² printing industry workers,³ and harbour workers (forwarding/shipping agents, harbour masters, and ferry and harbour service assistants).⁴

To investigate the relation between occupation and cancer incidence, we retrospectively studied cancer incidence in 4993 longshoremen ever employed at the dock of Genoa, Italy, between 1933 and 1980. They were employed at two dockyard trading companies: the “Stefano Canzio” and “San Giorgio”. Although men employed at the two dockyard companies loaded and unloaded the same products, employees of the former company performed their job mainly outdoors, and employees of the latter mostly inside the ships. They were categorised a priori according to their prevalent pattern of occupational sunlight exposure as indoor (2707) and outdoor workers (2286). Cancer frequency was established by record linkage with the Genova Cancer Registry, for 1986–96 (the interval for which incidence data are available). The vital status of each man was ascertained from the demographic registry of his place of residence until 31 December 1996. Those who died (409) or emigrated (32) before 1986 (that is, the starting date of follow up), were excluded from the analysis. Thus 2451 indoor and 2101 outdoor dockyard workers were eligible for statistical analysis.

Standardised incidence ratios (SIRs) were calculated as the ratio of observed to expected site specific cancer cases (external comparison). Expected cases were calculated by applying quinquennial age specific cancer incidence for the male population of the City of Genova (reference population) to the person-years of observation accumulated in each subgroup (24 364 and 21 087 for indoor and outdoor workers, respectively). Two sided 95% confidence intervals (CIs) for the SIRs were calculated on the assumption of a Poisson distribution of the observed cases. In addition, the site specific cancer incidence experienced by outdoor and indoor workers was contrasted by computing relative risk point estimates and their 95% CIs.⁵

Table 1 shows the results of the cohort study. All cancers incidence was similar in both subgroups and did not differ from that of the reference population. The excess incidence for larynx cancer detected in both subgroups was statistically significant only in indoor workers (24 cases, SIR = 213, 95% CI 136 to 316, p = 0.001). SIR for lung cancer was similar in the two subcohorts and did not differ from that of the reference population. Significantly increased SIRs were observed for pleural mesotheliomas in indoor (16 cases, SIR = 1362, 95% CI 778 to 2211, p < 0.0001) and outdoor dockyard workers (7 cases, SIR = 751, 95% CI 302 to 1547, p = 0.0001).

Eight cases of CMM were observed in outdoor workers (SIR = 288, 95% CI 125 to 568, p = 0.015), and three cases among indoor workers (SIR = 97, 95% CI 20 to 284, p = 0.99). Table 2 shows main anatomical site, age at diagnosis, ICD-9 code, and job type for each incident case of CMM.

SIR for skin cancer (other than CMM) was decreased (not significantly) in both indoor and outdoor longshoremen (table 1). A moderately increased SIR for bladder cancer was observed among outdoor workers (33 cases, SIR = 135, 95% CI 93 to 189, p = 0.118).

Internal comparison (data not shown), revealed a similar incidence for all cancers (RR = 0.97, 95% CI 0.81 to 1.16), digestive tract (RR = 0.99, 95% CI 0.67 to 1.49), and skin cancer (RR = 0.93, 95% CI 0.43 to 1.95) in the two subgroups. A higher incidence for CMM (RR = 2.97, 95% CI 0.71 to 17.41) and bladder cancer (RR = 1.66, 95% CI 0.96 to 2.91), and lower incidence for pleural mesotheliomas (RR = 0.56, 95% CI 0.19 to 1.42) were detected in outdoor compared to indoor workers. None of the RR values were statistically significant, a finding that is due to similar exposures shared by the two subgroups and the lower statistical power achieved by internal than external comparison.

The threefold increased risk that was detected for CMM only among outdoor workers supports the causal role of exposure to sunlight, and is apparently in contrast with the previously reported evidence of a potential association with occupational exposure among chemists, telecommunications and electronics, and printing industry workers.^2–4

This, together with a lower than expected incidence of skin cancer (other than CMM) observed in both indoor and outdoor workers, suggests that exposure to sunlight and to carcinogenic agents that were present in the dockyard environment are required in the development of CMM.