Objective: To study mortality of male workers employed in the petrochemical plant located in Gela, Sicily, Italy.
Design: A cohort of 6458 workers employed in the period 1960-1993 was enumerated from company payrolls, follow-up was between November 1, 1960 and December 31, 2002. Ascertainment of vital status was completed through Registrar's office at the place of residence/birth and linkage to the Internal Revenue Service database. Causes of death were retrieved from Registrar's office at the place of death and the National Death Index and coded using the ICD Revision at the time of death. The cause specific expected mortality was computed relative to Sicily Region, specific for gender, 5-year age groups and calendar year.
Setting: Petrochemical plant in Gela, Italy.
Main outcome measures: Cause specific Standardized Mortality Ratios (SMRs) and 90% confidence interval (CI).
Results: Observed mortality was below expected for all causes (662 cases, SMR 0.70 90% CI 0.66-0.74) and all malignant tumours (210 cases, SMR 0.71 90% CI 0.63-0.79). For the cancer sites of a priori interest in the petrochemical industry (skin, lymphatyic and hematopoietic tissues, lung, bladder, kidney liver, brain, pleura) the lower confidence interval of the SMR is never upper than unity.
Conclusions: In the cohort the Healthy Worker Effect, the lack of information on exposure and the high percentage of missing causes of deaths (7.6%) can result in an underestimation of mortality specifically for rare diseases as the ones of a priori interest. Information and classification of exposure over time are envisaged developments to better describe the mortality pattern of the Gela cohort.