The objective of this study was to test the hypothesis of a lower mortality from cancer and cardiovascular diseases among men expressing glucose-6-phosphate dehydrogenase (G6PD) deficiency. We designed a mortality study based on death certificates from January 1, 1982 through December 31, 1992 in a cohort of G6PD-deficient men. Cohort members were 1,756 men, identified as expressing the G6PD-deficient phenotype during a 1981 population screening of the G6PD polymorphism. The setting was the island of Sardinia, Italy. Outcome measures were cause-specific standardized mortality ratios (SMRs), which were computed as 100 times the observed/expected ratio, with the general Sardinian male population as the reference. Deaths from all causes were significantly less than expected due to decreased SMRs for ischemic heart disease (SMR, 28; 95% confidence interval [CI], 10 to 62), cerebrovascular disease (SMR, 22; 95% CI, 6 to 55), and liver cirrhosis (SMR, 12; 95% CI, 0 to 66), which explained 95.6% of the deficit in total mortality. All cancer mortality was close to the expectation, with a significant increase in the SMR for non-Hodgkin's lymphoma (SMR, 545; 95% CI, 147 to 1,395). A decrease in mortality from cardiovascular diseases was one of the study hypotheses, based on an earlier human report and experimental evidence. However, selection bias is also a likely explanation. Further analytic studies are warranted to confirm whether subjects expressing the G6PD-deficient phenotype are protected against ischemic heart disease and cerebrovascular disease. This cohort study is consistent with more recent case-control studies in rejecting the hypothesis of a decreased cancer risk among G6PD-deficient subjects. The observed increase in mortality from non-Hodgkin's lymphoma and decrease in mortality from liver cirrhosis were not previously reported.

Occupational associations with gastric cancer were investigated in a multicenter case-control study in Italy involving interviews with 640 histologically confirmed male cases and 959 controls, randomly selected from the resident populations of the study areas. From information on the three jobs each person held the longest, risks were evaluated according to employment in 35 occupations (ever or 21+ years) and to estimated exposure (ever or 21+ years) to six chemicals using a job-exposure matrix. All risk estimates were adjusted by personal, demographic, and dietary variables identified as gastric-cancer risk factors in previous analyses. The only significantly increased risk was observed for sailors, seamen, and allied groups (ever employed: odds ratio [OR] = 2.9; 95 percent confidence interval [CI] = 1.1-8.0; 21+ years: OR = 3.1, CI = 0.8-13). Nonsignificant increases after 21+ years of employment were observed for forestry workers, miners, and janitors and cleaners. Crude ORs were elevated significantly among farmers, but adjusting for demographic and lifestyle factors largely eliminated the association: a nonsignificant 30 percent excess risk remained for farm laborers, but there was no rise in risk among long-term farm laborers and no excess among farm owners. Application of the job-exposure matrix revealed excess risks of borderline significance associated with potential exposure to mineral dusts and nitrogen oxides. For subjects with 21+ years of potential exposure, nonsignificantly increased risks were related to mineral dusts, asbestos, fertilizers, and nitrosamines.(ABSTRACT TRUNCATED AT 250 WORDS)

Objectives-To evaluate the association between silica, silicosis and lung cancer, the mortality of 724 patients with silicosis, first diagnosed by standard chest x ray film between 1964 and 1970, has been analysed by a cohort study extended to 31 December 1997. Methods-Smoking and detailed occupational histories were available for each member of the cohort as well as the estimated lifetime exposure to respirable silica dust and radon daughters. Two independent readers blindly classified standard radiographs according to the 12 point International Labour Organisation (ILO) scale. Lung function tests meeting the American Thoracic Society's criteria were available for 665 patients. Standardised mortality ratios (SMRs) for selected causes of death were based on the age specific Sardinian regional death rates. Results-The mortality for all causes was significantly higher than expected (SMR 1.35, 95% confidence interval (95% CI) 1.24 to 1.46) mainly due to tuberculosis (SMR 22.0) and to non-malignant chronic respiratory diseases (NMCRD) (SMR 6.03). All cancer deaths were within the expected numbers (SMR 0.93; 95% CI 0.76 to 1.14). The SMR for lung cancer was 1.37 (95% CI 0.98 to 1.91, 34 observed), increasing to 1.65 (95% CI 0.98 to 2.77) allowing for 20 years of latency since the first diagnosis of silicosis. Although mortality from NMCRD was strongly associated to the severity of radiological silicosis and to the extent of the cumulative exposure to silica, SMR for lung cancer was weakly related to the ILO categories and to the cumulative exposure to silica dust only after 20 years of lag interval. A significant excess of deaths from lung cancer (SMR 2.35) was found among silicotic patients previously employed in underground metal mines characterised by a relatively high airborne concentration of radon daughters and among ever smokers who showed an airflow obstruction at the time of the first diagnosis of silicosis (SMR 3.29). Mortality for lung cancer related to exposure was evaluated with both the Cox's proportional hazards modelling within the entire cohort and a nested case-control study (34 cases of lung cancer and 136 matched controls). Both multivariate analyses did not show any significant association with cumulative exposure to silica or severity of silicosis, but confirmed the association between mortality for lung cancer and relatively high exposure to radon, smoking, and airflow obstruction as significant covariates. Conclusions-The findings indicate that the slightly increased mortality for lung cancer in this cohort of silicotic patients was significantly associated with other risk factors-such as cigarette smoking, airflow obstruction, and estimated exposure to radon daughters in underground mines-rather than to the severity of radiological silicosis or to the cumulative exposure to crystalline silica dust itself. (Occup Environ Med 2001;58:786-793)

Checking the resident 1,868 males found to be G6PD deficient out of 15,964 examined, we observed the percentage of G6PD deficiency in male population largely varying among the various Sardinian towns ranging from 1 to over 50%. As preliminary analysis, in 24 towns out of 39 in which the sample was sufficiently suggestive of their whole resident male population, we calculated Standardized Mortality Rates (SMR) for all cancers and linear regression between SMRs and frequency ofG6PD deficiency in the same towns. Previous observations suggested the hypothesis of a negative correlation between frequency of G6PD deficiency and incidence of cancer, as we found in the preliminary analysis of this study. The results are not sufficient either to accept or to refuse this hypothesis.

The objective of this study was to test the hypothesis of a lower mortality from cancer and cardiovascular diseases among men expressing glucose-6-phosphate dehydrogenase (G6PD) deficiency. We designed a mortality study based on death certificates from January 1, 1982 through December 31, 1992 in a cohort of G6PD-deficient men. Cohort members were 1,756 men, identified as expressing the G6PD-deficient phenotype during a 1981 population screening of the G6PD polymorphism. The setting was the island of Sardinia, Italy. Outcome measures were cause-specific standardized mortality ratios (SMRs), which were computed as 100 times the observed/expected ratio, with the general Sardinian male population as the reference. Deaths from all causes were significantly less than expected due to decreased SMRs for ischemic heart disease (SMR, 28; 95% confidence interval [CI], 10 to 62), cerebrovascular disease (SMR, 22; 95% CI, 6 to 55), and liver cirrhosis (SMR, 12; 95% CI, 0 to 66), which explained 95.6% of the deficit in total mortality. All cancer mortality was close to the expectation, with a significant increase in the SMR for non-Hodgkin's lymphoma (SMR, 545; 95% CI, 147 to 1,395). A decrease in mortality from cardiovascular diseases was one of the study hypotheses, based on an earlier human report and experimental evidence. However, selection bias is also a likely explanation. Further analytic studies are warranted to confirm whether subjects expressing the G6PD-deficient phenotype are protected against ischemic heart disease and cerebrovascular disease. This cohort study is consistent with more recent case-control studies in rejecting the hypothesis of a decreased cancer risk among G6PD-deficient subjects. The observed increase in mortality from non-Hodgkin's lymphoma and decrease in mortality from liver cirrhosis were not previously reported.