Parkinson's disease and occupation: differences in associations by case identification method suggest referral bias.

METHODS: We used a population-based sample of 403 Parkinson's disease cases and 405 controls to examine risks by occupation. Results were compared to a previous clinic-based analysis. RESULTS: With censoring of jobs held within 10 years of diagnosis, the following had significantly or strongly increased risks: social science, law and library jobs (OR = 1.8); farming and horticulture jobs (OR = 2.0); gas station jobs (OR = 2.6); and welders (OR = 3.0). The following had significantly decreased risks: management and administration jobs (OR = 0.70); and other health care jobs (OR = 0.44). CONCLUSIONS: These results were consistent with other findings for social science and farming occupations. Risks for teaching, medicine and health occupations were not elevated, unlike our previous clinic-based study. This underscores the value of population-based over clinic-based samples. Occupational studies may be particularly susceptible to referral bias because social networks may spread preferentially via jobs.

Strong mediators of social inequalities in risk of ischaemic heart disease: a six-year follow-up in the Copenhagen Male Study.

OBJECTIVE: Large social inequalities exist in risk of ischaemic heart disease (IHD) in Western populations; inequalities which are only little accounted for by established risk factors. We wished to find out if some newly identified cardiovascular risk factors in concert with established factors might contribute further to the explanation. DESIGN AND SETTING: A 6-year follow-up in the Copenhagen Male Study. SUBJECTS: Some 2974 males aged 53-75 years (mean 63) without overt cardiovascular disease were included in the study. Potential confounders included were: alcohol, physical activity, smoking, serum lipids, serum cotinine, serum selenium, lifetime occupational exposure to soldering fumes and organic solvents, body mass index, blood pressure, hypertension, use of sugar in hot beverages, use of diuretics, and Lewis phenotypes. MAIN OUTCOME MEASURES: During the 6-year follow-up period (1985/1986-1991), 184 men (6.2%) had a first IHD event. Compared to higher social classes (classes I, II and III), lower classes (classes IV and V) had a significantly (P < 0.05) increased risk of IHD; age-adjusted relative risk (RR) with 95% confidence limits was 1.44 (1.1-1.9), P = 0.02. After multivariate adjustment for age, blood pressure, serum lipids, physical activity, and smoking, the RR dropped to 1.38 (1.0-1.9), P = 0.05. Some newly identified risk factors were significantly associated with increased risk of IHD as well as with low social class: a low serum selenium concentration, a low level of leisure time physical activity in midlife, long-term exposure to soldering fumes, and abstention from or a low consumption of wine and strong spirits. After adjustment for these factors also, the RR dropped to 1.12 (P = 0.54). CONCLUSIONS: The results of this study suggest that potentially modifiable risk factors associated with lifestyle and working environment are strong mediators of social inequalities in risk of ischaemic heart disease.