Bioindicator and exposure data for a population based study of manganese

Exposure data and bioindicators were obtained for a study whose objective w as detection of early manifestations of manganese (Mn) neurotoxicity in a p opulation with potential environmental exposure. The study included persons with no history of neurotoxic workplace exposure in Southwest Quebec, draw n from seven postal code regions, defining a set of geographically contiguo us zones. Blood samples were analyzed for total Mn (MnB) lead (PbB) total m ercury (HgT) and serum iron (FeS). Drinking water samples from participants ' residences were analyzed for manganese (MnW). At 4 sites, limited 24-hour high volume air samples for total particulates (TP) and PM10, were analyze d for Mn and Pb. Sociodemographic and dietary information was obtained by s elf-administered questionnaire. The geometric mean (GM) for MnB values (n=2 97) was 7.14 mu g/L. Levels of MnB in women (n=156); GM 7.50 mu g/L) were s ignificantly higher than in men (n=141; GM 6.75 mu g/L). No relationship wa s found between MnB and PbB or HgT: FeS was significantly higher in men (GM 18.38 mu mol/L) than women (GM 15.0 mu mol/L). For women, MnB was correlat ed to FeS, with a tendency to decrease with increasing age. For men, no rel ationship was found between MnB levels and either FeS or age, although FeS showed a strong inverse relationship with age. The 24-hour mean levels of M nTP at the 4 sites varied between 0.009 mu g/m(3) and 0.035 mu g/m(3); inte rsite differences were not significant. For Mn in PM10 (MnPM10) mean values ranged from 0.007 mu g/m(3) to 0.019 mu g/m(3); intersite differences were significant. A total of 278 MnW samples were obtained, 16 from residences served by wells. The GM for MnW was 4.11 mu g/L (range: 0.50-71.1 mu g/L, e xcluding wells; MnW for wells ranged from non-detectable to 158.9 mu g/L. I ndividually, there was no relation between MnW and MnB. Geographic analysis of the MnB and MnW data by an algorithm grouping contiguous postal code zo nes, combined with air data, lead to definition of a geographic parameter, distinguishing two regions relative to a former manganese alloy plant, whic h contributed significantly to MnB. A multiple regression model was develop ed, explaining 6.7% of the variability in MnB (F=5.12; p<0.001); when contr olling for gender, geographic region with higher levels of airborne mangane se and the frequency of consumption of cereals and leaf vegetables contribu ted positively to MnB levels, while serum iron was negatively related. (C) 1999 Inter Press, Inc.