This study aimed to investigate the relation between environmental influenc
es such as arsenic, cadmium, lead, and mercury, as well as environmental to
bacco smoke, pet keeping, and heating systems on the prevalence of atopic e
czema. Therefore, a multicenter cross-sectional study of school children ag
ed 5-14 years, including a standardized questionnaire, blood and urine anal
yses, and a dermatological examination, was undertaken. A cases-control app
roach was chosen in order to identify relevant risk factors. A total of 220
0 school children (response 79.1%) of two areas (Bitterfeld, Hettstedt) pol
luted by industrial activities and an agricultural control region (Zerbst)
of the former German Democratic Republic were examined. Atopic eczema as id
entified by dermatological examination and history was the outcome variable
of interest. Body burden of arsenic and heavy metals and questionnaire dat
a on environmental tobacco smoke exposure, pet keeping, and heating system
were investigated as potential risk factors. The overall prevalence of atop
ic eczema was 2.6%, with higher prevalences in the industrial areas (2.5 an
d 2.9%) compared to the control area (1.6%, not significant). Bivariate ana
lyses did not reveal statistically significant associations between atopic
eczema and tobacco smoke exposure or the body burden of arsenic and heavy m
etals. According to multiple logistic regression analysis, atopic eczema wa
s significantly more frequent in predisposed families and those who reporte
d keeping guinea pigs (OR 4.37, CI 2.15-8.91), but not other pets, like dog
s, cats, and hamsters. In comparison to a distant heating system, a decreas
ed risk was observed in households with central heating system (OR = 0.30,
CI 0.10-0.90), whereas the presence of a gas heater with an exhaust pipe co
nnection to the wall was associated with a significantly elevated risk for
eczema (OR = 8.22, CI 2.44-27.66). The heating system and exposure to certa
in animal allergens are related to the manifestation of atopic eczema. Furt
her studies are needed to clarify how far a causal relationship is reflecte
d by these findings. (C) 1999 Academic Press.