RESULTS
Demographic and environmental characteristics of the 6 Kosovar centres
have been already described15. Table 1 shows the
prevalence of current eczema symptoms, severe eczema symptoms and eczema
ever overall, and with and without wheeze, together with the respective
95% confidence intervals. The lowest prevalence of current eczema
symptoms occurred in Ferizaj while the highest was in Gjakova. Eczema
ever was lowest in Ferizaj again and highest in Gjakova. Overall, the
proportion of adolescents with eczema and wheezing symptoms was much
lower than that with only eczema symptoms. The prevalence of severe
eczema symptoms was very low and similar among the centres.
Web tables 1 and 2 include the associations found in the univariate and
multivariate logistic regressions of the different factors with current
eczema symptoms and eczema ever, respectively, per centre. Those factors
that were consistently associated with current eczema symptoms or eczema
ever in most centres were sex, exercise, and paracetamol intake. Forest
plots of those three factors showed a trend that male sex (figure 1) was
associated to lower prevalence of eczema: pooled aOR=0.49; 95%CI
0.37-0.66, for current eczema; and pooled aOR=0.68; 96%CI 0.44-1.06,
for eczema ever.
Figure 2 shows that exercise was associated to higher prevalence of
current eczema symptoms and eczema ever, with no apparent dose-effect
relationship. The pooled aOR and 95%CI for any frequency of exercise
was 2.79 (1.89-4.10) and 2.07 (1.48-2.90) respectively for current
eczema symptoms and eczema ever.
Contrary to exercise, paracetamol intake seemed to have a dose-response
pattern of association, especially for current eczema symptoms. The
pooled aOR and 95%CI were 1.33 (0.81-2.16) and 2.64 (1.61-4.33) for the
intake of “at least once a year” and of “at least once a month”,
respectively. The corresponding figures for eczema ever were: 1.09
(0.75-1.58) and 1.37 (0.85-2.22) (figure 3).
According to the results of the multilevel analyses, stratification by
wheeze did not change the patterns of associations, although in those
children who suffered from skin plus respiratory symptoms those
associations seemed to be stronger. However, the low number of children
in the eczema and wheeze group was considerably lower than in the group
with eczema without wheeze, making comparisons problematic (table 2).