The concentration of DEHP in indoor dust was in about the same concentration range as in several other studies, including the Swedish DBH study (Bornehag et al. 2004
) (). The concentration of BBzP was found in somewhat higher concentration in Bulgaria compared with other studies. Large differences were observed for DnBP, which was found in concentrations up to 40 times higher in Bulgaria than in other countries (), and for DEP and DMP, which were also found in much higher concentrations in the present study than in other studies (Bornehag et al. 2004
; Fromme et al. 2004
). None of the main sources identified (balatum
flooring, polishing products) were correlated with the concentration of DnBP, and only a slight, nonsignificant correlation was found between DEP and use of polishing products. No information on other possible sources of these phthalates (e.g. cosmetics, soft toys, plastic covers on furniture) was obtained, so we are unable to put forward any hypotheses that might explain the high concentrations of DnBP and DEP that were found in this study.
Measurements of the concentration of phthalates in dust in different countries.
The results support the finding from the Swedish DBH study regarding an association between DEHP and asthmatic and allergic symptoms among children (Bornehag et al. 2004
). Although the Swedish findings regarding an association of the dust concentration of BBzP with rhinitis and eczema were not replicated in a significant manner in the current study, the results from Bulgaria point to the same conclusion as in Sweden.
In the present study, each child was treated as separate observation, because in homes with multiple children participating in the study, each child had a separate bedroom, so a separate dust sample was collected. However, all analyses performed on only one child per home (177 children) gave corresponding results, with even stronger significance (data not shown).
In contrast to the results of the Swedish DBH study (Bornehag et al. 2005a
), we found no correlation between concentrations of DEHP and BBzP in dust and the presence of balatum
in children’s sleeping rooms. One explanation can be that in Bulgaria the word balatum
has two meanings: linoleum or PVC. The fact that some of the “plastic” flooring was in fact linoleum can therefore have weakened or even removed any possible correlation between PVC flooring and phthalate concentration. For both studies, however, an association between health and the presence of PVC/balatum
flooring in the child’s room was observed. In the present study, of 60 homes with balatum
flooring (34% of investigated homes), 41 were found in the homes of case children ().
Because the use of polishing agents and the concentration of phthalates are correlated with each other and both of them were associated with symptoms, the association between health outcomes and DEHP concentration was tested in both adjusted and stratified analyses. We found a dose–response relationship between DEHP in dust and symptoms both in buildings with high and low polish use, with ORs in the highest quartiles (compared with the lowest quartile group): OR = 4.20 (95% CI, 0.96–18.33) for case status, and OR = 3.73 (95% CI, 0.85–16.44) for wheezing in the group of homes with high polish use; and OR = 1.93 (95% CI, 0.62–5.98) for case status and OR = 2.67 (95% CI, 0.79–8.95) for wheezing in the low polish use group. Also, with adjusting for polish use, the significant association between the concentration of DEHP and case status/ wheezing were found for the highest quartile (compared with the lowest quartile group): OR = 2.61 (95% CI, 1.07–6.37) for case status, and OR = 3.08 (95% CI, 1.21–7.83) for wheezing in preceding 12 months. This means that the association between phthalates and health cannot be explained by different use of polishing agents only; however, the use of such compounds seems to reinforce the association between health and the concentration of selected phthalates.
It is reasonable to expect that case families with asthma and allergy clean more frequently than control families, which could increase the amount and age of the indoor dust and thus the content of phthalates in control homes. In this study we found no difference in cleaning frequency in homes of cases and controls, and we found the opposite in dusting furniture—namely, that controls clean more frequently than case families. Parents of 44% of controls and 32% of cases reported that they customarily clean their furniture more often than once a week. However, adjusting for dusting frequency did not change the results in this study. Additionally, the only phthalate significantly associated with low dusting frequency was BBzP, which was found in significantly higher concentrations in homes where dusting was carried out once a week or less often, compared with more frequent dusting (p = 0.007). DEHP and DnOP were also found in higher concentrations in homes where dusting was not as frequent, compared with homes with more frequent dusting, although this difference was not significant.
Compared with Sweden, there were numerous problems in conducting the study in Bulgaria. The response rate in the cross-sectional study was very low (34.5%) compared with Sweden (almost 80%). In the current case–control study, several families refused to allow inspections or dust sampling, despite having given prior consent. These differences are most probably attributed to political and socioeconomic factors that are outside the scope of this study.
A low response rate is always a problem in epidemiologic investigations. However, such problems mainly involve representativity and may not introduce bias when analyzing associations between exposures and health effects. The baseline study was used mainly to select possible cases and controls (sick and healthy children).
Because it is known that families with allergic diseases are more prone to participate in epidemiologic studies, potential selection bias in the baseline questionnaire was investigated. A total of 240 children (78 in Burgas and 162 in Sofia) were randomly selected among nonresponders; they were contacted by phone and asked to answer all questions used in the questionnaire. Excluding eczema in preceding 12 months and asthma/allergic symptoms among family members, there was no difference between the children whose parents took part in the ALLHOME-1 study and those whose parents refused to do so, limiting the risk for serious selection bias regarding allergic and asthmatic symptoms and diagnoses (Naydenov 2007
Among children whose parents agreed to participate in the ALLHOME-2 study, the prevalence of current wheezing, rhinitis, and eczema was higher compared with children whose parents did not agree to participate in the ALLHOME-2 study or did not reply in the follow-up study (Naydenov 2007
). More health problems in the case families was one of the selection factors for participation in the Swedish DBH study (Bornehag et al. 2006
). However, such selection bias results in a greater contrast in health status between cases and controls, and hence a greater possibility of identifying differences in health-relevant exposures (Bornehag et al. 2006
In this study, we analyzed associations between children’s health (as reported by parents), and measured or inspected indoor environmental factors. A bias in the baseline study could be introduced if parents with sick children knew the risk factors for their child’s illness and reported more of such factors. However, such risk factors were measured or observed by “blinded” inspectors in the second phase. Also, the main potential risk factor studied—the concentration of specific phthalates—is not known to be a risk factor by the general public in Bulgaria, Sweden, or elsewhere. A strong source of such compounds, PVC flooring, could be recognized and reported, and thus be a proxy for phthalates. This was not possible in Bulgaria, where there is not even a word for PVC flooring, because balatum is used for both PVC and linoleum flooring. The same discussion is valid with regard to use of polish.
This study and the DBH study in Sweden both show an association between DEHP in indoor dust and airway symptoms in preschool children. The two studies were performed in very different regions of Europe with regard to building type, climate, and political and socioeconomic factors. The mechanisms behind these results are not known. Some toxicologic studies have indicated that DEHP may act as an adjuvant factor (Chalubinski and Kowalski 2006
; Nielsen et al. 2007
). However, it is impossible to determine whether the important exposure is during childhood or during pregnancy.