This is the first epidemiologic study that evaluated the risk of adult asthma related to nonoccupational use of common household cleaning products. We found an association between the use of products in spray form and the incidence of asthma according to either more sensitive or more specific definitions. This association was linked predominantly to the most commonly used air fresheners, glass cleaners, and furniture cleaning sprays; was consistent for various subgroups and not dependent on atopic status; and the risk increased when frequency of use or number of different sprays increased. A relevant number of adult asthma cases may be related to the use of household cleaning sprays, indicating an important public health issue.
The use of sprays during the 1990s was very common in all countries of our study. Market trends from household cleaners' manufacturers show a general increase of aerosolized applications in Europe (17
). Sprays and more conventional liquid cleaners contain similar active ingredients, including alcohols, ammonia, chlorine-releasing agents, glycols and glycol ethers, sodium hydroxide, acryl polymers, and terpenes (18
). The application through spraying is likely to facilitate respiratory exposure to these components, explaining why we have observed associations with the use of sprays but not liquid cleaners. The latter will give off volatile components but relevant inhalatory exposure will depend on the dilution used, the surface to which they are applied, and the ambient temperature, among other factors. We may have missed an association of asthma with liquid cleaner use by not being able to account for these exposure-modifying factors. It is likely that the application of a spray typically leads to some degree of relevant inhalatory exposure, and this may have resulted in less exposure misclassification than for liquid cleaners. However, there are few data available to describe the exposure patterns associated with use of different cleaning products, and there are few experimental studies on emissions and exposures and they have mainly focused on volatile components after
application of cleaning products (18
). Thus, although correlation between the use of sprays and other cleaning products was in general low, it is not unlikely that our findings reflect a risk of broader use of home cleaning products.
Not many studies have evaluated adverse respiratory effects of cleaning products. Our findings are consistent with occupational epidemiologic studies in which an increased asthma risk was related to professional use of sprays among both domestic (5
) and nondomestic (6
) cleaning women. The observed associations may be (partly) due to chance, to confounding by a third variable, or may reflect a true adverse effect on new-onset asthma. Although chance can never be excluded in observational studies, this is highly unlikely here given the robust associations that were consistent for various subgroups based on host factors and country of residence, and the observed dose–response relationships. Confounding is possible if the use of sprays was associated with host or environmental risk factors of asthma. We controlled for potential host confounders, such as sex, age, and smoking status. In addition, we evaluated potential confounding effects of occupational exposures, and of socioeconomic status according to two definitions that can be regarded as reflecting a variety of housing and lifestyle factors (21
). It is difficult to hypothesize other possible host or environmental factors that could have confounded the observed association between spray use and asthma.
Our study design precludes strong conclusions regarding the responsible effect mechanisms. Given the fact that asthma was related to several types of sprays with different chemical composition, and that the risk was not dependent on atopic status, we speculate that asthma could have been at least partly irritant induced. Cleaning sprays may contain sensitizers such as disinfectants, amines, pinene, or limonene (18
), and therefore a role of specific sensitization resulting in asthma is also plausible. From occupational settings, asthma can follow one-time intense irritant exposure, and there is increasing acceptance of the possibility that recurrent low-grade exposures to respiratory tract irritants can result in asthma as well (22
). The underlying mechanisms are largely unknown, but a localized airway inflammatory response is likely involved. A similar phenomenon for repeated household exposures to irritants seems plausible, despite the fact that frequency of exposure in nonprofessional home cleaning is generally lower than in professional domestic or nondomestic cleaning.
There are a number of limitations in our study that need to be considered. First, data on both product use and health outcome were based on questionnaire information at follow-up, introducing the possibility of differential misclassification and a bias away from the null. This would be the case if participants with new-onset asthma reported more use and/or recalled better their use of cleaning sprays. However, this is unlikely to be a major explanation given the fact that during the 1990s there was not much public awareness of adverse respiratory effects of domestic cleaning activities. Data in this study were collected before 2003, the year in which an article was published on associations between domestic cleaning work and asthma (4
), which received much media attention worldwide and likely initiated public awareness.
Second, scented products are widely reported by individuals with asthma to trigger symptoms (23
). Although it is possible that the asthma in this study is due to the scented component of cleaning agents, it is more likely that those with asthma avoided such products and therefore could have biased associations toward the null. Analysis of the specific question covering all types of perfumed and scented cleaning products showed that the frequency of use was not associated with asthma ().
Third, results using the objective outcome BHR were not consistent with the main findings using the three a priori definitions of asthma. Only a minority of the participants with asthma showed BHR, and the vast majority (>80%) of participants with BHR did not have asthma. Thus, despite using a definition with relatively high specificity (a 20% fall in FEV1 using 1 mg methacholine as a cutoff), BHR was not particularly specific for asthma. With the combination of temporal variability and a generally moderate reproducibility of methacholine challenge testing, it is difficult to judge to which extent the lack of association of spray use with BHR contradicts the overall positive findings for asthma. An additional limitation was that current asthma and wheeze were defined as the occurrence in the previous year as reported at the follow-up interview. Although we used a prospective study design, it is possible that in the analyses of cumulative incidence (symptoms in the last year), the time order of exposure and effect was confused. In other words, individuals who developed asthma during follow-up could tend to clean their homes more thoroughly. Albeit with less statistical power, findings for the more specific asthma definition based on diagnosis using conventional survival analysis of incidence were consistent, and they therefore do not support this possibility.
Finally, although not statistically significant, there appeared to be a certain degree of heterogeneity in the association between spray use and asthma among countries. Unrecognized confounding could have been different for different countries, creating false-positive associations in some countries and/or hiding true positive associations in others. The qualitative differences in the use of sprays as outlined in did not provide a clear hypothesis for the observed differences in the risk of any spray use among countries. Nevertheless, associations between the use of specific sprays and asthma incidence were more homogeneous across countries (results not presented). Whether chemical composition of cleaning sprays differs among countries, possibly related to the predominant brands sold on the local markets, remains unclear, and justifies more specific investigation.
Findings of our study may have significant implications for public health. Relative risks of 1.3 to 1.5 in combination with an overall proportion of 42% of weekly spray users suggest a population attributable fraction of about 15%. In other words, one in seven adult asthma cases could be attributed to common spray use. This indicates a relevant contribution of spray use to the burden of asthma in adults who do the cleaning in their homes. In addition, passive exposure might be relevant for individuals present in environments where sprays are being or have just been applied. One study even suggested that the use of cleaning and other household chemicals by the mother during pregnancy was related to wheeze in young children (24
We conclude that frequent use of household cleaning sprays may be an important risk factor for adult asthma. This finding needs to be confirmed in future studies, with a particular emphasis on chemical composition and other exposure determinants, and on the effect mechanisms involved, including sensitization and inflammatory reactions.