Measles-Mumps-Rubella Vaccination and Asthma-like Disease in Early Childhood

Anders Hviid; Mads Melbye


Am J Epidemiol. 2008;168(11):1277-1283. 

In This Article


We evaluated the association between MMR vaccination and asthma-like disease in early childhood. Our results show that MMR-vaccinated children are less often hospitalized with asthma diagnoses and use less anti-asthma medication than unvaccinated children. This effect was most pronounced among the youngest children and children who were vaccinated age-appropriately.

At birth, the immune system is immature. The "hygiene hypothesis" states that infections in early childhood play an important role in maturation of the immune system and consequently the risk of developing atopy.[1,2] The hygiene hypothesis is well supported epidemiologically, but the immunologic framework remains to be established.[16,17] Measles infection has been associated with reduced risk of atopy, although findings have been contrasting.[18,19] The MMR vaccine contains live attenuated measles virus and elicits an immune response which is similar to that of natural measles infection, with both cellular and humoral immune responses.

A limited number of studies have evaluated the association between MMR vaccination and asthma and other atopic diseases. The majority of these investigators have reported no association, and a few have reported a negative association.[7,8,20] One of the previously most extensive analytical studies was conducted by DeStefano et al..[21] In a US cohort including 18,407 cases, the authors reported a relative risk of asthma of 0.97 (95% CI: 0.91, 1.04) for MMR vaccination. In a subanalysis, the authors took into account the possible influence of medical-care utilization bias and reported a relative risk of 0.80 (95% CI: 0.61, 1.04) for children with at least 2 medical-care encounters during the first year of life. Even though statistical significance was not achieved, this reduced relative risk compares well with our results. Roost et al.[22] conducted a cross-sectional study among approximately 1,500 Swiss children with self-reported information on asthma. Both natural measles and measles vaccination were reported to be negatively associated with asthma; odds ratios were 0.36 (95% CI: 0.14, 0.91) and 0.45 (95% CI: 0.21, 0.98), respectively. These results were supported by a negative association between asthma and a positive measles-specific immunoglobulin G titer (odds ratio = 0.65, 95% CI: 0.35, 1.20). Jedrychowski et al.[23] conducted a small cohort study of approximately 1,000 Polish schoolchildren. The authors reported a negative association between measles vaccination and physician-diagnosed asthma (odds ratio = 0.50, 95% CI: 0.24, 1.00).

In the context of a nationwide cohort study, the ascertainment of asthma cases is feasible almost only through nationwide registries as opposed to actual clinical review. This invites speculation on the specificity and completeness of the methods used for outcome identification. We used information on hospitalization with asthma and use of anti-asthma medication. Instead of combining this information and using it algorithmically to identify asthma cases—for example, based on number of uses of both steroid inhalants and β2-agonist inhalants within a specified time period—we chose to analyze directly and independently hospitalizations and uses of different types of anti-asthma medication. The algorithmic approach has some drawbacks, at least in this study setting, which includes the possibility of survival bias and the identification of actual asthma cases being dependent on the definition of the algorithm. While our outcomes will typically have lower specificity than those derived from an algorithmic approach, this is offset by the more detailed results with respect to asthma phenotypes obtained in our analysis. Asthma is difficult to diagnose in early childhood, and a number of different phenotypes and conditions produce asthmatic symptoms in early childhood: transient wheezing coinciding with infections in infancy and among toddlers, more persistent nonatopic wheezing in childhood with later remission, and atopic wheezing with a more chronic course of illness. We found that the MMR effect was more pronounced for hospitalization with severe asthmatic symptoms and for anti-asthma medication used for asthmatic conditions not easily controlled by more standard treatment regimens (long-acting β2-agonist inhalants and "other anti-asthma medication"). Furthermore, there was no protective effect of MMR vaccination on the use of systemic β2-agonists. Systemic β2-agonists are mainly mixtures and are commonly used to treat transient wheezing in early childhood—for example, in the case of acute bronchitis. This minimizes the concern that the MMR effect alone relates to transient wheezing illness in early childhood instead of asthma.

In any observational study of routine vaccination, the group of unvaccinated children is a more or less selected group. We took this into account by adjusting for a wide range of possible confounding factors and by conducting a number of supplemental analyses. The uptake of MMR vaccine in our cohorts was approximately 85%, and consequently selection is less likely than it would be with a very high uptake. We specifically adjusted for receipt of infant vaccines and found that the MMR effect persisted among children vaccinated in the first year of life, further minimizing concern over selection of MMR-unvaccinated children. Maternal smoking and parental asthma/atopy are known risk factors for asthma and could, for different reasons, be associated with avoidance of vaccination. In subanalyses, we found the potential confounding effect of these factors to be negligible. We took the possibility of health-care-seeking bias and confounding by personal disease history (both all diseases and asthma alone) into account by including adjustment for hospital use in the first year of life (both all diseases and asthma alone); there was no effect on our hospitalization results. In general, health-care-seeking bias—for example, situations in which vaccinees were more likely to use anti-asthma medication—would tend to produce increased risks of asthma associated with vaccination, contrary to what we observed. While it is not possible to completely rule out remaining bias or confounding, we have examined all obvious sources and have found no indication that our findings were a result of such errors.

We identified a number of factors which modified the MMR effect. Overall, the modest differences in MMR effect and the lack of consistency across our different asthma outcomes limit the interpretability of these results. Asthma consists of a number of phenotypes, each with a highly complex genetic and environmental etiology, which we believe these results reflect.

Among children with no receipt of infant vaccines, a number of anti-asthma medication rate ratios were significantly increased. An explanation is that children with no infant vaccinations and no MMR vaccination are a highly selected group whose parents generally avoid medication. In contrast, children receiving no infant vaccines but at least 1 MMR vaccine are a group of children whose parents are less likely to be antivaccine and consequently less likely to be antimedication, hence the increased rate ratio. This is supported by the lack of a similar effect for hospitalizations, where no significant differences were found (see results for infant vaccine compliance in Table 1 ).

The strength of our study is the use of a nationwide cohort with prospective and independent ascertainment of exposure and outcomes, reducing the concern over selection and recall bias that is commonly found in other vaccine safety studies. Our results do not support the hypothesis of an adverse effect of MMR vaccination on asthma-like disease. Rather, they support the hypothesis that MMR vaccination is associated with a reduced risk of asthma-like disease in young children.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.