As the social determinants of health improved,1,2 so exposure to microbiological diversity reduced3 and asthma prevalence increased.4,5 The Hygiene Hypothesis suggests a causal relationship between these trends, through an effect of exposure of environmental factors (including invasive and non-invasive infections) on T-helper cells.4,6 Children with greater exposure to infections and a wider array of microbes early on in life, according to the hypothesis, can expect to have lower rates of asthma and better balanced immune function.7

In this issue of the PCRJ, Patel and colleagues studied 89 Somali immigrants living in Olmsted County in Minnesota in the United States.8 Firstly, they found that 10 (16% of the 62 children included in the analysis) had asthma, and 22 (35%) had other atopic conditions, similar prevalences to the general population of children in Rochester, Minnesota.9 Secondly, they found lower levels of mumps virus specific antibody levels in those Somali immigrants with asthma compared to those without asthma. The authors therefore conclude that their results may not support the hygiene hypothesis.

The hygiene hypothesis is appealing on many levels, not least to those who are keen to reduce antibiotic prescribing for self-limiting conditions. Evidence supporting the hypothesis may justify comments in primary care consultations to parents of young children with probable viral illnesses such as, “Fighting this illness without antibiotics will make your child's immune system stronger.”10–12 Society in general needs to shift its mindset from a superficial concept of bacteria as being bad and needing to be destroyed (“our cleaner kills 99.9% of all household germs”) to an ecological view of the world where we find better ways of living symbiotically with the micro-organisms we depend on for our well-being and immune competence. Indeed, our microbiome's relationship with immune status is a major developing research area,13 and some have come to think of our commensal bacteria as one of our major organs, essential for health and for life itself.

So what can be said about the applicability of Patel and colleague's findings? Study design issues limit our ability to interpret their findings both on asthma prevalence and on humoral immune response to MMR vaccine viruses. The authors describe their study design as “a retrospective cohort study with a cross sectional study component.” Their “convenience” sample, a secondary study base of 89 children, was obtained through their participation in a previous study of MMR vaccine response.14 Assessing the population prevalence of asthma in Somali immigrants would ideally have relied on systematic ascertainment from actual source population. What proportion of eligible Somali immigrants participated in the original vaccine study? And were there any important differences between those that participated in the original study and those that did not, as well as between the 62 that participated in the present study and the 27 that were excluded from the analyses? No specific hypothesis is tested in the comparison of asthma prevalence in the 62 participants of this study with a much larger, school-based, longitudinal study of physician-diagnosed asthma prevalence in Rochester, Minnesota, that is mentioned in the discussion.9 Therefore, we do not know how confident we can be that there is no difference in asthma prevalence between the two samples. The authors do consider the possibility of a Type II error because of the small numbers in their study. However, socio-economic status and other possible confounders were not taken into account when comparing asthma prevalence.

The authors have previously found that Caucasian children with asthma have suboptimal immune responses to the MMR vaccine,15 but whether this holds for children who have had increased exposure to a wide array of microbes early on in life is unclear. What can be said for the finding in this study that mumps specific IgG levels were lower in those Somali immigrants with asthma but no difference was found for measles and rubella IgG levels? The authors did adjust for the duration between the MMR vaccination and the date the blood sample was taken, but do not seem to have controlled for other factors such as age.

It would be interesting to know how other factors — such as time in the US, and intensity and duration of exposure to the “Somali environment” pre-natally and after birth — influenced risk of asthma and immune response to vaccinations. Are the early experiences of Somali immigrants in relation to microbiological and other relevant environmental exposures typical of other Somali children, or could it be that there is something about those that end up emigrating that also makes them more similar in this regard to children in urban Minnesota?

The authors are undoubtedly correct in concluding that a straightforward interpretation of the hygiene hypothesis should be made with caution. Potential confounders, as the authors point out, including genetic and epigenetic influences, are likely to apply.6 While the concept of this ‘natural experiment’ study is intriguing, study design issues make it hard to know what the real implications are for the longevity of the hygiene hypothesis…