Cohort profile: The Prevention and Incidence of Asthma and Mite Allergy (PIAMA) birth cohort

The Prevention and Incidence of Asthma and Mite Allergy (PIAMA) birth cohort was set up in The Netherlands with two main aims: to investigate the effect of mite-allergen avoidance on the incidence of childhood asthma (Intervention study) and to assess lifestyle and environmental risk factors for childhood asthma (Natural History study). The baseline study population consisted of 3963 children born in 1996–1997. Questionnaires were completed by parents during pregnancy, when the child was 3 months old, then annually from 1 up to 8 years and by parents as well as children at 11 and 14 years. Data were obtained on child and family characteristics, a wide range of environmental and lifestyle exposures and on asthma and other allergic and respiratory outcomes. Home visits and clinical examinations were conducted in sub-groups at the ages of 1, 4, 8 and 12 years to obtain objective measures. These included anthropometric measures, measurements of lung function, bronchial hyper-responsiveness and blood pressure, house dust and blood samples, which were used to determine total and specific IgEs, cholesterol and HbA1c and to extract DNA. More information on the data available in the PIAMA study can be found at http://www.birthcohorts.net/ and http://www.enrieco.dk/Default.asp and researchers interested in collaboration are invited to contact the principal investigators via alet.wijga@rivm.nl. In the 19880s and early 1990s, the prevalence of asthma in The Netherlands substantially increased, especially in children. This rising trend and the lack of understanding of its causes were reasons for major concern among public health policymakers. In 1992, the Public Health Council of The Netherlands issued a report entitled Indoor Allergens, Allergy and the Development of Chronic Respiratory Diseases. The report addressed the question whether reduction of exposure to indoor allergens early in life would prevent or delay the development of inhalant allergies and asthma in children. The Council recommended that an intervention trial should be set up to evaluate the effectiveness of measures to reduce early postnatal exposure to indoor allergens. In conjunction with the proposed Intervention study, a Natural History study was set up to investigate the many questions surrounding the development of inhalant allergy and asthma in the general population. The intervention part and the natural history part were integrated into one study: the PIAMA birth cohort study with two main aims: (i) to investigate the effect of mite-allergen avoidance on the incidence of asthma and allergy in childhood (Intervention study); and (ii) to investigate lifestyle and environmental (indoor and outdoor) risk factors for childhood asthma and allergy (Natural History study). The intervention consisted of application of mattress and pillow covers oto the bed of the infant and the parents’ bed. Participants of the intervention study were randomly allocated to receive either ‘active’ (mite-impermeable) covers or ‘placebo’ (standard cotton) covers. No intervention was applied in the Natural History study. At the start of the study, the length of follow-up was planned to be 8 years to allow sufficient time for allergic sensitization and symptoms to develop. The original design of the PIAMA study has been described in detail in 2002.1 In the course of the study genetic and genomic studies were incorporated also. Already in the early years of the study it became apparent that not only asthma, but also overweight was increasing in children in The Netherlands and childhood overweight rapidly developed into a major public health problem. In the PIAMA study, weight and height data had been collected annually since birth together with a wide range of environmental and lifestyle factors. Thus, a substantial number of overweight-related study questions could be addressed in addition to the questions that the study was originally designed to answer. The study was highly successful both in terms of follow-up rates and in terms of the wealth of data obtained. These success factors in combination with new research questions that came up during the course of the study were reason for the investigators to extend the follow-up beyond the age of 8 years and also to broaden the scope of the health outcomes investigated. Three years after the originally planned 8 years of follow-up had been completed, the remaining participants were invited to participate in 2 more waves of data collection at the age of 11–12 years and at the age of 14–15 years. When the study was thus extended to follow the children up into early adolescence, early markers of cardiometabolic risk, such as blood pressure, serum cholesterol concentrations and HbA1c, were included in the study as health outcomes besides asthma and allergy. The planned sample size of the intervention arm of the cohort was 800 ‘high risk’ children (children of allergic mothers, defined as mothers with at least one of the following: asthma ever, pet allergy, house dust mite allergy or nasal allergy such as hayfever). Of these 800 high risk children, 400 were to be allocated to the intervention group and 400 to the placebo group. The Natural History arm of the cohort was planned to consist of 400 high risk children, who received neither an active nor a placebo intervention, and of 2500 low risk children (children of non-allergic mothers). A validated screening questionnaire was used to identify allergic and non-allergic mothers.2 Recruitment of participants from the general population took place in 1996–1997 through 52 antenatal clinics in three different regions of The Netherlands. A total number of 10 819 pregnant women completed screening questionnaires of which 10 232 were usable; 29% of these women were classified as ‘allergic’. For the Intervention study, 2015 allergic women were invited to participate in the study and 855 (42%) agreed. For the Natural History study, 764 allergic and 5083 non-allergic women were invited and 472 (62%) and 2819 (55%) agreed to participate, respectively. All participants gave written informed consent. When recruitment was completed, the total study population of the PIAMA cohort thus consisted of 4146 pregnant women (31% allergic); the total number of children at baseline was 3963. The difference between the number of pregnant women and the number of children in the study (n = 183; 4.4%) was partly due to miscarriages and stillbirths and partly due to mothers being lost to follow-up without ever having completed a postnatal questionnaire. Characteristics of the study population are shown in Table 1. Characteristics of the study population aNorth: provinces Groningen, Friesland, Drenthe; Central: Utrecht, Gelderland; West: Rotterdam and surrounding municipalities. bParental allergy defined as ‘asthma ever and/or current house dust mite allergy and/or pet allergy and/or hay fever’. cParental education level: Low,primary school, lower vocational training or lower secondary education; intermediate, intermediate vocational education or intermediate/higher secondary education; high, higher vocational education or university degree. dMother smoked during at least the first 4 weeks of pregnancy. Characteristics of the study population aNorth: provinces Groningen, Friesland, Drenthe; Central: Utrecht, Gelderland; West: Rotterdam and surrounding municipalities. bParental allergy defined as ‘asthma ever and/or current house dust mite allergy and/or pet allergy and/or hay fever’. cParental education level: Low,primary school, lower vocational training or lower secondary education; intermediate, intermediate vocational education or intermediate/higher secondary education; high, higher vocational education or university degree. dMother smoked during at least the first 4 weeks of pregnancy. Participants were recruited from the general population. We compared the baseline characteristics of the study population with available information on the general population of The Netherlands. According to data from Statistics Netherlands, 22% of adults in The Netherlands were highly educated in 20013 and 15% of children born in 1997 in The Netherlands had at least one non-Western parent.4 In the PIAMA population, more parents are highly educated and fewer parents are from non-Western countries (see Table 1), reflecting the common phenomenon that in research involving lengthy questionnaires, more highly educated native speakers are likely to participate than the lower educated and non-native speakers. The PIAMA population was similar to the general population in the mid-nineties with respect to maternal age at childbirth5 and with respect to breastfeeding,6 but the prevalence of low birthweight (<2500 g) was lower in PIAMA than in the general population (7%).7 No comparison could be made for smoking prevalence as the PIAMA questionnaires specifically asked about parental smoking in the home. Maternal allergy was used to define children at high risk for asthma and allergy and to allocate children to different study arms (active intervention, placebo intervention and natural history). Questionnaires for parental completion were administered during pregnancy, at the child’s ages of 3 months and 1 year and then annually up to the child’s age of 8 years. At the child’s age of 11 years parents were again asked to complete a questionnaire and for the first time the children themselves completed a questionnaire also. Collection of questionnaire data from parents and children at the age of 14 years is currently (2011–2012) ongoing. Home visits and clinical examinations were conducted at the child’s ages of 3 months, 1 year, 4 years, 8 years and 12 years to collect house dust samples, blood samples and to take anthropometric and lung function measurements. These data collections were conducted in sub-groups of the total population. Clinical examination at the age of 16 years, including genomic characterization, is currently being prepared. The baseline study population consisted of 3963 children, and 92% of them were still in the study 8 years later. When it was decided to extend the study with a new wave of data collection at the age of 11 years, families who had not responded to the last three questionnaires ages 6, 7 and 8 years) or who had not responded to the last two questionnaires and had not participated in the clinical examination at age 8 years (n = 69) were excluded from further follow-up. The remaining 3541 families were invited to participate in the further follow-up of the PIAMA study. Table 2 shows how many participants were still in the study and how many questionnaires were returned at each wave of data collection. Participants still in the study and questionnaires completed per wave of data collection aParticipants were considered to be ‘in the study’ as long as they were not excluded (lost to follow-up) from the study population. Exclusion took place if participants themselves indicated that they were no longer able or willing to participate; when they moved and could not be traced; and when they failed to complete 2 or more subsequent questionnaires. b4146 pregnant women gave informed consent to participate; the number of children at baseline was 3963. Participants still in the study and questionnaires completed per wave of data collection aParticipants were considered to be ‘in the study’ as long as they were not excluded (lost to follow-up) from the study population. Exclusion took place if participants themselves indicated that they were no longer able or willing to participate; when they moved and could not be traced; and when they failed to complete 2 or more subsequent questionnaires. b4146 pregnant women gave informed consent to participate; the number of children at baseline was 3963. For the objective measurements 765, 1808, 3518 and 3202 children were invited at the age of 1 year, 4 years, 8 years and 12 years, respectively. Between 97% (at age 1 year) and 47% (at age 12 years) of those invited were willing to participate. However, not all children who participated in the clinical examinations were able and willing to give a blood sample and/or perform the lung function measurements, so that the number of successful measurements of some markers was lower than the total number of participants. Loss-to-follow-up and non-response to individual questionnaires were very low during the first 8 years of the study, but increased when parents and children were invited for a new wave of data collection 3 years after they had completed the study period that they had originally agreed to. Among the families who did not complete the questionnaires administered at the child’s age of 11 years, low parental education, parental smoking, maternal allergy and breastfeeding for less than 12 weeks were more prevalent than among the families who did participate again when the child was 11 years old (see Table 3). Characteristics of children/families that completed the child or the parent questionnaire at the child’s age of 11 years, and of those who did not complete these questionnaires aNorth: provinces Groningen, Friesland, Drenthe; Central: Utrecht, Gelderland; West: Rotterdam and surrounding municipalities. bParental allergy defined as ‘asthma ever and/or current house dust mite allergy and/or pet allergy and/or hay fever’. cParental education level: Low, primary school, lower vocational training or lower secondary education; intermediate, intermediate vocational education or intermediate/higher secondary education; high, higher vocational education or university degree. dMother smoked during at least the first 4 weeks of pregnancy. eDifferences between the 2 groups were statistically significant for all characteristics shown except for the presence of older siblings, paternal allergy and prematurity. Characteristics of children/families that completed the child or the parent questionnaire at the child’s age of 11 years, and of those who did not complete these questionnaires aNorth: provinces Groningen, Friesland, Drenthe; Central: Utrecht, Gelderland; West: Rotterdam and surrounding municipalities. bParental allergy defined as ‘asthma ever and/or current house dust mite allergy and/or pet allergy and/or hay fever’. cParental education level: Low, primary school, lower vocational training or lower secondary education; intermediate, intermediate vocational education or intermediate/higher secondary education; high, higher vocational education or university degree. dMother smoked during at least the first 4 weeks of pregnancy. eDifferences between the 2 groups were statistically significant for all characteristics shown except for the presence of older siblings, paternal allergy and prematurity. Missing of data due to loss-to-follow-up or non-response to individual questionnaires or questions was not completely at random (‘MCAR’), but strongly associated with participants’ characteristics collected in the early years of the study. This enabled us to use multiple imputation in order to avoid bias that might result from complete case analysis and to make efficient use of the available data. Table 4 shows the main topics that have been addressed in the questionnaires and Table 5 shows the most important materials and measurements that were obtained during home visits and clinical examinations. Details of the data available from the PIAMA cohort can also be found at http://www.birthcohorts.net/ and http://www.enrieco.dk/Default.asp. The ENRIECO website also includes the numbers available for many of the measures. Topics covered by the PIAMA questionnaires Topics covered by the PIAMA questionnaires Sub-groups invited for clinical examination and materials and measurements collected at different agesa aIS, Intervention Study participants; EPS, extracellular polysaccharides; Rint, interrupter resistance measurement; BHR, bronchial hyper-responsiveness; NO, Nitric Oxide. b‘Intensive follow up’ covers all high risk children (n = 1245) and a random selection of low risk children (n = 663), total n = 1935. ‘All’ covers all children except those that have moved outside the range of the study centres (incl. abroad) and children that lived in an institution. Sub-groups invited for clinical examination and materials and measurements collected at different agesa aIS, Intervention Study participants; EPS, extracellular polysaccharides; Rint, interrupter resistance measurement; BHR, bronchial hyper-responsiveness; NO, Nitric Oxide. b‘Intensive follow up’ covers all high risk children (n = 1245) and a random selection of low risk children (n = 663), total n = 1935. ‘All’ covers all children except those that have moved outside the range of the study centres (incl. abroad) and children that lived in an institution. Apart from the serum markers of primary interest, shown in Table 5, a number of other markers have been measured in serum that remained. These include vitamin D, CRP, erythrocyte fatty acids, visfatin, leptin, IL1RL1-a and helicobacter antibodies. The main focus of the study has been on asthma and allergy. With respect to the first of the original research questions, the PIAMA study found that early intervention with mite-impermeable mattress covers did not result in a substantial reduction in mite allergen exposure. It only temporarily reduced the risk of asthma symptoms and did not reduce the risk of hayfever, eczema aor allergic sensitization to house dust mite up to the age of 8 years.8 These results indicate that there is no reason to advise parents of high risk children to use mite-impermeable mattress covers on the child’s bed as the only intervention, in order to prevent the development of asthma and allergy. With respect to the second research question, we observed that several lifestyle and environmental factors early in life played a role in the development of asthma. For example, we reported an almost 2-fold increased risk of asthma at the age of 8 years in children who were born after caesarean section as compared with vaginal delivery, in particular in children with allergic parents.9 We observed a protective effect of breastfeeding for more than 16 weeks on the risk to develop asthma during childhood up to the age of 8 years for children of allergic as well as children of non-allergic mothers.10 Overweight at 6–7 years was shown to be a risk factor for asthma and bronchial hyper-responsiveness (BHR) at the age of 8 years. Children who had a high body mass index (BMI) early in life, but a normal BMI at 6–7 years, however, did not have an increased risk of asthma or BHR at 8 years.11 Also, we reported that children of mothers with overweight before pregnancy had a higher risk to develop asthma, irrespective of the BMI of the child.12 Maternal smoking during pregnancy is known to increase the child’s risk of respiratory symptoms. One of the genetic studies conducted in the PIAMA cohort suggested that interaction of variants of ‘asthma and COPD gene’ ADAM33 with in utero cigarette smoke exposure (but not with postnatal exposure) resulted in reduced lung function and the development of BHR.13 Early daycare attendance was found to be associated with an increase in airway symptoms until the age of 4 years followed by fewer symptoms between the ages of 4 and 8 years, but we observed no association with asthma symptoms, hyper-responsiveness or allergic sensitization at the age of 8 years.14 Also the presence of pets in the household was not associated with asthma development. Children with a cat or dog at home at 3 months of age had a significantly lower prevalence of sensitization to inhalant allergens at age 8, but not of asthma.15 The PIAMA cohort has participated in the European FP6 and FP7 programmes on genetics of asthma and allergy. Here the first genome-wide association (GWA) study on asthma showed that several genes were related to asthma and some genes in the 17q regions specifically to early onset asthma.16 We extended the observations and demonstrated that IL1RL1 polymorphisms that are associated with asthma in GWA studies were associated with serum IL1RL1-a levels, the transcript of this gene, with blood numbers and with asthma in The PIAMA study has several new into the role of in the development of asthma. concentrations at the birth by were shown to be associated with incidence and prevalence of asthma during the first 8 years of It is known that exposure to can asthma, but results that exposure also to the of asthma in children. of and genes the to of on childhood are applied to make use of the design of the study. analysis was used to identify on the of from to 8 years. the observed and an onset was with onset of after 2 years of in with similar observations in the was used to develop a clinical risk to the risk of asthma at years of age for children who with symptoms for the first In this population of children with asthma symptoms at the prevalence of asthma at years was clinical asthma at 7 to 8 years of delivery, parental education, parental from respiratory and In of the the between and children. A clinical risk was developed children with a of less than 10 had a risk, children with a of or more had a risk of asthma at In we demonstrated and the of objective measures of lung function and airway for the development of A of all PIAMA including those on health outcomes other than asthma and allergy, is available at of the study are the high follow-up the number of measurements up to the age of years with questionnaires in the first 8 years, the range of different exposures and health outcomes data from questionnaires, objective measurements, genetic information and exposure data to the participants’ The main results from its main to the size of the study population which for clinical examinations in all children at all is the of data on the period and of measurements, such as lung function, after At the start of the study, the of Health and had only to in to the development of asthma and allergy, and recruitment was not common we have collected information on exposures during pregnancy, such as parental smoking and the presence of pets in the with we would have more on exposure information on maternal weight during pregnancy, pregnancy The PIAMA study group was or is in a number of European with other birth including and interested in collaboration or further information are invited to contact the principal investigators via alet.wijga@rivm.nl. on the data available in the PIAMA study can also be found at and http://www.enrieco.dk/Default.asp The PIAMA study has been with from The Netherlands for Health and The Netherlands for The Netherlands Asthma the European the for Public Health and the The Netherlands of and the The Netherlands of and and We all the children and parents for in the study. we the to the PIAMA study of the and of of mite-impermeable mattress covers by high risk was associated with a risk of asthma symptoms at age 2 years, but not at other ages up to the age of 8 years, and did not reduce the risk of hayfever, eczema and allergic sensitization up to the age of 8 years. to can asthma, but the PIAMA results that exposure also to the development of asthma in children. Maternal by caesarean of breastfeeding and childhood overweight increased the risk of asthma, early daycare attendance and presence of pets in the household were not associated with asthma development. The presence of asthma at years of age for children who with symptoms for the first time could be available clinical