Study population and sampling
Children in this research project have been part of a study conducted earlier to evaluate the impact of the sanitation programme on the occurrence of childhood diarrhoea. This first study was originally designed to enrol three separate cohorts of children aged 0–3 years recruited from 24 small geographical areas selected to represent the population without sanitation in Salvador, called sentinel-areas [
16,
17] These three cohorts were recruited in 1997, 2001 and 2003, respectively. Within each selected area, a survey was conducted at baseline and the cohort was then followed up. Sampling and data collection procedures and instruments were the same in the 3 surveys and follow up making the study populations comparable. In each survey, children were randomly selected among those living in each area. Standardised questionnaires were applied during the baseline surveys to obtain data on demographic characteristics and living conditions of the families, sanitation, water supply and other characteristics of the households, pre-natal care, weight-at-birth, breastfeeding, attendance to kindergarten and vaccination status of children, nutritional status (anthropometry), characteristics of the neighbourhood and indoor environment. Children were then followed-up regularly for a period of one year from recruitment, forming thus the 3 cohort populations. During follow-up, children were visited every 2 weeks and data on occurrence of diarrhoea, cough, shortness of breath and fever as reported by the guardian was collected. Weight and height were measured using standard techniques and z-scores for height-for-age, weight-for-age and weight-for-height were calculated. A stool sample was collected and examined for parasitic infections. Details of the methodology and results of the first survey and cohort study have been published elsewhere [
18‐
20].
To take advantage of these three cohorts (since detailed information for them had been collected on early life factors proposed as potential risk factors for asthma and other allergic diseases), the children who were aged 4–11 years and had complete follow-up information (e.g. stool examination conducted at 3 different time points) were selected for the present study.
Data collection and instruments
The field work of the surveys in 2005 and 2008 involves 7 different activities, each conducted by different field teams and supervised by at least one of the co-investigators. Application of the main questionnaire, anthropometric survey, and stool and dust sample collection were completed in 2005. The activities were:
1. Questionnaire: mostly based on the ISAAC Phase II questionnaire [
7], translated into Brazilian Portuguese. Some additional questions were included on environment, housing conditions, allergic diseases and potential risk factors, as described below. Questionnaire application was done by trained field workers, after piloting the questionnaire in children outside the study population but in the same target age. Medical assistance by the project in a referral outpatient care unit and by project doctors will be offered for all study children identified as having severe asthma.
2. An anthropometric survey with two independent measures of height and weight collected in a standardised way conducted by trained nutritionists, taking the mean value as the final measure, as recommended by WHO: z-scores for weight-for-age, weight-for-height and height-for-age were calculated using the EPINUT program (Epi Info 6.0; CDC, Atlanta, GA, USA) (details published elsewhere [
20]).
3. Stool samples collected for helminths and parasites with two different samples for each child, 2 days apart. Stools were analysed using the gravitational sedimentation technique of Hoffman, Pons & Janner to detect helminth eggs, protozoan cysts and oocysts. Two slides were examined for each stool sample. Quantification of helminth eggs was performed using the Kato-Katz technique [
22]. All children with positive results were treated with appropriate antiparasitic drugs (Albendazole and Tinidazole).
4. Dust samples collected using a residential vacuum cleaner (Eletrolux Professional, 1220 watts) containing a nylon 25 um micromesh sock filter [
23]. The childrens' beds were aspirated for two minutes over a 1 m
2 area, adjacent to the head side. The filters were weighed before and after collection of the dust samples. Fibres and large particles were removed from the dust with forceps and the fine particulated dust samples were weighed, aliquotted as 100 mg samples and cryo-preserved at -20°C. Temperature and air humidity were recorded at the bedroom with a thermohygrometer (Minipa Ind & Co, São Paulo, Brazil).
5. Blood samples: An aliquot of 10 mL collected to measure (i) total and allergen specific IgE (anti-mite, and anti-cockroach), (ii) IgG antibodies to Hepatitis A virus, Herpes simplex, Herpes zoster, and Epstein-Barr viruses, A. lumbricoides, T. trichura, Toxoplasma gondii and Toxocara canis, and (iii) cytokines IFN-gamma, IL-13 and IL-10 and TGF-β in supernatant fluids collected from antigen-stimulated whole blood cultures.
6. Skin prick test (SPT) carried out by allergen skin prick testing of the right forearm of each children using extracts (ALK-Abello, São Paulo, Brazil) of Dermatophagoides pteronyssinus, Blomia tropicalis, Blattella germanica, Periplaneta americana, fungi, dog and cat epithelia. Saline and histamine was used as negative and positive controls, respectively. Wheal sizes will be read after 15 minutes and reactions will be considered positive if the mean of the two larger perpendicular diameters of the reaction is at least three millimeters superior to the mean of the two larger perpendicular diameters of the negative control reaction area.
7. Skin inspection for flexural dermatitis performed by trained observers using the ISAAC phase II protocol [
24] will complement data on atopic eczema collected by questionnaire.
Laboratory techniques
Aeroallergens will be quantified from dust samples by capture ELISA using commercially available kits (Indoor Biotechnologies, Virginia, USA) for the following aeroallergens: Blomia tropicalis Blo t 5, Dermatophagoides pteronyssinus Der p 1, Blatella germanica Bla g 2, Cat Fel d1, Dog Can f1Bacterial endotoxin and fungal b-glucan will be measured using the Limulus Amebocyte Lysate (LAL) assay (Cambrex Bio-Ciência do Brasil Ltda) following the manufacturers' instructions.
Heparinised whole blood will be cultivated at a dilution at 1:4 in RPMI (Gibco, Auckland, NZ) containing 10 mM glutamine (Sigma, St. Louis, MO, USA), 100 μg/ml of gentamicyn (Sigma, St. Louis, MO, USA), and stimulated with the following antigens: Ascaris lumbricoides (10 μg/ml); Blomia tropicalis (40 μg/ml); Dermatophagoides pteronyssinus (5 μg/ml). Pookweed mitogen (Gibco, Auckland, NZ) diluted at 1/1000 will be used as a positive control and media alone as negative control. House dust mites cultivated in fish food will be purified, lysed in pH 7.4 phosphate saline (PBS) with the use of a blender (Blender 51BL30; Waring Commercial, Torrington, Connecticut, U.S.A.). The A. lumbricoides antigen will be obtained by trituration of liquid nitrogen frozen adult worms obtained from a child treated with albendazole, in a blender as stated above in the presence of PBS. After centrifugation, the supernatants will be cryopreserved for storage before use. All antigens will be depleted of endotoxin by treatment with triton-114 (Sigma, St. Louis, MO, USA) and the protein contents will be determined by the Lowry method. Whole blood cultures will be cultivated at 5% CO2, for 24 hours for detection of IL-10 and for 5 days for the detection of IL-13, TGF-β and IFN-gamma. Cytokines in supernatant fluids will be measured using Pharmigen BD antibody pairs and recombinant standards (Pharmigen, San Diego, Ca, USA) by capture ELISA following the manufacturer's instructions. Cytokines will be estimated by interpolation of standard curves of recombinant standards.
Total IgE will be measured as follows: COSTAR high binding microassay plates (COSTAR, Cambridge, Me, U.S.A.) will be coated with 4 μg/ml of an anti- human -IgE antibody (PHARMIGEM, San Diego, CA, USA) overnight at 4C. Plates will be blocked with 0.15 M phosphate-buffered saline, pH 7.2 (PBS), containing 20% of dried skimmed milk (DSM) and 0.05% of Tween 20 (Sigma, St Louis, MO, USA) overnight, at 4°C. Sera to be tested and IgE antibody standards (RESEARCH DIAGNOSTICS INC, Flanders, NJ, USA). will be diluted 1:10 in PBS containing 10% of DSM and 0.05% of Tween 20 and incubated overnight at 4°C. A goat anti-human IgE-peroxidase conjugate (SIGMA, St Louis, MO, USA), and an anti-goat immunoglobulin-peroxidase conjugate (DAKO A/S, Glostrup, Denmark), will be diluted at 1:2000 and 1:10.000, respectively, and incubated for 60 minutes at room temperature. The results will be expressed in International Units (IU). A pool of parasite infected patients' sera will be used as positive control. Umbilical cord serum from a non-atopic and non-parasited mother will be used as negative control.
Determination of specific IgE serum concentration will be done for Dermatophagoides pteronyssinus, Blomia tropicalis, Blatella germanica, dog and cat epithelia main allergens using the RAST system (Pharmacia Diagnostics AB, Upsala Sweden).
Detection of anti-parasite antibodies. Anti-
A. lumbricoides IgG4 will be detected by an indirect ELISA using wells of high binding microassay plates (COSTAR, Cambridge, Me, U.S.A.), sensitized with 20 μg/ml, of the parasite antigen as stated above, diluted in carbonate-bicarbonate pH 9,6 buffer. The sera will be diluted at 1:50 in PBS containing 10% skimmed milk and 0,1 % tween 20 (PBS/SM/T). The reaction will be detected using a biotinylated anti-human-IgG4/(SIGMA Chemical Co., San Louis, MO, USA), streptoavidin/peroxidase (PHARMIGEN., San Jose, CA, EUA) and H
2O
2 e OPD (MERCK & Co., Inc., White house Station, NJ, USA). The assay cut-off will be the mean plus 3 SD of negative controls (sera from people with 3 negative stool samples collected serially). Anti-
Toxocara IgG antibodies will be detected using 20 μg/ml of larvae excretorial-secretorial antigen obtained by cultivation of the larva as described by De Savigny et al. [
30] (The sera will be diluted at 1:100 in PBS/SM/T and pre-absorbed with
A. lumbricoides; the development of the reaction will be done as described for
A. lumbricoides IgG4 except for the conjugate that will be a biotinylated anti-human IgG. The assay cut of will be the mean plus 3 SD of negative controls (sera from upper middle class people, without history of dog and cat contact). Anti-
Toxoplama gondii IgG and IgM will be detected using commercially available ELISA kits (Diamedix Corporation, Miami, USA), following the manufacturer's instructions. IgG antibody against the following virus will measured: Herpes simplex, Herpes zoster, Epstein-Bahr and Hepatitis A virus. Detection will be done using commercially available ELISA kits (Diamedix Corporation, Miami, USA), following the manufacturers' instructions.
Overview of plan of analysis
Statistical analysis will be conducted according to a conceptual framework defining a proposed causal pathway. Bivariate association analysis will be carried out by calculating prevalence ratios and 95% confidence intervals to measure the strength of the association between each potential risk factor and the outcomes of interest. At his stage, the main outcomes of interest are atopy (expressed as skin test positivity and level of IgE to allergen), asthma and allergy). Multivariate analysis will be taken out in several steps. First, within each level, multivariate data reduction techniques (e.g. principal component analysis) will be applied to summarise highly correlated variables to index variables (e.g. socio economic status, Th1- related immune response). Secondly, for each level, multivariate regression models will be fitted to estimate association parameters adjusted for confounding factors of the same level. In addition, multi-level modelling and robust variance estimation techniques will be used to adjust for intra-subject correlation due to repeated measures on the same individual and/or geographical areas. Finally, to address the complex inter-relationships between risk factors of different levels and outcomes according to our pre-defined conceptual framework, we will apply the following analysis strategies: i) A hierarchical effect decomposition strategy fitting a sequence of regression models each including different blocks of covariates, similar to an approach suggested by other authors [
25]; ii) Log-linear models, an association analysis technique for multi dimensional contingency tables to estimate conditional association parameters (relative risks), conditional on the distribution of the risk factors on the higher levels [
26,
27]. For example, the probability of developing asthma will be estimated, conditional on a positive skin prick test, a high IgE level and environmental exposures. In addition, interaction parameters will be introduced to model how the risk factors on different levels influence each other (e.g. modelling the impact of environmental exposures on the effect of high IgE level); iii) Finally, sophisticated statistical techniques such as structural equation modelling and path analysis [
28,
29] will be used to model the multiple relationships defined in the conceptual framework simultaneously aimed to quantify direct and mediated effect components of the risk factors [
30].