The study is a birth cohort of 2,403 children recruited from around the time of birth in HPAB over the period November 2005 to December 2009. The recruitment area for the cohort is defined geographically by the boundaries of the District of Quinindé, Esmeraldas Province in Northern coastal Ecuador (Figure ). The primary exposures are maternal and infant infections with geohelminths and the primary outcomes are vaccine immunity, atopy and allergic disease.
Figure 1 Study site. A. Map of Ecuador showing location of District of Quinindé, Esmeraldas Province (black oval) (Courtesy of The General Libraries, The University of Texas at Austin). The recruitment area for the cohort was defined by the geographic (more ...)
Study population and area
The Province is one of the poorest regions of Ecuador, with a per capita income of less than US$2,000 in 2005. Quinindé is a rural District that covers an area of 3,471 km2 and includes 1 urban (the town of Quinindé) and 6 rural parishes with an estimated population of 150,000. The District is located in an area of (formerly) equatorial rainforest that has largely been cleared, at an altitude of approximately 100 m and average annual day temperature of 30°C and 75% humidity. The District has an ethnically mixed population of mestizos (90%), Afro-Ecuadorians (7%), and Amerindians (3%). Twenty-two percent of the population is estimated to be urban (Quinindé town) and 78% rural. In the town of Quinindé, approximately 90% of the population have access to electricity, 60% to treated drinking water, 40% to sanitation; 60% to solid waste disposal services. In contrast, in the rural areas, 10% have access to electricity and none have access to other services. The main sources of income are derived from African palm oil and fruit cultivation, cattle, and extraction of timber. The study is based at the Hospital "Padre Alberto Buffoni" (HPAB) (H in Figure ) in Quinindé town. HPAB is the only Hospital serving the District and the only health centre with maternity services.
Although assessments and sampling are conducted during pregnancy, only a proportion of pregnant mothers attend antenatal clinics, and formal recruitment into the cohort occurs around the time of birth. Entry criteria into the study are: 1) healthy normal baby less than 14 days old; 2) at least one stool sample collected from the mother; 3) the family has lived in the District for the last 2 years and does not plan to move out of the District over the following 3 years; 4) the home is accessible; and 5) the mother is 17 years or older.
Sampling and evaluations
The sampling and examination schedule for follow-up is shown in Table . Evaluations are conducted at birth, 2 weeks, 3, 7, 13, 18, 24 and 30 months, and at 3, 5, and 8 years through home visits and scheduled visits at the ECUAVIDA outpatient clinic at HPAB. More detailed immunologic evaluations are performed in an immunology sub-cohort (ISC) of the last 295 infants recruited using blood samples collected at birth, 7, 13 and 24 months, and at 3, 5 and 8 years. A surveillance sub-cohort of 195 newborns living within the town of Quinindé are being followed-up actively and sampled for viral respiratory tract infections (influenza viruses, respiratory syncytial viruses, adenoviruses, and rhinoviruses) using nasal swabs and diarrheal illnesses (rotavirus and norovirus) using stool samples during the first 2 years of life. Aliquots of all stool samples collected and hypopharyngeal swabs (at routine sampling times) are being stored for future analyses of bacterial microbiota.
Sampling schedule for ECUAVIDA cohort
All vaccines are provided free of charge by the Ministry of Public Health at the vaccination clinic at HPAB. The vaccination schedule is: birth-BCG; 2 months and 4 months-pentavalent (Quinvaxem [DPT-HepB-Hib], Novartis), trivalent oral poliovirus vaccine (Chiron), rotavirus (Rotarix, GSK); 6 months-pentavalent, OPV; 12 months-measles-mumps-rubella (Serum Institute of India [SII]); and 18 months-DT (SII) and OPV.
Measurement and definition of geohelminth exposures
Stool samples to measure geohelminth infections have been collected from mothers in the third trimester of pregnancy (stool samples collected after birth and before the child is 14 days old are considered to represent a third trimester stool sample where the mother has not received anthelmintic treatment) and from infants at 3, 7, 13, 18, 24, and 30 months, and at 3, 5, and 8 years. Stool samples are analysed using a combination of standard methods including direct saline mounts, the modified Kato-Katz method, formol-ether concentration, and carbon-coproculture methods [45
]. Geohelminth infectious exposures will be defined as follows: 1) maternal geohelminth infections-presence of any geohelminth infection detected during the 3rd
trimester of pregnancy; and 2) infant geohelminth infections-presence of any geohelminth infections detected during the first 2 years of life.
Measurement and definitions of study outcomes
Primary study outcomes
1. Vaccine immune responses:
vaccine immunity will be measured by the presence of protective antibody levels measured as follows-1) IgG to tetanus toxoid as described previously [46
]. Immune protection will be defined > 0.15 IU/ml; 2) titer of neutralizing antibodies to OPV type 3 using the microneutralization assay at the Health Protection Agency (HPA), London, UK. Protection will be defined as OPV type 3 virus neutralization at 100TCID50 at serum dilution ≥ 1:8 dilution; 3) IgA to rotavirus (HPA); protection indicated by titers ≥ 20 U/mL; 4) IgG to polyribosylribitol phosphate of Haemophilus influenzae
type B (Binding site, Birmingham, UK); protection > 0.15 mg/mL; 5) anti-HBS IgG antibodies (Architect, Abbott Diagnostics, Sligo, Ireland), protection > 10 IU/mL IgG. Plasma samples collected at 24 months and 5 years will be used to measure short-term (24 months) and long-term (5 years) antibody-mediated protective immunity. Vaccine responses to Hib and rotavirus will be measured only at 7 months to distinguish vaccine-mediated responses from those acquired from natural exposures to these infections.
2. Atopy: defined by the presence of allergen skin test reactivity at 5 years to any of the following aeroallergens: Dermatophagoides pteronyssinus/farinae mix, American cockroach (Periplaneta americana), fungi mix, dog, cat, and mixed grass pollen. A positive test is defined by the presence of a skin wheal ≥ 3 mm greater than saline control. All skin testing will be conducted by trained physicians and will be supervised by a highly experienced clinical investigator (MEC).
Eczema has been assessed by questionnaire and physical examination using standardized instruments based on the United Kingdom Working Party (UKWP) criteria/Nottingham protocol [47
]. Cases of flexural dermatitis are assessed further for severity using the SCORAD protocol [48
]. Evaluations for eczema are performed at 7, 13, 24, and 36 months. Infants presenting to the ECUAVIDA outpatient clinic with skin complaints are also assessed for eczema. All clinical evaluations are done by trained physicians using standardized protocols. Eczema will be defined by at least one presentation during the first 3 years of life of an itchy skin condition plus 3 or more of the following: i) history of involvement of flexural sites including cheeks; ii) history of atopic disease in first degree relative: iii) a history of generally dry skin in last year; and iv) visible flexural dermatitis or dermatitis affecting cheeks/forehead and outer limbs.
Asthma will be measured at 5 and 8 years using the ISAAC phase II questionnaire [49
] and defined by the presence of wheeze within the previous 12 months plus: either
a previous history of wheeze within the previous 12 months (at 13, 24, and 36 months) or
a previous ECUAVIDA clinic diagnosis of asthma. Non-atopic and atopic asthma will be defined by the presence or absence, respectively, of allergen skin test reactivity. Measurement of pulmonary function and reversibility with a short-acting β2
-agonist will be done at 8 years of age. Wheezing illness will be defined as wheeze in the previous 12 months.
Secondary study outcomes-immunological outcomes
Five possible mechanisms have been defined that could mediate the putative effects of geohelminth infections on the primary study outcomes. A justification for choosing these mechanisms is provided in the Discussion. The mechanisms are: 1) Th2 polarization-measured by the ratio of IL-5 to IFN-γ protein produced by peripheral blood leukocytes (PBLs); 2) immune homeostasis-production of IL-10 spontaneously in 5-day PBL cultures [50
]; 3) immune suppression-antigen-specific suppression measured by frequencies of IL-10+
T cells or IL-10 production by antigen-stimulated PBLs [51
] or 'bystander suppression' measured by IL-10 production by A. lumbricoides
antigen-stimulated PBLs [54
]; 4) immune maturation-capacity of PBLs to produce IFN-γ to sub-optimal stimulus with Staphylococcal enterotoxin B (SEB); and 5) pro-inflammatory responses-IL-8 production by PBLs stimulated with LPS [56
], and quantities of IL-17, IFN-γ, and IL-5 produced by SEB-stimulated PBLs. Flow cytometry experiments will be performed using cryopreserved and fixed peripheral blood leukocytes using standardized protocols [41
For each of the outcomes, the following mechanisms will be evaluated
1. Vaccine immunity: Immunological outcomes will be measured at 24 months (short-term immunity) and 5 years (long-term immunity) for Th2 polarization (tetanus toxoid [TT] and tuberculin [PPD]-stimulated PBLs), immune homeostasis, and immune suppression (IL-10 production by PBLs stimulated with TT, PPD, and A. lumbricoides antigen).
2. Atopy: Immunological outcomes will be measured at 5 years for immune homeostasis and immune suppression (frequencies of IL-10+CD4+ T cells [ISC only] or IL-10 production by PBLs stimulated with D. pteronyssinus, P. americana, and A. lumbricoides antigen).
3. Eczema: Immunological outcomes will be measured up to 3 years for markers of immune maturation, immune homeostasis, immune suppression (frequencies of IL-10+CD4+ T cells (ISC only) or IL-10 production by A. lumbricoides-stimulated PBLs), and pro-inflammatory responses (ISC only).
4. Asthma: Immunological outcomes will be measured at 5 years for markers of immune homeostasis, immune regulation/suppression (frequencies of IL-10+CD4+ T cells [ISC only] and IL-10 production by PBLs stimulated with D. pteronyssinus and A. lumbricoides antigen), Th2 polarization (ratio of IL-5 to IFN-γ to sub-optimal SEB stimulus), and pro-inflammatory responses (ISC only).
We expect losses to follow-up to be up to 10% in the first year (i.e. 2,163/2,403 followed up) and 5% annually to age 3 and a further 5% between ages 3 and 5. Estimates for the number of children followed-up are: 2 years-2,055, 3 years-1,952, and 5 years-1,854. Losses are due to the mobility of the study population.
Potential sources of biases are selection bias caused by losses to follow up and information bias caused by systematic misclassification of outcomes. The following methods are being used to minimize these biases: Achieving high rates of follow-up - a) we have a well-trained study team employed full-time for the cohort; b) we conduct regular home visits to maintain contact with study mothers; c) the universal use of mobile phones allows us to maintain regular contact with study mothers; d) the mother of each child is given a cohort identification card with a phone number to make appointments at the ECUAVIDA outpatient Clinic; e) changes of address are followed up actively with home visits to repeat environmental sampling. Information bias - a) repeated measurements of outcomes (for asthma, eczema, and allergen skin test reactivity) should reduce misclassification and recall bias; b) clear definitions for primary outcomes and exposures; c) evaluation of outcomes is being performed blind to exposure status. Detailed information on potential confounding factors is being collected.
The study sample size is fixed at 2,403 newborns. Primary outcome variables will be treated as statistically independent. Power calculations for primary study questions are shown in Table [58
]. We expect a high degree of power for the estimation of most primary and secondary study outcomes (≥ 80% depending on immunological mechanism at α = 0.01) that will allow for the loss of power expected by controlling for confounding.
Power for analysis of primary exposure-outcome associations
Statistical analysis will be guided by the conceptual frameworks presented for each of the study outcomes (Figures , , and ). The analysis of the primary study outcomes will use multiple logistic regression to estimate the Odds Ratios for exposure-outcome associations with adjustment for appropriate confounding factors. For example, a priori
confounders for the association between maternal geohelminth infections and eczema by 3 years of age are gender, maternal educational level, household crowding, and infant geohelminth infections. The primary analyses will evaluate the effects of maternal geohelminth infections or infant geohelminth infections on: 1) protective immunity to common childhood vaccines: 2) the presence of allergen skin test reactivity at 5 years of age; 3) the development of eczema by 3 years of age; and 4) the development of asthma at 5 and 8 years of age. Immunological variables will be defined as binary variables with a cut-off defined by the limit of detection of the assays (i.e. IL-10) or median level (i.e. IL-8) or as loge
-transformed continuous variables. Because immunological variables are considered as intermediate in the causal pathways between exposures and outcomes, these variables will be evaluated in hierarchical analyses [60
Figure 2 Conceptual model for effects of maternal and infant geohelminth infections on vaccine immune responses measured at 7 months, 2 years or 5 years. Maternal infections would be expected to mediate effects on vaccine responses at 7 months and 2 years, although (more ...)
Figure 3 Conceptual model of effects of maternal and infant geohelminth infections on allergen skin test reactivity. The effects of both exposures are presumed to occur via the development of a modified Th2 (mTh2) response during chronic geohelminth infection. (more ...)
Figure 4 Conceptual model for the effects of maternal and infant geohelminth infections on asthma in children. The effects of these exposures on asthma are considered to occur via effects on bronchial hyperreactivity. The protective effects of chronic helminth (more ...)
Figure 5 Conceptual model for the effects of maternal geohelminth infections on eczema. Effects of maternal geohelminth exposures on eczema are considered to occur through the capacity of the immune system to regulate inflammation of the skin caused by external (more ...)
Information, ethical approval, consent, and ethical considerations
The Ministry of Health in Ecuador was informed of the study and gave its support for the study in Esmeraldas Province, Ecuador (Subsecretaria General de Salud, reference SSG-10-000285). Formal recruitment into study occurred at the first home visit done during before the child was 14 days old. Informed consent for the collection of samples during the study before this time was done at 2 different time points: 1) antenatal clinic visit for collection of a stool and blood samples: 2) in the maternity department of HPAB for collection of maternal stool and blood samples, cord blood and meconium. At each of the three times at which written consent was obtained, the mother received both written and verbal information in Spanish on the aims of the study and the reasons for collection of study samples. Results are provided to the mother for all clinical samples collected. Where necessary, the child will be evaluated by a study physician, and appropriate medication and nutritional supplements provided as required through the ECUAVIDA study clinic at HPAB. Appropriate antiparasite treatment is given for all stool samples with documented geohelminth infections. Benzimidazole drugs are not given to pregnant women or children before 2 years of age following the recommendations of the Ecuadorian Ministry of Health. Infected mothers are offered albendazole when breast-feeding has finished. Infected children aged below 2 years are treated with pyrantel and after 2 years with albendazole. Infections with Giardia intestinalis and Entamoeba histolytica are treated with a standard course of metronidazole if characteristic trophozoites are observed. The study protocol was approved by the Ethics Committees of the Hospital Pedro Vicente Maldonado and the Universidad San Francisco de Quito, and is registered as an observational study (ISRCTN 41239086).
Although anthelmintic treatment is a highly popular intervention, clear health benefits from the treatment of pregnant women have not been unequivocally demonstrated, and such benefits could be argued to be at best marginal in most endemic regions where hookworm anaemia is relatively infrequent and geohelminth infections are low intensity [61
]. For this reason, we opted for an observational rather than an interventional design that would allow us to understand the natural history of the interaction between early geohelminth exposures and the study outcomes, but provide the opportunity to identify potential interventions that could be beneficial and be evaluated in future intervention studies.