Setting
The study was conducted in the provinces of San Marcos and Cajabamba in the Cajamarca region, northern Peru. Both provinces are located in the rural Andes at 1900-3900 metres above sea level (masl) with a population of more than 134,000 inhabitants (32). The majority of the households are made of adobe walls and earthen floors, and local trade is the main source of income, most important agricultural products are potatoes, beans, manioc, rice and wheat. Families raise chickens, pigs, ducks, guinea pigs, sheep and goats for consumption. These animals are kept in dens but are often also kept in small numbers as cohabitants. Cattle are used for agricultural activity and milk production. Communities also receive aid from government supplementary feeding, cash transfer and other social programmes.
Study Design
We conducted a quasi-experimental study with the mothers and the fathers of children who were previously participating in a community-randomised controlled trial (c-RCT, parent study) (33).The parent study used a covariate-base constrained randomisation (33).The aim of the parent study was to implement an integrated home-based intervention package (IHIP) in resource-limited Andean communities and assess the health impact on children under 36 months of age. The parent study intervened with two home-environmental improvements comprising (i) an ICS, a kitchen sink, tap water and a kitchen hygiene education, and, (ii) an early child development intervention (ECD). The c-RCT used a 2x2 factorial design and the homes of participating children were allocated in four arms namely three single intervention arms (ICS, ECD and Control) and one combined arm (ICS–ECD). The trial assessed the impact on child diarrhoeal and respiratory morbidity, and early child development improvements.
In our study, all parents of the 317 children in the parent study were eligible to participate. Pregnant women were excluded given pregnancy-related changes in blood pressure or glucose levels. Because our hypothesis tested the effect of ICS, we allocated participants in two study arms: ICS (ICS and ICS-ECD in parent study) and control (ECD and control in parent study). We thus, report following CONSORT reporting guidelines and provide a completed CONSORT checklist in Additional file 3.
Enrolment and study participants
Participants were recruited between January and May 2017; corresponding to 6-10 months after the parent study interventions were implemented. Fieldworkers visited participants at their homes the day before the physical examination to invite them to participate. They requested them to fast overnight prior to the assessments between 5:30-11:00 AM the next day. Home visits were only done once due to time constraints. The physical examinations were carried out at a central location (i.e. schools, community hall), the closest health post in the community or occasionally in a participants household.
Data collection
Physical examination
The physical examination comprised measurements of weight, height, waist circumference, blood pressure (systolic and diastolic), HDL cholesterol, triglycerides and glucose levels. Standardised techniques were used for anthropometric examinations. Individuals were weighed using a floor scale, height was measured using a portable stadiometer, and waist circumference was measured using a tape measure (34). Blood pressure was measured with the subject seated and in repose, using calibrated automatic blood pressure meter (OMRON HEM-712C). Participants with high blood pressure (≥130 mm Hg systolic or ≥85 mm Hg diastolic) had a second measurement taken. We averaged the results for the MetS diagnosis. Capillary blood samples collected from a finger prick were used to assess triglycerides, HDL cholesterol and glucose levels with CardioChek PA, a point of care testing (POCT) device validated for reliably measuring lipid and glucose levels (35). Lipid values were used as presented by the POCT device. For glucose assessments, we applied a conversion factor of 1.11 to transform whole blood glucose samples to plasma glucose values (36). Internal quality control testing was performed as recommended by the manufacturer.
Dietary recall
A 24-hour food recall questionnaire was administered at the time of the physical examination. Participants were asked to recall all foods and beverages they had consumed the previous day, from the moment they woke up until they went to bed. The questionnaire was adapted from the guideline for Documenting Traditional Food Systems of the Centre for Indigenous Peoples’ Nutrition and Environment (CINE-McGill University) (37). We categorised foods in six groups (carbohydrates (cereals, grains and tubers), vegetables, fruits, vegetable protein (legumes), animal protein, dairy products and fats). We calculated the Diet Diversity Score (DDS), defined as the number of food groups consumed by each participant (38). The DDS ranges from 0 to 7.
Data abstracted from the parent study
Information on the type of stove use for cooking and heating was taken from the parent study baseline, between 6-10 months before the physical examination. Additionally, demographic data such as households’ characteristics, education and main economic activity were used to calculate the Peruvian index of Unsatisfied Basic Needs (NBI, Spanish abbreviation) according to the Peruvian National Institute of Statistics and Informatics (39).
Data analysis
Metabolic Syndrome
We analysed MetS as an exploratory outcome of the parent study, which focused on child health outcomes (33). MetS is an exploratory outcome studied in the adult population of the parent trial and was not previously defined in the trial registration. We hypothesised that household environmental improvements, such as ICS interventions likely impact also on adult health.
We used the JIS diagnostic criteria to define MetS (7). Participants were diagnosed with MetS if they presented at least three of the following five risk factors: elevated triglycerides (≥150 mg/dl), reduced high-density lipoprotein (HDL) cholesterol (≤40 mg/dl for men and ≤50 mg/dl for women), elevated blood pressure (BP) (systolic BP ≥130 mm Hg and/or diastolic BP ≥85 mm Hg), elevated fasting plasma glucose (≥100 mg/dl) and elevated waist circumference population specific for Ethnic Central and South Americans (90cm≥ for men and 80cm≥ for women).
Statistical analysis
Data was entered in the Census and Survey Processing System (CS Pro 6.3) data file and analysis was performed using STATA 14 Statistical software (STATA CORP, College Station, Texas, USA). Descriptive analysis was carried out for demographic, socioeconomic, health determinants and dietary characteristics by study arm and gender. During the physical examination, some subjects declined a second blood pressure measurement. In cases where high blood pressure was the defining criteria to diagnose MetS, the subject was excluded from the analysis.
Because ICS interventions aim at improving women and children health due to the considerable amount of time they spend in the kitchen environment, we tested the interaction between ICS intervention and gender. Using the Mantel-Haenszel method we obtained a p=0.09, therefore, all the analysis was stratified by gender. Mixed effect Poisson regression with robust variance was used for the univariable and multivariable analysis to estimate the prevalence ratio (PR) of MetS.(40). We included the following independent variables in the univariable analysis: cookstove used (ICS intervention and control), age, farming as main activity, secondary education, living at high altitude (≥2500 masl), NBI, DDS and household as place of physical examination. The variable ‘being a farmer’ was used as a proxy for the most demanding physical activity. Household as place of physical examination was included in the model to account for its possible influence on the outcome measurement. Variables with a p-value <0.20 in the univariable analysis were considered in the multivariable model, those with p<0.1 remained in the final model.
Ethics
The study was approved by the Universidad Peruana Cayetano Heredia (UPCH) Ethical Review Board (N° 192-08-16) and the Cajamarca Regional Health Authority. All participants signed a written informed consent form. Participants received their results after the physical examination. If their results indicated abnormal glucose, lipids and/or blood pressure, they were referred to the local health establishment. This sub-study was part of and used data from its parent trial, which received ethical clearance earlier (N°R74-15-16 and trial registration number ISRCTN26548981).