This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Research Protocols, is properly cited. The complete bibliographic information, a link to the original publication on http://www.researchprotocols.org, as well as this copyright and license information must be included.
Pesticide use is increasing in low- and middle-income countries (LMICs) including Costa Rica. This increase poses health risks to farm owners, farm workers, and communities living near agricultural farms.
We aimed to examine the health effects associated with occupational pesticide exposure in farm owners and workers from conventional and organic smallholder farms in Costa Rica.
We conducted a cross-sectional study involving 300 owners and workers from organic and conventional horticultural smallholder farms in Zarcero County, Costa Rica. During the baseline study visit, we administered a structured, tablet-based questionnaire to collect data on sociodemographic characteristics, pesticide exposure, and health conditions (eg, respiratory and allergic outcomes and acute pesticide intoxication symptoms) and administered a neurobehavioral test battery (eg, Finger Tapping Test and Purdue Pegboard); we measured blood pressure, anthropometry (height, weight, and waist circumference), and erythrocytic acetylcholinesterase activity and also collected urine samples. In addition, a functional neuroimaging assessment using near-infrared spectroscopy was conducted with a subset of 50 study participants. During the follow-up study visit (~2-4 weeks after the baseline), we administered participants a short questionnaire on recent pesticide exposure and farming practices and collected hair, toenail, and urine samples. Urine samples will be analyzed for various pesticide metabolites, whereas toenails and hair will be analyzed for manganese (Mn), a biomarker of exposure to Mn-containing fungicides. Self-reported pesticide exposure data will be used to develop exposure intensity scores using an exposure algorithm. Furthermore, exposure-outcome associations will be examined using linear and logistic mixed-effects regression models.
Fieldwork for our study was conducted between May 2016 and August 2016. In total, 113 farm owners and 187 workers from 9 organic and 83 conventional horticultural smallholder farms were enrolled. Data analyses are ongoing and expected to be published between 2019 and 2020.
This study is one of the first to examine differences in health effects due to pesticide exposure between farm owners and workers from organic and conventional smallholder farms in an LMIC. We expect that this study will provide critical data on farming practices, exposure pathways, and how occupational exposure to pesticides may affect farm owners and workers’ health. Finally, we hope that this study will allow us to identify strategies to reduce pesticide exposure in farm owners and workers and will potentially lay the groundwork for a future longitudinal study of health outcomes in farm owners and workers exposed to pesticides.
DERR1-10.2196/10914
Pesticides are extensively used in agriculture and for control of vector-borne diseases across the globe [
Pesticide applicators (either farm owners or workers) in smallholder farms are often exposed to these chemicals at different stages of the process (eg, storage, mixing, preparation, and application) [
The characterization of pesticide exposures in LMICs is challenging due to the short half-lives of most of these chemicals in the human body, limited availability of biomarkers of exposure, and lack of epidemiological data [
Several studies have examined the health effects of occupational pesticide exposure in tropical settings [
We conducted a cross-sectional study of 300 farm owners and workers with repeated exposure assessment at two time points (to study the variability of pesticide exposure over time) between May 2016 and August 2016 (rainy season, during which the highest pesticide application is expected). The specific objectives of the project are as follows (
To assess occupational pesticide exposure in owners and workers of conventional and organic farms, using self-reported pesticide use data and biomarkers of pesticide exposure.
To evaluate the association of occupational pesticide exposure (determined using a pesticide exposure matrix and also biomarkers of exposure) with self-reported symptoms of acute intoxication in the last 12 months.
To examine the association of occupational pesticide exposure with self-reported respiratory and allergic outcomes in the last 12 months.
To evaluate the association between occupational pesticide exposure and cardiometabolic effects, such as adiposity and high blood pressure.
To assess the association of occupational exposure to organophosphates and carbamates with erythrocytic acetylcholinesterase (AChE) activity.
To assess the association of occupational pesticide exposure with neurobehavioral outcomes, such as working memory, visual perception, and fine motor function.
To examine the association of occupational pesticide exposure with changes in brain activity, assessed using functional near-infrared spectroscopy (fNIRS).
This study is part of the Pesticide Use in Tropical Settings (PESTROP) Project, which aims to deepen our understanding of the environmental, health, and regulatory dimensions of pesticide use in conventional and organic agriculture in LMICs. The design of our research was partly informed by a study conducted in the same study area between 2014 and 2016 [
All study materials and procedures were approved by the human subjects committee of the Universidad Nacional in Costa Rica (UNA-CECUNA-ACUE-04-2016) and Ethical Board of the Ethikkommission Nordwest- und Zentralschweiz in Switzerland (EKNZ-UBE 2016-00771). Written informed consent was obtained from all study participants at enrollment. Study results will be communicated back to participants and stakeholders at restitution workshops (
The study was conducted in the Tapezco river catchment in the Zarcero County, Costa Rica (
Aims (bold text), research design details and tools used (tick) in the study conducted in Zarcero County, Costa Rica, 2016.
Diagram of the fieldwork setup in the Zarcero study, Costa Rica, 2016.
Study area (Tapezco river catchment) with global positioning system locations of 92 farms that were included in our study conducted in Zarcero County, Costa Rica, 2016. Shape files provided by Moraga (2015).
Conventional farms in the study area were identified using random global positioning system (GPS) points generated on the basis of smallholder land use data from 2015 (Moraga G, unpublished data, 2000) [
After organic and conventional farms in the study area were identified, farm owners were briefly informed about the study aims and procedures (initial visits to the farms are, henceforth, called Information campaign;
Participants were visited twice during the study duration, either at the farms where they worked or at their homes (
The follow-up study visits (duration ~15 minutes, 2-4 weeks after the first visit) were conducted by two trained research assistants and included the administration of a short questionnaire on recent pesticide exposure (administered to all participants) and farming practices (administered only to farm owners) and collection of toenail and hair samples and a second urine specimen.
All study instruments were pilot-tested in 10 farmers. Study protocols were administered by trained research assistants in Spanish, and data were entered directly into tablets (Samsung Galaxy Note 10.1 N8010) via an entry mask using Open Data Kit [
We based our sample size calculation on the difference in the erythrocytic AChE activity between organic and conventional farm owners and workers [
Information on sociodemographic variables and occupational exposure to pesticides was collected using structured questionnaires (
Information collected during the baseline study visit using structured questionnaires.
Questionnaire section | Data collected |
Sociodemographic characteristics | Age (years), sex, country of birth, years living in Costa Rica, years living in the Zarcero County, education level (last grade completed), handedness (right or left), smoking (number of cigarettes per day), alcohol consumption (number of glasses), family income (Costa Rican colones per household), computer literacy |
Housing characteristics | Years living in the current house, number of bedrooms, number of people living in the house, type of water source, pet ownership (number and type), and farm animals living in or next to the house (number and type) |
Work history | Age when started working in agriculture (years), age at first contact with pesticides (years), past jobs, and current jobs in agriculture |
Pesticide use | Pesticide use ( |
Residential pesticide use | Indoor or outdoor pesticide use in the home (type) |
Medical history | Respiratory and allergic symptoms, acute pesticide intoxications during the last 12 months (number of times), and other illnesses (type) |
Most frequently used pesticides (active ingredients; ordered from most frequently use to the least frequently used) in agricultural farms in the Zarcero County, Costa Ricaa.
Active ingredient | Commercial names | Chemical subgroup | Group of action |
Chlorothalonil | Bravo, Bravonil, Knight, Talonil, Thalonex, Folio Gold, Odeon | Chloronitriles | Fungicide |
Benfuracarb | Oncol | Carbamates | Insecticide |
Mancozeb | Dithane, Mancol, Ridomil, Titan | Dithiocarbamates | Fungicide |
Boscalid | Bellis, Endura | Pyridinecarboxamids | Fungicide |
Carbendazim | Afin, Cozaid, Crotonox, Carbendazina | Benzimidazoles | Fungicide |
Acephate | Acefate, Orthene, Yucal | Organophosphates | Insecticide |
Phorate | Forato, Thimet, Thimetoato, Timefor | Organophosphates | Insecticide |
Fenamiphos | Fenemiphos, Nemacur | Organophosphates | Insecticide |
Chlorpyrifos | Agromil, Batazo, Baygon, Lorsban, Solver, Terminator, Swat | Organophosphates | Insecticide |
Glyphosate | Atila, Evigras, Ranger, Round Up | Glycine derivatives | Herbicide |
Carbofuran | Carbodan, Curator, Furadan | Carbamates | Insecticide |
Cypermethrin | Best, Cascabel, Cipermetrina, Combat, Cruz Verde, Tigre, Excalibur | Pyrethroids | Insecticide |
Propamocarb | Acrobat CT, Previcur, Proplant, Prevalor | Carbamates | Fungicide |
Paraquat | Gramoxone, Preglone, Rafaga | Bipyridiliums | Herbicide |
Propineb | Antracol, Inicol, Taifen | Dithiocarbamates | Fungicide |
aModified from: Ramírez et al, 2016 [
Spot urine samples were collected during the baseline and follow-up visits in plastic containers of 100 mL
Urine samples will be analyzed for multiple pesticide metabolites including, but not limited to, ethylenethiourea (ETU, a metabolite of mancozeb); propylenethiourea (PTU, a metabolite of propineb); 3,5,6-trichloropyridinol (TCP, a metabolite of chlorpyrifos); 3-phenoxybenzoic acid (3-PBA, a metabolite of pyrethroids permethrin, cypermethrin, deltamethrin, and cyfluthrin); and hydroxy pyrimethanil (a metabolite of pyrimethanil). These pesticides and their metabolites were selected because they are among the most commonly used in Zarcero [
Hair samples (~20-30 strands) were collected from the occipital region, within 2 mm from the scalp, using stainless-steel scissors during the follow-up study visit. The samples were then stored at room temperature in sterile plastic bags until their shipment to the Federal University of Bahia, Brazil.
Hair specimens will be analyzed for manganese (Mn), which is contained in ethylene bisdithiocarbamate fungicides, such as mancozeb. Briefly, the nearest centimeter scalp (proximal end) of hair will be sonicated for 20 min in 0.5% Triton, rinsed 5 times with ultrapure water, sonicated for 10 min in 1-N nitric acid, rinsed once with 1-N nitric acid, and then rinsed 5 times with ultrapure water [
Toenail samples were collected during the follow-up study visit. Participants were asked to cut their toenails with clean stainless-steel nail clippers and put them inside of a sterile plastic bag. Toenail specimens were stored at room temperature until their shipment to Federal University of Bahia, Brazil.
Toenail samples will be analyzed for Mn using the same procedure described above for hair samples. Briefly, nails will be washed in a Triton X-100 solution, put in acetone, repeatedly rinsed with ultrapure water, and then dried in an oven. Later, the dried nails will be digested with spectroscopic-grade acid in a microwave digestion oven (Mars-Express6, CEM), diluted to 10 mL with ultrapure water, and analyzed electrothermal atomic absorption spectrometry with Zeeman background correction [
Participants were administered a checklist of symptoms of acute organophosphate and carbamate poisoning (eg, excessive salivation, lacrimation, vomiting, and diarrhea) during the 12-month period before the baseline study visit. This checklist has been previously used in studies of Latin American farm workers [
A short version of the European Community Respiratory Health Study II questionnaire [
Study participants were administered the following eight neurobehavioral tests: Animals and words with initial letter “F” (to assess semantic and phonemic verbal fluency) [
These tests were selected on the basis of previous studies of Latin American populations exposed to pesticides [
We used fNIRS (NIRSport, NIRx Medical Technologies, Los Angeles, CA, USA) to assess the cortical function associated with pesticide exposure in a random subsample of 50 study participants; a detailed description of our study methods can be found elsewhere [
Trained research assistants measured systolic and diastolic blood pressure at two different time points (ie, at the beginning of “
Capillary blood samples were collected at the baseline visit according to the manual of the Test-mate ChE Cholinesterase Test System (Model 400; EQM Research Inc, Cincinnati, OH, USA). Briefly, a small lancet (size 30) was used to collect a small sample of 10 µm from the tip of the index finger of each study participant and placed into a capillary tube. Blood samples were analyzed on site for the erythrocytic AChE activity and hemoglobin levels using the same collection instrument [
We will explore differences in self-reported pesticide exposures and health outcomes (ie, respiratory and allergic as well as neurobehavioral and cardiometabolic outcomes, brain activity, symptoms of acute pesticide poisoning, and erythrocytic AChE activity) in farm workers from conventional and organic farms. Cumulative lifetime pesticide exposure, exposure during the last 12 months, and during the last week will be estimated using exposure intensity scores derived from a semiquantitative exposure algorithm (based on self-reported data on pesticides, personal protective equipment use, and personal hygiene habits) [
Associations of pesticide exposure with health outcomes of interest will be examined using both exposure scores and biomarkers of exposure. More specifically, we will fit linear or logistic (depending on the outcome) mixed-effects regression models (with the variable “participant” as random effect and other variables, such as outcome and covariates, as fixed effects) to explore the association of pesticide exposure with the following: (1) self-reported symptoms of acute pesticide poisoning in the last 12 months; (2) self-reported symptoms of respiratory and allergic symptoms and outcomes; (3) neurobehavioral outcomes and brain activity; (4) cardiometabolic effects (ie, obesity and hypertension); and (5) erythrocytic AChE activity. These mixed-effects models will allow us to take into account the correlation between and within farms as well as between and within farm workers. Furthermore, we will fit generalized additive models to examine the nonlinearity of the exposure-outcome associations.
We will identify potential confounders and known predictors of the health outcomes of interest (eg, age and education level for neurobehavioral outcomes) using directed acyclic graphs and will include them
The fieldwork for this study was conducted between May 2016 and August 2016. A total of 300 participants, including 113 farm owners and 187 workers from nine organic (48 participants) and 83 conventional (252 participants) horticultural smallholder farms from Zarcero County, Costa Rica, were enrolled. We had a 6.3% (281/300) loss to follow-up of study participants between the baseline and follow-up study visits (conducted ~2-4 weeks apart). During the study implementation, we observed that farms were, on average, smaller than we expected (about three owners or workers per farm) and that there were not as many organic farms in the study area as anticipated (only 10 farms out of the 25 that we expected). Interviews to farm owners and key community actors conducted during this study revealed that many organic farm owners had recently started using synthetic pesticides due to the increasing costs of growing organic produce and getting certified as organic producers. Given the limited number of organic farms in the study area, we decided to include all farm owners and workers from organic farms located in the study area or within 5 km from it. In addition, to reach the targeted sample of 300 participants, we had to enroll more conventional farm owners and workers than what we had anticipated. Data analyses are ongoing and expected to be published between 2019 and 2021.
This study has several limitations. First, its cross-sectional design will prevent us from identifying causal associations of pesticide exposure with health effects of interest. Second, we will not be able to exclude the possibility that there is recall or information bias, especially when relying on self-reported exposure to pesticides. In our study population, this bias may have worsened under the following conditions: (1) most study participants had a relatively low educational level; (2) pesticide use varied by crop and season; and (3) some farm owners did not communicate the specific pesticides that were used in their farms by their employees. Therefore, the validation of self-reported pesticide exposure data against biomarkers of exposure is crucial. Third, given that we had to enroll more conventional farm owners and workers than what we expected, and this could potentially threaten the internal validity of our study, we will assess participants’ pesticide exposures using exposure intensity scores and biomarker concentrations in addition to their farming practices. We will also run analyses using only data from the conventional farms to evaluate the effects of the miscalculated sample size on our exposure-outcome associations. Fourth, we observed some significant differences between owners and workers from conventional and organic farms (eg, seasonal workers, differences in country of origin, and education level) that could potentially confound the exposure-outcome associations. Hence, it was important to also collect detailed information on confounding factors and predictors of the outcomes of interest.
The limitations of this study are offset by notable strengths, including (1) the quantification of pesticide metabolites and Mn concentrations in different biological matrices, which will allow us to validate the exposure information collected via questionnaires (or at least data from recent exposures given the relatively short half-lives of some of the biomarkers of exposure); (2) the comparison of workers and owners from conventional and organic farms using comprehensive questionnaires on occupational pesticide exposure; and (3) the assessment of health outcomes using internationally standardized tests that will allow for direct comparison of the results from this study to those from studies of other populations.
This study is one of the first studies to examine the health effects of exposure to a wide range of pesticides on Latin American workers from conventional and organic farms. In addition, this is one the first epidemiological studies to examine the association of pesticide exposure with brain cortical activity in farm workers. We expect that this study will provide critical data on how occupational exposure to common pesticides may affect farm owners’ and workers’ health. Finally, we hope that this study will allow us to identify strategies to reduce pesticide exposure in farm workers and will lay the groundwork for a future longitudinal study of health outcomes in farm owners and workers exposed to pesticides.
Peer-review reports.
3-phenoxybenzoic acid
acetylcholinesterase
functional near-infrared spectroscopy
global positioning system
intraclass correlation coefficient
limit of detection
low- and middle-income countries
manganese
3,5,6-trichloropyridinol
We gratefully acknowledge the study participants and staff (A Alfaro, C Hyland, A Campos, S Colombari, M Quirós-Lépiz, G Rodríguez, A Ulloa, and H Wey). We would also like to thank D Rojas for his support with the fNIRS assessments, C Wesseling and B van Wendel de Joode for their assistance with the questionnaire design, B Eskenazi for her insights on the neurobehavioral test selection, and Christian Schindler for his statistical support. This work was made possible by research supported by the Swiss Federal Institute of Aquatic Science and Technology (EAWAG), Universidad Nacional in Costa Rica, Forschungsfonds of the University of Basel, and the Swiss Network for International Studies (SNIS).
MSW is the principal investigator of the PESTROP project, responsible for the overall study coordination. AMM is the project leader of the PESTROP project in Costa Rica and, together with SF, is responsible for the study design, fieldwork supervision, and data analyses. JMB and RGV designed and implemented the fNIRS protocol. PS, FTW, CS, RILE, and FRM contributed to the development of the study protocols and questionnaires. PS also contributed to the fieldwork supervision. CHL and JAMF contributed to the development of the study protocols and will be responsible for the analyses of the biological samples. SF, in close collaboration with AMM and MSW, drafted this manuscript.
None declared.