Introduction We conducted a biomarker study to characterise exposure to pesticides among farmers and their families in Thailand to assess the relative importance of the dermal exposure route and to identify important factors that determine exposure levels within farmers' families.
Methods Sixteen farmers' families (eight vegetable and eight fruit farmers) participated in the study. Three morning spot urine samples were collected during a pesticide spraying week. Spot samples were grouped by individual and analysed for dialkylphosphate (DAP) metabolites and creatinine. Additional information on exposure and lifestyle was collected by means of questionnaires. Dermal exposure was assessed using a semi-quantitative observational method (DREAM).
Results Urinary DAP levels varied 20-fold between farmers, with average (geometric mean) levels of 51.1 μg/g for vegetable and 122.2 μg/g for fruit farmers. A moderate correlation (rs∼0.45) was found between loge-transformed DREAM scores and DAP levels. Farmers' urinary metabolite levels were not correlated with those of their spouses (rs∼−0.30) or children (rs∼−0.00) collected on the same days. Detectable spouses' DAP levels were on average (geometric mean) 13.0 μg/g and those of children 7.6 μg/g.
Discussion Farmers in Thailand as well as their families are exposed to pesticides in the spraying season and dermal exposure is an important route. The main route of exposure for farmers' families seems to be through transfer from the farmer to family members or contamination of the home environment, rather than family members helping or playing on the farm. Showering or washing immediately after pesticide spraying greatly reduces the potential exposure of family members to pesticide residues.
- Biological monitoring
- dermal exposure
- exposure monitoring
- international occupational health
Statistics from Altmetric.com
What this paper adds
As in other countries, the agricultural use of pesticides in Thailand has risen over the last 10 years, resulting in increased exposure of farmers in the spraying season, which could result in adverse health effects.
Data on work practices and resulting potential contamination of the home environment in this population are limited.
This study shows that farmers can expose their families to pesticide residues, most likely through direct transfer from the farmers to their families or by contamination of the home environment.
Washing or showering directly after work can reduce the exposure of family members.
Following trends in other countries, agricultural pesticide use in Thailand has risen over the last decade, resulting in increased agricultural production.1 Around 64% of Thai vegetable farmers regularly use pesticides, mostly organophosphates and pyrethroid.2 These chemicals are readily available and widely used in crop production.3
The adverse health effects of pesticide exposure include neurological abnormalities, increased cancer risk, respiratory problems and reproductive, endocrine and skin problems.4 In Thailand, farmers reported acute symptoms after working with pesticides, including headaches, dizziness, nausea and abdominal pain.5 Despite the health risks associated with pesticides, Thai farmers seem to have inadequate knowledge about their safe use.1 6 Most farmers apply cocktails of pesticides indiscriminately and use higher than recommended doses.3 5 7–9 Over- or under-dosing for a particular level of pest infestation is common and most farmers do not use protective equipment.3 5 7–9
Assessment of farmers' pesticide-handling practices in Thailand also reveals the potential for contamination of living areas, and consequent exposure of their families. For example, the majority of farmers do not discard pesticides safely,3 while about 25% of cut-flower farmers store pesticides in their living rooms.7 8 Biomarker data from farmers' children in Thailand show significantly higher urinary pyrethroid10 and dialkylphosphate (DAP)11 metabolite concentrations than children of parents in non-agricultural professions.
We conducted a biomarker study to characterise pesticide exposure among farmers and their families during the spraying season in Thailand, specifically aimed at characterising the relative importance of the dermal route of exposure and identifying important factors that determine exposure levels in farmers' families.
Sixteen households (the farmer, their spouse and one or more children) were randomly selected, based on a set of a priori inclusion criteria, from a total of 110 eligible families in the Nong-sarai and Wangkata subdistricts (Nakhon-ratchasima province) in Thailand. Families were eligible if they lived in either subdistrict, consisted of a farmer or farm worker and his/her spouse (both between 18 and 65 years of age), were employed in vegetable or fruit farming, were involved in the spraying season during the study period and had at least one child aged between 2 and 12 years. If there was more than one child in that age range, the child was selected at random. To minimise bias we contacted eight families randomly selected from lists of eligible families available from the vegetable farming community and eight from the list of eligible farming families from the fruit farming community, who agreed to participate.
Two questionnaires, one each for the farmer and their spouse, developed by the research team were administered by trained research assistants. Both questionnaires sought information on demographics, lifestyle, medical history and current health. The farmers' questionnaires also covered pesticide use and storage, work practices, and hygienic behaviour, while the spouses' questionnaires covered hygienic behaviour at home, occasional work on the farm or with pesticides, and behaviour of the children. Questionnaires were translated from English to Thai and back into English by two separate bi-lingual researchers to ensure accuracy and consistency.
Farmers were observed by one researcher (CH) during pesticide spraying on one working day. Potential and actual dermal exposure was assessed using the structured semi-quantitative observational dermal exposure method (DREAM), which has been described in detail elsewhere.12 Briefly, DREAM is based on a validated conceptual model for dermal exposure13 providing reliable14 and accurate15 semi-quantitative estimates. It comprises a structured multiple-choice questionnaire assessing exposure determinants. Dermal exposure is evaluated systematically through the most important exposure routes (emission, transfer and deposition). Exposure levels on the outside of clothing (potential dermal exposure) and, after accounting for the effects of clothing on the skin, exposure on the skin (actual dermal exposure) are estimated. The principal investigator was trained in DREAM methodology by a researcher experienced in its use (FdV).
Three morning spot urine samples were collected on three occasions during a spraying week, the day of collection depending on the spraying day. Approximately 50 ml of urine were collected and stored in a freezer at −20°C before urinary analysis for DAP metabolites (diethylphosphate (DEP), diethylthiophosphate (DETP), diethyldithiophosphate (DEDTP), dimethylphosphate (DMP) and dimethylthiophosphate (DMTP)) and creatinine. Spot urine samples were grouped by individual before laboratory analysis. Individual biomarker concentrations were summed to obtain a single summary score for total pesticide exposure for each subject.
Results were analysed using R (v 2.9.2). DREAM scores14 15 and metabolite levels were loge-transformed to summarise the data appropriately. The geometric mean is an estimator for the population median and confidence intervals based on the geometric mean makes inference about the population median.16 Univariate least-squares regression was employed. With 16 families in the study, only linear (in log-space) associations were assessed. Multivariate modelling and specific assessment of interactions were not possible. The level of statistical significance used was 5%, with p values between 5% and 10% classified as borderline significant results because of the small sample size.
All vegetable farmers and 75% of fruit farmers (6/8) were male. Vegetable farmers had worked, on average, for 8 years (median; range <1 year to 30 years) on the farm (average size 10 704 m2). Produce (50% white radish) was harvested 4–5 times per year. Fruit (mainly mango) farmers had, on average, worked for 16 years (median: range <1 year to 30 years) on their farms (average size 99 808 m2) and harvested twice a year. Most spouses occasionally helped on the farm, but whereas seven of eight vegetable farmers' spouses did not use pesticides, four of eight of fruit farmers' spouses sometimes used pesticides. Only two families occasionally brought their children to the farm, where they helped out.
Urinary DAP levels of farmers ranged from 17.4 to 350.6 μg/g creatinine, with a geometric mean level of 79.0 μg/g creatinine (51.1 μg/g creatinine for vegetable farmers and 122.2 μg/g creatinine for fruit farmers; p∼0.03) (figure 1A). There was moderate correlation (rs∼0.45) between loge-transformed potential dermal exposure scores and urinary DAP levels (rs∼0.35 for vegetable farmers and rs∼0.25 for fruit farmers) (figure 1B). Urinary DAP levels (geometric mean) were 2.4 times higher in fruit than in vegetable farmers and increased by 21% (95% CI 2% to 40%) per hour spraying (range 2–7 h). All fruit farmers used booms to apply pesticides (geometric mean 122.2 μg/g), while vegetable farmers used either a backpack (geometric mean ∼45.2 μg/g) or mist sprayer (geometric mean ∼73.8 μg/g); there was no significant difference between the latter two. Three farmers (two vegetable and one fruit farmer) who wore cotton gloves during spraying had on average 60% lower DAP metabolite levels (geometric mean ∼37.8 μg/g) than the other farmers (geometric mean ∼93.6 μg/g) (p∼0.08).
Farmers' geometric mean urinary DAP levels were not correlated with those of their spouses (rs∼−0.30) or children (rs∼−0.00) collected on the same days. Geometric mean DAP levels in the farmers' spouses were on average (median) 3.9% of farmers' levels but were only detected in 11 spouses, with an average (geometric mean) level (n=11) of 13.0 μg/g creatinine (range: limit of detection to 64.0 μg/g) (figure 1C). Seven of 16 children had detectable levels ranging from the limit of detection to 30.2 μg/g. DAP metabolite levels in children were on average (median) 0.34% of farmers' levels and 38% of spouses' levels and were on average (geometric mean of detectable samples) 7.6 μg/g (figure 1D).
Spouses' and children's loge-transformed DAP metabolite levels were not correlated with time spent on the farm, place of pesticide storage, or method of washing farmers' clothes (ie, washed with other clothes or not). The main determinant of spouse's urinary DAP metabolite levels was whether the farmer had showered immediately after work (geometric mean ∼1.1 μg/g) or not (geometric mean ∼15.9 μg/g) (p∼0.05). Although the correlation between farm size (where their father or mother worked) and children's loge-transformed DAP metabolite levels was borderline significant, the main determinant of the child's metabolite levels seemed to be whether their parent washed him/herself at the farm before returning home or at home. If the farmer washed him/herself at home, children's geometric mean metabolite levels were 22-fold (95% CI 2 to 337) higher than if the farmer washed him/herself at work (p∼0.04). Children's urinary DAP levels were not associated with age, or with whether their mother helped out on the farm or worked with pesticides.
Due to the relatively small sample size, more detailed assessment of the interaction of specific exposure determinants or pesticide-related health effects was not possible. For example, the significant difference in exposure levels between vegetable and fruit farmers might be a proxy for differences in products, methods of applying pesticides or handling concentrates, or equipment and layout of the farms between both types of farming (eg, all fruit farmers used a spraying boom but no vegetable farmer did).
Dermal exposure appears to be an important exposure route according to the moderate correlation between the DREAM dermal scores and urinary DAP levels. However, the DREAM methodology, integrating chemical properties, operator handling, hygienic behaviour and the use of protective clothing and equipment, estimates dermal exposure based on observation, not direct measurements.
Nonetheless, in this small study Thai farmers were exposed to pesticides in the spraying season and dermal exposure was important. We found evidence, in agreement with previous data,5 to suggest that wearing gloves lessens exposure, presumably through reducing direct transfer from contaminated equipment or pesticides on plants. Farmers' families were also exposed to pesticides during the spraying season, mainly through contamination of the home environment or direct transfer from the farmer to family members. As such, these data suggest that showering or washing immediately after pesticide spraying, preferably on the farm instead of at home, greatly reduces family members' potential exposure to pesticide residues.
We would like to thank all farmers and their families for participating in this study. Thanks are also due to the health officers of the Pak-Chong Health Care Centre in Thailand.
Funding This study is part of a PhD project sponsored by the Royal Thai Government.
Competing interests None.
Ethics approval This study was conducted with the approval of the University of Manchester's Committee on the Ethics of Research on Human Beings (ref TPCS/ethics/09225) and the Ethics Committee of the Suranaree University of Technology (ref 5621/0734).
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.