Association between Organophosphate Pesticide Exposure and Insulin Resistance in Pesticide Sprayers and Nonfarmworkers

Insulin resistance is a risk factor for various diseases. Chronic organophosphate exposure has been reported to be a cause of insulin resistance in animal models. This cross-sectional study aimed to evaluate the association between organophosphate exposure and insulin resistance in pesticide sprayers and nonfarmworkers. Participants aged 40–60 years, consisting of 150 pesticide sprayers and 150 nonfarmworkers, were interviewed and assessed for their homeostatic model assessment of insulin resistance (HOMA-IR) level. Organophosphate (OP) exposure was measured in 37 sprayers and 46 nonfarmworkers by first morning urinary dialkyl phosphate (DAP) metabolites. The DAP metabolite levels were not different in either group except for diethylthiophosphate (DETP; p = 0.03), which was higher in sprayers. No significant association was observed between DAP metabolite levels and HOMA-IR. Wearing a mask while handling pesticides was associated with lower dimethyl metabolites (95% CI = −11.10, −0.17). Work practices of reading pesticide labels (95% CI = −81.47, −14.99) and washing hands after mixing pesticide (95% CI = −39.97, −3.35) correlated with lower diethyl alkylphosphate level. Overall, we did not observe any association between OP exposure and insulin resistance in pesticide sprayers and the general population. However, personal protective equipment (PPE) utilization and work practice were associated with OP exposure level in sprayers.


Introduction
Type 2 diabetes is one of the major global health problems and was reported by the World Health Organization (WHO) in 2018 as the seventh leading cause of death worldwide. Diabetes also leads to

Study Design and Participants
A cross-sectional study was conducted in Chiang Mai province, Thailand, between May and July 2019, which is the period when the most intensive insecticide application occurs in agricultural villages. A community partnership that included village health volunteers, village headmen, and farmworker leaders was set up. People meeting the study criteria were recruited by the partnership. Participants aged 40-60 years and consisting of 150 pesticide sprayers from four agricultural villages and 150 subjects with no involvement in agricultural activity from three nonagricultural villages were enrolled. The inclusion criteria for the sprayers were having been pesticide sprayers for at least one year and having sprayed OP pesticides. The exclusion criteria were diagnosis of diabetes or consumption of antidiabetic medication, end-stage renal disease, cirrhosis, or consumption of HIV protease inhibitor. In addition, participants that were initially selected but were observed to have fasting plasma glucose of 126 mg/dL and higher were excluded.
Participants were requested to sign the consent form after full explanation of the study aims and methodology. After at least 8 h of fasting, each participant was interviewed and provided a blood sample to measure fasting blood glucose (FBG) and insulin levels.
A subset of 37 pesticide sprayers who had sprayed OPs in that season and planned to spray OPs within a month after blood collection as well as 46 nonfarmworkers were selected by convenience sampling for the measurement of urinary dialkyl phosphate (DAP) metabolites and lipid panel.
To collect single morning urine sample, each participant was provided with a urine container and instructed for first morning urine collection in the following morning after OP spraying. Blood and urine collection as well as interviews were conducted for both farmworkers and nonfarmworkers at village temples.
Several subjects were excluded from the enrollment group. This included four farmers that were not pesticide sprayers, a sprayer with a small blood sample, and three sprayers and seven nonsprayers who had high FBG levels. The final group to be analyzed consisted of 142 sprayers and 143 nonsprayers, including 36 sprayers and 42 nonsprayers to be analyzed for urinary DAP metabolites.

Questionnaire
The questionnaire consisted of questions on sociodemographic and health information, 24 h dietary recall, physical activity, fresh produce consumption behavior, household pesticide use, work practice, and trade name of pesticides used (for pesticide sprayers).
A chart providing pictures of OPs and other pesticide containers that are generally used throughout Thailand was used for the interviews. Pesticide sprayers were asked to indicate the pesticide they have used in agricultural activities, while both groups of participants were asked to indicate household pesticide use.
Participants were asked to select the frequency of each activity regarding fresh produce consumption behavior. Pesticide sprayers were asked about frequency of use of personal protective equipment (PPE), which are gloves, boots, long-sleeved shirt, goggles, mask, and hat, and frequency of performing each activity while mixing or spraying pesticides.
Data from 24 h dietary recall was used to calculated daily calories by a nutritionist using the specialized Thai software NutriFact, which belongs to the Research Institute for Health Science, Chiang Mai University.
Information on duration and frequency of occupational activities, daily duration of walking or bicycling to work, and leisure-time activity was gathered. Participants who had done at least 150 min per week of moderate-intensity aerobic activity or 75 min per week of vigorous aerobic activity were classified as adequate physical activity.
The questionnaire was tested through pilot interview and refined by the research team and our community partnership. The local research team was trained by experts from the Research Institute for Health Science, Chiang Mai University. Questionnaires were administered through a face-to face interview by trained interviewers and research assistants from the Department of Community Medicine, Faculty of Medicine, Chiang Mai University.

Anthropometric Measurement
Waist circumference was measured at the midpoint between the lower margin of the lowest rib and the top of the iliac crest. Weight was measured with participants wearing light clothes and without shoes using a digital weighing scale. Height was measured without shoes using a stadiometer. The body mass index (BMI) was calculated as body weight (kg) divide by height squared (m 2 ). Overweight and obesity was defined as having a BMI of 25 to < 30 and 30 and higher, respectively.

Exposure Measurement
First morning urine was collected and placed in zip-lock plastic bags, stored in cold boxes with dry ice, shipped to the laboratory at the Research Institute for Health Science, Chiang Mai University, and refrigerated at −20 • C until extraction within seven days and analyzed afterward.

Outcome Measurement
Venipuncture was done after 8 h of fasting by nurses at village centers and transferred to laboratories at the Faculty of Associated Medical Sciences, Chiang Mai University, and the Maharaj Nakorn Chiang Mai Hospital for analysis. Abbott ARCHITECT c8000 analyzer was used to analyze plasma glucose, cholesterol, high-density lipoprotein (HDL) cholesterol, and triglyceride level by the hexokinase method, enzymatic method, accelerator selective detergent, and glycerol phosphate oxidase, respectively. Insulin level was evaluated by electrochemiluminescence immunoassay and Cobas e411. Insulin resistance was measured by homeostatic model assessment of insulin resistance (HOMA-IR), which was calculated with the following equation: HOMA-IR = (fasting glucose (mg/dL) × fasting insulin (µIU/mL))/405 [34]. Low-density lipoprotein (LDL) cholesterol was calculated using the Friedewald equation [35].

Statistical Analysis
Chi-square test for categorical variables and t-test for continuous variables were used to examine the demographic characteristics of participants. DAP metabolite levels in both pesticide sprayers and nonfarmworkers were analyzed using the Mann-Whitney U test. Agricultural product consumption behavior between pesticide sprayers and nonfarmworkers were compared using chi-square test. Multiple linear regression analysis was performed to examine the association between DAP metabolites and blood chemistry levels as well as DAP metabolites and agricultural work practice in farmworkers. p-value < 0.05 was considered statistically significant.

Ethical Consideration
All participants gave their informed consent for inclusion before participating in the study. The research procedure was approved by the Research Ethics Committee, Faculty of Medicine, Chiang Mai University, Thailand (Certificate Ethical Clearance No. 090/2019, 29 April 2019)

Demographic Information, HOMA-IR, and Glucose Level of Pesticide Sprayers and Nonfarmworkers
The characteristics of participants by occupational status are presented in Table 1. Overall, pesticide sprayers were significantly older (p = 0.03), had a higher proportion of male (p = 0.02), and had lower education (p < 0.01) compared to nonfarmworkers. In addition, the sprayer group showed higher cardiometabolic risk factors, which were family history of diabetes (p < 0.01), smoking (p = 0.03), and excessive calorie (p < 0.01) and carbohydrate (p < 0.01) intake. Moreover, they reported higher hours of organophosphate pesticide exposure (p < 0.01), while no nonfarmworker reported OP exposure.  Table 2 illustrates urinary DAP metabolites in the two groups. There was no significant difference in the proportion of participants who had DAP metabolite levels above and below LOD. In addition, DAP levels of both groups were similar except for DETP, which was higher in pesticide sprayers (p = 0.03).

Assessment of Agricultural Product Consumption Behavior
Assessment of agricultural product consumption behavior is presented in Table 3. A higher proportion of pesticide sprayers reported always consuming organic vegetables (p < 0.01), fruits (p < 0.01), and rice (p < 0.01). Moreover, a greater percentage of sprayers reported that they always have food cooked by themselves or their family members.

Factors Associated with HOMA-IR, Glucose, and Lipid Levels
Factors associated with HOMA-IR, fasting plasma glucose, and lipid levels in sprayers and the general population are shown in Tables 4 and 5. HOMAR-IR had significant correlation with only waist circumference (95% CI = 0.03, 0.09). Participants who reported to have family history of diabetes showed lower glucose level (95% CI = −13.67, −1.87). For lipid panel, high amount of alcohol consumption correlated with high triglyceride (95% CI = 0.03, 0.12). Triglyceride level was also higher in subjects who had excessive calorie intake (95% CI = 30.84, 183.53). Furthermore, female tended to have higher HDL (95% CI = 2.86, 11.77). However, no significant correlation was observed between DAP and HOMAR-IR, glucose, triglyceride, LDL, and HDL cholesterol.   Table 6 demonstrates factors associated with DAP metabolites in pesticide sprayers. For PPE utilization, gloves and boots were not analyzed as all sprayers reported to usually wear them. The association between higher frequency of wearing a mask while handling pesticide and DMAP level was observed (95% CI = −11.10, −0.17). Participants who reported lower frequency of reading pesticide labels (95% CI = −81.47, −14.99) and washing hands after mixing (95% CI = −39.97, −3.35) exhibited significant higher DEAP level.

Discussion
To the best of our knowledge, this is one of the first studies to determine the association between organophosphate exposure and insulin resistance in agricultural and nonagricultural workers. However, we could not demonstrate the correlation between organophosphate exposure and insulin resistance, fasting plasma glucose, triglyceride, HDL, or LDL cholesterol levels.
DAP metabolite levels of pesticide sprayers in this study were lower than previous studies, which could be from several possible causes. First, as different techniques were used to analyze DAP metabolites, LOD of our metabolites was higher than other studies. This was especially the case for DMP, which was 2.5 µg/L in our study compared to 0.26 µg/L in a study in the Netherlands [36]. This also led to the low percentage of DMP detection in our study. Another explanation could be that some farmworkers might not have followed the urine collecting instruction because of communication barriers as half of them were from the hill tribes.
DAP metabolites in the current study were also lower than that reported in a study of pesticide sprayers in Greece [37], which could be due to the lower proportion of PPE use. In the study, the percentage of sprayers who reported wearing gloves, hat, and mask was 35.5, 52, and 50%, respectively, whereas it was 94.4, 94.4, and 88.9%, respectively, in our study (data not shown). Similar to our result, the study pointed out that DETP level in sprayers was significantly higher than nonfarmworkers.
The average diethyl metabolite levels in our study were close to those in chili farmers in Thailand [38], where first morning urine was collected after spraying OPs. This could be the result of a similar DAP analyzing technique. Moreover, chlorpyrifos was commonly used among chili farmers in the study site.
Our study indicated that pesticide sprayers had significantly higher DETP metabolites than nonfarmworkers. This result is related to chlorpyrifos and ethion, the two most popular OPs used by pesticide sprayers, which can metabolize to DETP [39]. However, the levels of the other five metabolites were similar in both groups. This could be explained by the fact that nonfarmworkers had higher dietary OP pesticide exposure compared to pesticide sprayers as sprayers reported significantly higher organic produce consumption. This result correlates with several studies on the general population that have reported an association between DAP metabolites and higher fruit and vegetable intake [36,40,41]. Consumption of fresh produce contaminated with OP residues could be pointed as the main route of OP exposure in the general population, while organic diet consumption could reduce OP metabolites [42,43]. In addition, a study in Chiang Mai, Thailand, reported that the most common OP residue found in fresh produce was chlorpyrifos, which was detected in all kinds of fruit and vegetable samples [44].
Our study did not show an association between DAP metabolites and insulin resistance. This is consistent with a study on the US general population, which reported the lack of association between urinary DAP metabolites and HOMA-IR and plasma glucose [45]. This study result is in contrast with a previous research on pesticide applicators, which indicated the positive correlation between blood malathion concentration and HOMA-IR [32]. This could be explained by the questionable conclusion of this study that the waist circumference and BMI, which are mandatory factors of insulin resistance [46,47], of the farmer group were significantly higher than the control group. In addition, an association between lipid levels and OP exposure was not observed, whereas several studies in animal models have reported that OPs lead to an increase in LDL and triglyceride and a decrease in HDL [48,49]. Moreover, a study on humans showed the correlation between DMPT metabolite and higher HDL, lower LDL, and lower total cholesterol [45]. Thus, further human studies to examine the effect of OPs on lipid metabolism are needed.
Our study showed the inverse correlation between family history of diabetes and fasting plasma glucose level, which contradicts a previous study on the general population [50]. According to the Thai National Health Examination Survey, the proportion of undiagnosed diabetes was 51.2% for men and 41.3% for women [51]. Furthermore, some farmworkers in our study were of the hill tribe ethnic group, in which barriers to accessing healthcare services has been reported [52]. A survey on elderly Thai population pointed out that hill tribes had higher prevalence of diabetes compared to the general Thai population [53]. Therefore, the number of participants who reported family history of diabetes in our study might be lower than reality, especially in farmworkers. Another possible explanation could be greater awareness and risk perception in individuals who have familial risk of diabetes [54]; consequently, they are more likely to adopt healthy behaviors [55].
A reason behind the lack of association between OP exposure and chronic outcomes, namely, insulin resistance, glucose, and lipid levels, in our study could be the limitation of DAP metabolites. Although first morning urine has been pointed out as the best predictor of daily DAP metabolite concentration [56], various studies have reported that single urine analysis may not indicate chronic exposure [57,58]. Another reason could be the high exposure in the general population as over 90% of our nonfarmworkers had detectable DAP metabolites. Compared to studies among the general population in Israel and the US, our study showed higher diethyl alkyl metabolite levels [40,59]. Our participant enrollment process, in which the frequency of OP exposure in sprayers was not specified, could have affected the results. The frequency of pesticide spraying has been identified as being correlated with adverse health effects [60]. However, we analyzed the association between cumulative OP exposure and insulin resistance, but an association was not observed. In addition, there is evidence that exposure to carbamate pesticides alters glucose homeostasis in animals [61]. The current study did not assess carbamate levels because the sprayers reported that the most recent carbamate pesticide spraying was longer than a month before enrollment. Therefore, the metabolites could not be captured. This could probably be a cause for our negative correlation.
Our study reported that wearing a mask while spraying pesticides was associated with lower dimethyl metabolites. This result is consistent with several studies that have indicated a correlation between PPE utilization and lower DAP levels [37,62]. In addition, a positive correlation was found between diethyl metabolite level and the behavior of pesticide sprayers, such as reading pesticide labels before mixing and washing hands after mixing pesticides. This finding corresponds with a previous study that pointed out the association between the working practices of farmers and pesticide exposure levels [63].
There are several limitations of this study. Due to the complicated data collection process, participants were not randomly selected. Therefore, healthy people tended to be recruited because of awareness about their health. Our sample size was smaller than expected because of the difficulty of contacting farmworkers who lived in mountain villages. Another point to note is that the ability to represent chronic OP exposure and variation of urinary DAP metabolites over time might have been limited due to single morning urine collection. As our DAP analytical method had lower sensitivity compared to other studies, detectable metabolites in our study might not represent the real values. In addition, farmworkers might overreport PPE utilization and appropriate work practices as they were not observed while working.

Conclusions
This study did not observe an association between OP pesticide exposure and insulin resistance in pesticide sprayers and nonfarmworkers. However, the results are inconclusive due to limitation in detecting chronic exposure from single urinary DAP metabolites. In addition, the general population had similar OP pesticide exposure level, which was associated with consumption of fruit and vegetable contaminated with OP. The results therefore need further longitudinal studies with adequate sample size and multiple urine samples. Although there was no evidence of a correlation between measurable DAP metabolites and adverse health effect, the DAP levels reflect high exposure in our nonfarmworker group. Measures to control pesticide exposure should therefore be taken.