Article Text

Download PDFPDF

Original article
Paraquat application and respiratory health effects among South Korean farmers
  1. Eun Shil Cha1,
  2. Yun Keun Lee2,
  3. Eun Kyeong Moon1,
  4. Yong Bae Kim3,
  5. Yong-Jin Lee3,
  6. Woo Chul Jeong4,
  7. Eun Young Cho4,
  8. Ik Jin Lee4,
  9. Jongil Hur4,
  10. Mina Ha5,
  11. Won Jin Lee1
  1. 1Department of Preventive Medicine, College of Medicine, Korea University, Seoul, South Korea
  2. 2Wonjin Institute for Occupational and Environmental Health, Seoul, South Korea
  3. 3Department of Preventive Medicine, College of Medicine, Soonchunhyang University, Cheonan, South Korea
  4. 4Environmental Health Center, Taean Public Health and Medical Center, Taean gun, South Korea
  5. 5Department of Preventive Medicine, Dankook University, Cheonan, South Korea
  1. Correspondence to Dr Won Jin Lee, Department of Preventive Medicine, College of Medicine, Korea University, Seoul 136-705, South Korea; leewj{at}korea.ac.kr

Abstract

Objectives Paraquat is commonly used worldwide as major herbicide. The objective of this study was to investigate the association among farmers between occupational paraquat exposure and respiratory health effects.

Methods A cross-sectional survey of health effects related to an oil spill was conducted in South Korea from 2008 to 2009. For this analysis, a total of 2882 full-time farmers were selected from the overall sample. Data collection included an interviewer-administered questionnaire and spirometry testing. Logistic regression analysis and linear regression analysis were performed to evaluate the relationship between paraquat exposure and respiratory health outcomes after adjustment for potential confounders.

Results The risks of self-reported physician-diagnosed asthma, chronic obstructive pulmonary disease and allergic rhinitis were non-significantly increased among paraquat-applying farmers compared with non-paraquat-applying farmers. Although the results of a pulmonary function test fell within normal limits, a decline in forced vital capacity and forced expiratory volume in one second was apparent among paraquat-applying farmers compared with non-paraquat-applying farmers. Forced vital capacity (β=−5.20, p<0.001) and forced expiratory volume in one second (β=−1.89, p=0.010) significantly decreased with each unit increase in years of paraquat application. Paraquat-applying farmers showed a significant exposure–response relationship between restrictive ventilatory defects and paraquat application years (p trend=0.015) or lifetime days of application (p trend=0.007).

Conclusions Our findings suggest a possible association between paraquat application and adverse respiratory health effects among farmers.

  • Agricultural workers
  • occupational exposure
  • paraquat
  • pulmonary function test
  • restrictive lung disease
  • ergonomics
  • hygiene/occupational hygiene
  • back disorders
  • epidemiology
  • public health
  • cancer
  • occupational health practice
  • preventive medicine

Statistics from Altmetric.com

Request Permissions

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.

What this paper adds

  • Paraquat is acutely toxic and causes a great number of health effects, especially in the case of suicidal ingestion, but the occupational hazards presented by chronic paraquat exposure have been less well characterised.

  • A decline in forced vital capacity and forced expiratory volume in one second was apparent among paraquat-applying farmers when compared with non-paraquat-applying farmers.

  • Paraquat-applying farmers showed a significant exposure–response relationship between restrictive ventilatory defects and paraquat application years or lifetime days of application.

  • Intensive intervention efforts to reduce occupational paraquat exposure are important for farmers considering the widespread use of paraquat around the globe.

Introduction

Paraquat (1,1'-dimethyl-4,4' bipyridinium dichloride) is a non-selective herbicide that has been applied in more than 100 developed and developing countries throughout the world.1 It is extensively used for weed control both on plantations and by small-scale farmers in hundreds of different crops, including corn, rice, orchards, soybeans, vegetables and many other crops. Paraquat has been ranked as one of the most commonly used pesticides and it is the main causative agent for acute poisoning in South Korea.2

Paraquat is acutely toxic and causes a great number of health effects. Over the preceding decades, numerous fatalities from paraquat exposure have been recorded, mainly as a result of accidental or voluntary ingestion. The major cause of death from paraquat poisoning is respiratory failure due to oxidative damage to the alveolar epithelium, with subsequent pulmonary fibrosis.3 Pulmonary function abnormalities, especially decreased diffusing capacity, were reported among survivors of acute paraquat poisoning in Japan.4

However, the occupational hazards presented by chronic paraquat exposure have been less well characterised, and a limited number of epidemiological studies on respiratory health effects among paraquat applicators have been performed.5–11 Chronic exposure to paraquat among farmers has been reported in certain studies to show an association with respiratory symptoms, such as wheeze,5 shortness of breath and chronic cough,6 7 subclinical gas exchange abnormalities,7–9 but not in others.10 11 The limitations of earlier epidemiological studies include small sample sizes and exposure to multiple pesticides.

This study, therefore, aimed to investigate the association of occupational paraquat exposure with respiratory health effects and lung function among South Korean farmers.

Methods

Study population

Data for this study were drawn from a cross-sectional survey of residents of Taean county in South Korea conducted in 2008–2009 in connection to the Hebei Spirit oil spill that occurred there in December 2007.12 Among the 2008 county population of 63 401, a total of 9246 residents living at or near the site of the oil spill enrolled in a voluntary health examination program study of potential health effects from the accident by completing a questionnaire and physical examination. A total of 2882 full-time farmers who apply pesticides (2508 paraquat-applying farmers and 374 non-paraquat-applying farmers) were selected for this analysis. They were self-employed farmers and approximately representative of full-time farmers in South Korea. All participants provided written informed consent, and the study was approved by the institutional review board of Dankook University Hospital.

Data collection

Data collection included an interviewer-administered questionnaire and spirometry testing. The questionnaire assessed basic demographic characteristics, pesticide use, smoking status, alcohol consumption, level of education, monthly income and the presence of any respiratory symptoms and diseases. Chronic obstructive pulmonary diseases (COPD) and asthma were defined in terms of the subject having ever been diagnosed with the diseases by a physician. Wheeze was defined as the occurrence over the past 12 months of any whistling sounds emanating from the chest, and exertional wheeze was defined as the occurrence over the same period of whistling sounds in the chest during or after exercise. Shortness of breath was defined as the occurrence over the past 12 months of waking in the night accompanied by a whistling sound.

Spirometry was performed by trained nurses blinded to paraquat exposure status following American Thoracic Society guidelines.13 The best forced vital capacity (FVC) and forced expiratory volume in one second (FEV1) among a minimum of three acceptable forced expirations were used as outcomes. Participants with an FEV1/FVC ratio of >0.7 and with an FVC of <80% of the predicted value were classified as having a restrictive ventilatory defect, and those with an FEV1/FVC ratio of <0.7 with an FEV1 of <80% of the predicted value were classified as having an obstructive ventilatory defect.14 Of the total 2882 full-time farmers applying pesticides, 97.3% (n=2805) showed an acceptable value on spirometry tests.

Exposure assessment

The interview questionnaire collected exposure data on 34 individual pesticides selected from information on pesticide usage in the study area, use of personal protective equipment and pesticide application method derived from a prior interview in a focus group. For this study, four paraquat exposure variables were determined: ever/never application, number of years of application, number of annual days of application and lifetime days of application. To evaluate cumulative paraquat exposure, the number of years of application among paraquat-applying farmers was categorised into tertiles (ie, <19, 19–30, >30 years). Lifetime application days were calculated based on the days and years of paraquat application (ie, application years × application days per year) and also categorised into tertiles (ie, <61, 61–150, >150 days).

Statistical analysis

Logistic regression analysis was used to evaluate the relationship between paraquat exposure and respiratory health effects after adjusting for potential confounders such as age (<50, 50–59, 60–69, ≥70 years), gender, smoking status (never, ≤25.5 pack-years former smoker, >25.5 pack-years former smoker, ≤28 pack-years current smoker, >28 pack-years current smoker), alcohol consumption (yes, no), distance from the oil spill site (>40, 20–40, <20 km) and level of education (<elementary, elementary, ≥middle). These variables were selected based on the difference between paraquat-applying farmers and non-paraquat-applying farmers and existing knowledge of potential factors that may relate with respiratory health outcomes. The model for wheeze was also examined with asthma (yes, no) additionally added as a covariate. For women, smoking status was classified into two categories (never-/former smoker, current smoker) since the majority of the women in our population have never been smokers. Because of potential concomitant exposure to additional pesticides, we included in the regression models the three pesticides (ie, carbofuran, imidacloprid, isoprothiolane) whose use was most highly correlated with paraquat. For the consideration of protective equipment and application methods, we included personal protective equipment variables indicated wearing or not wearing protective jacket, protective pants, goggles, gloves, boots and mask. Application method variables as covariates in the models were indicated as yes or no for each method including backpack spray, hand spray, speed spray, banded/in-furrow and others. The results with these variables included remained similar to those without adjustment for personal protective equipment and application method; therefore, the simpler results are presented in this study. To investigate exposure–response associations, we evaluated respiratory health effects by cumulative paraquat exposure indices. Trend tests were performed for categorical variables by assigning scores to categories using the median value of each category of the exposure indices and treating the scores as continuous variables in the regression analysis.

The results of spirometric measurements were analysed by applying a multiple regression analysis of FVC, FEV1 and FEV1/FVC ratio as dependent variables after adjusting for potential confounding factors such as age (continuous), gender, smoking status (never, ≤25.5 pack-years former smoker, >25.5 pack-years former smoker, ≤28 pack-years current smoker, >28 pack-years current smoker), alcohol consumption (yes, no), distance from the oil spill site (>40, 20–40, <20 km), level of education (<elementary, elementary, ≥middle) and the cumulative exposure (no exposure, low, high) of three highly correlated pesticides (carbofuran, imidacloprid, isoprothiolane) with paraquat, and height (continuous) as covariates. In regression analyses, the exposure indices of lifetime application days were log transformed to improve the fit of the regression models. In order to investigate the consistency of the associations, further analyses were conducted stratifying by smoking status, gender and distance from the oil spill site. We also have investigated the interaction between age and paraquat application years and smoking and paraquat application years. All tests of statistical significance were two-sided. Statistical analyses were performed using Stata V.11.0 (StataCorp.).

Results

Of the total 2882 full-time farmers applying pesticides, 2508 (87.0%) farmers reported applying paraquat; the majority of them were men and older than 60 years (table 1). The proportions of smokers and of consumers of alcohol among paraquat-applying farmers were higher than those among non-paraquat-applying farmers, but that of income level was similar. Most of the farmers in this study engaged in dry-field farming, followed by rice farming, and applied paraquat by using a backpack sprayer and primarily wore rubber boots and gloves as personal protective equipment. The median application years and lifetime application days were 25 years and 90 days among paraquat-applying farmers, respectively.

Table 1

Selected characteristics of subjects by paraquat application status among South Korean farmers in 2008

The risks of self-reported physician-diagnosed asthma, COPD and allergic rhinitis were non-significantly increased among paraquat-applying farmers compared with non-paraquat-applying farmers after adjusting for potential confounders (table 2). The results were generally similar for both male and female farmers, although those of asthma and COPD were unstable due to the small number of cases. These associations also varied little after stratifying by smoking status or distance from the oil spill site (data not shown).

Table 2

ORs and 95% CIs for selected respiratory health effects by paraquat application status and gender among South Korean farmers in 2008

Non-significant declines in FVC and FEV1 were observed among paraquat-applying farmers after adjusting for potential confounders (table 3). For male current smokers, however, FVC and FEV1 were significantly lower among paraquat-applying farmers than non-paraquat-applying farmers. Although they were not significant, FVC and FEV1 were also lower among former- and never-smoker paraquat-applying farmers. The means of percentage of predicted values were within the normal range for all spirometry measures in both groups.

Table 3

Adjusted mean* and SEs of pulmonary function measures by paraquat application status and gender among South Korean farmers in 2008

A statistically significant decrease in FVC was shown with an increase in paraquat application years (β=−5.20, p<0.001) and lifetime application days (β=−22.19, p=0.009) (table 4). FEV1 also decreased with an increase in paraquat application years (β=−1.89, p=0.010). These findings were more pronounced among male farmers than among female farmers.

Table 4

Multiple regression analyses on paraquat exposure indices and pulmonary function measures by gender among South Korean farmers in 2008

Significant exposure–response relationships between the restrictive ventilatory defect with paraquat application years (p trend=0.015) and lifetime paraquat application days (p trend=0.007) were observed (table 5). For obstructive ventilatory defect, however, there was no significant association with paraquat application years or days. The results did not differ fundamentally from the stratified analysis by smoking status or distance from the oil spill site (data not shown).

Table 5

ORs and 95% CIs for abnormal spirometry by paraquat exposure indices among South Korean farmers in 2008

Discussion

We examined the association between occupational exposure to paraquat and respiratory health effects using a large cross-sectional data set. Significant declines in FVC and FEV1 with paraquat application years and an exposure–response association with restrictive ventilatory defects were observed. Although the observed effects were small and residual confounding could explain these associations, the results remained consistent when the analyses were repeated stratified by smoking status and distance from the oil spill site.

Our findings of respiratory defects among paraquat-exposed farmers are consistent with previous studies6–9 which showed the restrictive type of lung disorder. One study from Nicaragua found a dose–response relationship between exertional dyspnoea and exposure to paraquat among 286 agricultural workers.6 A South African study reported a significant association between paraquat exposure and arterial oxygen desaturation during exercise among 126 male agricultural workers.8 More recent studies have also reported that with increased cumulative paraquat exposure, the ventilator equivalent for carbon dioxide was higher among 338 Costa Rican farmers,7 and a decrease was shown in the diffusing capacity of the lungs among 114 Spanish sprayers.9

However, previous studies did not invariably show consistent results with parameters of the type observed. The differences of characteristics of study subjects and methods of exposure assessment among studies may cause differences in estimating the risk. A possible reason for the significant association observed with spirometry in our study may be related with the long-term exposure to paraquat of our subjects (median of 25 years) compared with the exposure of other study populations of 8.5,7 11.58 and 12.0 years.11 It may also be explained as being due to an older population (mean age of 64.4 years) and higher level of smoking (median of 28 pack-years) than those populations.

Restrictive lung defects from other studies and from our findings may be comparable to biological mechanisms. The primary biochemical mechanism of paraquat toxicity is currently believed to occur through the cyclic redox reactions of paraquat with reducing equivalents and oxygen in the cell, causing toxicity and fibrosis in lung tissue.3 Paraquat is also known to increase the connective tissue growth factor and collagen expression by activating angiotensin signalling pathways in human fibroblasts.15 Pulmonary fibrosis induced by paraquat increases oxygen consumption, especially during exercise, and reduces aerobic performance in rats.16 Previous clinical findings among patients with paraquat poisoning have demonstrated a decrease in carbon monoxide diffusing capacity and restriction manifestation of FVC and FEV1 after intoxication.4

We also found a non-significant increase in the prevalence of self-reported asthma, COPD and allergic rhinitis among paraquat-applying farmers compared to non-paraquat-applying farmers. In the Agricultural Health Study conducted in the USA, paraquat was also significantly associated with wheeze5 and chronic bronchitis.17 Paraquat was also reported to associate the allergic rhinitis among farmers in Greece18 and to exacerbate the allergic reaction in animal study.19 Some asthma patients were reported to have restrictive impairment20 as consistent with our findings. In addition, subjects reported having respiratory problems would likely limit those occupational paraquat exposures that exacerbate respiratory symptoms; therefore, this would lead to an underestimation of the relations of interest.

In this study, men were more likely to experience lung function decrease associated with exposure than were female workers. This may point to either a difference in workplace exposures or a biological gender difference.21 In South Korea, most female farmers have performed as assistants helping the main applicators. In our study, female farmers reported applying paraquat fewer hours per day than male farmers (mean hours paraquat applied per day: 2.5 h for male farmers, 1.9 h for female farmers) even if they had applied pesticide for the same number of years or days. In addition, since most of our female subjects were non-smokers, the negative effect between exposure to paraquat and lung function aggravated by smoking may have been lessened.

One potential limitation of paraquat exposure misclassification should be considered. Since skin exposure is known to be the main route of absorption of paraquat into the body by farmers and using a backpack is more likely to lead to high levels of exposure to paraquat,22 the use of personal protective equipment and method of application may modify the associations. However, the majority of subjects in this study reported not wearing effective protection against skin absorption, such as protective clothing, and performing application using a backpack sprayer. Thus, they may have been exposed to similar level. We also adjusted for use of personal protective equipment and application method and found similar results. Although paraquat exposure was reported to be associated with the type of crop among Costa Rican farm workers,23 the majority of farming type were limited to dry-field and rice farming and a combination of the two was common. Therefore, we were unable to examine association by farming type in this study.

Another possible limitation is that exposure to additional hazardous agents may be related with respiratory health effects. Exposure to organophosphate pesticides24 25 and neonicotinoid pesticides9 has been shown to be associated with restrictive lung dysfunction and the development of respiratory symptoms. To address potential confounding from multiple pesticide exposures, we constructed models that included those three pesticides which are most highly correlated with the use of paraquat and found similar results. Since chemicals from the oil spill may also relate with respiratory effects,12 we also conducted analyses stratified by distance from the oil spill site and found similar results in each area.

Additionally, the self-reported data on medical diagnosis and symptoms used to assess respiratory health status may have resulted in a misclassification of outcome. However, for respiratory symptoms, self-administered questionnaires have been shown to be reliable and reproducible, particularly with wheeze, and diagnosis of asthma.26 Most other previous studies of farming populations have been successfully conducted in a similar manner. Spirometry also relies on the cooperation of the subject of investigation, as much as on the investigator's skills and approach. Nevertheless, spirometry plays an important role in the diagnosis and management of respiratory diseases.27 In addition, the sensitivity and specificity of spirometry for pulmonary function have shown high accuracy among the population aged 50–69 years in South Korea,28 the majority age group in this study. The sensitivity and specificity for restrictive lung function were reported as 72.4% and 87.1% among US adults, respectively.29 However, there may be possible non-differential misclassification of disease status, which leads to towards the null.

In conclusion, our findings show an association between occupational paraquat exposure and adverse respiratory health effects, especially restrictive pulmonary defects, among paraquat-applying farmers. Intensive intervention efforts to reduce paraquat exposure are important given the consistency with the toxicology and epidemiology literature and considering the widespread use of paraquat in the world.

References

Footnotes

  • Funding This work was carried out with the support of the Cooperative Research Program for Agriculture Science and Technology Development (Project No. PJ007455), Rural Development Administration and Ministry of Environment, Republic of Korea.

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval Ethics approval was provided by the institutional review board of Dankook University Hospital.

  • Provenance and peer review Not commissioned; externally peer reviewed.