Exposure to pesticides in present-day use, diabetes mellitus and lung function impairment

Abstract

Background and aims

Epidemiological studies have associated exposure to pesticides with increased risk of diabetes mellitus and decreased lung function, but many of the previous studies are cross-sectional and with inadequate confounder control. We aimed to investigate the correlations between exposure to cholinesterase inhibitor insecticides (organophosphates and carbamates), blood sugar levels and lung function while accounting for important confounders.

Methods

From October 2016 to April 2017, we conducted a nested cross-sectional study as part of the Community Based Intervention for Management of Diabetes in Nepal (COBIN-D) trial. We measured fasting plasma glucose among 2,310 persons from the general population of a semi-urban area of Nepal, and collected questionnaire-based subjective information on pesticide exposure and confounders. Odds of diabetes mellitus were analyzed in logistic regression models.

Next, we conducted the "Pesticide Exposure, Asthma and Diabetes in Uganda" (PEXADU) study: a short-term follow-up study among 364 smallholder farmers in Uganda. At baseline in September-October 2018 and at follow-up in November-December 2018 and January-February 2019, we measured each participant's glycated hemoglobin A (HbA_1c), a measure of average blood sugar level in the last 8-12 weeks. We quantified exposure to cholinesterase inhibitor insecticides using red blood cell acetylcholinesterase activity normalized by hemoglobin concentration (AChE/Hb). Lung function was quantified by spirometry, and spirometric indices converted to Z-scores using the Global Lung Function Initiative equations. Information on confounders was collected by questionnaire. We analyzed data in linear mixed effect models accounting for family relationships and repeated measurements.

Results and discussion

In the COBIN-D population, the risk of diabetes mellitus was lower among subjects who reported that they had ever used pesticides – adjusted OR 0.68 [0.52; 0.90], but exposure levels were relatively low, and there were no clear exposure-response relationships. Large demographic differences between the exposed and non-exposed groups mean that residual confounding is likely.

In the PEXADU study, low acetylcholinesterase activity (indicating exposure to organophosphate and carbamate insecticides) was associated with decreased HbA_1c. Compared to reference subjects with AChE/Hb 25.8 U/g (50^th percentile), subjects with AChE/Hb 24.3 U/g (35^th percentile) had HbA_1c that was 0.74 [0.17; 1.31] mmol/mol lower in the adjusted analysis, and subjects with ACHE/Hb 27.1 U/g (65^th percentile) had HbA_1c 0.63 [0.12; 1.14] mmol/mol higher than the reference. Low acetylcholinesterase activity was also associated with decreased pulmonary function. In the adjusted analysis, Z-score for FEV₁ (forced expiratory volume in 1 second) was 0.045 [0.003; 0.087] lower for subjects with AChE/Hb 24.5 U/g (35^th percentile) than for reference subjects with AChE/Hb 25.9 U/g (50^th percentile), and subjects with AChE/Hb 27.3 U/g (65^th percentile) had FEV₁ Z-score 0.043 [-.002; 0.087] higher than the reference. Demographic variables were similar between subjects with AChE/Hb below and above the median, making residual confounding less likely.

Conclusion

Our results do not support a causal link between exposure to cholinesterase inhibitor insecticides and diabetes mellitus. The association between low acetylcholinesterase activity and decreased HbA_1c in the PEXADU study may be due to reverse causality. We demonstrated an association between low acetylcholinesterase activity and decreased lung function. Evidence from human exposure studies with less toxic cholinesterase inhibitors indicate that the association between acetylcholinesterase and pulmonary function may represent a causal effect of exposure to cholinesterase inhibitor insecticides, and efforts to limit exposure should be strengthened.

Note

The thesis has been assessed to be acceptable for oral public defense. When this document was published on-line (April 29, 2020), the oral public defense had not yet taken place, and the PhD degree had not yet been awarded.