The Journal of Toxicological Sciences
Online ISSN : 1880-3989
Print ISSN : 0388-1350
ISSN-L : 0388-1350
Original Article
Pyrethroid pesticide exposure during early pregnancy and birth outcomes in southwest China: a birth cohort study
Qinghua XuBaosheng ZhuXudong DongSuyun LiXiaoxiao SongXia XiaoChao ZhangYan LvXiong ZhangYan Li
Author information
Supplementary material

2020 Volume 45 Issue 5 Pages 281-291


Despite the developmental toxicity reported in animals, few epidemiologic studies have investigated the potential effects of prenatal exposure to pyrethroid pesticides (PYRs) on fetal growth. A birth cohort study was conducted to examine the association between prenatal exposure to PYRs and birth outcomes, and a nested case-control study was conducted in this cohort to evaluate the effects of PYR on congenital defects. The assessment of PYR exposure was based on self-reported household pesticide use and urinary PYR metabolite levels. We found that pregnant women in this region were ubiquitously exposed to low-level PYRs, although few reported household pesticide use. Women who often ate bananas or cantaloupes had a higher level of urinary 3-(2,2-dibromovinyl)-2,2-dimethylcyclopropane-1-carboxylic acid (DBCA), and the number of fruit types consumed by pregnant women was positively related to the concentrations of 3-phenoxybenzoic acid (3PBA) and total PYR metabolites (P < 0.05). Increased urinary 4-fluoro-3-phenoxybenzoic acid (4F3PBA), DBCA, and total PYR metabolites were associated with increased birth weight, length, and gestational age, and with decreased risk of small for gestational age (SGA) and/or premature birth. However, maternal household pesticides use was related to congenital anomalies. Thus, although prenatal exposure to low-dose PYRs promoted the fetal growth, the beneficial effects of fruit intake may outweigh the adverse effects of pesticide exposure. This study provided us an insight into the biological mechanisms for the effect of prenatal PYR exposure on fetal development, and suggested that further investigations in a larger study population with low-dose PYR exposure is needed.


Pyrethroid pesticides (PYRs) are widely used in farms, orchards, and homes due to their great insecticidal activity and relatively low toxicity. These are gradually replacing more toxic organochlorine (OC), organophosphorus (OP), and carbamate insecticides, since the strengthening of pesticide regulations (Ding et al., 2015; Viel et al., 2015; Horton et al., 2011). Studies in animals showed that PYRs are developmental toxic agents which affect embryonic growth and development by inducing oxidative stress (Shi et al., 2011; Khatab et al., 2016; Guo et al., 2019) or altering the expression of genes involved in signaling transduction (Wang et al., 2017; Chueh et al., 2017). And exposure to PYRs during embryogenesis may result in embryo loss (Khatab et al., 2016; Ratnasooriya et al., 2003), fetal intrauterine growth restriction (IUGR) (Wang et al., 2017; Chueh et al., 2017), and embryo deformities (Uggini et al., 2012; Toumi et al., 2013).

The major pathways of exposure to PYRs in the general population are dietary intake, and the use of household insecticides or mosquito repellents (Viel et al., 2015; Chiu et al., 2018). During pregnancy, women generally increase their awareness of protection for themselves and reduce their exposure to household pesticides and insect repellents. However, dietary intake is imperceptible. Pyrethroid residues have been detected in various kinds of vegetables and fruits (Chauhan et al., 2014; Qu et al., 2017; Yu and Yang, 2017; Torbati et al., 2018). Results from epidemiology studies suggest that pregnant women in Chinese and French are ubiquitously exposed to low-level PYRs (Dereumeaux et al., 2018; Qi et al., 2012). As pyrethroids are rapidly metabolized in the human body, the metabolites are excreted in urine within a few days after exposure. Urinary concentrations of non-specific PYR metabolites are typically used as indicators of recent exposure (Le Grand et al., 2012; Viel et al., 2015).

So far, limited information is available concerning the developmental toxicity of PYRs to the fetus in early human life. Several studies have focused on the associations between PYR exposure during pregnancy and birth outcomes, but the results are less consistent (Ding et al., 2015; Dalsager et al., 2018; Berkowitz et al., 2004; Zhang et al., 2014). A recent review cited 8 epidemiology studies on birth outcomes and PYRs exposure. However, the exposure metrics among general population were primarily based on a single sample analysis at third trimester or delivery (Burns and Pastoor, 2018). As the first trimester is a critical period of fetal organogenesis, it is necessary to evaluate the association between pesticide exposure during this time and fetal development.

We investigated the associations between maternal exposure to PYRs and fetal growth in a cohort study. The biomarker tests and questionnaires were combined to assess the level of pregnant women exposure to PYRs. A single biomarker test does not reflect previous transient exposures since PYRs are rapidly metabolized in the human body. We regarded the exposure of household insecticides as exposure to pyrethroid pesticides because PYRs are currently the most widely used household insecticides since the standards for safety application of public health insecticides-pyrethroid insecticides (GB/T 27779-2011) issued by the Ministry of Health of the People’s Republic of China (Pesticide registration data on the China Pesticide Information Network).

We hypothesized that prenatal exposure to PYRs, as indicated by self-reported household insecticide use and/or analysis of urine samples collected during early pregnancy, was associated with reduced birth length, birth weight, gestational duration, or fetal congenital defect.


Study design and subject recruitment

This study was a prospective birth cohort study regarding the role of pesticide exposure during early pregnancy on fetal development. Pregnant women were recruited when they took an early pregnancy (all of the gestational ages were 11-14 weeks except for one: 40 days) prenatal care check at a general hospital in Kunming City from February 2017 to March 2017. The inclusion criteria were healthy pregnant women with a live singleton fetus, no known factors that cause miscarriage or fetal malformation such as trauma, uterine malformation, and abdominal CT or x-ray examination (within 6 months before enrollment for mother, and within 1 month before conception for father). Women were informed of the objectives of the study and provided signed written consent before they voluntarily participated in the project. The protocols used in this study were approved by the Ethics Committees of Kunming Medical University.

A total of 512 pregnant women were recruited and donated available urine samples (the percent participation of more than 80%). All of the recruited women were followed up to the fetus birth or the end of this pregnancy. Of these, some women were excluded because of not completing the questionnaires (n = 7), lost to follow-up (n = 28), twin pregnancies (n = 29), and miscarriage caused by trauma (n = 1). Thus, 447 women were analyzed in the final sample including women with live births (n = 433), stillbirth/fetal death (n = 4), spontaneous abortions (n = 2), fetuses with congenital malformation (n = 4), and chromosomal abnormalities (n = 4).

To evaluate the effects of prenatal exposure to PYR on congenital anomalies (i.e. congenital malformations and chromosomal abnormalities), a nested case-control study (Ernster, 1994) was conducted in this cohort. For each case, five controls were selected randomly from the cohort who were matched to case on folic acid supplement (Chevrier et al., 2011), adverse pregnancy history, age, premature rupture of membranes (PROM), and gestational hypertension/(pre-)eclampsia.

Data on birth outcomes

Fetal growth indicators such as birth weight and birth length were obtained from the subject’s medical records. Gestational age was estimated by the obstetrician or midwives based on the date of the last menstrual period (LMP) or the ultrasound results examined during early pregnancy (Ding et al., 2015). Premature delivery was defined by a gestational age at birth of less than 37 completed weeks. Low birth weight (LBW) was defined by a birth weight below 2500 g, and macrosomia was defined by a birth weight no less than 4000 g. Small for gestational age (SGA) was defined as a birth weight less than 10% of the average birth weight for their gestational age (Information on the Cooperative Group of Newborn Physical Development Research in 15 Cities in China, 1989).

Non-live birth fetuses were also included in this study. Congenital malformations in intrauterine fetus were diagnosed by ultrasonographers, and fetal chromosome aberrations were diagnosed by professionals in the hospital’s genetic diagnosis center.

PYR exposure assessments

PYR exposure in pregnant women was assessed by self-reported household insecticide use (within 6 months before enrollment), and the detection of PYR metabolites in urine. Maternal interviews were conducted by researchers based on a structural questionnaire when women visited the hospital for their first-trimester prenatal care. The questionnaire included sociodemographic characteristics, lifestyle, history of pregnancy, and exposure to household pesticides. Maternal urine samples were collected when women were fasting for blood collection (to screen Down syndrome in early pregnancy), and stored at –80°C until analysis. The urine sample was thawed, mixed thoroughly, and centrifuged. An aliquot of a 1.0 mL supernatant sample was extracted with 1.0 mL ethyl acetate, and the extraction process was conducted three times. The organic phase was gathered and evaporated to dryness using a rotary evaporator (ZLNS-1, Beijing, China).

The non-specific PYR metabolites 3-phenoxybenzoic acid (3PBA), 4-fluoro-3-phenoxybenzoic acid (4F3PBA), and 3-(2,2-dibromovinyl)-2,2-dimethylcyclopropane-1-carboxylic acid (DBCA) were detected using a sensitive ultra-performance liquid chromatography system (UPLC) (Ekspert ultra LC-100XL) coupled with a tandem mass spectrometry detector (MS/MS) (3200 Q Trap, ABSciex, Framingham, MA, USA). The assay method for quantification of the PYR metabolites was derived from a previous method and modified (Le Grand et al., 2012). Reference standards for the metabolites were purchased from Sigma Aldrich (St. Louis, MO, USA), and Dr. Ehrenstorfer GmbH (Augsburg, Germany). The limit of detection (LOD) for all analytes was defined as the lowest concentration of PYR metabolites giving a response of signal-to-noise ratio at least three times. The LOD was 0.02 ng/mL for 3PBA and 4F3PBA, and 0.09 ng/mL for DBCA. The accuracy and precision of the method were assessed by repeatedly analyzing fortified urine samples at three concentration levels, i.e. quality control urine. The relative standard deviations (RSDs) for the metabolites were less than 15%. Metabolite concentrations below the LOD were assigned a value of LOD divided by √2 (Lewis et al., 2014). Details of quality control and urine analysis approach are described in previous articles (Qi et al., 2012; Le Grand et al., 2012).

To correct for variable dilutions in the urine samples, the PYR metabolite concentrations were adjusted based on creatinine levels. The urinary creatinine concentrations were determined using a colorimetric method on a visible spectrometer at 510 nm wavelength (MODEL 722, Xianke Spectrophotometer Co. Ltd, Shanghai, China).

Potential covariates

Factors known to or potentially influence fetal development were considered as covariates or confounding factors. Covariates were evaluated through self-reporting on the questionnaire including maternal age, education, body mass index (BMI), folic acid supplementation, parity, and adverse pregnancy history (including fetal malformations or chromosomal abnormalities, stillbirths, and spontaneous abortion). The mother’s environmental exposure factors (within 6 months before enrollment) included passive smoking, alcohol use, noise, paint, antibiotics use, plastic containers use (to hold hot food or heat food), and lampblack (from the kitchen or automobile exhaust). In addition, the father’s risk factors (within 1 month before conception) were also taken into account, including alcohol use, active smoking, exposed to paint or lampblack.

The mother’s pregnancy-associated diseases such as gestational diabetes mellitus (GDM), PROM, and hypertensive disorder complicating pregnancy (including gestational hypertension, pre-eclampsia, eclampsia) were obtained from hospital records.

Statistical analysis

SPSS version 17.0 was used to perform the statistical analysis. Missing values (all < 5.0%) were replaced by the median of the respective data. Skewed values of metabolite concentrations were log-transformed for statistical analysis. Correlations between urinary metabolite levels were calculated by Spearman’s correlation tests, and associations between urinary metabolite concentrations and maternal characteristics were estimated by the independent-samples t test. Multivariate linear regression was used to estimate the associations between the maternal or paternal factors and birth outcomes, including birth weight, birth length, and gestation age. Initially, maternal factors such as age (continuous), pregnancy BMI (continuous), and adverse pregnancy history (no, yes) were forced into all models. Then, GDM (no, yes), PROM (no, yes), and hypertensive disorder complicating pregnancy (no, yes) were also induced into all models as they are risk factors that alter fetal growth (Seyom et al., 2015; Vääräsmäki et al., 2000; Pristauz et al., 2009). The other covariates such as place of residence (urban, suburban & rural), ethnicity (Han Chinese, ethnic minority), passive smoking (no, yes), and education (≤ middle school, ≥ college) etc., were all included in the models, and selected with a stepwise strategy, enter at P < 0.20 (Glorennec et al., 2017). The selected covariates were then introduced into the models to analyze the associations between the urinary PYR metabolite levels and birth outcomes. The confounders were selected using the same strategy in multiple logistic regression models.

Chi-square test and Cox regression analysis were used to determine the risk factors that may predispose a fetus to congenital malformations or chromosomal abnormalities.


Characteristics of the population

The characteristics of the study subjects are presented in Table 1 (variables selected or forced into the multivariate linear regression models), and Supplementary Table 1 (variables excluded in the multiple linear regression models). The average age of the women was 30.5 years, the majority of them were of Han ethnicity, 82.6% lived in urban areas, and most had received some college education or higher. At enrollment, 61.7% of the women were multiparous, and two-thirds had normal BMI during early pregnancy. Among them, 62.0% reported exposure to secondhand smoke, 6.9% had drunk some alcohol, 6.9% had been exposed to perceived noises, 12.1% used to eat fried foods, a few (n = 11, 2.5%) worked in a restaurant, and a few (n = 17, 3.8%) had used household pesticides, recently.

Table 1. Characteristics of the population (N = 447) and associations between the covariates and birth outcomes a.

Overall, 96.9% of the pregnancy outcomes were live births. The average gestational age was 38.9 (SD = 1.5, range, 28-42) weeks, and 6.5% of infants were preterm births. The average birth weight was 3216.2 (SD = 424.7, range, 1470-4530) (g), 3.9% of infants were born of LBW, and 3.9% were born in the SGA group. The mean of birth length was 50.0 (SD = 2.0, range, 36-55) (cm). A few (n = 14, 3.1%) pregnancy outcomes were non-live birth. Details of non-live birth were presented in Supplementary Table 2.

The distributions of PYR metabolites in urine samples

The distributions of the urinary PYR metabolites among the pregnant women are shown in Table 2. In this study, most of the urine samples contained one or more analyte(s). The detection frequency was 89.2% for 3PBA, 95.5% for 4F3PBA, and 72.9% for DBCA, respectively. The three metabolites were significantly intercorrelated (3PBA vs. 4F3PBA: Spearman’s r = 0.46; P < 0.01, and 3PBA vs. DBCA: r = 0.35; P < 0.01).

Table 2. Concentrations of PYR metabolites in urine of early pregnancy women.

The urinary concentrations of 4F3PBA were positively associated with indoor insecticide use (P < 0.05), and the DBCA levels were highly correlated with regularly eating bananas or cantaloupe (P < 0.05) (Table 3). No significant differences were observed between PYR metabolites and vegetable intake or eating other fruits such as pears, oranges, watermelons and cherry tomatoes (data not shown). The number of fruit types consumed by pregnant women was positively related to the concentrations of 3PBA and total PYR metabolites in urine (P < 0.05) (Table 3). The women who washed fruits or vegetables with soda, flour, rice-rinsing water or other detergent reduced the urinary level of total PYR metabolites (P < 0.05). The mothers who gave birth to SGA infants had a lower level of DBCA or total PYR metabolites (P < 0.05).

Table 3. Associations between maternal characteristics and urinary concentrations of PYR metabolites a.

Association between birth outcomes and prenatal exposure to PYRs or other risk factors

In the multiple linear regression analyses, we introduced maternal age, pregnancy BMI, adverse pregnancy history, GDM, PROM, and hypertensive disorder complicating pregnancy into all models. Potential covariates were all included, and entered into the models with a stepwise strategy at P < 0.20 (Table 1). Then, the selected covariates were forced into the corresponding regression models to analyze the associations between birth outcomes and prenatal exposure to PYRs. The results showed that urinary level of total PYR metabolites were positively associated with birth outcomes (birth weight: β = 125.5; 95% CI, 1.6 to 249.5, P = 0.047; birth length: β = 0.7; 95% CI, 0.09 to 1.2, P = 0.023; gestational age: β = 0.5; 95% CI, 0.05 to 0.9, P = 0.030), after adjusting for potential covariates (Table 4). In addition, we also observed a positive association between birth length and urinary 4F3PBA levels (β = 0.8; 95% CI, 0.2 to 1.5, P = 0.015), and positive associations between gestational age and DBCA (β = 0.3; 95% CI, 0.03 to 0.5, P = 0.028), or 4F3PBA (β = 0.6; 95% CI, 0.1 to 1.1, P = 0.019). The covariates introduced in the multiple linear regression models are shown in Table 1. Maternal PROM and hypertensive disorder complicating pregnancy were the risk factors for reduced birth weight, length, and gestational age. Worked in a restaurant was correlated to decreased birth weight. Perceived noise was associated with reduced birth length and gestational age.

Table 4. Multiple linear regression for analysis of associations between urinary PYR metabolites and birth outcomes a.

Multiple logistic regression analyses showed that the risk of SGA was negatively associated with increasing level of 4F3PBA (OR = 0.07; 95% CI, 0.01-0.7, P < 0.05), and increasing level of total PYR (OR = 0.08; 95% CI, 0.01-0.7, P < 0.05). Association between premature birth and 4F3PBA concentration was also negative (OR = 0.17; 95% CI, 0.04-0.8, P < 0.05) (Table 5). There was no evidence that showed an association between prenatal exposure to PYR and LBW. Our results showed that prenatal exposure to environmental factors such as noise, paint, and air pollution (passive smoke and work in a restaurant), were harmful to fetal growth. In addition, maternal hypertensive disorder complicating pregnancy and PROM were also the risk factors for LBW and premature birth. SGA was more common in the babies of mothers who had experienced adverse birth outcomes or hypertensive disorder (Supplementary Table 3).

Table 5. Multiple logistic regression testing effects of prenatal exposure to PYRs on the odds of adverse birth outcomes a.

During our follow-up, 8 fetuses were diagnosed with birth defects (Supplementary Table 2). Fisher’s exact test showed that women who had used household pesticides (within 6 months before enrollment) were more likely to have a fetus with congenital defect (P = 0.025). Multivariate analysis by Cox regression identified that household pesticides use was the independent risk factor for congenital defect. Congenital anomalies were not associated with urinary PYR metabolite levels (data not shown).


The results showed a wide exposure to low-level PYRs among pregnant women in this area, which was in accordance with previous reports (Ding et al., 2015; Dereumeaux et al., 2018; Qi et al., 2012). A comparison of these results to those from previous studies conducted among pregnant women, showed that the median level of 3PBA in the present study was close to that in France (Dereumeaux et al., 2018) and lower than that in Puerto Rico, and another two places in China (Ding et al., 2015; Qi et al., 2012, Lewis et al., 2014), while the median levels of DBCA and 4F3PBA in this study were higher than those in recent studies (Lewis et al., 2014; Dereumeaux et al., 2018; Ding et al., 2015). Urinary 3PBA is a non-specific metabolite of most PYRs, whereas DBCA and 4F3PBA are the specific metabolites of deltamethrin and cyfluthrin, respectively (Barr et al., 2010). These observations suggest that the overall PYR exposure levels in this region were similar to those observed in mainland France and Shandong (China) and lower than those in Jiangsu (China), although the types of PYRs used in this region were different from those in recent studies.

Pregnant women living in non-agricultural areas are more likely to be exposed to pesticides through diet. The concentrations of DBCA and total PYRs metabolites in urine of pregnant women who ate bananas regularly were significantly higher than those of women who did not. This was consistent with previous report that banana cultivation relies on heavy pesticide use (Mendez et al., 2018). And the urinary DBCA level in pregnant women who regularly ate cantaloupes was significantly higher than that of those who did not eat regularly. The reasons would be the same.

This prospective study indicated that increased urinary levels of total PYRs were significantly associated with increased birth weight, birth length and gestational age, and with decreased risk of SGA, after adjusting for potential covariates. These findings were different from previous studies performed in the USA (Berkowitz et al., 2004; Neta et al., 2011), Canada (Kennedy et al., 2005), Thailand (Mytton et al., 2007), the Philippines (Ostrea et al., 2012), Poland (Hanke et al., 2003) and northern China (Ding et al., 2015). Studies on permethrin use (for lice or scabies) showed no adverse effects on pregnancy outcome (Kennedy et al., 2005; Mytton et al., 2007). Similarly, it was reported that pyrethroids or its metabolites were not associated with gestational age or birth weight, length (Berkowitz et al., 2004; Neta et al., 2011). However, other researchers found that maternal exposure to PYRs was associated with a decrease in birth weight (Hanke et al., 2003; Ding et al., 2015). A study performed in Japan reported positive correlations between maternal 3PBA levels and birth weight or head circumference (Zhang et al., 2014), which is similar to our results. The variable effects of PYR exposure during pregnancy on fetal growth may be due to different levels of PYR exposure in various regions. The median level of 3PBA in the present study was close to that in Japan (Zhang et al., 2014), while lower than that in USA (Berkowitz et al., 2004) and northern China (Ding et al., 2015). This indicated that the beneficial effects of fruit intake may outweigh the adverse effects of PYR among low-dose exposure women, because we found the urinary concentrations of DBCA and/or total PYRs were positively correlated to maternal fruit intake, but inversely associated with the risk of SGA. This was consistent with that of Pedersen et al., who found maternal vegetable and fruit intake may be protective against the adverse effect of genotoxic agents on birth weight (Pedersen et al., 2013).

Pesticides may act as estrogen to promote fetal growth. The result in the present study showed urinary 4F3PBA levels were not associated with maternal fruit intake, but positively correlated to birth weight, length, and gestational age, and inversely related to the risk of LBW, SGA, and premature birth. Chevrier et al. observed that prenatal exposure to dichlorodiphenyltrichloroethane (DDT), but not pyrethroid, might have the estrogenic properties to accelerate fetal growth (Chevrier et al., 2019). PYR is a non-persistent endocrine-disrupting chemical (EDC) which may affect reproductive hormones produced in men and women (Rattan et al., 2017; Dziewirska et al., 2018). A study conducted in fish showed that exposure to nonlethal concentrations of bifenthrin increased the expression of the estrogen and glucocorticoid receptors in brain tissue (Ligocki et al., 2019).

The urinary 3PBA, a nonspecific metabolite of many PYR, was positively associated with the number of fruit types consumed by the mother. However, all βs for 3PBA were negative but not significant for birth weight, length, and gestational age (P > 0.05). This suggested that the PYR metabolite itself may have a different effect on fetal growth.

The results of the present study showed positive correlations between concentrations of total PYR and birth weight or gestational age, but no associations between total PYR levels and LBW or premature birth. The reason may be due to the small sample size (LBW: n = 17; premature birth: n = 28), or other confounding such as interactions between metabolites.

In addition, we found that women with GDM or gestational hypertension had lower DBCA levels in urine than those without (P < 0.1), and maternal PROM or hypertensive disorder complicating pregnancy were consistently associated with decreased birth weight, birth length, and gestational duration. This suggested that DBCA may increase the fetal growth indicators and gestational duration by reducing the risk of mother’s pregnancy-related diseases.

A positive correlation was observed between household insecticide use (within 6 months before enrollment) and birth defects. According to our market research and pesticide registration data on the China Pesticide Information Network, pyrethroids are the main component of household insecticides. Women who used household pesticides would be instantly exposed to high concentrations of PYRs. Furthermore, women may have used household pesticides exactly during the critical period of oocyte development and maturation or fetal organogenesis. This result suggested that maternal exposure to PYRs during this period may increase the risk of birth defect. This was consistent with the reports that PYRs could induce chromosomal aberrations in human or animal cells (Radwan et al., 2015; Bhunya and Pati, 1990; Muranli, 2013). Experiments in rats suggested that exposure to PYRs during early pregnancy induced embryo loss, but did not cause any detectable effect on developmental indicators (Ratnasooriya et al., 2003). Until now, few epidemiology studies examined the teratogenic effects of maternal exposure to PYRs on fetal development. Studies on permethrin products use showed no evidence of adverse effects of it on pregnancy outcome (Kennedy et al., 2005; Mytton et al., 2007).

Finally, our study suggested that maternal exposure to low level PRYs was associated with increased fetal growth indicators or gestational age. However, the beneficial effects of fruit intake may outweigh the adverse effects of pesticide exposure. Pregnancy-related disease and prenatal exposure to other environmental factors were also significant predictors of birth outcomes. In addition, maternal exposure to household pesticides may increase the risk of birth defects, which is consistent with an “all-or-nothing” pattern in maternal exposure to harmful substances during early pregnancy (Yang et al., 2018; Czeizel and Mosonyi, 1997).

The strength of the present study was the prospective cohort design (all subjects were recruited during early pregnancy). Meanwhile, the assessment of exposure was based on biomarker test and source of exposure which were characterized by neutral exposure metrics (Burns and Pastoor, 2018). However, the study needs to be replicated. Firstly, although we took into account several confounders in the final analysis, there may still be some potential confounding factors that we did not adjust for. The positive correlations between PYR metabolite concentrations in maternal urine and neonatal body size or gestational age may be derived from other underlying factors, such as nutrients that mothers take from other foods. Overall, fruits are more expensive than vegetables. Pregnant women who eat more fruits may have the opportunity to consume more nutrients.

Secondly, the sample size of congenital defect was small in our study (n =8), and few women reported exposure to household pesticides. Additionally, household insecticides may contain other insecticides, and the use of household insecticides cannot be considered as pure PYR exposure. Pregnant women might also be exposed to other types of pesticides such as organophosphorus pesticides, carbamate pesticides, and neonicotinoids, which are the most widely used agricultural pesticides in this region. It is too early to suggest that maternal exposure to PYRs had a teratogenic effect of the fetus. In future studies, it is necessary to take measurement of more types of widely used pesticides in urine samples from a larger general population.

Finally, one spot urine test in the first trimester does not represent PYR exposure throughout pregnancy. In future studies, measurement of PYR metabolites in urine samples collected in each trimester is desirable for an accurate evaluation of prenatal exposure.

In conclusions, this study showed that pregnant women in this region were exposed to low-dose PYRs, which is similar with those in other settings. Prenatal exposure to PYRs was associated with increased birth weight, length, and gestational age, and with decreased risk of SGA or preterm birth, however, the beneficial effects of fruit intake may outweigh the adverse effects of pesticide exposure. In addition, pesticides may act as estrogen to promote fetal growth, and the PYR metabolite itself may have a different effect on fetal growth. Yet, the use of household pesticides during or before pregnancy may increase the risk of birth defects. The mother’s pregnancy-associated disease and exposure to other environmental factors could also do harm to the fetal growth. Further study is needed to explore the biological mechanisms underlying the associations between prenatal exposure to PYR and fetal development.


The analysis of PYR metabolites was carried out in the School of Pharmaceutical Sciences and Yunnan Provincial Key Laboratory of Pharmacology for Natural Products, Kunming Medical University. The author would like to thank the teachers who gave guidance and help, and the pregnant women from Yunnan, China, who took part in this study.

This work was supported by the National Natural Science Foundation of China, 81673186, and Yunnan Provincial Collaborative Innovation Center for Public Health and Disease Prevention and Control grant, 2015YNPHXT01.

Conflict of interest

The authors declare that there is no conflict of interest.

© 2020 The Japanese Society of Toxicology