Abstract
Purpose:
The World Health Organization’s International Agency on Research for Cancer has determined that glyphosate is “probably carcinogenic to humans.” There is a great public interest to investigate whether glyphosate are detected in breast milk. Thus, the goal of this study was to assess the concentration of glyphosate and its main metabolite in breast milk.
Materials and Methods:
Liquid chromatography was performed at 25°C using a Luna NH2, 50 × 2 mm, 3⎛ m (Phenomenex) analytical column. Electrospray ionization mass spectrometry was collected using negative ionization mode. The calibration curve for glyphosate ranged from 10 to 250 ng/mL. The detection limit was 1 ng/mL.
Results:
Breast milk samples were collected from 74 women, which included vegans (n = 26), vegetarians (n = 22), and nonvegetarians (n = 26). One of the 74 milk samples contained a detectable concentration of glyphosate and an additional 7 were found to contain aminomethylphosphonic acid.
Conclusions:
In breast milk samples collected mainly from women residing in urban regions of the United States, glyphosate detection was rare. Consistently, breastfed infants have a low or minimal risk of being exposed to glyphosate through ingestion of mother’s milk. It is possible that the presence/absence and/or level of concentration of milk glyphosate depend on a place of residency and time of breastfeeding vis-à-vis time of its agricultural application.
Introduction
Glyphosate (N-phosphomethyl glycine) is the most widely used herbicide in the agricultural sector in the United States and the second most used weed killer in the home and garden sector. 1 The use of glyphosate is so widespread that it is found in the soil, water, air, and food. Consequently, exposure to glyphosate is extensive. In the National Health and Nutrition Examination Survey 2013–2014, glyphosate was found in 81% of 2,310 urine samples analyzed among Americans ≥6 years of age. 2
The United States Environmental Protection Agency stated that glyphosate “shows evidence of noncarcinogenicity for humans.” 3 Nonetheless, the World Health Organization’s International Agency on Research for Cancer has determined that glyphosate is “probably carcinogenic to humans.” 4 In addition, the Committee on Nutrition Council on Environmental Health and Climate Change at the American Academy of Pediatrics stated that, “Prenatal exposures to glyphosate are reported to be associated with increased risk of preterm birth and in utero endocrine disruption in children, but the impact of glyphosate residues on child health remains complex and incompletely understood.” 5 Likewise, studies with animals showed it to be an endocrine disrupter. 6
Aminomethylphosphonic acid (AMPA) is the primary degradation product of glyphosate. AMPA is considered to have toxicological properties similar to glyphosate.
Since many toxic compounds can accumulate in breast milk, breastfed infants may be detrimentally affected by them. To date, only a few studies have been published on glyphosate content in breast milk, only one of which was from the United States. Breastfeeding is the preferred method of feeding infants. The American Academy of Pediatrics and other professional organizations advocate exclusive breastfeeding for the first 6 months and continued breastfeeding along with complementary feeding for up to 2 years. 7
Considering the above and that there is a great public interest to investigate whether glyphosate or AMPA is detected in breast milk, the goal of this study was to assess glyphosate and AMPA breast milk concentration using a liquid chromatography-mass spectrometry (LC-MS) collected from American women.
Materials and Methods
Recruitment
Institutional Review Board approval for the study was obtained. Women of all ethnic backgrounds were eligible as long as they were between18 and 46 years of age, gave birth to a healthy infant at term, delivered at least 2 weeks prior to collection of milk, and were willing to collect one milk sample per the collection protocol. Exclusion criteria included: methylene tetrahydrofolate reductase gene mutation; health conditions that affect vitamin B12 status (e.g., patients with known intrinsic factor deficiency, history of bariatric surgery, inflammatory bowel disease, Celiac disease, etc.); pregnant; hypo or hyperthyroidism; advanced liver disease (cirrhosis, hepatitis, etc.); or myeloproliferative disorders. Recruitment and milk sample collection took place between November 2016 and April 2017.
Sample and survey collection
Each participant was asked to provide a breast milk sample, which was either collected in person or received via a shipment on dry ice to our laboratory. Participants were instructed to fully express milk from one breast and to collect the sample in the morning, during the first or second feeding of the day, and at least 2 hours after the previous feeding. Milk was to be collected in a dimly lit room to help protect light-sensitive nutrients. Women were instructed to completely express the content of one breast using the expression method of their choice. Expressed milk was to be transferred to a storage bag appropriate for freezing breast milk, labeled with the collection date, wrapped in aluminum foil for further protection, and stored in a home freezer until samples were collected or shipped.
Sample analysis
Chemicals and reagents
An analytical standard of glyphosate was obtained from Sigma Aldrich (Saint Louis, MO) and the internal standard (ISTD) AMPA-13C,15N,2D was purchased from Cambridge Isotope Laboratories (Tewksbury, MA). Optima grade water, acetonitrile, methanol, and isopropanol were purchased from Fisher Scientific (Pittsburgh, PA). Ammonium bicarbonate was purchased from Sigma Aldrich (Saint Louis, MO). Ultrafree-MC centrifugal filter, 0.45⎛ m pore size, hydrophilic PVDF, 0.5 mL volume was purchased from Millipore Sigma (Burlington, MA).
Sample preparation
Stock solutions of the analytical reference standards for glyphosate were made from a 100 µg/mL stock solution to a 10 µg/mL working solution prepared in water. The calibration curve for glyphosate ranged from 10 to 250 ng/mL. The detection limit was 1 ng/mL. The ISTD stock concentration of 50 µg/mL and 10 µL were spiked into the milk samples and the calibration concentrations.
Prior to extraction, the breast milk samples were stored at −80°C in microcentrifuge tubes in volumes of 1 mL each. A 500 µL aliquot of breast milk was put into a microcentrifuge tube and 1 mL of methanol and 10 µL of the 50 µg/mL of ISTD were added. The sample was then vortexed well and incubated at −20°C for an hour. The samples were centrifuged at 13.5 kRPM for 30 minutes at 4°C and the supernatant was placed in a new microcentrifuge tube where it was evaporated to dryness under nitrogen at 50°C. Breast milk samples were dried, reconstituted with 300⎛L water, vortexed, and transferred to a filtered centrifuge tube. The sample was centrifuged at 10 kRPM for 2 minutes to remove particulates, and an aliquot was transferred into autosampler vials for analysis.
Pooled samples
A quality control sample was generated by pooling 100 µL aliquots from each breast milk sample. Prior to using the pool sample, we tested to see if it was positive for our analytes of interest. Two separate 500 µL portions of the sample were spiked with a low and high glyphosate standard, 0.01 and 0.5 µL, respectively. The ISTD was also added to the sample, and the QC samples were extracted along with the individual breast milk samples.
Instrumental analysis
A Sciex Exion 100 high-pressure liquid chromatography/3200 triple quadrupole mass spectrometer (LC-MS) was used for this study. Liquid chromatography was performed at 25°C using a Luna NH2, 50 × 2 mm, 3⎛ m analytical column (Phenomenex). The mobile phase consisted of 10 mM ammonium bicarbonate (A) and acetonitrile (B) delivered as a gradient as follows: 0–2 minutes 0% B, 2–6 minutes 90% B, 6–8 minutes 95% B, and 8–15 minutes 0% B. The flow rate was 0.3 mL/min and 50 µL of the sample was injected. Electrospray ionization mass spectrometry was collected using negative ionization mode. The instrument parameters were as follows: ion source voltage was 4500 V, source temperature was 120°C, gas sources 1 and 2 were set to 50, and curtain gas was set to 25. Mass spectra were collected utilizing multiple reaction monitoring. Injection volumes for both standards and breast milk samples were 20 µL.
Results
A total of 371 individuals completed BSQ of whom 229 were excluded based on our inclusion/exclusion criteria. Of the 142 women invited to participate, milk samples were collected from 74. The remaining 68 either did not respond to our invitation, could not donate milk, or asked to be excluded. The 74 included women adhered to 3 diet types: vegan (n = 26), vegetarian (n = 22), and nonvegetarian (n = 26). Demographic and socioeconomic data are found in Table 1.
Characteristics of Study Participants
BMI, body mass index; GED, General Education Diploma; HS, high school.
Indicates no significance difference between vegan and vegetarian.
Indicates no significance difference between vegetarian and non-vegetarian.
Indicates no significance difference between vegan and vegetarian.
Indicates means ± standard deviations. Data were evaluated for differences between groups using one-way ANOVA and Tukey test was used for multiple comparisons. Groups with a common letter in the superscript are not significantly different.
Data represent number (percentage) and were evaluated for differences between groups using Fisher’s exact test.
Results indicate that only 1 of the 74 milk samples contained a detectable concentration of glyphosate and an additional 7 were found to contain AMPA. The concentration for the positive sample was 13 ng/mL. The sample was collected from a Caucasian nonvegetarian female with a doctorate degree, residing in small town in eastern North Carolina.
Discussion
The present study analyzed breast milk for glyphosate in the United States where this herbicide is widely used in agriculture along with homes and gardens. Results indicate that only one of the 74 milk samples contained a detectable concentration of glyphosate and an additional 7 were found to contain AMPA.
Our findings are consistent with those reported by two other studies published to date, one from the United States (n = 41) and the other one from Germany based on 114 samples, none of which contained glyphosate above the detection concentration.8,9 On the contrary, the third study published to date, which was conducted in Brazil, found all samples (n = 67) contained a detectable concentration of glyphosate (mean concentration = 1.45 µg/L). 10
What distinguished the Brazilian study from the two previously published studies and ours was the time when samples were obtained “at the peak of glyphosate application in corn and soy crops in the region.” The second difference was the place of residence of the study participants. In the Brazilian study, 72% of the participating women lived close to where crops were cultivated. Samples included in our analysis were collected from women from several US states, study participants mainly resided in urban regions, and sample collection took place between November and April. There is little agricultural activity in terms of pesticide application during the months of our data collection. The samples analyzed in the German study were collected in August and September, also a time during which glyphosate is not widely used, as most of the grain harvest takes place from late June to early August. 11
It should also be pointed out that the mean stage of lactation in our study was 37, 55, and 28 weeks for vegan, vegetarian, and nonvegetarian women, respectively. The German study included samples from two regions of the country, Bavaria (n = 17) and lower Saxony (n = 97). The median lactation period for the first group was 18 week and 11 weeks for the second group. This means that at least some of the women in these two studies breastfed during glyphosate application periods. These findings seem to imply that although milk can be contaminated by glyphosate, this occurs mainly during glyphosate application and that this herbicide is not stored in the body for long periods of time. This conclusion is consistent with the reported half-life of glyphosate in humans of 5.5 to 10 hours. 12 The conclusion is also consistent with findings from a study which assess glyphosate exposure of members of farm and nonfarm families. The former had twice as high exposure of AMPA compared with the latter (61% versus 32%) and more than double the maximum concentration of 7.24 µg/L versus 3.21 µg/L. 13
Most of the samples in this study were collected from women with 4-year or graduate college degrees. Thus, the findings reflect mainly breast milk of highly educated women and may not represent all women.
Conclusions
In breast milk samples collected mainly from women residing in urban regions of the United States, glyphosate detection was rare. Consistently, breastfed infants have a low or minimal risk of being exposed to glyphosate through ingestion of mother’s milk. It is possible that the presence/absence and/or level of concentration of milk glyphosate depend on a place of residency and time of breastfeeding vis-à-vis time of its agricultural application.
Footnotes
Authors’ Contributions
R.P.: Funding acquisition, conceptualization, writing, and project administration. A.W.: Data curation, formal analysis, and writing. M.S.: Data curation, formal analysis, and writing. K.K.: Data curation, formal analysis, and writing.
Author Disclosure Statement
The authors have no conflict of interest.
Funding Information
Supported by grants from the Academy of Nutrition and Dietetics Foundation, Vegetarian Nutrition Dietary Practice Group, and East Carolina University Faculty Senate.