The Centers for Disease Control (CDC) conducts periodic biomonitoring to track levels of chemicals in human urine and blood.
Over the years CDC has compiled detailed data on several, including chlorinated hydrocarbon insecticides widely used from the 1960s through the 1970s, and , many of which are still widely applied. A few herbicides, including 2,4-dichlorophenoxyacetic acid (2,4-D), have been assessed in recent CDC testing.
The National Health and Nutrition Examination Survey (NHANES) is the major biomonitoring program run by the CDC. In the program’s 2009-2010 sampling of 2,747 people, the CDC reported a contemporary baseline for 2,4-D levels in human urine. The geometric mean of urinary 2,4-D levels across all people tested was 0.308 ug/l ( per liter, or ). The 50th percentile level was 0.28 ug/l and the 95th level was 1.43 ug/l, about five-times higher than the individual at the 50th percentile level of the distribution. 2,4-D levels in males exceeded those in females by 30%-50% (CDC, 2015).
The CDC also reports creatinine corrected 2,4-D in urine. Levels of chemicals in urine vary as a function of an individual’s fluid balance and volume of urine. A widely used method has been developed to compensate for this source of variability. It measures the creatinine concentration in a urine sample, and expresses the level of the analyte (in this case, 2,4-D) as a ratio of the creatinine concentration (Cocker et al., 2011).
In 2009-2010 CDC testing, 2,4-D levels averaged 0.309 ug/g ((CDC, 2015). Across all ages and both sexes, 95th percentile level concentrations (1.55 ug/g of creatinine) exceeded the 50th percentile level (0.301 ug/g) by 5-fold.per gram, or ) of creatinine in males (n=1,343) and 0.334 ug/g of creatinine in females (n=1,404)
CDC also measured levels of a major 2,4-D— 2,4-dichlorophenol. In 2009-2010, the levels averaged 0.788 ug/g of creatinine in males and 0.89 ug/g in females, nearly three-times higher than the average levels of 2,4-D.
Metabolite levels in Mexican Americans and non-Hispanic blacks were over 50% higher than in non-Hispanic whites in 2009-2010 (1.24 and 1.11 ug/g compared to 0.731 ug/g in non-Hispanic whites) (CDC, 2013).
Glyphosate biomonitoring data are limited, although many new efforts are underway to better understand levels in people and recent trends.
A team based at U.C. San Francisco tested the urine of 131 individuals in 2015 from across the U.S. Detectable residues were found in 93% (Limit of Detection = 0.2 ug/l) (Adams et al,. 2016). The average level of glyphosate in the 75 women in this study was 2.9 ug/l, while 56 men had an average level of 3.3 ug/l Seven children had an average of 3.6 ug/l in their urine.
In preliminary results for a targeted study, glyphosate was found in the urine of 63 of 69 (91%) pregnant women receiving prenatal care at an Indiana obstetric practice, at a mean level of 3.44 ug/l (Winchester et al., 2017). Surprisingly, the final study as published reported that higher glyphosate levels were associated with reductions in gestational length (r’s= -0.30, p=0.01). Women living in rural areas had higher mean glyphosate levels than women in urban/suburban regions (mean glyphosate level of 4.19 vs. 3.3, p=0.02), suggesting an additional route of exposure associated with proximity to corn and soybean production fields (Parvez e al., 2018).
Trends in glyphosate and AMPA (aminomethylphosphonic acid, glyphosate’s primary metabolite) levels were measured from 2001-2015 in the urine of 20-29 year old males living in a Northeast German city. Out of 399 samples, 31.8% contained glyphosate levels above the 0.1 ug/l (ppb) Limit of Quantification (LOQ) and 40.1% contained AMPA (Conrad et al., 2017). In 2015 samples, the team reported 0.16 ug/l at the 75th level of the distribution, and 0.45 at the 95th. Accordingly, levels in these German citizens were about an order of magnitude lower than levels among the Americans tested by Adams et al. (2016).
An industry-funded study measured urinary glyphosate levels in 48 farmers, their spouses, and 79 children the day before, day of, and three days after a glyphosate application. A geometric mean level of 3 ug/l, and a high level of 233 ug/l were reported on the day of application, with 60% of farmer-applicators testing positive (Acquavella et al., 2004). Glyphosate levels in the urine of a French farm family ranged from 9.5 ug/l the day after a spray application, to 2 ug/l two days after the application in both the farmer/applicator and one of his children (Mesnage et al., 2012b).
Most recently published glyphosate biomonitoring studies report average levels in urine that correspond to exposures that are, at most, just a few percent of the (Niemann et al., 2015). However, estimates of contemporary urinary levels among people living in areas like the rural Midwest cannot be made from past studies, because herbicide use patterns and intensity have changed so dramatically. This is especially true in areas like the rural Midwest where over one-half the surface area is typically planted to -HR crops that are sprayed one to three times annually with glyphosate.-set chronic Reference Dose (1.75 mg/kg/day)
Key Insights on Trends in Exposure
A number of recent papers and datasets point to some critical conclusions regarding trends in glyphosate and AMPA exposure levels. The Herbicides in Humans, a dynamic Prezi presentation, summarizes one of the important new studies done at the the Rancho Bernardo community. Click the thumbnail below to open.
The Rancho Bernado Study
The Rancho Bernado Study (RBS) of Aging is an Mills et al. (2017) report a significant increase in both mean glyphosate and aminomethylphosphonic acid (AMPA, glyphosate’s major metabolite) levels in the urine of 100 RBS residents over the course of five sampling periods beginning in 1993-1996 and ending in 2014-2016.-funded, longitudinal that has been underway since 1972 in a residential Southern California community. A team led by
The mean age of the 100 study participants when the last urine samples were drawn in 2014-2016 was 77.7 years. Highly sensitive analytical methods were used to quantify glyphosate and AMPA levels. The glyphosate limit of detection was 0.03 µg/l (ppb) and 0.04 µg/l for AMPA. All samples were corrected for differences in water consumption and urine volume across the subjects.
Twelve participants (12%) had a detectable level of glyphosate in the 1993-1996 sampling. In 1999-2000 samples, 30 participants tested positive for glyphosate, with steady increases in subsequent sampling periods: 43% in 2001-2002, 38% in 2004-2005, and 70% in 2014-2016 (see Figure 2C.1).
The glyphosate metabolite AMPA was found in slightly more participants in the last two sampling periods, but in only about one-half as many participants in the first two sampling periods (i.e., 5 were positive for AMPA in the first sampling, whereas 12 were positive for glyphosate). The percent of samples positive for AMPA were 5%, 15%, 43%, 40%, and 71% across the five sampling periods.
Mean levels of glyphosate and AMPA among the participants testing positive also rose markedly from the 1993-1996 sampling to the 2014-2016 sampling. The mean glyphosate level in 2014-2016 was 0.449 µg/l and 0.401 µg/l for AMPA, compared to 0.203 µg/l for glyphosate and 0.168 ug/l for AMPA in 1993-1996.
So, glyphosate levels among those testing positive rose 2.2-fold and AMPA levels rose 2.4-fold from the 1993-1996 sampling to the 2014-2016 sampling.
Plantdata on glyphosate and AMPA show that over a period spannning a few to several months, residues of glyphosate break down incrementally to AMPA. As this process occurs, the mean level of glyphosate residues falls and the mean of AMPA residues rises.
Consider this example (see Figure 2C.2). Glyphosate was applied to most soybeans grown in the U.S. in 2011. The’s Pesticide Data Program ( ) conducted a one-time, special sampling of glyphosate and AMPA residues in soybeans. Out of 300 samples, 90.3% tested positive for glyphosate at a mean level of 1.94 ppm, while 95.7% were positive for AMPA with a mean level of 2.28 ppm
The soybeans in the 2011 PDP sampling had probably been harvested 5 to 8 months earlier. Had residue levels been checked on the day of harvest, the percent positive for glyphosate would have been around 97%, with a much lower percent positive for AMPA. If the testing were done after the soybeans were in storage for another year, little glyphosate would be found relative to AMPA.
A large body of data on the metabolism of glyphosate and AMPA in mammals shows that most of the glyphosate and AMPA residues ingested during any given day pass quickly through the GI track and are excreted within 1-2 days. Hence, whenever either glyphosate or AMPA are found in a person’s urine, the individual was likely exposed to the chemical within the last few days, and in the same form in which the chemical was excreted in urine.
Figures 2C.3 and 2C.4 below show the trend in mean glyphosate and AMPA levels reported in the RBS. The figures plot both mean levels among all study participants and those with detected levels; in both figures, the line covering the trend among individuals testing positive rises more sharply than the line where the mean includes participants lacking detectable residues.
It is clear in these figures that the frequency of positive samples for AMPA went up faster than in the case of glyphosate. It takes considerable time (i.e., weeks to months) for glyphosate to breakdown to AMPA. Accordingly, the data for the first two sampling periods suggest that some Rancho Bernado residents in this study may have been exposed to recently applied glyphosate-based herbicides, leading to a higher frequency of positives for glyphosate, but not for AMPA.
The trend in glyphosate and AMPA mean levels among those testing positive across the five sampling periods suggest a sharp increase in exposures between the 2004-2005 and 2014-2016 sampling periods. Figures 3 and 4 show that mean levels rose modestly from the 1993-1996 sampling through the 2004-2005 sampling, but then about doubled from the 2004-2006 sampling to the 2014-2016 sampling.
(For an interactive version of this data, see here).
What might have accounted for this big change, and might this trend continue?
Glyphosate Metabolism in Plants
Plant metabolism data on glyphosate and AMPA show that over a period of a few to several months, residues of glyphosate break down incrementally to AMPA. As this process occurs, the mean level of glyphosate residues falls, while the mean of AMPA residue levels rises.
Consider the 2011 soybean residues shown in Figure 2C.2 above. The soybeans in the 2011 PDP sampling had probably been harvested 5 to 8 months earlier. Had residue levels been checked on the day of harvest, the percent positive for glyphosate would have been around 97% (the percent of the soybean crop planted to Roundup Ready seeds), with a lower percent positive for AMPA. If the testing had been done after the soybeans were in storage for another year, little glyphosate would be found, but AMPA would be present in most samples.
A large body of data on the metabolism of glyphosate and AMPA in mammals shows that glyphosate and AMPA residues pass quickly through the GI track and are excreted within 1-2 days. Hence, whenever levels of either chemical are found in human urine, the individual was likely exposed to the chemical within the last few days, and in the same form in which the chemical was excreted in urine.
It takes considerable time for glyphosate to breakdown to AMPA. Accordingly, most people with AMPA residues in their urine likely ingested AMPA residues in food or beverages in the previous 24-48 hours.
A number of studies and datasets show that both glyphosate and AMPA residue levels in human urine rose sharply from the mid-2000s through ~2016 (date of last testing results, not necessarily the end of the upward trend in residue levels).
What might have accounted for this change in the trend of glyphosate and AMPA residues in human urine, and might the upward trend continue?
Pre-Harvest, Green Burndown GBH Applications
It all likelihood, most of the increases in the levels of glyphosate and AMPA reported in recent studies were caused by pre-harvest applications of glyphosate-based herbicides (GBHs) on grain and other crops.
Such applications are made about two to three weeks before normal harvest operations in order to prematurely kill mother plants. The goal is to speed up the drying process, so that the grain, beans, or other crop can be harvested a week or 10 days earlier than otherwise would be possible. “Harvest aid” applications of GBHs are also referred to as “green burn down” sprays, and GBHs are one of several options farmers have for chemical desiccation of a crop.
Wheat farmers in northern Idaho, Montana, and the Dakotas were the first to adopt this new use pattern for GBH use in the U.S. Small grain farmers in northern Europe, and especially the U.K., had been applying GBHs as a harvest aid since the early 1980s, especially in wet years when it is hard to cure (dry) and harvest a wheat, oat or barley crop before fall rains begin. In recent years, between 30% and 60% of the U.K. wheat crop is sprayed pre-harvest with a GBH.
This little-noticed change in how some U.S. farmers were using GBHs began around 2004. A few years later, most glyphosate patents had expired and a number of lower-cost generic brands of GBHs were widely available and heavily promoted. Once the cost of the GBH needed to spray an acre fell to $5.00 or less, a growing number of wheat and small grain farmers decided to adopt this method of expanding the harvest window.
As a result around 2005, glyphosate and AMPA residues began finding their way into a variety of grain-based products, especially those made from whole wheat, or containing a significant amount of wheat germ. Unfortunately, the USDA’s PDP has only tested wheat flour twice for pesticide residues in 2003 and 2004, and when they did test wheat, they did not include a method capable of detecting glyphosate and AMPA because of the added cost. Wheat grain has been tested in six years, five between 1995 and 2006, and one since (2012). But just as the case with wheat flour, the methods used could not detect glyphosate or AMPA.
But routine, annual residue testing conducted by the U.K.’s Food Standards Agency has measured glyphosate levels in wheat and grain-based products for nearly 20 years. The frequency of positive samples in U.K.-grown wheat and grain-based products roughly track the percent of the wheat crop sprayed each year with GBHs as harvest aids. Mean levels of detection have risen modestly since 1999.
Might the upward trend in glyphosate and AMPA residues in urine continue?
Mean residues levels in food, and human urine, will continue to rise if steps are not taken to prevent that outcome. Regulation is unlikely to restrict harvest aid uses, and indeed recent actions taken by EPA have markedly expanded the scope of higher-residue GBH uses, including potential harvest aid applications.
Current GBH product labels would accommodate substantial growth in the number of crops and share of acres sprayed with GBHs as a harvest aid. Tolerances already approved by EPA will cover substantial increases in glyphosate and AMPA residues in food. It is also possible the EPA will approve increases in allowable tolerances on other foods, to accommodate more harvest aid applications and/or newly introduced GE-HR crops.
Some food companies and retailers, on the other hand, are considering the incorporation in supply chain purchase agreements a provision that prohibits GBH uses as a harvest aid. As more consumers learn about the role pre-harvest applications of GBHs play in their personal exposure profile, demand may also grow for organic grain-based products.