Association between Glyphosate Exposure and Erythrograms in a Representative Sample of US Adults: NHANES 2013–2014

Glyphosate is the most commonly utilized herbicide globally, and a growing body of experimental research has linked its exposure to red blood cell damage. However, the potential toxicity of glyphosate exposure on erythrocytes in the general population remains poorly understood. Therefore, we analyzed data from the 2013–2014 National Health and Nutrition Examination Survey (NHANES) of 1466 adults (≥ 18 years) to explore the potential relationship between glyphosate exposure and erythrocyte profiles. Our results indicated a significant negative association between urinary glyphosate levels and hemoglobin (Hb) and hematocrit (Hct) in multiple regression analysis, with ß coefficients of -0.157 (S.E. = 0.055, P = 0.012) and -0.431 (S.E. = 0.195, P = 0.043), respectively. Additionally, the odds ratio showed a significant increase in individuals with anemia with a one-unit increase in ln-glyphosate levels (odds ratio = 1.523 (95% CI = 1.301 – 1.783), P < 0.001 in the final model). The negative correlation between glyphosate and Hb was more pronounced in subjects older than 60 years, non-Hispanic white ethnicity, lower income, and those with a body mass index (BMI) < 25 and ≥ 30. In conclusion, our results provide preliminary evidence of a plausible association between glyphosate exposure and anemia in a subset of the adult population in the United States. However, further research is necessary to understand the underlying mechanisms and clinical implications of this association.


Introduction
Glyphosate acts as a competitive inhibitor for the 5-enolpyruvylshikimate-3-phosphate synthase, which is crucial for the biosynthesis of aromatic amino acids in plants and microorganisms (Peillex and Pelletier 2020). Since 1974, glyphosate has been used as the primary active ingredient in herbicides. Glyphosate based herbicides (GBH) are a mixture of glyphosate and various adjuvants, mainly surfactants, that help in the herbicide penetrate the plant and increase its efficacy (Williams et al. 2000). Due to their high effectiveness in controlling weed growth, glyphosate and GBH have become the most widely used herbicides in the world (Benbrook 2016). Humans can potentially be exposed to glyphosate and GBH through multiple pathways, including dermal contact, inhalation, and oral ingestion (Ospina et al. 2022). The elimination half-life of glyphosate in humans typically falls within the range of 5.5 to 10 h. Notably, the primary routes of excretion are fecal excretion and urinary excretion, accounting for 90% and 6% of elimination, respectively, and occurring within a 48-h timeframe (Zoller et al. 2020). Available evidence suggests that the frequency and magnitude of glyphosate exposure in the United States has increased significantly over time. (Mills et al. 2017).
Glyphosate has historically been considered safe because it targets an enzyme that is not present in animals. In recent years, however, there has been growing concern regarding the potential adverse health impacts of glyphosate and its derivatives. Numerous studies have linked glyphosate exposure to a variety of health hazards, including cancer, birth defects, and reproductive problems (de Araujo et al. 2016). These findings prompted the International Agency for Research on Cancer to classify glyphosate as a probable human carcinogen in 2015 (Davoren and Schiestl 2018). Glyphosate and its formulations have been found to have varying cellular effects depending on the cell type and concentrations tested (Peillex and Pelletier 2020). Studies have shown that glyphosate and GBH exhibit dose-dependent cytotoxicity and genotoxicity at environmentally-relevant concentrations in human cell culture (Vanlaeys et al. 2018). Several experimental investigations have explored the potential toxic effects of glyphosate and GBH on erythrocytes and hematological parameters. In vitro studies indicate that these compounds may cause damage to erythrocytes, inducing oxidative stress, hemolysis, and hemoglobin oxidation (Kwiatkowska et al. 2014;Pieniążek et al. 2004;Rodrigues et al. 2011). Exposure to glyphosate and GBH can lead to decreased red blood cell (RBC) counts, altered erythrogram parameters, increased inflammation, and bone marrow toxicity in animal models (Bojarski et al. 2022;Mzengereza et al. 2020;Prasad et al. 2009;Rodrigues et al. 2011).
Although experimental studies suggest that glyphosate and GBH may have adverse effects on erythrocytes, there is limited understanding of their association with erythrocyte parameters in humans. In an occupational study, a significant difference in RBC count was observed in glyphosate-exposed workers compared to unexposed workers when the airborne glyphosate concentration in the air was below 9.40 mg/m 3 (Zhang et al. 2019). However, there is no research on the relationship between glyphosate and erythrocyte profiles in the general population. Erythrograms are standard clinical tools used to assess the overall RBC health and identify a spectrum of erythrocyte disorders.
To augment our comprehension of the correlation between glyphosate levels and erythrograms, we recruited participants from the 2013-2014 National Health and Nutrition Examination Survey (NHANES), which includes extensive data on urinary glyphosate levels and erythrograms.

Study population
The NHANES is a biennial national survey that selects a representative sample of U.S. citizens. The survey methodology and consent forms can be accessed on the NHANES website (CDC 2016a). For our study, we selected individuals aged 18 years and older who had measurements of glyphosate exposure and erythrograms, along with necessary demographic data from NHANES 2013-2014. We included a total of 1466 participants aged between 18-80 years in our analysis.

Measurement of urinary glyphosate levels
In the NHANES 2013-2014 survey, urinary glyphosate concentrations were quantified in a representative sample of onethird of participants aged 6 years and older. For our analysis, we obtained data from individuals 18 years of age and older. The urinary samples were analyzed using a highly sensitive and specific two-dimensional on-line ion chromatography coupled with tandem mass spectrometry method with isotope dilution for quantification (Schütze et al. 2021). Quality control and quality assurance measures were strictly adhered to during the analytical measurements. For concentrations below the limits of detection (LOD), a value was imputed using the square root of 2 times the LOD value, according to NHANES protocol. The detailed analytical method is available on the official NHANES website (CDC 2022).

Erythrograms
RBC counts indicate the quantity of RBC in a given volume of blood, while hemoglobin (Hb) levels measure the weight of Hb in a corresponding volume of blood. Hematocrit (Hct), expressed as a percentage, reflects the volume of RBC in the blood. Mean corpuscular volume (MCV) is used to determine the average size of RBC by dividing HCT by RBC counts. Mean corpuscular hemoglobin (MCH) reflects the average weight of Hb in each RBC and is calculated by dividing Hb by the number of RBC in a given volume of blood. Mean corpuscular hemoglobin concentration (MCHC) is used to determine the concentration of Hb in each RBC by dividing Hb by HCT. Red cell distribution width (RDW) shows the deviation in the volume of RBC and 1 3 is calculated by dividing the standard deviation of MCV by MCV and multiplying by 100, expressed as a percentage. The detailed analytical method is available on the official NHANES website (CDC 2016b).

Covariates
According to the NHANES website, skilled personnel utilizing standardized procedures to collect data at all study sites. Sociodemographic information, such as age, sex, and race/ ethnicity, was gathered during the household interview. Body mass index (BMI) was calculated by dividing weight in kilograms was divided by height in meters squared. Participants were categorized as active smokers, environmentally tobacco smoke (ETS) exposed, or non-smokers based on the results of a smoking questionnaire (CDC 2016c). An alcohol consumption questionnaire was utilized to assess alcohol consumption. The unit of measurement for alcohol consumption was defined as number of standard drinks. Participants were queried whether they had consumed at least 12 standard alcoholic beverages in the preceding year, and their responses were then dichotomized. During NHANES 2013-2014, individuals who were at least 1 year old were questioned whether they had received treatment for anemia in the previous 3 months. Chronic kidney disease is defined as an estimated glomerular filtration rate of less than 60 ml/min per 1.73 square meters (Yan et al. 2021). The criteria used to define anemia are Hb concentrations of less than 12 gm/dL for women and less than 13 gm/dL for men in whole blood.

Statistics
In this study, measurements of glyphosate levels were obtained in units of µg/L or µg/g creatinine. The natural logarithm of glyphosate was used to determine the exponential mean and standard deviation for various subgroups. Statistical analyses were performed using a two-tailed Student's t-test and one-way analysis of variance. To evaluate the relationship between the erythrograms, which were used as the dependent variable, and ln-glyphosate, the independent variable, a complex sample of the general linear model was utilized. Two models were used for covariate adjustment. Model 1 adjusted for several demographic factors including age, sex, ethnicity, BMI, smoking, alcohol consumption, household income, and urinary creatinine. Model 2 included additional adjustments for chronic renal failure and treatment for anemia in the previous three months. Urinary creatinine was treated as a separate independent variable based on prior research, instead of being adjusted for hydration (Barr et al. 2005;O'Brien et al. 2016). To minimize the potential for model-dependent associations, a result was considered statistically significant only if it was significant in both Models 1 and 2 (Lang et al. 2008;Lin et al. 2009). For logistic regression analyses examining the potential relationship between anemia and glyphosate, model 2 was used for covariate adjustment in the complex sample. In our analysis, we reported the population size using both the unweighted numbers, which represent the raw count of participants, and the weighted numbers, which estimate the population size. This approach is essential to account for the complex sampling design employed in the study and to ensure that the findings can be appropriately generalized to the larger population (CDC 2005). The natural logarithm of glyphosate and urinary creatinine were used in the analysis due to their non-normal distributions. All statistical analyses were conducted using SPSS version 20 (SPSS Inc., Chicago, Illinois, USA). The significance level was set at P < 0.05.

Results
The study population had a mean age (SD) of 48.16 (18.31) years, with 80.2% of individuals having detectable levels of glyphosate. The mean levels (SD) of glyphosate were found to be 0.55 (0.54) µg/L. As presented in Table 1, the geometric means of glyphosate in different subgroups indicate that urinary glyphosate levels were higher in men, older individuals, non-Hispanic blacks, and individuals with a higher BMI. Moreover, creatinine-corrected glyphosate levels were higher in women, older individuals, non-Hispanic whites, and nonsmokers.
In the complex multiple linear regression analysis, a total of 1465 unweighted participants were enrolled, while the estimated population size for the complex multiple linear regression analysis was 209,123,002. Regression analyses reported in Table 2 suggest that a one-unit increase in ln-glyphosate levels was negatively correlated with both Hb and Hct, with ß coefficients of -0.157 (S.E. = 0.055, P = 0.012) and -0.431 (S.E. = 0.195, P = 0.043), respectively. An overview of the Hb and Hct across quartiles of urinary glyphosate in multiple linear regression models is presented in Fig. 1. The findings of the study reveal that in model 2, both Hb and Hct show a statistically significant decrease with increasing glyphosate quartiles, with P values of 0.028 and 0.036, respectively.
The logistic regression analyses reported in Table 3 assess the association between glyphosate and anemia. The odds ratios (ORs) of anemia were significantly lower with a one-unit increase in ln-transformed glyphosate in both models (OR = 1.523, 95% CI = 1.301 -1.783, P < 0.001 in model 2). Table 4 further illustrates the negative correlation between glyphosate and Hb in different subgroups of study subjects. This correlation was found to be more pronounced in subjects aged over 60 years of age, those of non-Hispanic white ethnicity, those with lower income, and those with a BMI < 25 and ≥ 30.

Discussion
Our study employed a subset of representative cohort of the general U.S. adult population and has revealed a significant negative correlation between urinary glyphosate levels and both Hb and Hct. Furthermore, our findings suggest a significant increase in the odds of anemia in individuals with higher glyphosate levels. The negative correlation between glyphosate and Hb was particularly pronounced in subgroups of subjects aged over 60 years, of ethnicities other than non-Hispanic white, with lower income, and those with BMI < 25 and ≥ 30. This study provides preliminary evidence of a possible association between glyphosate exposure and anemia in the general adult population. The possible establishment of a causal relationship is of great concern because it highlights the potential adverse impact of glyphosate exposure on erythrocytes in American adults. The significance of this research is attributed to its use of reliable and comprehensive data from the NHANES database and the inclusion of a representative sample of American adults aged 18 years and older.
In this study of 1466 individuals aged 18-80 years, detectable levels of glyphosate were found in 80.2% of participants. Similarly, another NHANES survey of individuals aged six years and older found glyphosate residues in 81.2% of participants (Ospina et al. 2022). These high detection rates suggest that exposure to glyphosate is unavoidable during routine daily activities. One study indicated that glyphosate concentrations vary with fasting status, while another found a positive correlation between whole-grain bread consumption and increased urinary glyphosate levels (Ashley-Martin et al. 2023;Ospina et al. 2022). These findings suggest that individuals with elevated glyphosate levels may have been exposed to a diet with frequently herbicide applications. In our present study, we found contradictory results in urinary glyphosate levels when corrected for urinary creatinine compared to when no correction was applied. The disparity between these two groups can be attributed to the distinct approaches utilized to account for urine dilution, with the creatinine correction method offering a more dependable and precise evaluation of glyphosate exposure levels among individuals and groups. Our findings revealed that creatinine corrected urinary glyphosate levels were significantly higher in women, older adults, non-Hispanic whites, and nonsmokers. It is well documented that demographic characteristics can influence an individual's exposure to environmental contaminants, including glyphosate (Shea 2003). In a study using data from NHANES 2013-2016, the amount of fruit and vegetable intake was higher among women, older adults, and nonsmokers, who are commonly exposed to glyphosate as a herbicide (Kolakowski et al. 2020). Further studies are warranted to determine the specific reasons for differences in glyphosate exposure among subpopulations. Numerous experimental studies have investigated the effects of glyphosate and GBH on erythrocytes. For instance, an in vitro study discovered that exposure to glyphosate and GBH at concentrations of 0.25 mmol results in the production of reactive oxygen species in human erythrocytes, without altering their size or shape (Kwiatkowska et al. 2014). Another study found that GBH induced an increase in methemoglobin levels, lipid peroxidation products, and hemolysis at concentrations of 500 ppm and 1500 ppm, whereas glyphosate caused an increase in methemoglobin levels and lipid peroxidation only at a higher dose of 1000 ppm, and hemolysis was observed at 1500 ppm. GBH were more likely to affect erythrocyte function than glyphosate, possibly due to the properties of the additives (Pieniążek et al. 2004). Moreover, one study demonstrated that GBH caused 100% hemolysis within the concentration limit recommended for agricultural purposes (approximately 200 ppm of Roundup) (Rodrigues et al. 2011).
Animal studies have also been conducted to investigate the effects of glyphosate and GBH on erythrocytes. One such study examined the sublethal toxicity of glyphosate herbicide on koi carp fingerlings at water concentrations of 3.6 mg/L and 6.6 mg/L, and found that exposure to glyphosate resulted in a significant decrease in RBC counts, Hb, Hct, and MCHC, along with an increase in MCV and MCH (Mzengereza et al. 2020). Moreover, a study investigating the effects of GBH on common carp by exposing them to water concentrations equivalent to 1 and 5 mg/L of glyphosate, found a decrease in RBC counts and an increase in inflammation in the exposed carp. The study suggests that hematological parameters are more sensitive and reliable indicators of Roundup exposure in common carp than other parameters tested (Bojarski et al. 2022). Another study reported that oral administration of 375 mg/kg of GBH for 56 days induced severe anemia, altered platelets and white blood cells, and caused inflammation and oxidative stress in mice. However, supplementation with vitamin B12 attenuated these effects, suggesting a potential role for antioxidants in counteracting GBH-induced toxicity (Ngatuni et al. 2022). Additionally, a study demonstrated that intraperitoneal administration of glyphosate at levels of 25 and 50 mg/kg was clastogenic and cytotoxic to mouse bone marrow, potentially leading to decreased erythrocyte production in the mice (Prasad et al. 2009). These findings provide evidence that glyphosate and GBH may have harmful effects on erythrocytes, which may lead to anemia and inflammation in animals.
Most research investigating the effects of glyphosate exposure on humans has focused on the consequences of intentional or unintentional poisoning. One case report highlighted a woman who suffered from hypoxemia, hyperkalemia, hypotension, and hemoconcentration after accidentally ingesting 100 mL of GBH (Ozaki et al. 2017). In 2015, the European Food Safety Authority conducted a review to evaluate the potential risks associated with glyphosate use and established several toxicological endpoints based on laboratory studies conducted in rabbits (Soares et al. 2021). Current epidemiologic evidence suggests that the use of glyphosate at recommended concentrations is improbable to result in detrimental health effects, including changes in hematologic parameters (Agostini et al. 2020). While there are a limited number of epidemiologic studies examining the potential negative impact of glyphosate/GBH exposure on RBC, one occupational study compared 178 workers Table 2 Linear regression coefficients (standard error) of erythrograms with one unit increase in ln-glyphosate (µg/L) in multiple linear regression models, with results weighted for sampling strategy Model 1 adjusted for age, gender, ethnicity, smoking status, drinking status, household income, BMI, and urinary creatinine Model 2 adjusted for model 1 plus chronic kidney disease, and treatment for anemia in three months, exposed to glyphosate with 203 nonexposed administrative workers. The study concluded that there were no significant differences in the rate of routine blood abnormalities between the two groups. However, when the time-weighted average concentration of glyphosate in the air was below 9.40 mg/m 3 , there was a notable difference in the platelet distribution width and RBC count results between the two groups (Zhang et al. 2019). Using data from NHANES, the present study found a negative correlation between glyphosate levels and Hb as well as Hct. Considering the widespread detection of urinary glyphosate and the routine application of herbicides to crops consumed by individuals (Ashley-Martin et al. 2023;Ospina et al. 2022), it is possible that the exposure dosage of glyphosate and GBH in U.S. adults may exceed expected levels and may lead to detrimental effects on erythrocytes. Exposure to glyphosate has been linked to the development of anemia through potential mechanisms such as hemolysis and bone marrow suppression in  animal studies (Kwiatkowska et al. 2014;Prasad et al. 2009). Moreover, glyphosate exposure may disrupt the gut microbiome, which has been linked to intestinal iron status (Hu et al. 2021;Yilmaz and Li 2018). Our investigation provides preliminary evidence that there may be an association between glyphosate exposure and erythrograms in adults.
Our study found that the relationship between glyphosate exposure and Hb was more pronounced in individuals who were older, had ethnic backgrounds other than non-Hispanic whites, had lower income, and had a BMI < 25 and ≥ 30. To our knowledge, no studies have been conducted to investigate potential differences in the effects of glyphosate among various ethnic groups and socioeconomic statuses. Nonetheless, discrepancies in genetics, lifestyle factors such as diet, occupation, and environment could all play a role in the varying levels of glyphosate toxicity observed among individuals from different ethnic backgrounds and income levels. Prior epidemiological investigations have demonstrated that obesity can disrupt iron homeostasis through chronic inflammation, leading to iron deficiency anemia (Alshwaiyat et al. 2021;Qin et al. 2013), while low body weight is also a risk factor for anemia (Sumarmi et al. 2016). Glyphosate has also been found to modify the gut microbiome, which can affect nutrient metabolism and energy expenditure (Aoun et al. 2020;Hu et al. 2021). It is plausible that there is a complex interplay and synergistic impact between glyphosate and BMI on anemia. Our research has identified demographic subgroups that may be vulnerable to the adverse effects of glyphosate exposure on erythrocytes. Further research is required to comprehend the potential mechanisms behind such differences.
There are several limitations that must be considered when interpreting the findings of this study. Firstly, it is important to note that the observed negative association between glyphosate levels and Hb/Hct does not imply causation. Other variables, such as dietary factors, iron status, or additional environmental exposures, may also influence this correlation. Secondly, since this study used a cross-sectional design, it provides only a snapshot of glyphosate exposure and hemoglobin levels at one point in time and cannot establish a causal relationship. Therefore, further research, including longitudinal studies that follow participants over time, is necessary to verify and expand our understanding of the potential relationship between glyphosate exposure and erythrocytes. Thirdly, the sample size was limited to NHANES 2013-2014 data, which may have restricted the scope of the analysis and could have affected the reliability of the findings. Lastly, the study only included only adult participants from the United States, which may limit the generalizability of the results to other populations and regions.

Conclusions
After conducting our research in a subset of representative cohort of adults in the United States, we found important evidence of a correlation between urinary glyphosate levels and erythrograms. Specifically, our study showed a significant negative association between glyphosate exposure and Hb and Hct levels. Furthermore, our results indicated that the likelihood of anemia was significantly higher in individuals with higher glyphosate levels. While further investigation is necessary to determine the causality and clinical significance of these findings, our study highlights the importance of continued research on the potential toxic effects of glyphosate on erythrocytes in adult populations. By informing public health policy and regulatory decisions related to glyphosate use, such research has the potential to contribute to the protection of human health.