Researchers Find “Compelling Link” Between Roundup Exposure and Non-Hodgkin Lymphoma

​In this 2019 study, researchers at the University of California, Berkeley, University of Washington and the Icahn School of Medicine at Mount Sinai, New York, reported what they call a “compelling link” between exposure to a glyphosate-based herbicide (GBH) like Roundup and non-Hodgkin lymphoma.

The scientists, led by Luoping Zhang of the UC Berkeley School of Public Health, combined data from several previous studies of GBH exposure in an analytical process known as meta-analysis. Meta-analysis increases the number of subjects being studied and thereby enhances the ability of the research to detect an effect that might otherwise remain hidden or unclear. The investigators concluded that exposure to a glyphosate-based herbicide increases the risk of non-Hodgkin lymphoma by 41%.

Zhang’s study was unique in that she and her colleagues focused on individuals in each study who were reported to have had the highest exposure to GBHs. Higher and longer exposures, they reasoned, would be more likely to reveal the true risk of GBH-associated NHL.

This is particularly true, they noted, in a world where most people have been exposed to glyphosate directly or indirectly. The reason is relatively simple: to determine if a substance has negative health consequences, scientists need to compare those exposed to the substance with a control group of individuals not exposed. But where do you find that control group if everyone is exposed?

The AHS was a cohort study. In cohort studies, individuals exposed to a substance thought to increase disease risk and a related group of people not exposed to the risk factor are followed for a defined period of time. Researchers then determine if the incidence of the disease was greater in the exposed group.

In the AHS, the exposed group consisted of licensed pesticide applicators who were enrolled in the study from 1993 – 1997. A 2005 study based on the AHS results found no statistically significant link between glyphosate exposure. A second study published in 2018, which followed some subjects up to 2013, also reported no link between glyphosate-based herbicide and cancer. Indeed, the second study suggested that GBHs decreased the risk of NHL.

However, several well-respected scientists—including scientists at Monsanto and one of the authors of the original 2005 AHS study (Aaron Blair, who worked for the National Cancer Institute at the time)—said the AHS study was fundamentally flawed due to non-differential exposure misclassification. Non-differential exposure misclassification occurs when individuals who have been exposed to the toxin being studied, in this case GBHs, are misclassified as non-exposed – or the reverse, non-exposed subjects being classified as exposed. Those who acquire the disease being studied and those who do not may both be misclassified. Non-differential exposure misclassification will bias the results of a study in the direction of no effect. It will appear that there is no association, even if one exists. In her testimony during the Pilliod trial, Dr. Beate Ritz, chair of the Department of Epidemiology at UCLA, provided detailed analysis of the problems with the Agricultural Health Study (beginning on pdf p. 100 of the Day 5 transcript.)

The Zhang study, by combining data from multiple studies, including the AHS, but focusing on the high-exposure groups, revealed the association that was obscured by the exposure misclassification of the AHS study.

Zhang’s results are supported by a new French study, published in March 2019 in the International Journal of Epidemiology. This study of multiple pesticides (herbicides like Roundup are categorized within the pesticide group) pooled data from three large cohort studies of agricultural workers, the AHS and two other studies, one in France and one in Norway. The authors of this study did not attempt to identify different levels of exposure. Exposure was categorized as either exposed or never exposed. In addition, pesticide use among the Norwegian and French cohorts (63% and 68 % respectively) was much lower than the AHS cohort (99%). Nevertheless, while not finding an association between glyphosate and NHL, authors report a statistically significant association between any exposure to glyphosate and risk for the most common type of NHL— diffuse large B-cell lymphoma (DLBCL). Glyphosate increased the risk by 36%. Out of 14 pesticides and 33 active ingredients analyzed by the French investigators, glyphosate was the only one to show a statistically significant association with diffuse large B-cell lymphoma.