Glyphosate Facts

Transparency on safety aspects and use of glyphosate-containing herbicides in Europe


... Meta-analysis connects glyphosate with non-Hodgkin lymphona

This is not the first claim of an epidemiological association between glyphosate and NHL. However, no plausible mechanisms for a causal effect have been postulated, and there has been no evidence of such a link from toxicological studies. Even in well designed epidemiology and meta-studies with statistical differences differentiating between “correlation” and “causal” is fraught with difficulties. The coupled role of epidemiology and toxicology in discerning human health effects by environmental agents is obvious, but there is currently no systematic and transparent way to bring the data and analysis of the two disciplines together in a way that provides a unified view on an adverse causal relationship between an agent and a disease (Adami et al 2011):

  • De Roos et al., 2003; Hardell et al., 2002; and McDuffie et al., 2001 all report an increased risk of NHL following glyphosate exposure. It is important to note that, because these studies are retrospective in design, they are susceptible to recall bias when reporting exposure to glyphosate. De Roos et al. 2005, state recall bias, characterized as inaccurate recollections by study participants, as a reason for their discrepant finding regarding no association between glyphosate and NHL in some earlier studies.
  • Additionally, the epidemiological studies used for the meta-analysis performed by Schinasi et al. associating glyphosate and NHL lack any direct toxicological evidence. Furthermore, while the meta-analysis may support the hypothesis that pesticides are associated with NHL; the studies lack sufficient qualitative and quantitative information on pesticide exposure and other information on risk factors including the possible influence of other occupational, environmental, lifestyle, or genetic factors for hematopoietic cancers to identify specific causes.
  • Finally, while it is reasonable to believe that this meta-analysis suffers from publication bias, characterized as undue influence by smaller studies included in the meta-analysis, no efforts were made by Schinasi et al. to assess the potential publication bias in this meta-analysis.
  • Eriksson et al., 2008 conducted a population-based case-control study of exposure to a variety of pesticides and NHL or several histopathological categories of NHL (e.g., B cell lymphoma). Remembering that odds ratios (OR) above 1 suggest an association but are not significant if the confidence intervals includes the null value of 1.0. Eriksson et al. report ORs for glyphosate exposure of <10 days and >10 days to be 1.69 (95% CI: 0.70-4.07) and 2.36 (1.04-5.37), respectively. The ORs for “latency” periods of 1-10 years and >10 years were 1.11 (95% CI: 0.24-5.08) and 2.26 (95% CI: 1.16-4.40), respectively. The odds ratios for the other types (total B-cell lymphomas, grade I-III follicular lymphoma, diffuse large B-cell lymphoma, other specified B-cell lymphoma, unspecified B-cell lymphoma, and T-cell lymphomas) were above 1.0, but were not statistically significant. From this the authors concluded “glyphosate was associated with a statistically significant increased OR for lymphoma in our study…” However, the interpretation of the results of this study is hindered by potential problems including referral, selection, or recall (other information) biases, evidence for a causal relationships is based on a very weak association, and confounding factors such as exposure to other pesticides (Mink Review).


Reference to the meta-study in question:

L. Schinasi and M.E. Leon (2014). Non-Hodgkin Lymphoma and Occupational Exposure to Agricultural Pesticide Chemical Groups and Active Ingredients: A Systematic Review and Meta-Analysis. Int J Environ Res Public Health; 11(4): 4449–4527.

Last update: 10 July 2014