Epidemioiogy The company does not accept there is credihie evidence that giyphosate use can cause NHL. The panel classified the epidemiological data as ?limited evidence? based on associations in some case- controi studies (studies of voiunteers where exposure data is coilected after NHL has been diagnosed for cases, but without such a precipitating event for controls). Such studies are weil known to be prone to a number of biases, especiaiiy enhanced recali of exposures by cases. The IARC panei makes ciear, in the iargest and singie most important study into the health of pesticide appiicators (the Agricuitural Health Study (AHS) or Ag Health Study), that there was no iink with NHL (De Roos et ai., 2005). The AHS is the iargest cohort study conducted of approximateiy 60,000 iicensed pesticide applicators that was set up in the 19905 to provide a iarge unbiased set of data to examine cancer and other heaith risks in pesticide appiicators. in the Ag Health Study, information on pesticide use was coilected before foiiow-up for cancer and other outcomes. In this study, there was no greater risk of NHL in applicators when their cancer experience was compared to State cancer incidence rates (Koutros et ai, 2010), no greater risk of NHL in giyphosate users compared to non-users (De Roos et al, 2005), and NHL risk did not increase with amount of giyphosate use (De Roos et ai, 2005). AHS AHS began in 1993. it is a coliaboration of the US EPA, the Nationai lnStitute of Environmentai Health Sciences the Natiunai Cancer institute and the Natipnai institute for Qccupationai Safety and Neaith The EPA plan to use the results from the AHS in their registration reviews. it is quite surprising that the resuits of the AHS findings of no excess of NHL as discussed in the De Roos et al., 2005 pubiication didn?t drive a conciusion of no evidence of NHL, particuiariy since Aaron Biair, the chair of the IARC panei and member of the Epidemioiogy Workgroup, was one ofthe (Io-Principal investigators that started the AHS study, is on its Executive Committee and was a co-author with De Roos on the 2005 pubiication. From an interview with Arron Biair: ?This is the work Dr. Blair is proudest of after decades of cancer research. This sequence of carefuiiy pianned studies eventually led to the extraordinarily productive Agriculturai Health Study, which foliowed more than 89,000 individuais iiving on farms or appiying pesticides commerciaily in North Carolina and Iowa, and resulted in dozens ofarticies pubiished in major scientificjournals. The great strength of this project is that uniike much other important research on cancer, it does not depend on individuals? to recail exposures. It surveys those with a high likeiihood of exposure and moves forward through time, asking questions and recording disease as it occurs. This type of experimental design is caiied a cohort study, and where it is possible, it is abie to eiiminate many types of bias and confounding factors that piague other experimental designs.? rchiid ren .Word pressco