Systematic review and meta-analysis of glyphosate exposure and risk of lymphohematopoietic cancers

Figures & data

Table 1. Design characteristics of studies of glyphosate exposure and risk of lymphohematopoietic cancer (LHC), including non-Hodgkin lymphoma (NHL), NHL subtypes, Hodgkin lymphoma (HL), multiple myeloma (MM), and leukemia.

Authors	Year	Outcomes studied	Study location	Study design	Study years	Source population	Subject identification	Subject participation	Subjects (n)	Proxy respondents
Brown et al.^[Citation35]	1990	Leukemia (including myelodysplasias)	United States (Iowa and Minnesota)	Population-based case-control	1980–1983	White men aged ≥ 30 years in Iowa and Minnesota, excluding Minneapolis, St. Paul, Duluth, and Rochester	Cases: Iowa Tumor Registry and special surveillance of Minnesota hospital and pathology laboratory records Controls: random-digit dialing if aged < 65 years, Medicare files if aged ≥ 65 years, state death certificate files if deceased	Cases: 86% Controls: 77% random digit dialing, 79% Medicare, 77% proxies for deceased Supplemental interview: 93% cases, 96% controls	Cases: 578 Controls: 1,245 Supplemental interview: 86 cases, 203 controls	Cases: 238 (41%) Controls: 425 (34%) Supplemental interview, 63 (73%) cases, 57 (28%) controls
Brown et al.^[Citation32]	1993	MM	United States (Iowa)	Population-based case-control	1981–1984	White men aged ≥ 30 years in Iowa	Cases: Iowa Health Registry Controls: random-digit dialing if aged < 65 years, Medicare files if aged ≥ 65 years, state death certificates if deceased	Cases: 84% Controls: 78% overall	Cases: 173 Controls: 650	Cases: 72 (42%) Controls: 198 (30%)
Cantor et al.^[Citation24]	1992	NHL	United States (Iowa and Minnesota)	Population-based case-control	1980–1983	White men aged ≥ 30 years in Iowa and Minnesota, excluding Minneapolis, St. Paul, Duluth, and Rochester	Cases: Iowa State Health Registry and special surveillance of Minnesota hospital and pathology laboratory records Controls: random-digit dialing if aged < 65 years, Medicare files if aged ≥ 65 years, state death certificate files if deceased	Cases: 89% Controls: 77% random-digit dialing, 79% Medicare, 77% proxies for deceased	Cases: 622 Controls: 1245	Cases: 184 (30%) Controls: 425 (34%)
Cocco et al.^[Citation18]	2013	B-cell NHL	Europe (Czech Republic, France, Germany, Ireland, Italy, and Spain)	Population- and hospital-based case-control	1998–2004	Persons aged ≥ 17 years in Germany and Italy general populations, and in referral areas of participating hospitals in Czech Republic, France, Ireland, and Spain	Cases: NR Controls: random sampling of population registers in Germany and Italy; recruitment from hospital departments for infectious and parasitic (17.6%), mental and nervous (14.6%), circulatory (8.7%), digestive (7.1%), endocrine and metabolic (4.1%), respiratory (3.9%), and several other conditions (33.2%), excluding cancer, in Czech Republic, France, Ireland, and Spain	Cases: 88% overall; 90% Czech Republic, 91% France, 87% Germany, 90% Ireland, 93% Italy, 82% Spain Controls: 69% overall, 81% hospital-based, 52% population-based; 60% Czech Republic, 74% France, 44% Germany, 75% Ireland, 66% Italy, 96% Spain	Cases: 2348 Controls: 2462	None
De Roos et al.^[Citation13]	2003	NHL	United States (Nebraska, Iowa, Minnesota, and Kansas)	Population-based case-control (pooled analysis of 3 studies)	1979–1986	White men aged ≥ 21 years in one of the 66 counties of eastern Nebraska; white men aged ≥ 30 years in Iowa and Minnesota, excluding Minneapolis, St. Paul, Duluth, and Rochester; white men aged ≥ 21 years in Kansas	Cases: Nebraska Lymphoma Study Group and area hospitals; Iowa State Health Registry; special surveillance of Minnesota hospital and pathology laboratory records; University of Kansas Cancer Data Service registry Controls: random-digit dialing if aged < 65 years, Medicare files if aged ≥ 65 years, state death certificate files if deceased	Cases: 91% Nebraska (93% living, 89% deceased); 89% Iowa and Minnesota; 96% Kansas Controls: 85% Nebraska; 77% random-digit dialing, 79% Medicare, 77% deceased (proxies) Iowa and Minnesota; 93% Kansas Analysis restricted to subjects who lived or worked on a farm before 18 years of age (% NR); analysis of multiple pesticides restricted to subjects with non-missing data (75% cases, 75% controls)	Cases: 650 (in analyses of multiple pesticides) Controls: 1933 (in analyses of multiple pesticides)	Cases: 201 (30.9%) (in analyses of multiple pesticides) Controls: 767 (39.7%) (in analyses of multiple pesticides)
De Roos et al.^[Citation12]	2005	LHC, NHL, MM, leukemia	United States (Iowa and North Carolina)	Prospective cohort	1993–1997 through 2001 Median = 6.7 years	Private and commercial pesticide applicators in Iowa and North Carolina who were licensed to apply restricted-use pesticides	Pesticide applicators identified when seeking a state-issued restricted-use pesticide license; invited to complete the enrollment questionnaire at the licensing facility	298 subjects (0.5%) lost to follow-up or with no person-time contributed > 80% of eligible pesticide applicators enrolled in study by completing on-site questionnaire 44% of applicators completed take-home questionnaire	Eligible cohort: 36,509–49,211 in analyses adjusted for demographics and lifestyle 30,613–40,719 in analyses additionally adjusted for other pesticides	None
Eriksson et al.^[Citation14]	2008	NHL, B-cell NHL, SLL/CLL, FL grades I-III, DLBCL, other specified B-cell NHL, unspecified B-cell NHL, T-cell NHL, unspecified NHL	Europe (Sweden)	Population-based case-control	1999–2002	Adults aged 18–74 years in 4 of 7 health service regions in Sweden associated with university hospitals in Lund, Linköping, Örebro, and Umeå	Cases: contact with treating physicians and pathologists Controls: national population registry	Cases: 81% Controls: 65% (92% of initially enrolled controls with 71% participation)	Cases: 995 Controls: 1016	None
Hardell and Eriksson^[27]	1999	NHL	Europe (Sweden)	Population-based case-control	1987–1990	Men aged ≥ 25 years in the four northernmost counties of Sweden and three counties in mid-Sweden	Cases: regional cancer registries Controls: national population registry if living, national registry for causes of death if deceased	Cases: 91% (91% living, 92% deceased) Controls: 84% (83% living, 85% deceased)	Cases: 404 Controls: 741	Cases: 177 (44%) Controls: NR (∼44%; matched to cases)
Hardell et al.^[Citation15]	2002	NHL including hairy-cell leukemia	Europe (Sweden)	Population-based case-control	1987–1990	Men aged ≥ 25 years in the four northernmost counties of Sweden and three counties in mid-Sweden (for NHL) or in the entire country of Sweden (for hairy-cell leukemia)	Cases: regional cancer registries for NHL, national cancer registry for hairy-cell leukemia Controls: national population registry, national registry for causes of death if deceased	Cases: 91% Controls: 84%	Cases: 515 Controls: 1141	Cases: ∼35% (NR) Controls: ∼29% (NR)
Hohenadel et al.^[Citation28]	2011	NHL	Canada (Alberta, British Columbia, Manitoba, Ontario, Quebec, and Saskatchewan)	Population-based case-control	1991–1994	Men aged ≥ 19 years in Alberta, British Columbia, Manitoba, Ontario, Quebec, and Saskatchewan	Cases: hospital records in Quebec, cancer registries in all other provinces Controls: provincial health insurance records in Alberta, Saskatchewan, Manitoba, and Quebec; computerized telephone listings in Ontario; voter lists in British Columbia	Cases: 67% Controls: 48% Based on postal codes, respondents were not more or less likely than non-respondents to live in a rural area.	Cases: 513 Controls: 1506	Cases: 110 (21%) Controls: 220 (15%)
Kachuri et al.^[Citation33]	2013	MM	Canada (Alberta, British Columbia, Manitoba, Ontario, Quebec, and Saskatchewan)	Population-based case-control	1991–1994	Men aged ≥ 19 years (≥ 30 years in analysis) in Alberta, British Columbia, Manitoba, Ontario, Quebec, and Saskatchewan	Cases: hospital records in Quebec, cancer registries in all other provinces Controls: provincial health insurance records in Alberta, Saskatchewan, Manitoba, and Quebec; computerized telephone listings in Ontario; voter lists in British Columbia	Cases: 58% Controls: 48% Based on postal codes, respondents were not more or less likely than non-respondents to live in a rural area.	Cases: 342 Controls: 1357	Cases: 103 (30%) Controls: 202 (15%)
Karunanayake et al.^[Citation31]	2012	HL	Canada (Alberta, British Columbia, Manitoba, Ontario, Quebec, and Saskatchewan)	Population-based case-control	1991–1994	Men aged ≥ 19 years in Alberta, British Columbia, Manitoba, Ontario, Quebec, and Saskatchewan	Cases: hospital records in Quebec, cancer registries in all other provinces Controls: provincial health insurance records in Alberta, Saskatchewan, Manitoba, and Quebec; computerized telephone listings in Ontario; voter lists in British Columbia	Cases: 68% Controls: 48% Based on postal codes, respondents were not more or less likely than non-respondents to live in a rural area.	Cases: 316 Controls: 1506	Cases: NR Controls: 220 (15%)
Kaufman et al.^[Citation36]	2009	Leukemia	Bangkok, Thailand	Hospital-based case-control	1997–2003	Patients aged ≥ 18 years residing in Bangkok proper and suburbs of Nonthaburi, Nakornpathom, Patumthani, Samutprakarn, and Samusakorn, admitted to Siriraj Hospital or Dhonburi Hospital	Cases: hospital records Controls: hospital records for acute infection or inflammation (33%), trauma (22%), acute abdominal emergencies such as appendicitis (27%), or various other diagnoses with elective admission, such as cataract, hernia repair, or cosmetic surgery (17%), excluding head trauma with loss of consciousness or cancer; controls at Dhonburi Hospital (a nearby private hospital) matched to 21 cases admitted to private wards for wealthy patients	Cases: 100% Controls: 100%	Cases: 180 Controls: 756	None
Lee et al.^[Citation29]	2004	NHL	United States (Nebraska, Iowa, and Minnesota)	Population-based case-control (pooled analysis of 2 studies)	1980–1986	White men and women aged ≥ 21 years in one of 45 counties in eastern Nebraska; white men aged ≥ 30 years in Iowa and Minnesota, excluding Minneapolis, St. Paul, Duluth, and Rochester	Cases: Nebraska Lymphoma Study Group and area hospitals; Iowa State Health Registry; special surveillance of Minnesota hospital and pathology laboratory records Controls: random-digit dialing if aged < 65 years, Medicare files if aged ≥ 65 years, state death certificate files if deceased	Cases: 91% Nebraska, 89% Iowa and Minnesota Controls: 85% Nebraska, 78% Iowa and Minnesota	Cases: 872 Controls: 2336	Cases: 266 (31%) Controls: 779 (33%)
McDuffie et al.^[Citation16]	2001	NHL	Canada (Alberta, British Columbia, Manitoba, Ontario, Quebec, and Saskatchewan)	Population-based case-control	1991–1994	Men aged ≥ 19 years in Alberta, British Columbia, Manitoba, Ontario, Quebec, and Saskatchewan	Cases: hospital records in Quebec, cancer registries in all other provinces Controls: provincial health insurance records in Alberta, Saskatchewan, Manitoba, and Quebec; computerized telephone listings in Ontario; voter lists in British Columbia	Cases: 67% Controls: 48% Based on postal codes, respondents were not more or less likely than non-respondents to live in a rural area.	Cases: 517 Controls: 1506	Cases: ∼21% (NR) Controls: 220 (15%)
Nordström et al.^[Citation30]	1998	Hairy-cell leukemia	Europe (Sweden)	Population-based case-control	1987–1992 (1993 for one case)	Men living in Sweden	Cases: national cancer registry Controls: national population registry	Cases: 91% Controls: 83%	Cases: 111 Controls: 400	Cases: 4 (4%) Controls: 5 (1%)
Orsi et al.^[Citation17]	2009	LHC, NHL, DLBCL, FL, LPS, CLL, hairy-cell leukemia, HL, MM	Europe (France)	Hospital-based case-control	2000–2004	Men aged 20–75 years living in the catchment areas of the main hospitals in Brest, Caen, Nantes, Lille, Toulouse, and Bordeaux, with no history of immunosuppression or taking immunosuppressant drugs	Cases: hospital records Controls: hospital records for orthopedic or rheumatological conditions (89.3%), gastrointestinal or genitourinary tract diseases (4.8%), cardiovascular diseases (1.1%), skin and subcutaneous tissue disease (1.8%), and infections (3.0%), excluding patients admitted for cancer or a disease directly related to occupation, smoking, or alcohol abuse	Cases: 95.7% Controls: 91.2%	Cases: 491 LHC, 244 NHL, 104 LPS, 87 HL, 56 MM Controls: 456	None
Pahwa et al.^[Citation34]	2012	MM	Canada (Alberta, British Columbia, Manitoba, Ontario, Quebec, and Saskatchewan)	Population-based case-control	1991–1994	Men aged ≥ 19 years in Alberta, British Columbia, Manitoba, Ontario, Quebec, and Saskatchewan	Cases: hospital records in Quebec, cancer registries in all other provinces Controls: provincial health insurance records in Alberta, Saskatchewan, Manitoba, and Quebec; computerized telephone listings in Ontario; voter lists in British Columbia	Cases: 58% Controls: 48% Based on postal codes, respondents were not more or less likely than non-respondents to live in a rural area.	Cases: 342 Controls: 1506	Cases: 103 (30%) Controls: 220 (15%)
Sorahan^[26]	2015	MM	United States (Iowa and North Carolina)	Prospective cohort	1993–1997 through 2001 Median = 6.7 years	Private and commercial pesticide applicators in Iowa and North Carolina who were licensed to apply restricted-use pesticides	Pesticide applicators identified when seeking a state-issued restricted-use pesticide license; invited to complete the enrollment questionnaire at the licensing facility	298 subjects (0.5%) lost to follow-up or with no person-time contributed > 80% of eligible pesticide applicators enrolled in study by completing on-site questionnaire 44% of applicators completed take-home questionnaire	Eligible cohort (1): 54,315 excluding subjects with cancer before enrollment, loss to follow-up, missing age at enrollment, or missing glyphosate use 49,211 also excluding missing education, smoking, or alcohol 40,719 excluding missing other pesticides Eligible cohort (2): 53,656 excluding subjects with cancer before enrollment, loss to follow-up, missing age at enrollment, missing glyphosate use, or missing cumulative exposure days of glyphosateuse 53,304 also excluding missing intensity of glyphosate use Eligible cohort (3): 55,934 excluding subjects with cancer before enrollment, loss to follow-up, or missing age at enrollment	None

Table 1. Continued (additional columns).

Authors	Year	Exposure assessment	Outcome assessment	Investigator blinding	Confounders considered or adjusted	Funding source	Overlap
Brown et al.^[Citation35]	1990	In-person structured interview, including detailed farming and pesticide use history For each pesticide, evaluated ever use, first and last year of use, and personal applying/mixing/handling In 1987, supplemental telephone interview to evaluate usual number of days of pesticide use per year among Iowa subjects who had reported agricultural use of specific pesticides	Diagnostic confirmation by regional pathologists; special review of myelodysplasias by one pathologist co-author	No	Adjusted: vital status, age, state, ever used tobacco daily, first-degree family history of LHC, non-farming job related to leukemia risk in this study, exposure to substances (benzene, naphtha, hair dyes) related to leukemia risk in this study	Partial support from National Institute of Environmental Health Sciences	Brown et al.^[Citation32]; Cantor et al.^[Citation24]; De Roos et al.^[Citation13]; Lee et al.^[Citation29]
Brown et al.^[Citation32]	1993	In-person structured interview, including detailed farming and pesticide use history For each pesticide, evaluated ever use, first and last year of use, personal applying/mixing/handling, and use of protective equipment	Diagnostic confirmation by an expert pathologist	No	Adjusted: vital status, age Considered: smoking, education, other factors found not to be confounders of agricultural risk factors	Partial support from National Institute of Environmental Health Sciences	Brown et al.^[Citation35]; Cantor et al.^[Citation24]; De Roos et al.^[Citation13]; Lee et al.^[Citation29]
Cantor et al.^[Citation24]	1992	In-person structured interview, including detailed farming and pesticide use history of all subjects who had worked on a farm for ≥ 6 months since age 18 years For each pesticide, evaluated ever use, first and last year of use, method of application, personal applying/mixing/handling, and use of protective equipment	Diagnostic confirmation and morphological classification by panel of 4 experienced regional pathologists	No	Adjusted: vital status, state, age, cigarette smoking status, first-degree family history of LHC, non-farming job related to NHL risk in this study, exposure to hair dyes, exposure to other substances associated with NHL risk in this study Considered: pesticides belonging to other chemical families	Partial support from National Institute of Environmental Health Sciences	Brown et al.^[Citation35]; Brown et al.^[Citation32]; De Roos et al.^[Citation13]; Lee et al.^[Citation29]
Cocco et al.^[Citation18]	2013	In-person structured interview, including detailed farming and pesticide use history for all subjects who reported having worked in agriculture For each agricultural job, reported tasks, crops, size of cultivated area, pests treated, pesticides used, crop treatment procedures, use of personal protective equipment, re-entry after treatment, and frequency of treatment in days per year	Histologically or cytologically confirmed cases with central review of slides of ∼20% by an international team of pathologists	No	Adjusted: age, gender, education, study center	European Commission, 5th and 6th Framework Programmes; Spanish Ministry of Health; German Federal Office for Radiation Protection; La Fondation de France; Italian Ministry for Education, University and Research; Italian Association for Cancer Research	None
De Roos et al.^[Citation13]	2003	Telephone interview in Nebraska and Kansas; in-person structured interview in Iowa and Minnesota Nebraska: Question about use of any pesticide, followed by prompting for specific selected pesticides, including years of use and average days per year Iowa and Minnesota: Direct question about a selected use of specific pesticides, including first and last years of use Kansas: Open-ended question about use of pesticides, followed by questions on duration of use and days per year for groups of pesticides but not individual pesticides (with validation study)	Nebraska: Pathology review with histological confirmation and classification including immunologic phenotyping Iowa and Minnesota: Diagnostic confirmation and morphological classification by panel of 4 experienced regional pathologists Kansas: Diagnostic confirmation and classification by panel of 3 pathologists	Yes in Nebraska; no in Iowa, Minnesota, and Kansas	Adjusted: age, study site, other individual pesticides with ≥ 20 users in full study Considered: first-degree family history of LHC, education, smoking	NR; assume National Cancer Institute	Brown et al.^[Citation35]; Brown et al.^[Citation32]; Cantor et al.^[Citation24]; Lee et al.^[Citation29] (also Hoar et al.^[Citation47]; Hoar Zahm et al.^[Citation48])
De Roos et al.^[Citation12]	2005	Self-administered written questionnaire (with validation study) evaluating detailed use of 22 pesticides for private applicators, 28 pesticides for commercial applicators (ever/never use, frequency, duration, and intensity of use, decade of first use), and ever/never use for additional pesticides up to total of 50, with general information on pesticide application methods, personal protective equipment, pesticide mixing, and equipment repair Additional self-administered take-home questionnaire with further questions on occupational exposures and lifestyle factors	Linkage to state cancer registry files, state death registries, and National Death Index	None	Adjusted: age at enrollment, education, cigarette smoking pack-years, alcohol consumption in past year, first-degree family history of cancer, state of residence Considered (adjusted for MM only): 5 pesticides for which cumulative exposure-days were most highly associated with those for glyphosate (i.e., 2,4-dichlorophenoxyacetic acid, alachlor, atrazine, metolachlor, trifluralin), 5 pesticides for which ever/never use was most highly associated with that for glyphosate (i.e., benomyl, maneb, paraquat, carbaryl, diazinon)	National Cancer Institute, National Institute of Environmental Health Sciences, Environmental Protection Agency, and National Institute for Occupational Safety and Health	Sorahan^[26]
Eriksson et al.^[Citation14]	2008	Self-administered mailed questionnaire with additional telephone interview for missing or unclear answers; evaluated occupational exposure to individual pesticides, including number of years, number of days per year, and approximate length of exposure per day	Diagnostic pathological specimens examined and classified by 1 of 5 Swedish expert lymphoma reference pathologists, if not already initially reviewed by one of them; panel review if classification differed from original report	Yes	Adjusted: age, sex, and year of diagnosis or enrollment; other associated agents (4-chloro-2-methyl phenoxyacetic acid, 2,4-dichlorophenoxyacetic acid and/or 2,4,5-trichlorophenoxyacetic acid, mercurial seed dressing, arsenic, creosote, tar) for NHL only	Swedish Council for Working Life and Social Research; Cancer and Allergy Fund; Key Fund; Örebro University Hospital Cancer Fund	None
Hardell and Eriksson^[27]	1999	Self-administered mailed questionnaire with supplemental telephone interview for unclear answers; assessed use of pesticides within different occupations, wet contact if not handling the sprayer, brand names of pesticides, years of exposure, and cumulative days of exposure Exposure excluded 1 year prior to diagnosis or index year	Histopathological diagnosis of NHL reported to regional cancer registries, confirmed by review of pathology reports	Yes	Adjusted: age, county, vital status, year of death if deceased, use of phenoxyacetic acids	Swedish Work Environment Fund, Swedish Medical Research Council, Örebro County Council Research Committee, Örebro Medical Center Research Foundation	Hardell et al.^[Citation15]
Hardell et al.^[Citation15]	2002	Self-administered mailed questionnaire with supplemental telephone interview for unclear answers; assessed years and total number of days of occupational exposure to various agents and names of agents Exposure defined as ≥ 1 working day with induction period of ≥ 1 year	Histologically verified NHL; confirmation of hairy-cell leukemia NR	Yes	Adjusted: study, study area, vital status, other associated pesticides (4-chloro-2-methyl phenoxyacetic acid, 2,4-dichlorophenoxyacetic acid + 2,4,5-trichlorophenoxyacetic acid, other herbicides)	Swedish Cancer Research Fund, Swedish Medical Research Council, Örebro County Council Research Committee, Örebro Medical Centre Research Foundation	Hardell and Eriksson^[27] Nordström et al.^[Citation30]
Hohenadel et al.^[Citation28]	2011	Telephone interview for detailed information on pesticide use in subjects who reported in a self-administered mail questionnaire that they had ≥ 10 hours of pesticide use during their lifetime, plus 15% random sample of subjects with < 10 hours Pesticide interview (with validation study) included a pre-mailed list of specific pesticides (chemical and trade names) with number of days used and number of hours per day at home or work for each pesticide	Diagnostic confirmation based on information, including pathology reports, from cancer registries and hospitals; pathological material reviewed and classified by a reference pathologist; subjects with unavailable pathological material retained in study	No	Adjusted: age, province, use of a proxy respondent Considered: diesel exhaust, ultraviolet radiation, farm animals, chemicals such as benzene, first-degree family history of cancer	Health Canada, British Columbia Health Research Foundation, Centre for Agricultural Medicine at University of Saskatchewan	Kachuri et al.^[Citation33]; Karunanayake et al.^[Citation31]; McDuffie et al.^[Citation16]; Pahwa et al.^[Citation34]
Kachuri et al.^[Citation33]	2013	Telephone interview for detailed information on pesticide use in subjects who reported in a self-administered mail questionnaire that they had ≥ 10 hours of pesticide use during their lifetime, plus 15% random sample of subjects with < 10 hours Pesticide interview (with validation study) included a pre-mailed list of specific pesticides (chemical and trade names) with number of days used and number of hours per day at home or work for each pesticide	Diagnostic confirmation based on information, including pathology reports, from cancer registries and hospitals; pathological material reviewed and classified by a reference pathologist (including pathology and tumor tissue slides for 125 [37%] of 342 cases); subjects with unavailable pathological material retained in study	No	Adjusted: age, province, use of a proxy respondent, smoking status, personal history of rheumatoid arthritis, allergies, measles, shingles, or cancer, family history of cancer	Occupational Cancer Research Centre; Cancer Care Ontario; Ontario Workplace Safety and Insurance Board; Canadian Cancer Society, Ontario Division, Mitacs-Accelerate Graduate Research Internship Program	Hohenadel et al.^[Citation28]; Karunanayake et al.^[Citation31]; McDuffie et al.^[Citation16]; Pahwa et al.^[Citation34]
Karunanayake et al.^[Citation31]	2012	Telephone interview for detailed information on pesticide use in subjects who reported in a self-administered mail questionnaire that they had ≥ 10 hours/year of cumulative exposure to any combination of herbicides, insecticides, fungicides, fumigants, and algicides Pesticide interview collected information on exposure to individual pesticides, place of pesticide use, year of first use, first year on market, number of years of use, and days per year of use [Note differences from related studies]	Initial diagnosis based on information from cancer registries and hospitals; pathology and tumor tissue slides for 155 of 316 cases reviewed by a reference pathologist who confirmed HL in 150/155 cases, plus 7 cases originally classified as NHL; subjects with unavailable pathological material retained in study	No	Adjusted: age, province, personal history of measles, acne, hay fever, or shingles, first-degree family history of cancer	NR; assume same as in related studies	Hohenadel et al.^[Citation28]; Kachuri et al.^[Citation33]; McDuffie et al.^[Citation16]; Pahwa et al.^[Citation34]
Kaufman et al.^[Citation36]	2009	Interview with nurse to assess occupational and non-occupational exposure to pesticides and other potential risk factors	Histologically confirmed leukemia diagnosed within 6 months before current hospital attendance or admission	No	Considered: age, sex, income, use of cellular telephones, benzene and other solvent exposure, occupational and non-occupational pesticide exposure, pesticides used near home, working with power lines, living near power lines, exposure to X-rays, exposure to certain types of electromagnetic fields, use of hair dyes	Thailand Research Fund and Commission on Higher Education	None
Lee et al.^[Citation29]	2004	Telephone interview in Nebraska; in-person structured interview in Iowa and Minnesota Questions included personal handling of groups of pesticides and individual pesticides used on crops or animals, with years of first and last use	Nebraska: Pathology review with histological confirmation and classification including immunologic phenotyping Iowa and Minnesota: Diagnostic confirmation and morphological classification by panel of 4 experienced regional pathologists	Yes in Nebraska; no in Iowa and Minnesota	Adjusted: age, state, vital status Considered: gender, smoking, first-degree family history of LHC, ever having a job correlated with risk of LHC (e.g., painting or welding), use of protective equipment	NR; assume National Cancer Institute	Brown et al.^[Citation35]; Brown et al.^[Citation32]; Cantor et al.^[Citation24]; De Roos et al.^[Citation13] (also Hoar Zahm et al.^[Citation48])
McDuffie et al.^[Citation16]	2001	Telephone interview for detailed information on pesticide use in subjects who reported in a self-administered mail questionnaire that they had ≥ 10 hours of pesticide use during their lifetime, plus 15% random sample of subjects with < 10 hours (total = 179 cases, 456 controls with telephone interview) Pesticide interview (with validation study) included a pre-mailed list of specific pesticides (chemical and trade names) with number of days used and number of hours per day at home or work for each pesticide	Diagnostic confirmation from cancer registries and hospitals; pathological material reviewed and classified by a reference pathologist; subjects with unavailable pathological material retained in study	No	Adjusted: age, province, personal history of measles, mumps, cancer, or allergy desensitization shots, first-degree family history of cancer Considered: pesticide exposure, smoking history	Health Canada, British Columbia Health Research Foundation, Centre for Agricultural Medicine at University of Saskatchewan	Hohenadel et al.^[Citation28]; Kachuri et al.^[Citation33]; Karunanayake et al.^[Citation31]; Pahwa et al.^[Citation34]
Nordström et al.^[Citation30]	1998	Self-administered mailed questionnaire with supplemental telephone interview for unclear or missing answers; assessed total number of days of occupational exposure to various agents Exposure defined as ≥ 1 working day with induction period of ≥ 1 year	Reported to national cancer registry; further confirmation not described	Yes	Adjusted: age Considered: exposure to animals, herbicides, insecticides, fungicides, impregnating agents, organic solvents, exhausts, or ultraviolet light	Swedish Work Environment Fund, Örebro County Council Research Committee, Örebro Medical Centre Research Foundation.	Hardell et al.^[Citation15]
Orsi et al.^[Citation17]	2009	Self-administered written questionnaire with lifetime occupational history, followed by in-person structured interview evaluating non-occupational exposure to pesticides and agricultural questionnaire for subjects who had worked as a farmer or gardener for ≥ 6 months during lifetime Agricultural questionnaire collected data on location of all farms where subject had worked for ≥ 6 months, period of occupation and area, farmer's status at each farm, crops and animal husbandry with mean sizes, all pesticides used on each crop during a given period, whether subject had personally prepared, mixed, or sprayed the pesticide, chemical used, brand name, main use, type of spraying equipment used, annual number and duration of applications, and use of pesticides in farm buildings for animals, grain, hay or straw, or to clear lanes and yards All questionnaires reviewed by an occupational hygienist and an agronomist; repeat telephone interviews conducted to clarify information from 95 (56.8%) of 158 subjects who completed the agricultural questionnaire, not completed by 35 (20.8%) who refused (n = 15), died/were in poor health (n = 10), or could not be contacted (n = 15); all chemicals coded using ad hoc system and classified as definite or possible exposure	All diagnoses cytologically or histologically confirmed and reviewed by a panel of pathologists and hematologists	Yes	Adjusted: age, study center, socioeconomic category Considered: all combinations of pesticide families associated with the LHC subtype considered with a p-value ≤ 0.10, rural/urban status, type of housing, educational level, history of mononucleosis, history of influenza immunization, family history of cancer, skin characteristics, smoking status, and alcohol drinking status	Association pour la Recherche contre le Cancer, Fondation de France, AFSSET, Faberge employees (donation)	None
Pahwa et al.^[Citation34]	2012	Telephone interview for detailed information on pesticide use in subjects who reported in a self-administered mail questionnaire that they had ≥ 10 h of pesticide use during their lifetime, plus 15% random sample of subjects with < 10 h Pesticide interview (with validation study) included a pre-mailed list of specific pesticides (chemical and trade names) with number of days used and number of hours per day at home or work for each pesticide	Diagnostic confirmation based on information, including pathology reports, from cancer registries and hospitals; pathological material reviewed and classified by a reference pathologist (including pathology and tumor tissue slides for 125 [37%] of 342 cases); subjects with unavailable pathological material retained in study	No	Adjusted: age, province, personal history of measles, mumps, allergies, arthritis, or shingles, first-degree family history of cancer	Occupational Cancer Research Centre; Cancer Care Ontario; Ontario Workplace Safety and Insurance Board; Canadian Cancer Society	Hohenadel et al.^[Citation28]; Kachuri et al.^[Citation33]; Karunanayake et al.^[Citation31]; McDuffie et al.^[Citation16]
Sorahan^[26]	2015	Self-administered written questionnaire (with validation study) evaluating detailed use of 22 pesticides for private applicators, 28 pesticides for commercial applicators (ever/never use, frequency, duration, and intensity of use, decade of first use), and ever/never use for additional pesticides up to total of 50, with general information on pesticide application methods, personal protective equipment, pesticide mixing, and equipment repair Additional self-administered take-home questionnaire with further questions on occupational exposures and lifestyle factors Missing data classified into “not known/missing” category, with unknown use of 2,4-dichlorophenoxyacetic acid classified with no use and unknown education classified with no education beyond high school due to lack of MM cases in unknown categories	Linkage to state cancer registry files, state death registries, and National Death Index	None	Fully adjusted: age, gender, smoking pack-years, alcohol use in year before enrollment, first-degree family history of cancer, education, use of 2,4-dichlorophenoxyacetic acid, alachlor, atrazine, metolachlor, or trifluralin, ever use of benomyl, maneb, paraquat, carbaryl, or diazinon Intermediate adjusted: age, gender, smoking, alcohol, family history of cancer, education Adjusted in full cohort: age, gender, family history of cancer, education	Monsanto Europe SA/NV	De Roos et al.^[Citation12]

CI: confidence interval; CLL: chronic lymphocytic leukemia; DLBCL: diffuse large B-cell lymphoma; FL: follicular lymphoma; HL: Hodgkin lymphoma; LHC: lymphohematopoietic cancer; LPS: lymphoproliferative syndrome; MM: multiple myeloma; NHL: non-Hodgkin lymphoma; NR: not reported; OR: odds ratio; SLL: small lymphocytic lymphoma.

CI: confidence interval; CLL: chronic lymphocytic leukemia; DLBCL: diffuse large B-cell lymphoma; FL: follicular lymphoma; HL: Hodgkin lymphoma; LHC: lymphohematopoietic cancer; LPS: lymphoproliferative syndrome; MM: multiple myeloma; NHL: non-Hodgkin lymphoma; NR: not reported; OR: odds ratio; SLL: small lymphocytic lymphoma.

Table 2. Estimated associations between glyphosate exposure and risk of lymphohematopoietic cancer (LHC), including non-Hodgkin lymphoma (NHL), NHL subtypes, Hodgkin lymphoma (HL), multiple myeloma (MM), and leukemia.

Authors	Year	Exposure groups and number of subjects	Relative risk	95% CI
Brown et al.^[Citation35]	1990	Non-farmers: 243 cases, 547 controls Ever mixed, handled, or applied glyphosate: 15 cases, 49 controls	Leukemia OR = 0.9	Leukemia 95% CI = 0.5–1.6
Brown et al.^[Citation32]	1993	Non-farmers: 62 cases, 272 controls Ever mixed, handled, or applied glyphosate: 11 cases, 40 controls	MM OR = 1.7 Among those who did not use protective equipment, MM OR = 1.9	MM 95% CI = 0.8–3.6 Among those who did not use protective equipment, MM 95% CI = NR
Cantor et al.^[Citation24]	1992	Non-farmers: 226 cases, 547 controls Ever handled, mixed, or applied glyphosate: 26 cases, 49 controls	NHL OR = 1.1	NHL 95% CI = 0.7–1.9
Cocco et al.^[Citation18]	2013	Unexposed to any pesticides: NR cases, 2262 controls Occupationally exposed to glyphosate: 4 cases (1 DLBCL, 1 CLL, 1 MM, 1 unspecified B-cell NHL), 2 controls	B-cell NHL OR = 3.1	B-cell NHL 95% CI = 0.6–17.1
De Roos et al.^[Citation13]	2003	Unexposed to glyphosate: 614 cases, 1892 controls Exposed to glyphosate: 36 cases, 61 controls	Hierarchical regression NHL OR = 1.6 Logistic regression NHL OR = 2.1	Hierarchical regression NHL 95% CI = 0.9–2.8 Logistic regression NHL 95% CI = 1.1–4.0
De Roos et al.^[Citation12]	2005	Never used glyphosate: 47 LHC, 21 NHL, 8 MM, 14 leukemia; 13,280 cohort members Ever used glyphosate: 143 LHC, 71 NHL, 24 MM, 43 leukemia; 41,035 cohort members	Fully adjusted LHC RR = 1.1 Age-adjusted LHC RR = 1.1 Fully adjusted NHL RR = 1.1 Age-adjusted NHL RR = 1.2 Fully adjusted MM RR = 2.6 (2.6 in Iowa, 2.7 in North Carolina) Age-adjusted MM RR = 1.1 Fully adjusted leukemia RR = 1.0 Age-adjusted leukemia RR = 1.1	Fully adjusted LHC 95% CI = 0.8–1.6 Age-adjusted LHC 95% CI = 0.8–1.5 Fully adjusted NHL 95% CI = 0.7–1.9 Age-adjusted NHL 95% CI = 0.7–1.9 Fully adjusted MM 95% CI = 0.7–9.4 Age-adjusted MM 95% CI = 0.5–2.4 Fully adjusted leukemia 95% CI = 0.5–1.9 Age-adjusted leukemia 95% CI = 0.6–2.0
		1–20 glyphosate exposure days: 48 LHC, 29 NHL, 8 MM, 9 leukemia 21–56 glyphosate exposure days: 38 LHC, 15 NHL, 5 MM, 14 leukemia 57–2,678 glyphosate exposure days: 36 LHC, 17 NHL, 6 MM, 9 leukemia 0.1–79.5 intensity-weighted glyphosate exposure days: 38 LHC, 24 NHL, 5 MM, 7 leukemia 79.6–337.1 intensity-weighted glyphosate exposure days: 40 LHC, 15 NHL, 6 MM, 17 leukemia 337.2–18,241 intensity-weighted glyphosate exposure days: 43 LHC, 22 NHL, 8 MM, 8 leukemia	Cumulative exposure days, tertiles 2 and 3 vs. 1 LHC RRs = 1.2, 1.2; p-trend = 0.69 NHL RRs = 0.7, 0.9; p-trend = 0.73 MM RRs = 1.1, 1.9; p-trend = 0.27 Leukemia RRs = 1.9, 1.0; p-trend = 0.61 > 108 vs. > 0–9 exposure days, NHL RR = 0.9 Intensity-weighted exposure days, tertiles 2 and 3 vs. 1 LHC RRs = 1.0, 1.0; p-trend = 0.90 NHL RRs = 0.6, 0.8; p-trend = 0.99 MM RRs = 1.2, 2.1; p-trend = 0.17 Leukemia RRs = 1.9, 0.7; p-trend = 0.11 Intensity tertile 3 vs. 1 MM RR = 0.6 Cumulative exposure days, tertiles 1, 2, and 3 vs. never MM RRs = 2.3, 2.6, 4.4; p-trend = 0.09 Cumulative exposure days, quartile 4 vs. never MM RR = 6.6; p-trend = 0.01	Cumulative exposure days, tertiles 2 and 3 vs. 1 LHC 95% CIs = 0.8–1.8, 0.8–1.8 NHL 95% CIs = 0.4–1.4, 0.5–1.6 MM 95% CIs = 0.4–3.5, 0.6–6.3 Leukemia 95% CIs = 0.8–4.5, 0.4–2.9 > 108 vs. > 0–9 exposure days, NHL 95% CI = 0.4–2.1 Intensity-weighted exposure days, tertiles 2 and 3 vs. 1 LHC 95% CIs = 0.6–1.5, 0.7–1.6 NHL 95% CIs = 0.3–1.1, 0.5–1.4 MM 95% CIs = 0.4–3.8, 0.6–7.0 Leukemia 95% CIs = 0.8–4.7, 0.2–2.1 Intensity tertile 3 vs. 1 MM 95% CI = 0.2–1.8 Cumulative exposure days, tertiles 1, 2, and 3 vs. never MM 95% CIs = 0.6–8.9, 0.6–11.5, 1.0–20.2 Cumulative exposure days, quartile 4 vs. never MM 95% CI = 1.4–30.6
Eriksson et al.^[Citation14]	2008	No pesticide exposure: NR Glyphosate exposure for ≥ 1 full working day, ≥ 1 calendar year prior to year of diagnosis or enrollment: 29 NHL cases, 18 controls (NHL subtypes NR) Glyphosate exposure for 1 to ≤ 10 days: 12NHL cases, 9 controls Glyphosate exposure for > 10 days: 17 NHL cases, 9 controls	NHL OR, any glyphosate, multivariate = 1.51 NHL OR, any glyphosate, univariate = 2.02 NHL OR, glyphosate 1 to ≤ 10 days = 1.69 NHL OR, glyphosate > 10 days = 2.36 NHL OR, any glyphosate, latency 1–10 years = 1.11 NHL OR, any glyphosate, latency > 10 years = 2.26	NHL 95% CI, any glyphosate, multivariate = 0.77–2.94 NHL 95% CI, any glyphosate, univariate = 1.10–3.71 NHL 95% CI, glyphosate 1 to ≤ 10 days = 0.70–4.07 NHL 95% CI, glyphosate > 10 days = 1.04–5.37 NHL 95% CI, any glyphosate, latency 1–10 years = 0.24–5.08 NHL 95% CI, any glyphosate, latency > 10 years = 1.16–4.40
			B-cell NHL OR, any glyphosate = 1.87 SLL/CLL OR, any glyphosate = 3.35 FL grades I–III OR, any glyphosate = 1.89 DLBCL OR, any glyphosate = 1.22 Other specified B-cell NHL OR, any glyphosate = 1.63 Unspecified B-cell NHL OR, any glyphosate = 1.47 T-cell NHL OR, any glyphosate = 2.29 Unspecified NHL OR, any glyphosate = 5.63	B-cell NHL 95% CI, any glyphosate = 0.998–3.51 SLL/CLL 95% CI, any glyphosate = 1.42–7.89 FL grades I–III 95% CI, any glyphosate = 0.62–5.79 DLBCL 95% CI, any glyphosate = 0.44–3.35 Other specified B-cell NHL 95% CI, any glyphosate = 0.53–4.96 Unspecified B-cell NHL 95% CI, any glyphosate = 0.33–6.61 T-cell NHL 95% CI, any glyphosate = 0.51–10.4 Unspecified NHL 95% CI, any glyphosate = 1.44–22.0
Hardell and Eriksson^[27]	1999	No pesticide exposure Glyphosate exposure ≥ 1 year prior to diagnosis or control index year: 4 cases, 3 controls	NHL OR adjusted for phenoxyacetic acids = 5.8 NHL OR unadjusted for phenoxyacetic acids = 2.3	NHL 95% CI adjusted for phenoxyacetic acids = 0.6–54 NHL 95% CI unadjusted for phenoxyacetic acids = 0.4–13
Hardell et al.^[Citation15]	2002	No pesticide exposure: NR Glyphosate exposure for ≥ 1 working day, ≥ 1 year prior to diagnosis or control index date: 8 cases, 8 controls	Multivariate NHL OR = 1.85 Univariate NHL OR = 3.04	Multivariate NHL 95% CI = 0.55–6.20 Univariate NHL 95% CI = 1.08–8.52
Hohenadel et al.^[Citation28]	2011	Use of neither glyphosate nor malathion: 422 cases, 1301 controls Use of glyphosate only: 19 cases, 78 controls Use of malathion only: 41 cases, 72 controls Use of glyphosate and malathion: 31 cases, 55 controls	NHL OR, glyphosate only = 0.92 NHL OR, malathion only = 1.95 NHL OR, glyphosate and malathion = 2.10 Interaction contrast ratio = 0.23, P-interaction = 0.69	NHL 95% CI, glyphosate only = 0.54–1.55 NHL 95% CI, malathion only = 1.29 –2.93 NHL 95% CI, glyphosate and malathion = 1.31–3.37
Kachuri et al.^[Citation33]	2013	Never used glyphosate: 310 cases, 1236 controls (216 cases, 1047 controls without proxy) Ever used glyphosate: 32 cases, 121 controls (23 cases, 108 controls without proxy) Used glyphosate for > 0 to ≤ 2 days per year: 15 cases, 88 controls (11 cases, 78 controls without proxy) Used glyphosate for > 2 days per year: 12 cases, 29 controls (10 cases, 26 controls without proxy)	MM OR, ever glyphosate = 1.19 MM OR, ever glyphosate, no proxies = 1.11 MM OR, glyphosate > 0 to ≤ 2 days per year = 0.72 MM OR, glyphosate > 0 to ≤ 2 days per year, no proxies = 0.70 MM OR, glyphosate > 2 days per year = 2.04 MM OR, glyphosate > 2 days per year, no proxies = 2.11	MM 95% CI, ever glyphosate = 0.76–1.87 MM 95% CI, ever glyphosate, no proxies = 0.66–1.86 MM 95% CI, glyphosate > 0 to ≤ 2 days per year = 0.39–1.32 MM 95% CI, glyphosate > 0 to ≤ 2 days per year, no proxies = 0.35–1.40 MM 95% CI, glyphosate > 2 days per year = 0.98–4.23 MM 95% CI, glyphosate > 2 days per year, no proxies = 0.95–4.70
Karunanayake et al.^[Citation31]	2012	Never used glyphosate: 278 cases, 1373 controls Ever used glyphosate: 38 cases, 133 controls	Fully adjusted HL OR = 0.99 Minimally adjusted (age, province) HL OR = 1.14	Fully adjusted HL 95% CI = 0.62–1.56 Minimally adjusted (age, province) HL 95% CI = 0.74–1.76
Kaufman et al.^[Citation36]	2009	No glyphosate use: 179 cases, 753 controls Glyphosate: 1 case, 3 controls	Crude leukemia OR = 1.40	Crude leukemia 95% CI = 0.15–13.56
Lee et al.^[Citation29]	2004	Non-farmers, non-asthmatics: 259 cases, 684 controls Non-farmers, asthmatics: 9 cases, 37 controls Exposed to glyphosate, non-asthmatics: 53 cases, 91 controls Exposed to glyphosate, asthmatics: 6 cases, 12 controls	NHL OR, non-farmers, asthmatics = 0.6 NHL OR, glyphosate, non-asthmatics = 1.4 NHL OR, glyphosate, asthmatics = 1.2	NHL 95% CI, non-farmers, asthmatics = 0.3–1.4 NHL 95% CI, glyphosate, non-asthmatics = 0.98–2.1 NHL 95% CI, glyphosate, asthmatics = 0.4–3.3
McDuffie et al.^[Citation16]	2001	Never used glyphosate: 466 cases, 1373 controls Ever used glyphosate: 51 cases, 1506 controls Glyphosate use for > 0 to ≤ 2 days per year Glyphosate use for > 2 days per year	Fully adjusted NHL OR, ever glyphosate = 1.20 Minimally adjusted (age, province) NHL OR, ever glyphosate = 1.26 Minimally adjusted NHL OR, glyphosate > 0 to ≤ 2 days per year = 1.00 Minimally adjusted NHL OR, glyphosate > 2 days per year = 2.12	Fully adjusted NHL 95% CI, ever glyphosate = 0.83–1.74 Minimally adjusted (age, province) NHL 95% CI, ever glyphosate = 0.87–1.80 Minimally adjusted NHL 95% CI, glyphosate > 0 to ≤ 2 days per year = 0.63–1.57 Minimally adjusted NHL 95% CI, glyphosate > 2 days per year = 1.20–3.73
Nordström et al.^[Citation30]	1998	No glyphosate exposure: 107 cases, 395 controls Glyphosate exposure for ≥ 1 working day, ≥ 1 year prior to diagnosis or control index date: 4 cases, 5 controls	Hairy-cell leukemia OR = 3.1	Hairy-cell leukemia 95% CI = 0.8–12
Orsi et al.^[Citation17]	2009	Never exposed to glyphosate: 464 LHC, 232 NHL, 102 DLBCL, 47 FL, 100 LPS, 75 CLL, 25 hairy-cell leukemia 81 HL, 51 MM, 432 controls Ever exposed to glyphosate: 27 LHC, 12 NHL, 5 DLBCL, 3 FL, 4 LPS, 2 CLL, 2 hairy-cell leukemia, 6 HL, 5 MM, 24 controls	LHC OR = 1.2 NHL OR = 1.0 DLBCL OR = 1.0 FL OR = 1.4 LPS OR = 0.6 CLL OR = 0.4 Hairy-cell leukemia OR = 1.8 HL OR = 1.7 MM OR = 2.4	LHC 95% CI = 0.6–2.1 NHL 95% CI = 0.5–2.2 DLBCL 95% CI = 0.3–2.7 FL 95% CI = 0.4–5.2 LPS 95% CI = 0.2–2.1 CLL 95% CI = 0.1–1.8 Hairy-cell leukemia 95% CI = 0.3–9.3 HL 95% CI = 0.6–5.0 MM 95% CI = 0.8–7.3
Pahwa et al.^[Citation34]	2012	Never used glyphosate: 310 cases, 1373 controls Ever used glyphosate: 32 cases, 133 controls	MM OR = 1.22	MM 95% CI = 0.77–1.93
Sorahan^[26]	2015	Never used glyphosate: 8 cases, 13,280 cohort members (of 54,315); 4 cases, 11,881 cohort members (of 49,211); 3 cases, 9809 cohort members (of 40,719) Ever used glyphosate: 24 cases, 41,035 cohort members (of 54,315); 22 cases, 37,330 cohort members (of 49,211); 19 cases, 30,910 cohort members (of 40,719)	Fully adjusted MM RR, cohort of 54,315 = 1.24 Age- and sex-adjusted MM RR, cohort of 54,315 = 1.12 Age-adjusted MM RR, cohort of 54,315 = 1.08 Age-adjusted MM RR, cohort of 49,211 = 1.91 Intermediate adjusted MM RR, cohort of 49,211 = 2.07 Age-adjusted MM RR, cohort of 40,719 = 2.21 Fully adjusted MM RR, cohort of 40,719 = 2.79	Fully adjusted MM 95% CI, cohort of 54,315 = 0.52–2.94 Age- and sex-adjusted MM 95% CI, cohort of 54,315 = 0.50–2.49 Age-adjusted MM 95% CI, cohort of 54,315 = 0.48–2.41 Age-adjusted MM 95% CI, cohort of 49,211 = 0.66–5.53 Intermediate adjusted MM 95% CI, cohort of 49,211 = 0.71–6.04 Age-adjusted MM 95% CI, cohort of 40,719 = 0.65–7.48 Fully adjusted MM 95% CI, cohort of 40,719 = 0.78–9.96
		1–20 glyphosate exposure days: 10 cases 21–56 glyphosate exposure days: 8 cases 57–2678 glyphosate exposure days: 6 cases 0.1–79.5 intensity-weighted glyphosate exposure days: 6 cases 79.6–337.1 intensity-weighted glyphosate exposure days: 8 cases 337.2–18,241 intensity-weighted glyphosate exposure days: 10 cases	Cumulative exposure days, tertiles 1, 2, and 3 vs. never Fully adjusted MM RRs = 1.14, 1.52, 1.38; p-trend = 0.48 using scores, > 0.50 using means Intermediate adjusted MM RRs = 1.13, 1.50, 1.23; p-trend > 0.50 using scores or means Age- and sex-adjusted MM RRs = 1.06, 1.34, 1.08; p-trend > 0.50 using scores or means Intensity-weighted exposure days, tertiles 1, 2, and 3 vs. never Fully adjusted MM RRs = 1.00, 1.27, 1.87; p-trend = 0.22 using scores, 0.18 using means Intermediate adjusted MM RRs = 0.99, 1.22, 1.65; p-trend = 0.27 using scores, 0.24 using means Age- and sex-adjusted MM RRs = 0.91, 1.12, 1.44; p-trend = 0.39 using scores, 0.33 using means	Cumulative exposure days, tertiles 1, 2, and 3 vs. never Fully adjusted MM 95% CIs = 0.43–3.03, 0.54–4.34, 0.42–4.45 Intermediate adjusted MM 95% CIs = 0.44–2.88, 0.56–4.05, 0.42–3.58 Age- and sex-adjusted MM 95% CIs = 0.42–2.70, 0.50–3.58, 0.37–3.11 Intensity-weighted exposure days, tertiles 1, 2, and 3 vs. never Fully adjusted MM 95% CIs = 0.33–3.00, 0.45–3.56, 0.67–5.27 Intermediate adjusted MM 95% CIs = 0.34–2.86, 0.45–3.28, 0.64–4.24 Age- and sex-adjusted MM 95% CIs = 0.31–2.62, 0.42–3.00, 0.57–3.67
		Never used glyphosate: 8 cases Ever used glyphosate: 24 Unknown glyphosate use: 2 cases	MM RR, ever glyphosate = 1.18 MM RR, unknown glyphosate = 1.71 Cumulative exposure days, tertiles 1, 2, 3, and unknown vs. never MM RRs = 1.11, 1.45, 1.17, 1.19; p-trend > 0.50 using scores or means, excluding unknown Intensity-weighted exposure days, tertiles 1, 2, 3, and unknown vs. never MM RRs = 0.95, 1.19, 1.58, 1.04; p-trend = 0.30 using scores, 0.26 using means, excluding unknown	MM 95% CI, ever glyphosate = 0.53–2.65 MM 95% CI, unknown glyphosate = 0.36–8.20 Cumulative exposure days, tertiles 1, 2, 3, and unknown vs. never MM 95% CIs = 0.44–2.83, 0.54–3.88, 0.40–3.41, 0.25–5.65 Intensity-weighted exposure days, tertiles 1, 2, 3, and unknown vs. never MM 95% CIs = 0.33–2.75, 0.44–3.19, 0.62–4.05, 0.22–4.92

CI: confidence interval; CLL: chronic lymphocytic leukemia; DLBCL: diffuse large B-cell lymphoma; FL: follicular lymphoma; HL: Hodgkin lymphoma; LHC: lymphohematopoietic cancer; LPS: lymphoproliferative syndrome; MM: multiple myeloma; NHL: non-Hodgkin lymphoma; NR: not reported; OR: odds ratio; RR: relative risk; SLL: small lymphocytic lymphoma.

Table 3. Selected estimates included in meta-analyses and calculated meta-analysis relative risks (meta-RRs) of the association between glyphosate exposure and risk of (LHC), including non-Hodgkin lymphoma (NHL), NHL subtypes, Hodgkin lymphoma (HL), multiple myeloma (MM), and leukemia.

Study ^#	Authors	Year	Outcome	Number of exposed subjects	RR	95% CI
1	De Roos et al.^[Citation13]	2003	Non-Hodgkin lymphoma	36 cases, 61 controls	a. 1.6 (hierarchical regression) b. 2.1 (logistic regression)	a. 0.9–2.8 (hierarchical regression) b. 1.1–4.0 (logistic regression)
2	De Roos et al.^[Citation12]	2005	Non-Hodgkin lymphoma	71 cases*	1.1	0.7–1.9
3	Eriksson et al.^[Citation14]	2008	Non-Hodgkin lymphoma	29 cases, 18 controls	1.51	0.77–2.94
4	Hardell et al.^[Citation15]	2002	Non-Hodgkin lymphoma	8 cases, 8 controls	1.85	0.55–6.20
5	Hohenadel et al.^[Citation28]	2011	Non-Hodgkin lymphoma	50 cases, 133 controls	1.40 (random effects meta-RR)	0.62–3.15 (random effects meta-CI)
6	McDuffie et al.^[Citation16]	2001	Non-Hodgkin lymphoma	51 cases, 133 controls	1.2	0.83–1.74
7	Orsi et al.^[Citation17]	2009	Non-Hodgkin lymphoma	12 cases, 24 controls	1.0	0.5–2.2
	Meta-analysis model		Outcome	Studies included	Meta-RR	95% CI	I^2	Pheterogeneity
	Model 1		Non-Hodgkin lymphoma	1a, 2, 3, 4, 6, 7	1.3	1.0–1.6	0.0%	0.84
	Model 2		Non-Hodgkin lymphoma	1b, 2, 3, 4, 6, 7	1.3	1.0–1.6	0.0%	0.59
	Model 3		Non-Hodgkin lymphoma	1a, 2, 3, 4, 5, 7	1.3	1.0–1.7	0.0%	0.85
	Model 4		Non-Hodgkin lymphoma	1b, 2, 3, 4, 5, 7	1.4	1.0–1.8	0.0%	0.63
3	Eriksson et al.^[Citation14]	2008	B-cell lymphoma	Not reported	1.87	0.998–3.51
8	Cocco et al.^[Citation18]	2013	B-cell lymphoma	4 cases, 2 controls	3.1	0.6–17.1
	Meta-analysis model		Outcome	Studies included	Meta-RR	95% CI	I^2	Pheterogeneity
	Model 1		B-cell lymphoma	3, 8	2.0	1.1–3.6	0.0%	0.58
3	Eriksson et al.^[Citation14]	2008	Diffuse large B-cell lymphoma	Not reported	1.22	0.44–3.35
7	Orsi et al.^[Citation17]	2009	Diffuse large B-cell lymphoma	5 cases, 24 controls	1.0	0.3–2.7
	Meta-analysis model		Outcome	Studies included	Meta-RR	95% CI	I^2	Pheterogeneity
	Model 1		Diffuse large B-cell lymphoma	3, 7	1.1	0.5–2.3	0.0%	0.79
3	Eriksson et al.^[Citation14]	2008	CLL/SLL	Not reported	3.35	1.42–7.89
7	Orsi et al.^[Citation17]	2009	CLL/SLL	2 cases, 18 controls	0.4	0.1–1.8
	Meta-analysis model		Outcome	Studies included	Meta-RR	95% CI	I^2	Pheterogeneity
	Model 1, random effects		CLL/SLL	3, 7	1.3	0.2–10.0	83.7%	0.01
	Model 1, fixed effects		CLL/SLL	3, 7	1.9	0.9–4.0
3	Eriksson et al.^[Citation14]	2008	Follicular lymphoma	Not reported	1.89	0.62–5.79
7	Orsi et al.^[Citation17]	2009	Follicular lymphoma	3 cases, 24 controls	1.4	0.4–5.2
	Meta-analysis model		Outcome	Studies included	Meta-RR	95% CI	I^2	Pheterogeneity
	Model 1		Follicular lymphoma	3, 7	1.7	0.7–3.9	0.0%	0.73
7	Orsi et al.^[Citation17]	2009	Hairy-cell leukemia	2 cases, 18 controls	1.8	0.3–9.3
9	Nordström et al.^[Citation30]	1998	Hairy-cell leukemia	4 cases, 5 controls	3.1	0.8–12
	Meta-analysis model		Outcome	Studies included	Meta-RR	95% CI	I^2	Pheterogeneity
	Model 1		Hairy-cell leukemia	7, 9	2.5	0.9–7.3	0.0%	0.63
7	Orsi et al.^[Citation17]	2009	Hodgkin lymphoma	6 cases, 24 controls	1.7	0.6–5.0
10	Karunanayake et al.^[Citation31]	2012	Hodgkin lymphoma	38 cases, 133 controls	0.99	0.62–1.56
	Meta-analysis model		Outcome	Studies included	Meta-RR	95% CI	I^2	Pheterogeneity
	Model 1		Hodgkin lymphoma	7, 10	1.1	0.7–1.6	0.0%	0.36
2	De Roos et al.^[Citation12]	2005	Multiple myeloma	19 cases	2.6	0.7–9.4
7	Orsi et al.^[Citation17]	2009	Multiple myeloma	5 cases, 24 controls	2.4	0.8–7.3
11	Brown et al.^[Citation32]	1993	Multiple myeloma	11 cases, 40 controls	1.7	0.8–3.6
12	Kachuri et al.^[Citation33]	2013	Multiple myeloma	32 cases, 121 controls	a. 1.19 (with proxies) b. 1.11 (without proxies)	a. 0.76–1.87 (with proxies) b. 0.66–1.86 (without proxies)
13	Pahwa et al.^[Citation34]	2012	Multiple myeloma	32 cases, 133 controls	1.22	0.77–1.93
14	Sorahan^[Citation26]	2015	Multiple myeloma	24 cases	1.24	0.52–2.94
	Meta-analysis model		Outcome	Studies included	Meta-RR	95% CI	I^2	Pheterogeneity
	Model 1		Multiple myeloma	7, 11, 12a, 14	1.4	1.0–1.9	0.0%	0.63
	Model 2		Multiple myeloma	2, 7, 11, 12a	1.5	1.0–2.1	0.0%	0.48
	Model 3		Multiple myeloma	7, 11, 12b, 14	1.4	0.9–1.9	0.0%	0.58
	Model 4		Multiple myeloma	7, 11, 13, 14	1.4	1.0–2.0	0.0%	0.66
	Model 5		Multiple myeloma	2, 7, 11, 13	1.5	1.0–2.1	0.0%	0.52
2	De Roos et al.^[Citation12]	2005	Leukemia	43 cases*	1.0	0.5–1.9
16	Brown et al.^[Citation35]	1990	Leukemia	15 cases, 49 controls	0.9	0.5–1.6
17	Kaufman et al.^[Citation36]	2009	Leukemia	1 case, 3 controls	1.4	0.15–13.56
	Meta-analysis model		Outcome	Studies included	Meta-RR	95% CI	I^2	Pheterogeneity
	Model 1		Leukemia	2, 16, 17	1.0	0.6–1.5	0.0%	0.92

*Number of exposed cases is provided for the total cohort of 54,315 subjects; the number of exposed cases in the analytic cohort of 49,211 subjects is not stated.

^† Number of exposed cases is provided for the analytic cohort of 40,719 subjects, as reported by Sorahan.^[26]

CI: confidence interval; CLL: chronic lymphocytic leukemia; RR: relative risk; SLL: small lymphocytic lymphoma.

Figure 1. Forest plots of relative risk (RR) estimates and 95% confidence intervals (CIs) for the association between glyphosate exposure and risk of non-Hodgkin lymphoma. Meta-RRs were identical in random-effects and fixed-effects models.

Figure 2. Forest plots of relative risk (RR) estimates and 95% confidence intervals (CIs) for the association between glyphosate exposure and risk of multiple myeloma. Meta-RRs were identical in random-effects and fixed-effects models.

Table 4. Sensitivity analysis of the association between glyphosate exposure and risk of non-Hodgkin lymphoma (NHL).

Stratum	Number of studies	Meta-RR*	95% CI
All	6	1.3	1.0–1.6
Case-control	5	1.3	1.0–1.7
Cohort	1	NR
Population controls	4	1.4	1.0–1.8
Hospital controls	1	NR
Males only	4	1.3	1.0–1.7
Males and females	2	1.2	0.8–1.8
North America	3	1.2	1.0–1.6
Europe	3	1.3	0.8–2.1
Sweden	2	1.6	0.9–2.8
Cases in 1980s	2	1.6	1.0–2.7
Cases in 1990s	4	1.2	1.0–1.6
Cases in 2000s	3	1.2	0.8–1.7

*All meta-RRs were identical in random-effects and fixed-effects models.

CI: confidence interval; meta-RR: meta-analysis relative risk; NR: not reported, when only one study was available.

Figure A1. Flow chart of literature identification and selection process.

Brown, L.M.; Blair, A.; Gibson, R.; Everett, G.D.; Cantor, K.P.; Schuman, L.M.; Burmeister, L.F.; Van Lier, S.F.; Dick, F. Pesticide exposures and other agricultural risk factors for leukemia among men in Iowa and Minnesota. Cancer Res. 1990, 50, 6585–6591.

 PubMed Web of Science ®Google Scholar

Brown, L.M.; Burmeister, L.F.; Everett, G.D.; Blair, A. Pesticide exposures and multiple myeloma in Iowa men. Cancer Causes Control. 1993, 4, 153–156.

 PubMed Web of Science ®Google Scholar

Cantor, K.P.; Blair, A.; Everett, G.; Gibson, R.; Burmeister, L.F.; Brown, L.M.; Schuman, L.; Dick, F.R. Pesticides and other agricultural risk factors for non-Hodgkin's lymphoma among men in Iowa and Minnesota. Cancer Res. 1992, 52, 2447–2455.

 PubMed Web of Science ®Google Scholar

Cocco, P.; Satta, G.; Dubois, S.; Pili, C.; Pilleri, M.; Zucca, M.; t Mannetje, A.M.; Becker, N.; Benavente, Y.; de Sanjose, S.; Foretova, L.; Staines, A.; Maynadie, M.; Nieters, A.; Brennan, P.; Miligi, L.; Ennas, M.G.; Boffetta, P. Lymphoma risk and occupational exposure to pesticides: results of the Epilymph study. Occup. Environ. Med. 2013, 70, 91–98.

 PubMed Web of Science ®Google Scholar

De Roos, A.J.; Zahm, S.H.; Cantor, K.P.; Weisenburger, D.D.; Holmes, F.F.; Burmeister, L.F.; Blair, A. Integrative assessment of multiple pesticides as risk factors for non-Hodgkin's lymphoma among men. Occup. Environ. Med. 2003, 60, E11.

 PubMed Web of Science ®Google Scholar

De Roos, A.J.; Blair, A.; Rusiecki, J.A.; Hoppin, J.A.; Svec, M.; Dosemeci, M.; Sandler, D.P.; Alavanja, M.C. Cancer incidence among glyphosate-exposed pesticide applicators in the Agricultural Health Study. Environ. Health Perspect. 2005, 113, 49–54.

 PubMed Web of Science ®Google Scholar

Eriksson, M.; Hardell, L.; Carlberg, M.; Akerman, M. Pesticide exposure as risk factor for non-Hodgkin lymphoma including histopathological subgroup analysis. Int. J. Cancer. 2008, 123, 1657–1663.

 PubMed Web of Science ®Google Scholar

Hardell, L.; Eriksson, M.; Nordstrom, M. Exposure to pesticides as risk factor for non-Hodgkin's lymphoma and hairy cell leukemia: pooled analysis of two Swedish case-control studies. Leuk. Lymphoma. 2002, 43, 1043–1049.

 PubMed Web of Science ®Google Scholar

Hohenadel, K.; Harris, S.A.; McLaughlin, J.R.; Spinelli, J.J.; Pahwa, P.; Dosman, J.A.; Demers, P.A.; Blair, A. Exposure to multiple pesticides and risk of non-Hodgkin lymphoma in men from six Canadian provinces. Int. J. Environ. Res. Public Health. 2011, 8, 2320–2330.

 PubMed Web of Science ®Google Scholar

Kachuri, L.; Demers, P.A.; Blair, A.; Spinelli, J.J.; Pahwa, M.; McLaughlin, J.R.; Pahwa, P.; Dosman, J.A.; Harris, S.A. Multiple pesticide exposures and the risk of multiple myeloma in Canadian men. Int. J. Cancer. 2013, 133, 1846–1858.

 PubMed Web of Science ®Google Scholar

Karunanayake, C.P.; Spinelli, J.J.; McLaughlin, J.R.; Dosman, J.A.; Pahwa, P.; McDuffie, H.H. Hodgkin lymphoma and pesticides exposure in men: a Canadian case-control study. J. Agromedicine. 2012, 17, 30–39.

 PubMed Web of Science ®Google Scholar

Kaufman, D.W.; Anderson, T.E.; Issaragrisil, S. Risk factors for leukemia in Thailand. Ann. Hematol. 2009, 88, 1079–1088.

 PubMed Web of Science ®Google Scholar

Lee, W.J.; Cantor, K.P.; Berzofsky, J.A.; Zahm, S.H.; Blair, A. Non-Hodgkin's lymphoma among asthmatics exposed to pesticides. Int. J. Cancer. 2004, 111, 298–302.

 PubMed Web of Science ®Google Scholar

McDuffie, H.H.; Pahwa, P.; McLaughlin, J.R.; Spinelli, J.J.; Fincham, S.; Dosman, J.A.; Robson, D.; Skinnider, L.F.; Choi, N.W. Non-Hodgkin's lymphoma and specific pesticide exposures in men: cross-Canada study of pesticides and health. Cancer Epidemiol. Biomarkers Prev. 2001, 10, 1155–1163.

 PubMed Web of Science ®Google Scholar

Nordström, M.; Hardell, L.; Magnuson, A.; Hagberg, H.; Rask-Andersen, A. Occupational exposures, animal exposure and smoking as risk factors for hairy cell leukaemia evaluated in a case-control study. Br. J. Cancer. 1998, 77, 2048–2052.

 PubMed Web of Science ®Google Scholar

Orsi, L.; Delabre, L.; Monnereau, A.; Delval, P.; Berthou, C.; Fenaux, P.; Marit, G.; Soubeyran, P.; Huguet, F.; Milpied, N.; Leporrier, M.; Hemon, D.; Troussard, X.; Clavel, J. Occupational exposure to pesticides and lymphoid neoplasms among men: results of a French case-control study. Occup. Environ. Med. 2009, 66, 291–298.

 PubMed Web of Science ®Google Scholar

Pahwa, P.; Karunanayake, C.P.; Dosman, J.A.; Spinelli, J.J.; McDuffie, H.H.; McLaughlin, J.R. Multiple myeloma and exposure to pesticides: a Canadian case-control study. J Agromedicine. 2012, 17, 40–50.

 PubMed Web of Science ®Google Scholar

Hoar, S.K.; Blair, A.; Holmes, F.F.; Boysen, C.D.; Robel, R.J.; Hoover, R.; Fraumeni, J.F., Jr. Agricultural herbicide use and risk of lymphoma and soft-tissue sarcoma. JAMA. 1986, 256, 1141–1147.

 PubMed Web of Science ®Google Scholar

Hoar Zahm, S.; Weisenburger, D.D.; Babbitt, P.A.; Saal, R.C.; Vaught, J.B.; Cantor, K.P.; Blair, A. A case-control study of non-Hodgkin's lymphoma and the herbicide 2,4-dichlorophenoxyacetic acid (2,4-D) in eastern Nebraska. Epidemiol. 1990, 1, 349–356.

 PubMedGoogle Scholar

Sorahan, T. Multiple myeloma and glyphosate use: a re-analysis of US Agricultural Health Study (AHS) data. Int. J. Environ. Res. Public Health. 2015, 12, 1548–1559.

 PubMed Web of Science ®Google Scholar