Abstract
Objectives:
Lower mortality rates compared to the general population have been reported for Agricultural Health Study (AHS) participants (enrolled 1993–1997) followed through 2007. We extended analysis of mortality among AHS participants (51,502 private pesticide applicators, their 31,867 spouses, and 4,677 commercial pesticide applicators from North Carolina and Iowa) through 2015 and compared results using several analytic approaches.
Methods:
We calculated standardized mortality ratios (SMRs), causal mortality ratios (CMRs), and relative SMRs using state-specific mortality rates of the general populations as the referent.
Results:
Over the average 16 years of follow-up (1999–2015), 9,305 private applicators, 3,384 spouses, and 415 commercial applicators died. SMRs and CMRs, with expected deaths calculated using the person-time among the cohort and the general population respectively, indicated lower overall mortality in all study subgroups (SMRs from 0.61 to 0.69 and CMRs from 0.74 to 0.89), although CMRs indicated elevated mortality in private applicators from North Carolina and in ever-smokers. In SMR analyses, there were fewer-than-expected deaths from many causes, but deaths from some external causes including transportation-related injuries and mechanical forces were elevated in private applicators. CMRs indicated higher-than-expected deaths from prostate cancer, lymphohematopoietic cancers, Parkinson’s and Alzheimer’s disease, and chronic glomerulonephritis in private applicators, and non-Hodgkin lymphoma in spouses (from 1.19 to 1.53). Relative SMR results were generally elevated, similar to CMR findings.
Conclusions:
AHS participants experienced lower overall mortality than the general population.
Mortality from a few specific causes was increased in private applicators, specifically when CMR and rSMR approaches were used.
Keywords: Standardized mortality ratio, Causal mortality ratio, Relative standardized mortality ratio, Farmers, Agricultural Health Study
INTRODUCTION
Farmers have been shown to experience lower overall mortality as compared to general populations, including deaths from major causes such as lung cancer and heart diseases, despite higher exposures to toxicants, such as pesticides, organic solvents, and dusts.[1–4] This finding may be attributed, in part, to healthier lifestyles among farmers, including low tobacco and alcohol consumption and high physical activity, as well as to selecting for healthier individuals who can tolerate the physical demands of farming. Conversely, mortality rates are higher among farmers for some types of cancers including skin, prostate, and lymphohematopoietic system, potentially due to factors such as excessive sunlight, pesticides, and other agricultural exposures.[1–4]
The Agricultural Health Study (AHS) is a prospective cohort study of licensed pesticide applicators and their spouses from North Carolina (NC) and Iowa (IA), enrolled between 1993 and 1997.[5] In prior AHS studies with approximately five[6] and twelve years of follow-up,[7] standardized mortality ratios (SMRs) were used to compare the observed mortality rates in the AHS with the mortality rates of the general populations of NC and IA. Findings indicated reduced mortality rates overall.
An SMR is the ratio of observed deaths to deaths estimated from the product of the observed person-time accrued by study participants and the referent mortality rates. While the denominator of the SMR is typically referred to as the “expected” number of deaths, it does not represent the expected number of deaths in the population under the referent mortality rates,[8,9] except when exposure does not affect lifespan (which generally would only occur under the null). The SMR may provide a biased estimate when the referent mortality rates do not represent mortality rates of study participants in the absence of exposure(s).[10,11] Richardson et al.[10] introduced the causal mortality ratio (CMR) that uses person-time that would have been expected had the study cohort actually had the mortality rates of the referent population to estimate expected deaths. The CMR may provide more reliable estimates of expected mortality than the SMR as it does not assume that occupational exposures do not influence the cohort’s person-time. In the most recent AHS mortality investigation, Waggoner et al.[7] also estimated relative SMRs (rSMRs, cause-specific SMR/SMR for all other causes omitting the cause of interest) and found elevated rSMRs for several causes. When a “healthy worker bias” is present, the rSMR may reduce bias of the true mortality rate ratio by such indirect adjustment (assuming comparable bias for all causes).[7,11] Here, we extended the mortality follow-up in the AHS using additional deaths through 2015 (with average follow-up of 16 years) and compared the results using SMR, CMR, and rSMR approaches. This study is one of the few studies of farming populations with such long-term mortality follow-up.[2]
METHODS
Study population
At pesticide licensing locations between 1993–1997, the AHS targeted enrollment of pesticide applicators applying for or renewing a license to use restricted pesticides, either for private or commercial purposes; 80% of licensed applicators enrolled [52,394 private pesticide applicators (mostly farmers, 97.4% male) from NC and IA and 4,916 commercial applicators (95.8% male) from IA].[5] Married spouses of private applicators were also invited to participate, and 75% (n=32,345) enrolled (99.3% female) by completing a questionnaire brought home by the applicator. We identified deaths through linkage with state death registries and the National Death Index through December 31, 2015. Cause of death was unknown for 16 participants, mostly due to their recency of death. All applicable Institutional Review Boards approved this study.
Reference population
We obtained mortality rates for NC and IA from the Centers for Disease Control and Prevention’s WONDER Online database.[12] The WONDER database uses the Tenth Revision of the International Classification of Diseases (ICD-10) to specify causes of death from 1999 onwards and the ninth revision (ICD-9) for deaths occurring before 1999. The database provides ‘detailed’ mortality rates for 113 selected causes for 1999 and beyond, and ‘compressed’ mortality rates for 72 selected causes before 1999. We considered mortality between 1999 through 2015 to focus on the 113 underlying causes of death available for that period, omitting deaths and person-time for 1993–1998. We also analyzed mortality by ICD subchapters, broader categories of individual ICD codes that represent similar conditions. However, to avoid redundancy, we present selected causes of deaths, mainly from the 113 causes, but include some from ICD subchapters when relevant and not overlapping. As the database suppresses data for population strata with fewer than nine deaths, we assumed that the suppressed values and, therefore, the corresponding mortality rates were zero for our analyses, although we performed sensitivity analyses assuming all suppressed cells contained 9 deaths.
Statistical analyses
We excluded 1,197 deaths occurring before 1999 (892 (1.7% of enrolled) private applicators), 268 (0.8%) spouses, and 37 (0.8%) commercial applicators). Further, we excluded 210 male spouses and 202 female commercial applicators due to the small number of male spouses and female commercial applicators, leaving 51,502 private applicators, 31,867 female spouses, and 4,677 male commercial applicators for the final analyses (total deaths=13,104). We conducted separate analyses for these three study subgroups. Except as indicated, we focused on underlying causes of death with more than five observed deaths over the course of follow-up.
For SMRs, we estimated the “expected” number of deaths as the product of observed person-time accrued by the participants (i.e., from January 1, 1999 until death or December 31, 2015) and mortality rates for the general populations stratified by 10-year age categories, calendar time (1999–2003, 2004–2008, and 2009–2015), sex, race, and state. We estimated CMRs for the 113 causes and ICD subchapters separately, as illustrated by Richardson et al.[10] We estimated 95% confidence intervals (CIs) for both SMRs and CMRs using Byar’s method.[13] We estimated rSMRs and 95% CIs using Poisson regression. We used 10-year age categories to reduce numbers of strata with suppressed death counts in the general population, but also performed sensitivity analyses using 5-year categories.
Additionally, for private applicators and spouses, we estimated SMRs and CMRs separately by state and by smoking status (defined as ever-smoked 100 cigarettes) at enrollment to examine potential heterogeneity by these factors. As the WONDER database does not provide smoking status-specific mortality rates, for smoking status-specific analysis, we used overall mortality rates (i.e., rates for both non-smokers and smokers combined) as the reference. Lastly, although we focused on underlying cause of death, we conducted secondary analyses estimating SMRs using underlying as well as contributing causes, using multiple cause of death mortality rates as the reference. We performed all analyses using SAS v9.4 (SAS Institute, Inc., Cary, NC).
RESULTS
Over the average 16 years of follow-up, 18% (n=9,305) of private applicators, 10% (n=3,384) of spouses, and 9% (n=415) of commercial applicators died (Table 1, Supplementary Table 1). Deaths occurred in 25% of NC and 13% of IA private applicators, and 15% of NC and 8% of IA spouses. NC participants were older and had higher ever-smoking prevalence and lower education than their IA counterparts. Pesticide use and farming practices reported at enrollment also differed between NC and IA.
Table 1:
Demographic characteristics of Agricultural Health Study participants from Iowa and North Carolina for the period 1999 to 2015 by study subgroups.
| Private applicators |
Spouses |
Commercial applicators |
|||||
|---|---|---|---|---|---|---|---|
| Characteristic | All (n=51,502) |
Iowa (n=31,543) |
North Carolina (n=19,959) |
All (n=31,867) |
Iowa (n=21,560) |
North Carolina (n=10,307) |
All (n=4,677) |
| Age at 1999 (years) | |||||||
| 15–24 | 607 (1.2) | 314 (1) | 293 (1.5) | 71 (0.2) | 37 (0.2) | 34 (0.3) | 125 (2.7) |
| 25–34 | 4,508 (8.8) | 2,751 (8.7) | 1,757 (8.8) | 2,760 (8.7) | 1,930 (9) | 830 (8.1) | 1,115 (23.8) |
| 35–44 | 13,021 (25.3) | 9,049 (28.7) | 3,972 (19.9) | 8,835 (27.7) | 6,615 (30.7) | 2,220 (21.5) | 1,566 (33.5) |
| 45–54 | 13,427 (26.1) | 8,551 (27.1) | 4,876 (24.4) | 8,608 (27.0) | 5,904 (27.4) | 2,704 (26.2) | 1,126 (24.1) |
| 55–64 | 10,851 (21.1) | 6,415 (20.3) | 4,436 (22.2) | 7,204 (22.6) | 4,667 (21.6) | 2,537 (24.6) | 506 (10.8) |
| 65–74 | 6,969 (13.5) | 3,691 (11.7) | 3,278 (16.4) | 3,661 (11.5) | 2,085 (9.7) | 1,576 (15.3) | 205 (4.4) |
| 75–84 | 1,975 (3.8) | 741 (2.3) | 1,234 (6.2) | 692 (2.2) | 306 (1.4) | 386 (3.7) | 33 (0.7) |
| 85+ | 144 (0.3) | 31 (0.1) | 113 (0.6) | 36 (0.1) | 16 (0.1) | 20 (0.2) | 1 (0) |
| Gender | |||||||
| Male | 50,157 (97.4) | 31,103 (98.6) | 19,054 (95.5) | 4,677 (100) | |||
| Female | 1,345 (2.6) | 440 (1.4) | 905 (4.5) | 31,867 (100) | 21,560 (100) | 10,307 (100) | |
| Education | |||||||
| ≤ High school | 28,686 (55.7) | 17,227 (54.6) | 11,459 (57.4) | 12,661 (39.7) | 8,020 (37.2) | 4,641 (45) | 2,131 (45.6) |
| 1–3 years beyond high school | 11,996 (23.3) | 8,302 (26.3) | 3,694 (18.5) | 8,415 (26.4) | 6,109 (28.3) | 2,306 (22.4) | 1,324 (28.3) |
| > College graduate | 8,511 (16.5) | 5,214 (16.5) | 3,297 (16.5) | 6,600 (20.7) | 4,613 (21.4) | 1,987 (19.3) | 1,074 (23) |
| Something else | 113 (0.2) | 57 (0.2) | 56 (0.3) | 2,867 (9.0) | 2,240 (10.4) | 627 (6.1) | 13 (0.3) |
| Missing | 2,196 (4.3) | 743 (2.4) | 1,453 (7.3) | 1,324 (4.2) | 578 (2.7) | 746 (7.2) | 135 (2.9) |
| Smoking status | |||||||
| Never Smoker | 26,667 (51.8) | 18,790 (59.6) | 7,877 (39.5) | 22,741 (71.4) | 16,097 (74.7) | 6,644 (64.5) | 2,191 (46.8) |
| Ever Smoker | 23,647 (45.9) | 12,423 (39.4) | 11,224 (56.2) | 8,540 (26.8) | 5,219 (24.2) | 3,321 (32.2) | 2,418 (51.7) |
| Missing | 1,188 (2.3) | 330 (1) | 858 (4.3) | 586 (1.8) | 244 (1.1) | 342 (3.3) | 68 (1.5) |
| Race | |||||||
| Other | 1,456 (2.8) | 50 (0.2) | 1406 (7) | 544 (1.7) | 52 (0.2) | 492 (4.8) | 25 (0.5) |
| White | 49,042 (95.2) | 31,200 (98.9) | 17,842 (89.4) | 30,510 (95.7) | 21,191 (98.3) | 9,319 (90.4) | 4,623 (98.8) |
| Missing | 1,004 (1.9) | 293 (0.9) | 711 (3.6) | 813 (2.6) | 317 (1.5) | 496 (4.8) | 29 (0.6) |
| Death | |||||||
| No | 42,197 (81.9) | 27,335 (86.7) | 14,862 (74.5) | 28,483 (89.4) | 19,741 (91.6) | 8,742 (84.8) | 4,262 (91.1) |
| Yes | 9,305 (18.1) | 4,208 (13.3) | 5,097 (25.5) | 3,384 (10.6) | 1,819 (8.4) | 1,565 (15.2) | 415 (8.9) |
| Ever insecticide use | |||||||
| No | 4,253 (8.3) | 1,982 (6.3) | 2,271 (11.4) | 18,730 (58.8) | 12,471 (57.8) | 6,259 (60.7) | 1,047 (22.4) |
| Yes | 46,561 (90.4) | 29,369 (93.1) | 17,192 (86.1) | 12,041 (37.8) | 8,596 (39.9) | 3,445 (33.4) | 3,616 (77.3) |
| Missing | 688 (1.3) | 192 (0.6) | 496 (2.5) | 1,096 (3.4) | 493 (2.3) | 603 (5.9) | 14 (0.3) |
| Ever organochlorine use | |||||||
| No | 23,459 (45.5) | 15,761 (50) | 7,698 (38.6) | 27,833 (87.3) | 18,891 (87.6) | 8,942 (86.8) | 3,220 (68.8) |
| Yes | 24,929 (48.4) | 15,136 (48) | 9,793 (49.1) | 2,280 (7.2) | 1,631 (7.6) | 649 (6.3) | 1,368 (29.2) |
| Missing | 3,114 (6) | 646 (2) | 2,468 (12.4) | 1,754 (5.5) | 1,038 (4.8) | 716 (6.9) | 89 (1.9) |
| Ever organophosphate use | |||||||
| No | 6,674 (13) | 2,681 (8.5) | 3,993 (20) | 22,938 (72) | 15,307 (71) | 7,631 (74) | 1,276 (27.3) |
| Yes | 44,115 (85.7) | 28,670 (90.9) | 15,445 (77.4) | 7,806 (24.5) | 5,740 (26.6) | 2,066 (20) | 3,387 (72.4) |
| Missing | 713 (1.4) | 192 (0.6) | 521 (2.6) | 1123 (3.5) | 513 (2.4) | 610 (5.9) | 14 (0.3) |
| Ever carbamate use | |||||||
| No | 16,706 (32.4) | 13,892 (44) | 2,814 (14.1) | 21,157 (66.4) | 14,616 (67.8) | 6,541 (63.5) | 2,195 (46.9) |
| Yes | 32,646 (63.4) | 17,201 (54.5) | 15,445 (77.4) | 9,590 (30.1) | 6,434 (29.8) | 3,156 (30.6) | 2,426 (51.9) |
| Missing | 2,150 (4.2) | 450 (1.4) | 1,700 (8.5) | 1,120 (3.5) | 510 (2.4) | 610 (5.9) | 56 (1.2) |
| Ever fungicide use | |||||||
| No | 31,968 (62.1) | 25,148 (79.7) | 6,820 (34.2) | 28,773 (90.3) | 19,990 (92.7) | 8,783 (85.2) | 3,637 (77.8) |
| Yes | 18,734 (36.4) | 6,135 (19.4) | 12,599 (63.1) | 1,452 (4.6) | 657 (3) | 795 (7.7) | 1,022 (21.9) |
| Missing | 800 (1.6) | 260 (0.8) | 540 (2.7) | 1,642 (5.2) | 913 (4.2) | 729 (7.1) | 18 (0.4) |
| Ever fumigant use | |||||||
| No | 38,732 (75.2) | 27,967 (88.7) | 10,765 (53.9) | 29,729 (93.3) | 20,473 (95) | 9,256 (89.8) | 3,794 (81.1) |
| Yes | 11,916 (23.1) | 3,306 (10.5) | 8,610 (43.1) | 537 (1.7) | 199 (0.9) | 338 (3.3) | 867 (18.5) |
| Missing | 854 (1.7) | 270 (0.9) | 584 (2.9) | 1601 (5) | 888 (4.1) | 713 (6.9) | 16 (0.3) |
| Ever herbicide use | |||||||
| No | 1,950 (3.8) | 664 (2.1) | 1,286 (6.4) | 19,271 (60.5) | 12,408 (57.6) | 6,863 (66.6) | 663 (14.2) |
| Yes | 48,894 (94.9) | 30,698 (97.3) | 18,196 (91.2) | 11,345 (35.6) | 8,572 (39.8) | 2,773 (26.9) | 4,002 (85.6) |
| Missing | 658 (1.3) | 181 (0.6) | 477 (2.4) | 1251 (3.9) | 580 (2.7) | 671 (6.5) | 12 (0.3) |
| Own or work on farm | |||||||
| No | 2,048 (4) | 370 (1.2) | 1,678 (8.4) | 3,134 (67) | |||
| Yes | 47,576 (92.4) | 30,753 (97.5) | 16,823 (84.3) | 14,75 (31.5) | |||
| Missing | 1,878 (3.6) | 420 (1.3) | 1,458 (7.3) | 68 (1.5) | |||
| Raise animals | |||||||
| No | 21,712 (42.2) | 8,849 (28.1) | 12,863 (64.4) | ||||
| Yes | 29,790 (57.8) | 22,694 (71.9) | 7,096 (35.6) | ||||
| Missing | |||||||
| Grow crops | |||||||
| No | 6,027 (11.7) | 1,375 (4.4) | 4,652 (23.3) | ||||
| Yes | 45,475 (88.3) | 30,168 (95.6) | 15,307 (76.7) | ||||
| Missing | |||||||
Private pesticide applicators
Among private applicators, in SMR analyses, overall mortality rates, as well as many cause-specific mortality rates, were less than expected (SMRoverall=0.69, 95%CI: 0.67–0.70) (Table 2). The SMR for mycoses was significantly elevated, but imprecise. The SMRs for ovarian cancer, multiple myeloma, acute and subacute endocarditis, asthma, some skin disorders, osteopathies and chondropathies, and chronic glomerulonephritis were modestly elevated (≥1.20). SMRs also indicated significant excess mortality from accidents associated with land transport and inanimate mechanical forces.
Table 2:
Standardized mortality ratios (SMRs), causal mortality ratios (CMRs), relative SMRs, and their 95% confidence intervals based on underlying cause of death for private pesticide applicators in the Agricultural Health Study for the period 1999–2015 (n=51,502).
| Causea | O | Es | SMR (95% CI)b | Ec | CMR (95% CI)b | rSMR (95% CI)b |
|---|---|---|---|---|---|---|
| All causes | 9,305 | 13,529 | 0.69 (0.67, 0.70) | 10,448 | 0.89 (0.87, 0.91) | - |
| Intestinal infectious diseases (A00–A09)c | 28 | 27 | 1.04 (0.69, 1.44) | 18 | 1.57 (1.04, 2.19) | 1.48 (1.02, 2.15) |
| Septicemia (A40–A41) | 106 | 154 | 0.69 (0.56, 0.82) | 117 | 0.91 (0.74, 1.08) | 0.99 (0.82, 1.20) |
| Mycoses (B35–B49)c | 10 | 4 | 2.47 (1.18, 4.13) | 3 | 3.08 (1.47, 5.15) | 3.53 (1.90, 6.57) |
| Malignant neoplasms (C00–C97) | 2,781 | 3,758 | 0.74 (0.71, 0.77) | 3,042 | 0.91 (0.88, 0.95) | 1.10 (1.05, 1.15) |
| Malignant neoplasms of lip, oral cavity and pharynx (C00–C14) | 31 | 64 | 0.48 (0.33, 0.66) | 54 | 0.57 (0.39, 0.78) | 0.69 (0.49, 0.98) |
| Malignant neoplasm of esophagus (C15) | 92 | 142 | 0.65 (0.52, 0.79) | 119 | 0.77 (0.62, 0.94) | 0.93 (0.76, 1.15) |
| Malignant neoplasm of stomach (C16) | 46 | 62 | 0.74 (0.54, 0.97) | 51 | 0.90 (0.66, 1.18) | 1.06 (0.80, 1.42) |
| Malignant neoplasms of colon, rectum and anus (C18–C21) | 275 | 330 | 0.83 (0.74, 0.93) | 269 | 1.02 (0.91, 1.15) | 1.20 (1.07, 1.36) |
| Malignant neoplasms of liver and intrahepatic bile ducts (C22) | 71 | 117 | 0.61 (0.47, 0.75) | 98 | 0.73 (0.57, 0.90) | 0.87 (0.69, 1.10) |
| Malignant neoplasm of pancreas (C25) | 194 | 217 | 0.89 (0.77, 1.02) | 178 | 1.09 (0.94, 1.25) | 1.29 (1.12, 1.49) |
| Malignant neoplasm of larynx (C32) | 9 | 36 | 0.25 (0.11, 0.43) | 30 | 0.30 (0.14, 0.51) | 0.36 (0.19, 0.69) |
| Malignant neoplasms of trachea, bronchus and lung (C33–C34) | 703 | 1,260 | 0.56 (0.52, 0.60) | 1,042 | 0.67 (0.63, 0.73) | 0.79 (0.73, 0.85) |
| Malignant melanoma of skin (C43) | 73 | 72 | 1.02 (0.80, 1.26) | 59 | 1.23 (0.97, 1.53) | 1.47 (1.17, 1.85) |
| Malignant neoplasm of breast (C50) | 13 | 12 | 1.09 (0.58, 1.73) | 10 | 1.26 (0.67, 2.00) | 1.57 (0.91, 2.71) |
| Malignant neoplasm of ovary (C56) | 6 | 4 | 1.46 (0.53, 2.73) | 4 | 1.67 (0.61, 3.12) | 2.11 (0.95, 4.70) |
| Malignant neoplasm of prostate (C61) | 308 | 318 | 0.97 (0.86, 1.08) | 230 | 1.34 (1.19, 1.49) | 1.41 (1.26, 1.58) |
| Malignant neoplasms of kidney and renal pelvis (C64–C65) | 108 | 120 | 0.90 (0.74, 1.08) | 99 | 1.09 (0.90, 1.31) | 1.30 (1.08, 1.58) |
| Malignant neoplasm of bladder (C67) | 80 | 116 | 0.69 (0.55, 0.85) | 86 | 0.93 (0.74, 1.15) | 1.00 (0.80, 1.24) |
| Malignant neoplasms of meninges, brain and other CNS parts (C70–C72) | 87 | 100 | 0.87 (0.69, 1.06) | 87 | 1.00 (0.80, 1.22) | 1.25 (1.01, 1.54) |
| Malignant neoplasms of lymphoid, hematopoietic and related tissue (C81–C96) | 408 | 389 | 1.05 (0.95, 1.15) | 307 | 1.33 (1.20, 1.46) | 1.54 (1.39, 1.70) |
| Non–Hodgkin lymphoma (C82–C85) | 139 | 146 | 0.95 (0.80, 1.12) | 116 | 1.19 (1.00, 1.40) | 1.38 (1.17, 1.63) |
| Leukemia (C91–C95) | 168 | 156 | 1.08 (0.92, 1.25) | 122 | 1.38 (1.18, 1.59) | 1.56 (1.34, 1.82) |
| Multiple myeloma and immunoproliferative neoplasms (C88, C90) | 95 | 79 | 1.20 (0.97, 1.45) | 62 | 1.53 (1.23, 1.85) | 1.73 (1.42, 2.12) |
| Diabetes mellitus (E10–E14) | 228 | 393 | 0.58 (0.51, 0.66) | 312 | 0.73 (0.64, 0.83) | 0.83 (0.73, 0.95) |
| Parkinson’s disease (G20–G21) | 153 | 156 | 0.98 (0.83, 1.14) | 104 | 1.47 (1.25, 1.71) | 1.42 (1.21, 1.66) |
| Alzheimer’s disease (G30) | 212 | 286 | 0.74 (0.64, 0.84) | 175 | 1.21 (1.06, 1.38) | 1.07 (0.93, 1.22) |
| Major CVDs (I00–I78) | 3,106 | 4,452 | 0.70 (0.67, 0.72) | 3,403 | 0.91 (0.88, 0.95) | 1.01 (0.97, 1.06) |
| Diseases of heart (I00–I09, I11, I13, I20–I51) | 2,431 | 3,517 | 0.69 (0.66, 0.72) | 2,712 | 0.90 (0.86, 0.93) | 1.00 (0.95, 1.04) |
| Acute rheumatic fever and chronic rheumatic heart diseases (I00–I09) | 14 | 13 | 1.05 (0.57, 1.64) | 10 | 1.44 (0.79, 2.25) | 1.51 (0.89, 2.55) |
| IHDs (I20–I25) | 1,800 | 2,666 | 0.68 (0.64, 0.71) | 2,091 | 0.86 (0.82, 0.90) | 0.97 (0.92, 1.02) |
| Acute myocardial infarction (I21–I22) | 666 | 931 | 0.72 (0.66, 0.77) | 751 | 0.89 (0.82, 0.96) | 1.03 (0.95, 1.12) |
| Other acute IHD (I24) | 14 | 13 | 1.09 (0.60, 1.71) | 10 | 1.42 (0.78, 2.23) | 1.57 (0.93, 2.66) |
| Other forms of chronic IHD (I20, I25) | 1120 | 1,721 | 0.65 (0.61, 0.69) | 1,329 | 0.84 (0.79, 0.89) | 0.93 (0.87, 0.99) |
| Atherosclerotic CVD, so described (I25.0) | 225 | 418 | 0.54 (0.47, 0.61) | 340 | 0.66 (0.58, 0.75) | 0.77 (0.67, 0.88) |
| All other forms of chronic IHD (I20, I25.1–I25.9) | 895 | 1,302 | 0.69 (0.64, 0.73) | 989 | 0.91 (0.85, 0.97) | 0.99 (0.92, 1.06) |
| Other heart diseases (I26–I51) | 552 | 715 | 0.77 (0.71, 0.84) | 519 | 1.06 (0.98, 1.15) | 1.12 (1.03, 1.22) |
| Acute and subacute endocarditis (I33) | 6 | 4 | 1.47 (0.54, 2.74) | 3 | 1.84 (0.67, 3.44) | 2.12 (0.95, 4.72) |
| Heart failure (I50) | 134 | 182 | 0.74 (0.62, 0.87) | 119 | 1.12 (0.94, 1.32) | 1.06 (0.89, 1.26) |
| All other forms of heart disease (I26–I28, I34–I38, I42–I49, I51) | 410 | 521 | 0.79 (0.71, 0.87) | 389 | 1.05 (0.95, 1.16) | 1.14 (1.03, 1.26) |
| Hypertensive disease (I10–I15) c | 124 | 210 | 0.59 (0.49, 0.70) | 156 | 0.80 (0.66, 0.94) | 0.84 (0.71, 1.01) |
| Cerebrovascular diseases (I60–I69) | 492 | 649 | 0.76 (0.69, 0.83) | 477 | 1.03 (0.94, 1.12) | 1.10 (1.00, 1.20) |
| Atherosclerosis (I70) | 27 | 49 | 0.55 (0.36, 0.77) | 36 | 0.74 (0.49, 1.04) | 0.79 (0.54, 1.15) |
| Other diseases of circulatory system (I71–I78) | 97 | 144 | 0.67 (0.54, 0.81) | 113 | 0.86 (0.70, 1.04) | 0.97 (0.79, 1.18) |
| Aortic aneurysm and dissection (I71) | 63 | 96 | 0.66 (0.51, 0.83) | 77 | 0.82 (0.63, 1.04) | 0.95 (0.74, 1.21) |
| Other disorders of circulatory system (I80–I99) | 18 | 18 | 1.01 (0.60, 1.52) | 14 | 1.25 (0.74, 1.87) | 1.46 (0.92, 2.32) |
| Influenza and pneumonia (J09–J18) | 159 | 276 | 0.58 (0.49, 0.67) | 194 | 0.82 (0.70, 0.95) | 0.83 (0.71, 0.97) |
| Chronic lower respiratory diseases (J40–J47) | 440 | 919 | 0.48 (0.44, 0.52) | 695 | 0.63 (0.58, 0.69) | 0.68 (0.62, 0.74) |
| Emphysema (J43) | 34 | 79 | 0.43 (0.30, 0.59) | 64 | 0.53 (0.37, 0.73) | 0.62 (0.44, 0.87) |
| Asthma (J45–J46) | 9 | 7 | 1.24 (0.57, 2.12) | 6 | 1.56 (0.71, 2.67) | 1.79 (0.93, 3.44) |
| Other chronic lower respiratory diseases (J44, J47) | 397 | 826 | 0.48 (0.43, 0.53) | 620 | 0.64 (0.58, 0.70) | 0.68 (0.61, 0.75) |
| Pneumonitis due to solids and liquids (J69) | 67 | 121 | 0.55 (0.43, 0.69) | 81 | 0.82 (0.64, 1.03) | 0.80 (0.63, 1.01) |
| Other diseases of respiratory system (J00–J06, J30– J39, J67, J70–J98) | 171 | 184 | 0.93 (0.80, 1.07) | 138 | 1.24 (1.06, 1.43) | 1.34 (1.16, 1.56) |
| Peptic ulcer (K25–K28) | 11 | 20 | 0.56 (0.28, 0.92) | 15 | 0.73 (0.36, 1.20) | 0.81 (0.45, 1.47) |
| Chronic liver disease and cirrhosis (K70, K73–K74) | 63 | 194 | 0.32 (0.25, 0.41) | 173 | 0.36 (0.28, 0.46) | 0.47 (0.36, 0.60) |
| Infections of the skin and subcutaneous tissue (L00–L08) c | 7 | 8 | 0.89 (0.36, 1.61) | 5 | 1.32 (0.53, 2.38) | 1.28 (0.61, 2.68) |
| Other disorders of the skin and subcutaneous tissue (L80–L98) c | 6 | 4 | 1.38 (0.50, 2.58) | 3 | 2.05 (0.75, 3.82) | 1.98 (0.89, 4.41) |
| Arthropathies (M00–M25) c | 16 | 16 | 1.03 (0.59, 1.57) | 11 | 1.40 (0.80, 2.13) | 1.48 (0.90, 2.41) |
| Systemic connective tissue disorders (M30–M35) c | 11 | 11 | 1.04 (0.52, 1.70) | 9 | 1.24 (0.62, 2.03) | 1.48 (0.82, 2.68) |
| Osteopathies and chondropathies (M80–M94)c | 11 | 9 | 1.25 (0.62, 2.05) | 6 | 1.91 (0.95, 3.13) | 1.79 (0.99, 3.24) |
| Nephritis, nephrotic syndrome and nephrosis (N00–N07, N17–N19, N25–N27) | 150 | 196 | 0.76 (0.65, 0.89) | 140 | 1.07 (0.91, 1.25) | 1.11 (0.94, 1.3) |
| Chronic glomerulonephritis, nephritis and nephropathy not specified as acute or chronic, and renal sclerosis unspecified (N02–N03, N05–N07, N26) | 11 | 7 | 1.50 (0.75, 2.46) | 5 | 2.37 (1.18, 3.88) | 2.16 (1.19, 3.9) |
| Renal failure (N17–N19) | 138 | 185 | 0.74 (0.63, 0.87) | 132 | 1.04 (0.88, 1.22) | 1.07 (0.91, 1.27) |
| Accidents (unintentional injuries) (V01–X59, Y85–Y86) | 555 | 579 | 0.96 (0.88, 1.04) | 482 | 1.15 (1.06, 1.25) | 1.41 (1.30, 1.54) |
| Transport accidents (V01–V99, Y85) | 227 | 203 | 1.12 (0.98, 1.27) | 183 | 1.24 (1.08, 1.41) | 1.63 (1.43, 1.86) |
| Motor vehicle accidents (V02–V04, V09.0, V09.2, V12–V14, V19.0–V19.2, V19.4–V19.6, V20–V79, V80.3–V80.5, V81.0–V81.1, V82.0–V82.1, V83–V86, V87.0–V87.8, V88.0–V88.8, V89.0, V89.2) | 211 | 188 | 1.12 (0.98, 1.28) | 169 | 1.25 (1.08, 1.42) | 1.63 (1.42, 1.87) |
| Other land transport accidents (V01, V05–V06, V09.1, V09.3–V09.9, V10–V11, V15–V18, V19.3, V19.8–V19.9, V80.0–V80.2, V80.6–V80.9, V81.2–V81.9, V82.2–V82.9, V87.9, V88.9, V89.1, V89.3, V89.9) | 6 | 1 | 4.24 (1.55, 7.92) | 1 | 4.47 (1.63, 8.35) | 6.11 (2.74, 13.61) |
| Water, air and space, and other and unspecified transport accidents and their sequelae (V90–V99, Y85) | 10 | 8 | 1.20 (0.58, 2.01) | 8 | 1.26 (0.60, 2.10) | 1.73 (0.93, 3.22) |
| Non-transport accidents (W00–X59, Y86) | 328 | 376 | 0.87 (0.78, 0.97) | 299 | 1.10 (0.98, 1.22) | 1.27 (1.14, 1.42) |
| Falls (W00–W19) | 117 | 159 | 0.74 (0.61, 0.87) | 112 | 1.05 (0.87, 1.24) | 1.06 (0.88, 1.27) |
| Accidental drowning and submersion (W65–W74) | 7 | 10 | 0.72 (0.29, 1.31) | 9 | 0.79 (0.32, 1.42) | 1.04 (0.50, 2.19) |
| Accidental exposure to smoke, fire and flames (X00–X09) | 20 | 17 | 1.19 (0.73, 1.76) | 14 | 1.39 (0.85, 2.04) | 1.72 (1.11, 2.67) |
| Accidental poisoning and exposure to noxious substances (X40–X49) | 19 | 71 | 0.27 (0.16, 0.40) | 68 | 0.28 (0.17, 0.42) | 0.38 (0.24, 0.60) |
| Exposure to inanimate mechanical forces (W20–W31, W35–W49) | 91 | 29 | 3.17 (2.55, 3.85) | 26 | 3.56 (2.87, 4.33) | 4.60 (3.74, 5.65) |
| Other and unspecified non-transport accidents and their sequelae (W50–W64, W75–W99, X10–X39, X50–X59, Y86) | 71 | 82 | 0.87 (0.68, 1.08) | 63 | 1.12 (0.87, 1.39) | 1.25 (0.99, 1.58) |
| Intentional self–harm (suicide) (U03, X60–X84, Y87.0) | 122 | 213 | 0.57 (0.48, 0.68) | 193 | 0.63 (0.52, 0.75) | 0.83 (0.69, 0.99) |
| Assault (homicide) (U01–U02, X85–Y09, Y87.1) | 15 | 26 | 0.58 (0.33, 0.90) | 24 | 0.62 (0.34, 0.95) | 0.84 (0.51, 1.40) |
| Complications of medical and surgical care (Y40–Y84, Y88) | 14 | 14 | 0.99 (0.54, 1.55) | 11 | 1.29 (0.70, 2.01) | 1.43 (0.85, 2.41) |
Abbreviations: CI, confidence intervals; CMR, causal mortality ratio; CNS, central nervous system; CVD, cardiovascular disease; Ec, expected number of deaths for CMR calculation; Es, expected number of deaths for SMR calculation; IHD, ischemic heart disease; O, observed number of deaths; rSMR, relative standardized mortality ratio; SMR, standardized mortality ratio.
International Classification of Disease - 10th revision codes presented in parenthesis can be found at https://wonder.cdc.gov/.
Adjusted for age, calendar year, sex, race, and state (Calculation of mortality ratio (O/E) using values from the table may not provide exact mortality ratio provided in the table because rounded E values are presented).
Obtained from ICD subchapter analysis.
The CMRs (i.e., when expected cases were calculated using the person-time among the general population) indicated deficits in overall mortality (CMRoverall=0.89, 95%CI: 0.87–0.91) among private applicators, but were significantly elevated for causes including intestinal infections, prostate cancer, lymphohematopoietic cancers, Parkinson’s disease, Alzheimer’s disease, other respiratory diseases (mainly interstitial pulmonary diseases with fibrosis, n=102), chronic glomerulonephritis, and some transport accidents. CMRs> 1 were also seen for skin cancer, breast and ovarian cancers in females, some specific cardiovascular diseases, asthma, and arthropathies to name a few. The rSMR results, though similar to the CMR results, also indicated elevated mortality for cancers of the colon, pancreas, kidney and renal pelvis, and brain; the rSMR for Alzheimer’s disease however was lower than the CMR. Results using 5-year age categories were generally similar but inflated for mycoses and land transport accidents (Supplementary Table 2). Results with all suppressed deaths assumed to be nine were attenuated for rare causes including mycoses, chronic glomerulonephritis, and arthropathies, and other land transport accidents (Supplementary Table 3). In the SMR analysis that used underlying and contributing causes, SMRs were largely below the null, but were elevated (≥1.20) for ovarian cancer and land transport and inanimate mechanical forces accidents (Supplementary Table 4).
Separate analyses by state revealed differences (Supplementary Table 5). The CMR for all causes combined was elevated in NC (CMR=1.10, 95%CI: 1.07–1.13), but showed a deficit in IA (CMR=0.72, 95%CI: 0.70–0.74), although the SMRs indicated reduced mortality rates for both states. The CMR for septicemia was elevated in NC but showed a significant deficit in IA. In NC, CMRs were also elevated for cancers of the colon, rectum, and anus; kidney and renal pelvis; and pancreas. Both SMR and CMR were elevated for skin cancer in IA but not in NC. CMRs for Alzheimer’s disease, many specific cardiovascular causes, other respiratory diseases, and renal diseases were modestly elevated in NC but not in IA.
Overall SMRs for never- and ever-smokers were less than 1 (Table 3); however, CMR results indicated slightly higher-than-expected overall mortality in smokers (CMR=1.03, 95%CI: 1.00–1.06)) and lower mortality in never-smokers (CMR=0.70, 95%CI: 0.68–0.72)). Never-smokers generally experienced larger deficits in deaths from cancers and cardiovascular causes than ever-smokers. The CMR for Parkinson’s disease, although elevated in both smoking categories, was higher in never-smokers than in ever-smokers. The CMR for other respiratory diseases was significantly elevated in ever-smokers. Both the SMR and CMR were elevated for chronic glomerulonephritis in ever-smokers, but few deaths (n=3) limited analysis among never-smokers.
Table 3:
Standardized mortality ratios (SMRs), causal mortality ratios (CMRs), and their 95% confidence intervals based on underlying cause of death for private pesticide applicators in the Agricultural Health Study for the period 1999–2015 stratified by smoking status.
| Never smoker (n=26,667) | Ever smoker (n=23,647) | |||||
|---|---|---|---|---|---|---|
| Causea | O | SMR (95% CI)b | CMR (95% CI)b | O | SMR (95% CI)b | CMR (95% CI)b |
| All causes | 3381 | 0.54 (0.52, 0.56) | 0.70 (0.68, 0.72) | 5519 | 0.79 (0.77, 0.82) | 1.03 (1.00, 1.06) |
| Intestinal infectious diseases c | 11 | 0.90 (0.45, 1.48) | 1.37 (0.68, 2.25) | 15 | 1.06 (0.59, 1.64) | 1.60 (0.89, 2.47) |
| Septicemia | 41 | 0.61 (0.44, 0.81) | 0.80 (0.57, 1.06) | 58 | 0.70 (0.53, 0.90) | 0.93 (0.70, 1.18) |
| Malignant neoplasms | 925 | 0.54 (0.51, 0.58) | 0.66 (0.62, 0.71) | 1758 | 0.90 (0.86, 0.94) | 1.11 (1.06, 1.17) |
| Malignant neoplasms of lip, oral cavity and pharynx | 12 | 0.40 (0.21, 0.64) | 0.47 (0.24, 0.75) | 19 | 0.58 (0.35, 0.86) | 0.69 (0.42, 1.02) |
| Malignant neoplasm of esophagus | 28 | 0.42 (0.28, 0.59) | 0.50 (0.33, 0.69) | 64 | 0.89 (0.68, 1.12) | 1.06 (0.82, 1.33) |
| Malignant neoplasm of stomach | 12 | 0.43 (0.22, 0.69) | 0.52 (0.27, 0.83) | 33 | 1.02 (0.70, 1.39) | 1.25 (0.86, 1.70) |
| Malignant neoplasms of colon, rectum and anus | 113 | 0.74 (0.61, 0.88) | 0.90 (0.74, 1.07) | 153 | 0.91 (0.77, 1.05) | 1.11 (0.94, 1.30) |
| Malignant neoplasms of liver and intrahepatic bile ducts | 31 | 0.58 (0.39, 0.79) | 0.68 (0.46, 0.94) | 39 | 0.64 (0.46, 0.86) | 0.78 (0.55, 1.04) |
| Malignant neoplasm of pancreas | 76 | 0.76 (0.60, 0.94) | 0.92 (0.72, 1.14) | 111 | 0.99 (0.82, 1.18) | 1.21 (1.00, 1.45) |
| Malignant neoplasm of larynx | - | - | - | 7 | 0.37 (0.15, 0.67) | 0.44 (0.18, 0.80) |
| Malignant neoplasms of trachea, bronchus and lung | 73 | 0.13 (0.10, 0.16) | 0.16 (0.12, 0.19) | 590 | 0.88 (0.81, 0.96) | 1.07 (0.99, 1.16) |
| Malignant melanoma of skin | 37 | 1.12 (0.79, 1.50) | 1.34 (0.94, 1.79) | 34 | 0.92 (0.64, 1.25) | 1.13 (0.78, 1.53) |
| Malignant neoplasm of breast | 7 | 0.94 (0.37, 1.69) | 1.08 (0.43, 1.94) | 6 | 1.49 (0.54, 2.77) | 1.70 (0.62, 3.16) |
| Malignant neoplasm of prostate | 138 | 0.97 (0.82, 1.14) | 1.34 (1.13, 1.58) | 159 | 0.95 (0.81, 1.11) | 1.30 (1.11, 1.51) |
| Malignant neoplasms of kidney and renal pelvis | 44 | 0.79 (0.57, 1.04) | 0.95 (0.69, 1.24) | 63 | 1.03 (0.79, 1.30) | 1.25 (0.96, 1.58) |
| Malignant neoplasm of bladder | 22 | 0.42 (0.27, 0.61) | 0.57 (0.36, 0.82) | 57 | 0.94 (0.71, 1.20) | 1.26 (0.96, 1.61) |
| Malignant neoplasms of meninges, brain and other parts of CNS | 38 | 0.79 (0.56, 1.05) | 0.90 (0.64, 1.20) | 46 | 0.92 (0.67, 1.20) | 1.08 (0.79, 1.41) |
| Malignant neoplasms of lymphoid, hematopoietic and related tissue | 190 | 1.07 (0.92, 1.23) | 1.35 (1.16, 1.55) | 208 | 1.03 (0.89, 1.17) | 1.30 (1.13, 1.49) |
| Non–Hodgkin lymphoma | 62 | 0.93 (0.71, 1.17) | 1.15 (0.88, 1.46) | 71 | 0.94 (0.74, 1.17) | 1.18 (0.92, 1.47) |
| Leukemia | 78 | 1.10 (0.87, 1.36) | 1.40 (1.10, 1.72) | 87 | 1.07 (0.86, 1.31) | 1.37 (1.10, 1.67) |
| Multiple myeloma and immunoproliferative neoplasms | 48 | 1.35 (0.99, 1.75) | 1.70 (1.26, 2.21) | 46 | 1.11 (0.81, 1.45) | 1.41 (1.03, 1.84) |
| Diabetes mellitus | 80 | 0.44 (0.35, 0.54) | 0.56 (0.44, 0.68) | 135 | 0.67 (0.56, 0.78) | 0.84 (0.71, 0.99) |
| Parkinson’s disease | 83 | 1.19 (0.94, 1.45) | 1.79 (1.42, 2.19) | 68 | 0.83 (0.64, 1.03) | 1.23 (0.96, 1.54) |
| Alzheimer’s disease | 95 | 0.72 (0.58, 0.87) | 1.18 (0.96, 1.43) | 107 | 0.74 (0.60, 0.88) | 1.19 (0.97, 1.42) |
| Major CVDs | 1183 | 0.58 (0.55, 0.61) | 0.76 (0.71, 0.80) | 1774 | 0.78 (0.74, 0.81) | 1.01 (0.97, 1.06) |
| Diseases of heart | 909 | 0.56 (0.53, 0.60) | 0.73 (0.68, 0.77) | 1405 | 0.78 (0.74, 0.82) | 1.01 (0.96, 1.06) |
| Acute rheumatic fever and chronic rheumatic heart diseases | 7 | 1.12 (0.45, 2.01) | 1.54 (0.62, 2.78) | 7 | 1.04 (0.41, 1.87) | 1.41 (0.56, 2.53) |
| IHDs | 643 | 0.52 (0.48, 0.56) | 0.66 (0.61, 0.71) | 1066 | 0.78 (0.74, 0.83) | 1.00 (0.94, 1.06) |
| Acute myocardial infarction | 237 | 0.55 (0.48, 0.62) | 0.68 (0.59, 0.77) | 389 | 0.82 (0.74, 0.90) | 1.02 (0.92, 1.12) |
| Other acute IHDs | 7 | 1.37 (0.55, 2.48) | 1.79 (0.72, 3.23) | 7 | 0.96 (0.39, 1.74) | 1.25 (0.50, 2.25) |
| Other forms of chronic IHD | 399 | 0.50 (0.45, 0.55) | 0.65 (0.59, 0.71) | 670 | 0.76 (0.71, 0.82) | 0.98 (0.91, 1.06) |
| Atherosclerotic CVD, so described | 83 | 0.39 (0.31, 0.48) | 0.48 (0.38, 0.59) | 136 | 0.68 (0.57, 0.80) | 0.84 (0.71, 0.99) |
| All other forms of chronic IHD | 316 | 0.54 (0.48, 0.60) | 0.71 (0.64, 0.79) | 534 | 0.79 (0.72, 0.85) | 1.03 (0.94, 1.12) |
| Other heart diseases | 235 | 0.74 (0.65, 0.83) | 1.01 (0.88, 1.14) | 291 | 0.78 (0.69, 0.87) | 1.07 (0.95, 1.20) |
| Heart failure | 49 | 0.63 (0.47, 0.82) | 0.97 (0.72, 1.26) | 76 | 0.77 (0.61, 0.95) | 1.16 (0.92, 1.44) |
| All other forms of heart disease | 183 | 0.78 (0.67, 0.89) | 1.03 (0.89, 1.19) | 210 | 0.78 (0.68, 0.89) | 1.04 (0.90, 1.18) |
| Hypertensive diseasesc | 53 | 0.54 (0.40, 0.69) | 0.72 (0.54, 0.93) | 66 | 0.63 (0.49, 0.79) | 0.84 (0.65, 1.06) |
| Cerebrovascular diseases | 210 | 0.72 (0.62, 0.82) | 0.97 (0.85, 1.11) | 259 | 0.77 (0.68, 0.87) | 1.04 (0.92, 1.17) |
| Atherosclerosis | 8 | 0.34 (0.14, 0.59) | 0.46 (0.20, 0.80) | 18 | 0.75 (0.44, 1.12) | 1.00 (0.59, 1.50) |
| Other diseases of circulatory system | 27 | 0.41 (0.27, 0.58) | 0.53 (0.35, 0.74) | 67 | 0.89 (0.69, 1.11) | 1.13 (0.88, 1.42) |
| Aortic aneurysm and dissection | 15 | 0.35 (0.19, 0.53) | 0.43 (0.24, 0.66) | 45 | 0.90 (0.66, 1.18) | 1.13 (0.82, 1.47) |
| Other disorders of circulatory system | 8 | 0.95 (0.41, 1.67) | 1.16 (0.50, 2.03) | 9 | 1.01 (0.46, 1.73) | 1.25 (0.57, 2.14) |
| Influenza and pneumonia | 66 | 0.52 (0.40, 0.65) | 0.75 (0.58, 0.94) | 85 | 0.60 (0.48, 0.74) | 0.85 (0.68, 1.04) |
| Chronic lower respiratory diseases | 57 | 0.14 (0.10, 0.18) | 0.18 (0.14, 0.23) | 355 | 0.73 (0.66, 0.81) | 0.97 (0.87, 1.07) |
| Emphysema | - | - | - | 30 | 0.70 (0.47, 0.97) | 0.86 (0.58, 1.19) |
| Asthma | 6 | 1.78 (0.65, 3.31) | 2.25 (0.82, 4.20) | - | - | - |
| Other chronic lower respiratory diseases | 51 | 0.14 (0.10, 0.18) | 0.18 (0.14, 0.24) | 322 | 0.74 (0.66, 0.83) | 0.99 (0.88, 1.10) |
| Pneumonitis due to solids and liquids | 32 | 0.59 (0.40, 0.80) | 0.87 (0.60, 1.20) | 32 | 0.51 (0.35, 0.70) | 0.75 (0.51, 1.03) |
| Other diseases of respiratory system | 58 | 0.71 (0.54, 0.91) | 0.95 (0.72, 1.21) | 107 | 1.11 (0.91, 1.32) | 1.47 (1.21, 1.76) |
| Peptic ulcer | - | - | - | 6 | 0.60 (0.22, 1.12) | 0.77 (0.28, 1.45) |
| Chronic liver disease and cirrhosis | 22 | 0.24 (0.15, 0.34) | 0.26 (0.16, 0.38) | 38 | 0.39 (0.28, 0.53) | 0.45 (0.32, 0.60) |
| Arthropathies c | - | - | - | 12 | 1.50 (0.77, 2.41) | 2.01 (1.04, 3.24) |
| Systemic connective tissue disorders c | 6 | 1.28 (0.47, 2.38) | 1.52 (0.55, 2.84) | - | - | - |
| Osteopathies and chondropathies c | 6 | 1.51 (0.55, 2.81) | 2.31 (0.84, 4.31) | - | - | - |
| Nephritis, nephrotic syndrome and nephrosis | 64 | 0.76 (0.59, 0.96) | 1.08 (0.83, 1.35) | 77 | 0.72 (0.57, 0.89) | 1.01 (0.80, 1.25) |
| Chronic glomerulonephritis, nephritis and nephropathy not specified as acute or chronic, renal sclerosis unspecified | - | - | - | 7 | 1.81 (0.73, 3.27) | 2.82 (1.13, 5.08) |
| Renal failure | 60 | 0.76 (0.58, 0.96) | 1.07 (0.82, 1.35) | 70 | 0.70 (0.54, 0.87) | 0.97 (0.76, 1.21) |
| Accidents (unintentional injuries) | 229 | 0.79 (0.69, 0.89) | 0.93 (0.81, 1.05) | 305 | 1.11 (0.99, 1.24) | 1.36 (1.21, 1.52) |
| Transport accidents | 89 | 0.84 (0.68, 1.03) | 0.92 (0.74, 1.12) | 131 | 1.41 (1.18, 1.66) | 1.58 (1.33, 1.87) |
| Motor vehicle accidents | 79 | 0.81 (0.64, 1.00) | 0.89 (0.70, 1.09) | 125 | 1.45 (1.21, 1.72) | 1.64 (1.36, 1.93) |
| Water, air and space, and other and unspecified transport accidents and their sequelae | 7 | 1.55 (0.62, 2.80) | 1.61 (0.64, 2.90) | - | - | - |
| Non-transport accidents | 140 | 0.75 (0.63, 0.88) | 0.93 (0.78, 1.09) | 174 | 0.96 (0.83, 1.11) | 1.23 (1.05, 1.42) |
| Falls | 41 | 0.54 (0.39, 0.72) | 0.76 (0.55, 1.01) | 74 | 0.93 (0.73, 1.15) | 1.33 (1.04, 1.65) |
| Accidental drowning and submersion | - | - | - | 6 | 1.39 (0.51, 2.59) | 1.54 (0.56, 2.87) |
| Accidental exposure to smoke, fire and flames | 8 | 1.00 (0.43, 1.75) | 1.15 (0.49, 2.01) | 11 | 1.32 (0.66, 2.16) | 1.55 (0.77, 2.54) |
| Accidental poisoning and exposure to noxious | - | - | - | 12 | 0.40 (0.21, 0.64) | 0.42 (0.22, 0.68) |
| Exposure to inanimate mechanical forces | 53 | 3.60 (2.69, 4.62) | 3.98 (2.98, 5.11) | 36 | 2.67 (1.87, 3.60) | 3.05 (2.14, 4.11) |
| Other and unspecified non-transport accidents and their sequelae | 32 | 0.83 (0.57, 1.14) | 1.06 (0.72, 1.45) | 34 | 0.83 (0.57, 1.12) | 1.07 (0.74, 1.46) |
| Intentional self–harm (suicide) | 48 | 0.43 (0.32, 0.56) | 0.47 (0.34, 0.61) | 66 | 0.69 (0.53, 0.86) | 0.77 (0.59, 0.96) |
| Assault (homicide) | - | - | - | 10 | 0.84 (0.40, 1.40) | 0.90 (0.43, 1.50) |
| Complications of medical and surgical care | - | - | - | 10 | 1.35 (0.64, 2.25) | 1.73 (0.83, 2.90) |
Note: CI, confidence intervals; CMR, causal mortality ratio; CNS, central nervous system; CVD, cardiovascular disease; IHD, ischemic heart disease; O, observed number of deaths; SMR, standardized mortality ratio.
International Classification of Disease - 10th revision codes can be found at https://wonder.cdc.gov/.
Adjusted for age, calendar year, sex, race, and state (calculation of mortality ratio (O/E) using values from the table may not provide exact mortality ratio provided in the table because rounded E values are presented).
Obtained from ICD subchapter analysis.
Female spouses
In spouses, overall deaths (SMRoverall=0.61, 95%CI: 0.59–0.63; CMRoverall=0.74, 95%CI: 0.71–0.76) as well as deaths from many underlying causes were lower than expected (Table 4). Both SMRs and CMRs were significantly elevated for certain immune-related disorders (7 of 8 deaths from sarcoidosis) and for exposure to inanimate forces, as was the CMR for non-Hodgkin lymphoma. CMRs for skin cancer and assault by firearm discharge were also non-significantly elevated. rSMRs were elevated for intestinal infections, most cancers, and Parkinson’s disease, and a few external causes. Estimates for immune-related disorders, and firearm discharge were stronger when 5-year age categories were used, but estimates could not be calculated for inanimate forces due to zero expected deaths; estimates for these causes were attenuated when suppressed values were replaced by nine (Supplementary Tables 6 and 7). Except for inanimate forces, SMRs for underlying and contributing causes combined indicated lower mortality rates for nearly all causes (Supplementary Table 8). Mortality deficits were generally larger in never-smokers than in ever-smokers and in IA than in NC (Supplementary Tables 9 and 10).
Table 4:
Standardized mortality ratios (SMRs), causal mortality ratios (CMRs), relative SMRs, and their 95% confidence intervals based on underlying cause of death for female spouses of private pesticide applicators in the Agricultural Health Study for the period 1999–2015 (n=31,867).
| Causea | O | Es | SMR (95% CI) b | Ec | CMR (95% CI) b | rSMR (95% CI) b |
|---|---|---|---|---|---|---|
| All causes | 3,384 | 5,589 | 0.61 (0.59, 0.63) | 4,582 | 0.74 (0.71, 0.76) | - |
| Intestinal infectious diseases (A00–A09)c | 15 | 18 | 0.83 (0.47, 1.29) | 13 | 1.11 (0.62, 1.72) | 1.35 (0.81, 2.25) |
| Septicemia (A40–A41) | 43 | 74 | 0.58 (0.42, 0.76) | 61 | 0.70 (0.51, 0.92) | 0.93 (0.69, 1.26) |
| Malignant neoplasms (C00–C97) | 1,176 | 1,625 | 0.72 (0.68, 0.77) | 1,431 | 0.82 (0.78, 0.87) | 1.29 (1.20, 1.38) |
| Malignant neoplasm of esophagus (C15) | 6 | 16 | 0.37 (0.13, 0.68) | 14 | 0.42 (0.15, 0.77) | 0.59 (0.26, 1.31) |
| Malignant neoplasm of stomach (C16) | 13 | 18 | 0.73 (0.39, 1.15) | 16 | 0.84 (0.44, 1.33) | 1.17 (0.68, 2.02) |
| Malignant neoplasms of colon, rectum and anus (C18–C21) | 125 | 143 | 0.88 (0.73, 1.04) | 124 | 1.01 (0.84, 1.19) | 1.43 (1.20, 1.71) |
| Malignant neoplasms of liver and intrahepatic bile ducts (C22) | 25 | 29 | 0.85 (0.55, 1.21) | 25 | 0.98 (0.64, 1.40) | 1.38 (0.93, 2.05) |
| Malignant neoplasm of pancreas (C25) | 79 | 98 | 0.81 (0.64, 0.99) | 85 | 0.93 (0.73, 1.14) | 1.31 (1.05, 1.63) |
| Malignant neoplasms of trachea, bronchus and lung (C33–C34) | 208 | 463 | 0.45 (0.39, 0.51) | 412 | 0.50 (0.44, 0.58) | 0.71 (0.61, 0.81) |
| Malignant melanoma of skin (C43) | 22 | 20 | 1.10 (0.69, 1.60) | 18 | 1.23 (0.77, 1.78) | 1.78 (1.17, 2.71) |
| Malignant neoplasm of breast (C50) | 198 | 235 | 0.84 (0.73, 0.96) | 212 | 0.94 (0.81, 1.07) | 1.38 (1.20, 1.60) |
| Malignant neoplasm of cervix uteri (C53) | 12 | 21 | 0.57 (0.29, 0.92) | 20 | 0.61 (0.31, 0.98) | 0.92 (0.52, 1.62) |
| Malignant neoplasms of corpus uteri and uterus, part unspecified (C54–C55) | 41 | 51 | 0.81 (0.58, 1.07) | 45 | 0.91 (0.65, 1.2) | 1.30 (0.96, 1.77) |
| Malignant neoplasm of ovary (C56) | 94 | 96 | 0.98 (0.80, 1.19) | 85 | 1.10 (0.89, 1.33) | 1.61 (1.31, 1.97) |
| Malignant neoplasms of kidney and renal pelvis (C64–C65) | 26 | 30 | 0.87 (0.57, 1.23) | 26 | 1.00 (0.65, 1.41) | 1.41 (0.96, 2.08) |
| Malignant neoplasm of bladder (C67) | 13 | 20 | 0.65 (0.34, 1.03) | 17 | 0.78 (0.42, 1.24) | 1.04 (0.61, 1.80) |
| Malignant neoplasms of meninges, brain and other parts of CNS (C70–C72) | 37 | 43 | 0.86 (0.61, 1.16) | 39 | 0.95 (0.67, 1.28) | 1.40 (1.01, 1.93) |
| Malignant neoplasms of lymphoid, hematopoietic and related tissue (C81–96) | 133 | 143 | 0.93 (0.78, 1.09) | 123 | 1.08 (0.91, 1.27) | 1.53 (1.28, 1.82) |
| Non–Hodgkin lymphoma (C82–C85) | 65 | 57 | 1.15 (0.88, 1.44) | 49 | 1.33 (1.03, 1.68) | 1.87 (1.46, 2.39) |
| Leukemia (C91–C95) | 47 | 51 | 0.91 (0.67, 1.19) | 44 | 1.07 (0.78, 1.39) | 1.48 (1.11, 1.98) |
| Multiple myeloma and immunoproliferative neoplasms (C88, C90) | 18 | 31 | 0.58 (0.34, 0.86) | 27 | 0.67 (0.40, 1.01) | 0.93 (0.59, 1.48) |
| Certain disorders involving the immune mechanism (D80–D89) | 8 | 2 | 3.73 (1.60, 6.53) | 2 | 4.38 (1.89, 7.67) | 6.06 (3.03, 12.14) |
| Diabetes mellitus (E10–E14) | 96 | 168 | 0.57 (0.46, 0.69) | 143 | 0.67 (0.54, 0.81) | 0.92 (0.75, 1.13) |
| Parkinson’s disease (G20–G21) | 32 | 39 | 0.83 (0.57, 1.13) | 29 | 1.01 (0.75, 1.51) | 1.34 (0.95, 1.90) |
| Alzheimer’s disease (G30) | 133 | 208 | 0.64 (0.54, 0.75) | 143 | 0.93 (0.78, 1.1) | 1.03 (0.87, 1.23) |
| Major CVDs (I00–I78) | 939 | 1,601 | 0.59 (0.55, 0.62) | 1,270 | 0.74 (0.69, 0.79) | 0.95 (0.88, 1.02) |
| Diseases of heart (I00–I09, I11, I13, I20–I51) | 663 | 1,154 | 0.57 (0.53, 0.62) | 923 | 0.72 (0.66, 0.77) | 0.92 (0.84, 1.00) |
| Acute rheumatic fever and chronic rheumatic heart diseases (I00–I09) | 12 | 13 | 0.94 (0.48, 1.51) | 11 | 1.13 (0.58, 1.83) | 1.52 (0.86, 2.67) |
| IHDs (I20–I25) | 414 | 762 | 0.54 (0.49, 0.60) | 620 | 0.67 (0.60, 0.73) | 0.86 (0.78, 0.95) |
| Acute myocardial infarction (I21–I22) | 165 | 274 | 0.60 (0.51, 0.70) | 227 | 0.73 (0.62, 0.84) | 0.97 (0.83, 1.14) |
| Other forms of chronic IHD (I20, I25) | 248 | 483 | 0.51 (0.45, 0.58) | 390 | 0.64 (0.56, 0.72) | 0.81 (0.72, 0.93) |
| Atherosclerotic CVD, so described (I25.0) | 46 | 126 | 0.37 (0.27, 0.48) | 105 | 0.44 (0.32, 0.57) | 0.59 (0.44, 0.78) |
| All other forms of chronic IHD (I20, I25.1–I25.9) | 202 | 358 | 0.56 (0.49, 0.64) | 285 | 0.71 (0.62, 0.81) | 0.91 (0.79, 1.04) |
| Other heart diseases (I26–I51) | 211 | 325 | 0.65 (0.56, 0.74) | 251 | 0.84 (0.73, 0.96) | 1.05 (0.91, 1.21) |
| Heart failure (I50) | 41 | 83 | 0.50 (0.36, 0.66) | 59 | 0.70 (0.50, 0.92) | 0.80 (0.59, 1.09) |
| All other forms of heart disease (I26–I28, I34–I38, I42–I49, I51) | 167 | 237 | 0.70 (0.60, 0.81) | 187 | 0.89 (0.76, 1.03) | 1.14 (0.98, 1.34) |
| Hypertensive heart and renal diseases and essential hypertension (I10–I15)c | 50 | 102 | 0.49 (0.36, 0.63) | 78 | 0.64 (0.48, 0.83) | 0.79 (0.60, 1.05) |
| Cerebrovascular diseases (I60–I69) | 214 | 330 | 0.65 (0.57, 0.74) | 258 | 0.83 (0.72, 0.94) | 1.05 (0.91, 1.21) |
| Atherosclerosis (I70) | 8 | 19 | 0.42 (0.18, 0.73) | 15 | 0.54 (0.23, 0.95) | 0.67 (0.34, 1.35) |
| Other diseases of circulatory system (I71–I78) | 30 | 44 | 0.68 (0.46, 0.93) | 36 | 0.83 (0.56, 1.14) | 1.10 (0.77, 1.58) |
| Aortic aneurysm and dissection (I71) | 15 | 23 | 0.64 (0.36, 0.99) | 19 | 0.77 (0.43, 1.19) | 1.03 (0.62, 1.72) |
| Influenza and pneumonia (J09–J18) | 53 | 118 | 0.45 (0.34, 0.58) | 91 | 0.59 (0.44, 0.75) | 0.73 (0.56, 0.96) |
| Chronic lower respiratory diseases (J40–J47) | 143 | 426 | 0.34 (0.28, 0.39) | 356 | 0.40 (0.34, 0.47) | 0.53 (0.45, 0.63) |
| Emphysema (J43) | 8 | 29 | 0.28 (0.12, 0.49) | 25 | 0.32 (0.14, 0.56) | 0.45 (0.23, 0.90) |
| Other chronic lower respiratory diseases (J44, J47) | 130 | 384 | 0.34 (0.28, 0.40) | 321 | 0.41 (0.34, 0.48) | 0.53 (0.44, 0.63) |
| Pneumonitis due to solids and liquids (J69) | 24 | 39 | 0.62 (0.39, 0.88) | 30 | 0.80 (0.51, 1.15) | 0.99 (0.67, 1.48) |
| Other diseases of respiratory system (J00–J06, J30– J39, J67, J70–J98) | 43 | 71 | 0.60 (0.44, 0.80) | 58 | 0.74 (0.54, 0.98) | 0.98 (0.72, 1.32) |
| Chronic liver disease and cirrhosis (K70, K73–K74) | 25 | 58 | 0.43 (0.28, 0.61) | 53 | 0.47 (0.31, 0.67) | 0.71 (0.48, 1.05) |
| Arthropathies (M00–M25)c | 13 | 16 | 0.82 (0.44, 1.30) | 13 | 1.00 (0.53, 1.59) | 1.33 (0.77, 2.30) |
| Systemic connective tissue disorders (M30–M35)c | 12 | 16 | 0.74 (0.38, 1.20) | 14 | 0.83 (0.43, 1.34) | 1.21 (0.68, 2.13) |
| Nephritis, nephrotic syndrome and nephrosis (N00–N07, N17–N19, N25–N27) | 47 | 74 | 0.63 (0.46, 0.82) | 59 | 0.80 (0.59, 1.04) | 1.04 (0.78, 1.38) |
| Renal failure (N17–N19) | 41 | 70 | 0.58 (0.42, 0.77) | 56 | 0.74 (0.53, 0.98) | 0.94 (0.69, 1.28) |
| Accidents (unintentional injuries) (V01–X59, Y85–Y86) | 106 | 180 | 0.59 (0.48, 0.71) | 153 | 0.69 (0.57, 0.83) | 0.96 (0.79, 1.17) |
| Transport accidents (V01–V99, Y85) | 34 | 54 | 0.63 (0.44, 0.86) | 50 | 0.68 (0.47, 0.92) | 1.02 (0.73, 1.43) |
| Motor vehicle accidents (V02–V04, V09.0, V09.2, V12–V14, V19.0–V19.2, V19.4–V19.6, V20–V79, V80.3–V80.5, V81.0–V81.1, V82.0–V82.1, V83–V86, V87.0–V87.8, V88.0–V88.8, V89.0, V89.2) | 33 | 52 | 0.64 (0.44, 0.87) | 48 | 0.69 (0.47, 0.94) | 1.03 (0.73, 1.45) |
| Non-transport accidents (W00–X59, Y86) | 72 | 126 | 0.57 (0.45, 0.71) | 103 | 0.7 (0.55, 0.87) | 0.93 (0.73, 1.17) |
| Falls (W00–W19) | 39 | 59 | 0.66 (0.47, 0.88) | 44 | 0.89 (0.64, 1.19) | 1.07 (0.78, 1.47) |
| Exposure to inanimate mechanical forces (W20–W31, W35–W49) | 7 | 0 | 19.77 (7.92, 35.64) | 0 | 25.10 (10.06, 45.26) | 31.96 (15.23, 67.09) |
| Other and unspecified non-transport accidents and their sequelae (W50–W64, W75–W99, X10–X39, X50–X59, Y86) | 17 | 28 | 0.62 (0.36, 0.93) | 22 | 0.77 (0.45, 1.17) | 1 (0.62, 1.60) |
| Intentional self–harm (suicide) (U03, X60–X84, Y87.0) | 12 | 35 | 0.34 (0.18, 0.56) | 34 | 0.36 (0.18, 0.58) | 0.56 (0.32, 0.99) |
| Assault (homicide) (U01–U02, X85–Y09, Y87.1) | 6 | 7 | 0.89 (0.33, 1.67) | 6 | 0.93 (0.34, 1.73) | 1.46 (0.66, 3.26) |
| Assault (homicide) by discharge of firearms (U01.4, X93–X95) | 6 | 2 | 2.42 (0.88, 4.52) | 2 | 2.51 (0.92, 4.68) | 3.91 (1.76, 8.71) |
| Complications of medical and surgical care (Y40–Y84, Y88) | 7 | 7 | 0.97 (0.39, 1.75) | 6 | 1.14 (0.46, 2.05) | 1.57 (0.75, 3.29) |
Abbreviations: CI, confidence interval; CMR, causal mortality ratio; CVD, cardiovascular disease; Ec, expected number of deaths for CMR calculation; Es, expected number of deaths for SMR calculation; IHD, ischemic heart disease; O, observed number of deaths; rSMR, relative standardized mortality ratio; SMR, standardized mortality ratio.
International Classification of Disease - 10th revision codes presented in parenthesis can be found at https://wonder.cdc.gov/.
Adjusted for age, calendar year, race, and state (calculation of mortality ratio (O/E) using values from the table may not provide exact mortality ratio provided in the table because rounded E values are presented).
Obtained from ICD subchapter analysis.
Male commercial applicators
Commercial applicators had lower-than-expected mortality rates in general (Supplementary Table 11).
DISCUSSION
In the AHS, using SMRs, mortality, overall and from many underlying causes, was generally lower than expected. These findings are consistent with prior AHS analyses with shorter follow-up.[6,7] However, when we estimated expected deaths using the CMR approach – which was not used in the earlier analyses – we detected higher-than-expected overall mortality for NC applicators and among smokers, as well as higher-than-expected cause-specific mortality for causes including prostate and lymphohematopoietic cancers, Parkinson’s disease, and Alzheimer’s disease in private applicators. CMR analysis also revealed deficits in mortality for several causes including oral cavity/pharynx cancers, and diabetes mellitus; however, these CMR deficits were smaller in magnitude than the corresponding SMR. Current rSMRs findings were similar to those reported previously[7] and, for some causes, comparable to CMR findings in the direction of association.
CMR results indicated that private applicators, specifically smokers and those from NC, experienced higher-than-expected overall mortality as well as mortality from several causes. Reasons for state-specific differences are less clear, although differences between states in occupational exposures, such as pesticides, endotoxins, and heat stress, and lifestyle factors, such as diet or smoking, may have contributed to these findings. NC had higher mortality than IA in the AHS, and overall age-adjusted mortality for the study period in the general population was also higher in NC than in IA.[12] Further, private applicators in NC had higher smoking prevalence and lower education than their IA counterparts and they also differed in their farm-related exposures, factors that may partly explain more unfavorable outcomes in NC.
We found significantly elevated CMRs and rSMRs for intestinal infections in private applicators. The most frequent cause of death in this category was Clostridium difficile-related enterocolitis, an illness that commonly affects immunocompromised individuals and that is often associated with exposures in health-care facilities,[14] although we did not find evidence of immune-suppression when we examined contributing causes of deaths for these individuals. Aging farmers may be at increased risk of contracting C. difficile infections due to age-related increases in susceptibility to infections together with frequent contact with livestock (a likely reservoir).[14,15]
Increased mortality from prostate and lymphohematopoietic cancers in private applicators, and from non-Hodgkin lymphoma in female spouses, as suggested by our CMR and rSMR results, is consistent with previously reported rSMR results[7] and, generally reflects cancer incidence patterns in the cohort.[16–18] Elevated mortality from these cancers in the AHS could be due to pesticide exposures,[19–21] although immunomodulatory responses elicited by infectious agents and endotoxins related to exposure to farm animals may also influence risk of lymphohematopoietic cancers.[22]
Although the number of breast and ovarian cancer deaths among private applicators was small, CMRs were modestly elevated. The CMR for skin cancer was also elevated for applicators and for female spouses. rSMRs were also elevated for these outcomes in the current, as well as, in the prior analyses.[7] Cancer incidence analyses in the AHS have consistently found elevated skin cancer (melanoma) incidence in female spouses and ovarian cancer in female private pesticide applicators.[16–18] Excess melanoma in farmers could be due to sunlight and pesticide exposures.[23,24] Not much evidence corroborates elevated incidence or mortality from female breast and ovarian cancers in farming populations,[1,3,25] although there have been some suggestive reports.[26,27]
We also observed elevated CMRs for neurodegenerative disorders, Parkinson’s disease and Alzheimer’s disease, both shown to be associated with pesticide use.[28,29] Further, we observed heterogeneity for Parkinson’s disease by smoking status with a higher CMR in never-smokers than in ever-smokers, consistent with prior observations that cigarette smoking may be protective for Parkinson’s disease.[30]
Although our findings indicated lower-than-expected overall mortality from cardiovascular diseases, we noted some interesting findings for private applicators in state-specific analyses. SMRs and CMRs in IA were generally below one, as were SMRs for many cardiovascular causes in NC; but mortality deficits in NC estimated using SMRs were smaller than those in IA. Further, CMRs for several cardiovascular causes in NC indicated more deaths than expected, which is contrary to the common finding that farmers are less likely to die from cardiovascular diseases than general populations due to their healthier lifestyles and the “healthy worker bias,” as higher physical demands required by farming may select for healthy individuals to become farmers. Although many studies support deficits in cardiovascular mortality in farmers, there are a few suggesting excess mortality from cardiovascular causes.[31,32] A few studies have explored associations between specific pesticides and cardiovascular diseases, but with limited findings.[33,34] Overall, increased CMRs in NC may simply be chance findings but deserve further evaluation.
Our findings show elevated mortality from chronic glomerulonephritis and unspecified nephritis in private applicators among NC participants and among smokers in particular. Interestingly, the CMRs for renal failure and kidney and renal pelvis cancer were also elevated in NC applicators. Several pesticides were also associated with end-stage renal disease in the AHS[35] and working in an agricultural crop production occupation was associated with kidney diseases in a NC-based case-control study,[36] consistent with our findings.
Private applicators, specifically smokers and those from NC, also experienced higher than expected mortality from other respiratory diseases, mostly interstitial pulmonary diseases with fibrosis. The higher risk of interstitial lung diseases in agricultural populations may be attributable to chronic exposures to airborne particles during farming.[37] Further, we saw some evidence of elevated mortality from causes such as sarcoidosis and arthropathies, although these estimates were based on few observed deaths and attenuated when nine observations in suppressed cells were assumed to represent deaths. These conditions potentially share a common underlying etiology involving immunopathogenic pathways, supporting hypotheses linking farm-related exposures and immunomodulation.[38]
Mortality from accidents, including motor-vehicle accidents, other land transport accidents, and machine- and animal-related accidents were higher than expected in private applicators. Motor-vehicle non-traffic accidents, machine-related deaths, and fatal collision with objects were also elevated in the prior analyses[7] and consistent with other US reports.[1,3,31]
The expected numbers of deaths estimated from the participants’ observed person-time using the SMR approach were consistently higher than those estimated using the CMR approach (i.e., using person-time expected in the absence of agricultural exposures).[10] This observation is a mathematical result of lower all-cause mortality rates in the study population relative to the general population. As a result, larger mortality deficits were suggested by SMRs as compared to CMRs and, for some causes, CMR findings suggested higher-than-expected mortality when SMRs were below the null. The CMR can estimate a causal effect when the referent rates accurately represent the rates in the study sample, had the key occupational exposures been set to the levels experienced by the referent population. In our case, however, this assumption is likely violated, given that the study may suffer from limitations such as uncontrolled and residual confounding and healthy worker biases. Still, CMR analysis may provide more reliable estimates of expected mortality than the SMR. Further, CMR findings are generally supported by rSMR findings, estimates that are likely less biased than SMR estimates[11].
An inferential challenge in the current context is what “exposure” means, since farming can entail a spectrum of exposures with deleterious effects (e.g., dusts, microbes/viruses, and pesticides), as well as with beneficial effects (e.g., physical activity). Further, an active farming cohort, such as the AHS, can have a strong “healthy worker” tendency. Consequently, mixing of effects from multiple exposures including beneficial factors could lead to results that varied by analytical approach. Further methodological studies under different exposure scenarios may help to evaluate our findings, in particular, whether confounding from unmeasured differences between the study population and general population has a differential impact on SMR, CMR, and rSMR.
Our study has several limitations. Death certificates may not always accurately identify causes of death.[39,40] We would expect that cause-of-death misclassification is approximately proportional between the general population used for comparison and our cohort, although given that AHS participants are from rural areas and that death certificate accuracy may differ by urbanicity, some systematic misclassification is possible. The death counts that were suppressed due to small numbers were assumed to be zero; when the highest possible value of nine was used instead of zero, estimates were attenuated for some rarer causes although robust for others. Causes of death were unknown for 16 participants, but this is very unlikely to have affected results. We performed stratified analyses by smoking status of participants to assess heterogeneity, although lack of smoking-specific data for the general population may have biased the mortality estimates. Healthy worker bias is inherent in occupational studies that compare mortality with that of the general population. Although we used rSMR to examine the relative importance of specific causes of mortality within the cohort, an implicit assumption of this approach is that, in the absence of farm-related exposures of interest experienced by the cohort, other factors such as confounding and healthy worker bias that contribute to cause-specific deaths also contribute to deaths from all other causes without any heterogeneity, though even without this assumption rSMRs may reduce bias.[11]
Our analysis discounted deaths that occurred shortly after enrollment to 1998. Compared with those who died after 1999, those who died before were older, from NC, and ever-smokers. We do not believe that this exclusion meaningfully impacted our results. Specifically, we carried out sensitivity analyses exploring SMRs and CMRs for all causes for the complete follow-up period (from enrollment to 2015; observed deaths=10,197) and the results were similar [SMR=0.66 (95%CI: 0.65–0.67) versus 0.69 (95%CI: 0.67–0.70) in this report and CMR=0.88 (95%CI: 0.87–0.90) versus 0.89 (95%CI: 0.87–0.91)] for private applicators; similar results were also observed for spouses (data not shown).
Overall, AHS participants experienced lower overall mortality as compared to the general population. Nevertheless, all three approaches indicated higher-than-expected mortality from some accidental causes. SMRs were also modestly elevated for some natural causes, especially among private applicators. In contrast to SMRs but in agreement with rSMRs, CMRs identified elevated mortality from additional causes.
Supplementary Material
KEY MESSAGES.
What is already known about this subject?
Studies have shown that farmers experience lower overall mortality from natural causes as compared to the general population. This finding has been attributed to farmers’ healthier lifestyle and a potential healthy worker effect.
Prior studies have generally used the standardized mortality ratio (SMR) to compare mortality rates observed in the cohort with that in the general/referent population, but this method assumes that exposures do not affect the person-time distribution of the cohort and thus may provide a biased estimate when this assumption is not met.
What are the new findings?
Although overall mortality in Agricultural Health Study participants (a cohort of farmers and their spouse and commercial pesticide applicators) was lower than expected, elevated mortality from some specific natural/non-accidental causes was observed using the causal mortality ratio (CMR) and relative SMR (rSMR) approaches.
How might this impact on policy or clinical practice in the foreseeable future?
Studies using the SMR approach may underestimate mortality in farming populations. Our findings support the value of using other complementary approaches, in addition to traditional SMR, when evaluating mortality in farmers or other occupational cohorts, and suggest the need for methodologic studies to evaluate bias in SMR, CMR, and rSMR analysis under different exposure scenarios.
ACKNOWLEDGEMENT
We would like to thank Stuart Long from Westat for help with data management. The study was approved by the Institutional review boards of the National Institute of Environmental Health Sciences (North Carolina, protocol number 11-E-N196), the National Cancer Institute (Maryland, protocol number OH93-NC-N013), Westat, Inc. (Maryland), the University of Iowa (Iowa), and Battelle Health Sciences (North Carolina). We used AHS data release AHSREL201706.00 for the analysis.
FUNDING
This work was supported by the Intramural Research Program of the National Institute of Health, National Institute of Environmental Health Sciences (Z01-ES-049030) and National Cancer Institute (Z01-CP-010119).
Footnotes
DATA SHARING STATEMENT
Requests for data, including the data used in this manuscript, are welcome as described on the Study Website (https://www.aghealth.nih.gov/collaboration/process.html). Data requests may be made directly at www.aghealthstars.com; registration is required. The Agricultural Health Study is an ongoing prospective study. The data sharing policy was developed to protect the privacy of study participants and is consistent with study informed consent documents as approved by the NIH Institutional Review Board. Dr. Dale Sandler is the NIEHS Principal Investigator of the Agricultural Health Study and is responsible for ensuring participant safety and privacy.
COMPETING FINANCIAL INTERESTS
The authors declare they have no actual or potential competing financial interests.
REFERENCES
- 1.Blair A, Dosemeci M, Heineman EF. Cancer and other causes of death among male and female farmers from twenty-three states. Am J Ind Med 1993;23:729–42.doi: 10.1002/ajim.4700230507 [DOI] [PubMed] [Google Scholar]
- 2.Armitage TL, Mitchell D, Schenker M. Mortality in the California Farmer Health Study cohort. J Agromedicine 2012;17:288–99.doi: 10.1080/1059924X.2012.687614. [DOI] [PubMed] [Google Scholar]
- 3.Frost G, Brown T, Harding AH. Mortality and cancer incidence among British agricultural pesticide users. Occup Med (Lond) 2011;61:303–10.doi: 10.1093/occmed/kqr067. [DOI] [PubMed] [Google Scholar]
- 4.Blair A, Zahm SH, Pearce NE et al. Clues to cancer etiology from studies of farmers. Scand J Work Environ Health 1992;18:209–15.doi: 10.5271/sjweh.1578. [DOI] [PubMed] [Google Scholar]
- 5.Alavanja MC, Sandler DP, McMaster SB et al. The Agricultural Health Study. Environ Health Perspect 1996;104:362–9.doi: 10.1289/ehp.96104362. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Blair A, Sandler DP, Tarone R et al. Mortality among participants in the Agricultural Health Study. Ann Epidemiol 2005;15:279–85.doi: 10.1016/j.annepidem.2004.08.008. [DOI] [PubMed] [Google Scholar]
- 7.Waggoner JK, Kullman GJ, Henneberger PK et al. Mortality in the Agricultural Health Study, 1993–2007. Am J Epidemiol 2011;173:71–83.doi: 10.1093/aje/kwq323. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Keiding N, Vaeth M. Calculating expected mortality. Stat Med 1986;5:327–34.doi: 10.1002/sim.4780050405 [DOI] [PubMed] [Google Scholar]
- 9.Hartz AJ, Giefer EE, Hoffmann RG. A comparison of two methods for calculating expected mortality. Stat Med 1983;2:381–6.doi: 10.1002/sim.4780020310 [DOI] [PubMed] [Google Scholar]
- 10.Richardson DB, Keil AP, Cole SR et al. Observed and Expected Mortality in Cohort Studies. Am J Epidemiol 2017;185:479–486.doi: 10.1093/aje/kww205. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Richardson DB, Keil AP, Tchetgen Tchetgen E et al. Negative Control Outcomes and the Analysis of Standardized Mortality Ratios. Epidemiology 2015;26:727–32.doi: 10.1097/EDE.0000000000000353. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Centers for Disease Control and Prevention, National Center for Health Statistics. Underlying Cause of Death 1999–2016 on CDC WONDER Online Database, released December 2017. [Google Scholar]
- 13.Breslow NE, Day NE. Statistical Methods in Cancer Research: Volume II: The Design and Analysis of Cohort Studies. Lyon, France: International Agency for Research on Cancer, 1987. [PubMed] [Google Scholar]
- 14.Martin JS, Monaghan TM, Wilcox MH. Clostridium difficile infection: epidemiology, diagnosis and understanding transmission. Nat Rev Gastroenterol Hepatol 2016;13:206–16.doi: 10.1038/nrgastro.2016.25. [DOI] [PubMed] [Google Scholar]
- 15.Keessen EC, Harmanus C, Dohmen W et al. Clostridium difficile infection associated with pig farms. Emerg Infect Dis 2013;19:1032–4.doi: 10.3201/eid1906.121645. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Alavanja MC, Sandler DP, Lynch CF et al. Cancer incidence in the Agricultural Health Study. Scand J Work Environ Health 2005;31 Suppl 1:39–45; discussion 5–7 [PubMed] [Google Scholar]
- 17.Koutros S, Alavanja MC, Lubin JH et al. An update of cancer incidence in the Agricultural Health Study. J Occup Environ Med 2010;52:1098–105.doi: 10.1097/JOM.0b013e3181f72b7c. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lerro CC, Koutros S, Andreotti G et al. Cancer incidence in the Agricultural Health Study after 20 years of follow-up. Cancer Causes Control 2019;30:311–322.doi: 10.1007/s10552-019-01140-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Koutros S, Beane Freeman LE, Lubin JH et al. Risk of total and aggressive prostate cancer and pesticide use in the Agricultural Health Study. Am J Epidemiol 2013;177:59–74.doi: 10.1093/aje/kws225. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Alavanja MC, Hofmann JN, Lynch CF et al. Non-hodgkin lymphoma risk and insecticide, fungicide and fumigant use in the Agricultural Health Study. PLoS One 2014;9:e109332. doi: 10.1371/journal.pone.0109332. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Schinasi L, Leon ME. 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 2014;11:4449–527.doi: 10.3390/ijerph110404449. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Beane Freeman LE, Deroos AJ, Koutros S et al. Poultry and livestock exposure and cancer risk among farmers in the Agricultural Health Study. Cancer Causes Control 2012;23:663–70.doi: 10.1007/s10552-012-9921-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Dennis LK, Lynch CF, Sandler DP et al. Pesticide use and cutaneous melanoma in pesticide applicators in the Agricultural Heath Study. Environ Health Perspect 2010;118:812–7.doi: 10.1289/ehp.0901518. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Fortes C, Mastroeni S, Melchi F et al. The association between residential pesticide use and cutaneous melanoma. Eur J Cancer 2007;43:1066–75.doi: 10.1016/j.ejca.2007.01.013. [DOI] [PubMed] [Google Scholar]
- 25.Kachuri L, Harris MA, MacLeod JS et al. Cancer risks in a population-based study of 70,570 agricultural workers: results from the Canadian census health and Environment cohort (CanCHEC). BMC Cancer 2017;17:343.doi: 10.1186/s12885-017-3346-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Engel LS, Werder E, Satagopan J et al. Insecticide Use and Breast Cancer Risk among Farmers’ Wives in the Agricultural Health Study. Environ Health Perspect 2017;125:097002.doi: 10.1289/ehp1295. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Lerro CC, Koutros S, Andreotti G et al. Organophosphate insecticide use and cancer incidence among spouses of pesticide applicators in the Agricultural Health Study. Occup Environ Med 2015;72:736–44.doi: 10.1136/oemed-2014-102798. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Pezzoli G, Cereda E. Exposure to pesticides or solvents and risk of Parkinson disease. Neurology 2013;80:2035–41.doi: 10.1212/WNL.0b013e318294b3c8. [DOI] [PubMed] [Google Scholar]
- 29.Hayden KM, Norton MC, Darcey D et al. Occupational exposure to pesticides increases the risk of incident AD: the Cache County study. Neurology 2010;74:1524–30.doi: 10.1212/WNL.0b013e3181dd4423. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Hernan MA, Takkouche B, Caamano-Isorna F et al. A meta-analysis of coffee drinking, cigarette smoking, and the risk of Parkinson’s disease. Ann Neurol 2002;52:276–84.doi: 10.1002/ana.10277. [DOI] [PubMed] [Google Scholar]
- 31.Lee E, Burnett CA, Lalich N et al. Proportionate mortality of crop and livestock farmers in the United States, 1984–1993. Am J Ind Med 2002;42:410–20.doi: 10.1002/ajim.10131. [DOI] [PubMed] [Google Scholar]
- 32.Robinson CF, Walker JT, Sweeney MH et al. Overview of the National Occupational Mortality Surveillance (NOMS) system: leukemia and acute myocardial infarction risk by industry and occupation in 30 US states 1985–1999, 2003–2004, and 2007. Am J Ind Med 2015;58:123–37.doi: 10.1002/ajim.22408. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Ha MH, Lee DH, Jacobs DR. Association between serum concentrations of persistent organic pollutants and self-reported cardiovascular disease prevalence: results from the National Health and Nutrition Examination Survey, 1999–2002. Environ Health Perspect 2007;115:1204–9.doi: 10.1289/ehp.10184. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Mills KT, Blair A, Freeman LE et al. Pesticides and myocardial infarction incidence and mortality among male pesticide applicators in the Agricultural Health Study. Am J Epidemiol 2009;170:892–900.doi: 10.1093/aje/kwp214. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Lebov JF, Engel LS, Richardson D et al. Pesticide use and risk of end-stage renal disease among licensed pesticide applicators in the Agricultural Health Study. Occup Environ Med 2016;73:3–12.doi: 10.1136/oemed-2014-102615. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Sponholtz TR, Sandler DP, Parks CG et al. Occupational exposures and chronic kidney disease: Possible associations with endotoxin and ultrafine particles. Am J Ind Med 2016;59:1–11.doi: 10.1002/ajim.22541. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Taskar VS, Coultas DB. Is idiopathic pulmonary fibrosis an environmental disease? Proc Am Thorac Soc 2006;3:293–8.doi: 10.1513/pats.200512-131TK. [DOI] [PubMed] [Google Scholar]
- 38.Korsten P, Tampe B, Konig MF et al. Sarcoidosis and autoimmune diseases: differences, similarities and overlaps. Curr Opin Pulm Med 2018;24:504–512.doi: 10.1097/MCP.0000000000000500. [DOI] [PubMed] [Google Scholar]
- 39.Coady SA, Sorlie PD, Cooper LS et al. Validation of death certificate diagnosis for coronary heart disease: the Atherosclerosis Risk in Communities (ARIC) Study. J Clin Epidemiol 2001;54:40–50.doi: 10.1016/S0895-4356(00)00272-9. [DOI] [PubMed] [Google Scholar]
- 40.Macera CA, Sun RK, Yeager KK et al. Sensitivity and specificity of death certificate diagnoses for dementing illnesses, 1988–1990. J Am Geriatr Soc 1992;40:479–81.doi: 10.1111/j.1532-5415.1992.tb02015.x [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.