Military environmental exposures and risk of breast cancer in active-duty personnel and veterans: a scoping review

Background The effects of military environmental exposures (MEE) such as volatile organic compounds (VOCs), endocrine-disrupting chemicals (EDCs), tactile herbicides, airborne hazards and open burn pits (AHOBP), and depleted uranium on health are salient concerns for service members and Veterans. However, little work has been done to investigate the relationship between MEE and risk of breast cancer. Data sources and methods We conducted a scoping review on MEE, military deployment/service, and risk of breast cancer among active-duty service members and Veterans. PRISMA was used. PubMed, Embase, and citations of included articles were searched, resulting in 4,364 articles to screen: 28 articles were included. Results Most papers on military deployment and military service found a lower/equivalent risk of breast cancer when comparing rates to those without deployment or civilians. Exposure to VOCs due to military occupation or contaminated groundwater was associated with a slightly higher risk of breast cancer. Exposure to Agent Orange was not associated with an increased risk of breast cancer. Evidence regarding EDCs was limited. No paper directly measured exposure to AHOBP or depleted uranium, but deployments with known exposures to AHOBP or depleted uranium were associated with an equivalent/lower risk of breast cancer. Conclusions Women are the fastest growing population within the military, and breast cancer poses a unique risk to women Veterans who were affected by MEE during their service. Unfortunately, the literature on MEE and breast cancer is mixed and limited, in part due to the Healthy Soldier Paradox and poor classification of exposure(s).


Introduction
The number of women Veterans served by the Department of Veterans Affairs -Veterans Health Administration (VA) more than quintupled between 2000 and 2021 (159,810 to 870,000+) (1, 2), while the number of men grew substantially slower over the same period (2, 3).In 2020, women comprised 19% of all military branches (2, 4), which highlights an ongoing need for the expansion of women-specific health services.The 2023 Office of Women's Health -State of Reproductive Health governmental report found that abnormal breast conditions were reported as one of the top five reproductive and sexual health concerns for women Veterans aged 45+ (5).As VA projects the resources needed to care for the expanding women Veteran population, clinical and educational efforts must consider the unique health concerns faced by women Veterans.
Breast cancer (BC) is the most prevalent cancer among women, with around 300,000 cases diagnosed in the United States (U.S.) annually (6).One out of every eight women will be diagnosed at least once in their lifetime (6).The incidence rate (IR) of BC peaks in the 60s and 70s for women and the mortality rate increases exponentially with age (7), with Black women having the highest risk of mortality out of all racial and ethnic groups in the U.S.While the IR of BC has increased over the past two decades, the mortality rate has lowered substantially following advancements in early detection and treatment (7).Conversely, less than 1% of all BC patients are men (8) but BC in men is deadlier than in women (8).Military men with BC tend to present at a higher stage and with a larger tumor size than military women with BC, though demographics or tumor characteristics do not fully explain the higher rate of mortality in men with BC (9).BC is of great concern to VA and is a presumptive condition under The Sergeant First Class Heath Robinson Honoring our Promise to Address Comprehensive Toxics (PACT) Act of 2022.Presumptive conditions allow Veterans to receive care for ongoing health concerns that are of unknown etiology, and can be presumed to be related to service (10).Cancer of any kind remains an ongoing concern for Veterans as they age, and especially among Veterans with military environmental exposures (MEE).
The rates of cancers differ among active-duty personnel and the general U.S. population (11).Over 800 active-duty personnel receive a cancer diagnosis yearly, and tumor etiology is often correlated with service characteristics and MEE (12).These exposures include, but are not limited to, airborne hazards and open burn pits (AHOBP), asbestos, biological and chemical warfare tests, contaminated water, chemical agent resistant coating paint, embedded substances such as depleted uranium and lead, fuels, industrial solvents, ionizing radiation, mefloquine for malaria, nerve agents, noise, pesticides, perfluoroalkyl and polyfluoroalkyl substances, pyridostigmine bromide pills for sarin gas exposure, tactile herbicides, and vaccines.Cancer among current and former military personnel with known MEE persists as a complex health concern (12)(13)(14)(15).
The current literature on MEE and cancer is limited.For example, the tactile herbicide Agent Orange was linked to an increased incidence of several cancers, including leukemia and cancers that start in soft tissues (16), and a slightly higher rate of BC was found among military personnel when compared to civilians (17).However, higher rates of BC may be tied to confounding risk factors in military personnel, such as delayed age of first childbirth or increased use of contraceptives.Additionally, military personnel often have greater access to routine screening, resulting in quicker identification of early-stage BCs (18,19).In other words, tying BC incidence to MEE rather than characteristics associated with service (i.e., confounding factors) is a difficult task.
Combat exposure has increased from 7% to 24% when comparing pre-1990 to post-1990 women Veterans, suggesting that MEE concerns may grow among women Veterans in the coming decades (20).However, few have investigated BC in association with specific MEE.Therefore, we conducted a scoping review to determine whether deployment/military service and MEE affect the risk of BC among active-duty personnel and Veterans.

Search strategy
Unlike systematic reviews that focus on a specific research question, scoping reviews ask broad research questions to characterize and understand a developing and heterogenous area within the literature (21).Search terms were compiled using PubMed's Medical Subject Headings (MeSH) trees and through consultation with the California War Related Illness and Injury Study Center (CA WRIISC), the Women's Operational Military Exposure Network Center of Excellence (WOMEN CoE) and Advisory Board, and staff oncologists at VA Palo Alto Health Care System.Relevant articles were searched for in PubMed and Embase and terms can be found in the notes of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart (Figure 1).Some articles broadly examined cancer incidence and did not mention BC in the title or abstract, but included estimates of BC within the results/tables.Therefore, four authors (AIB, DJJ, DKG, MTA) screened citations from the included articles to find these additional manuscripts.

Inclusion criteria
To be included, studies had to: (1) enroll active-duty personnel, Reservists, or Veterans, (2) measure MEE or military service/ deployment, (3) concern BC risk (i.e., papers on BC mortality were excluded), and (4) have an English full-text.

Study selection
Covidence software was used to collate and screen the articles.Four authors (AIB, DJJ, DKG, MTA) screened titles/abstracts and full-texts and met weekly to resolve disagreements through discussion.The database search was conducted on June 16, 2023, and resulted in a total of 4,510 articles.After the removal of 1,482 duplicates, 3,028 titles and abstracts were screened and 2,865 were excluded.A total of 163 full-texts were assessed, of which 11 were included.After screening an additional 1,336 citations from the included articles, 17 were retained for a final total of 28 articles.

Data extraction
Data extraction was completed by four authors (AIB, DJJ, DKG, MTA) with each paper receiving at least two checks for accuracy and included the following headings: author/publication year, sample characteristics, sample size, exposure, results, warfare era/ service years, and diagnosis years.See Table 1 for the characteristics of each study.

Results
In total, 28 papers were synthesized.Sample size ranged from 64 to millions.Several military conflicts were included: Malayan Emergency, Vietnam War, Israel-Lebanon Conflicts, Persian Gulf War, Kosovo War, Bosnian War, Croatian War of Independence, and post-9/11 conflicts (Operation Enduring Freedom [OEF], Operation Iraqi Freedom [OIF], Operation New Dawn [OND]).More than half of the studies used a case-control or cohort study design.Several MEE were examined: military service/deployment, volatile organic compounds (VOCs), endocrine-disrupting chemicals (EDCs), Agent Orange, and ultraviolet B radiation (Vitamin D synthesis).

Risk compared to civilians & standardized rates
Nine papers compared risk of BC in military personnel compared to civilians or standardized national rates.Zhu and colleagues (2009) conducted a cohort study and compared military and civilian cancer surveillance data.They found a slightly higher IRR for Black military women 1.37 [1.21, 1.55] and for White military women 1.19 [1.09, 1.30] when compared to civilians (17).Katuwal (41).Yi (42) included 185,265 male Vietnam Veterans from Korea (n=8 cases), but found a statistically equivalent rate of BC when compared to the general PRISMA flow chart.MeSH Trees used: Diseases Category ==> Neoplasms ==> Neoplasms by Site ==> Breast Neoplasms.Titles and abstracts were searched using the terms: ("breast cancer*" OR "breast neoplasm*" OR "breast tumor*" OR "breast metas*" OR "mammary metas*" OR "mammary cancer*" OR "malignant neoplasm* of the breast*" OR "malignant neoplasm* of breast*" OR "breast malignant neoplasm*" OR "malignant tumor* of breast*" OR "malignant tumor* of the breast*" OR "breast malignant tumor*" OR "cancer of breast*" OR "cancer of the breast*" OR "mammary carcinoma*" OR "mammary neoplasm*" OR "breast carcinoma*" OR "mastect*" OR "lumpect*" OR "mammogr*") AND (Veteran* OR military OR combat OR deploy* OR undeploy* OR soldier* OR war OR wars OR warzone OR "department of defense" OR DOD OR front-line* OR duty OR enlist*).Asterisk wildcards were used to find word endings.Terms were left purposefully broad to examine the largest possible selection of the literature.Prospective or retrospective cohort, case-control, cross-sectional, ecological, or related study designs (e.g., case-cohort) were included.Case studies, case series, reviews and meta-analyses, book chapters, theses, and dissertations were excluded.In the adjusted models, no statistically significant relationship was found between serum Vitamin D levels and odds of breast cancer.Inverse trends were present among women with a blood draw within 90 days of their breast cancer diagnosis, where women in the lowest quintile of 25(OH)D had a higher estimated risk of breast cancer (OR=3.3

Demographics & work characteristics
Six papers provided estimates that differed by demographic or work characteristics.Hansen and Lassen (26) studied 218 women with BC and 899 age-matched controls from a nested cohort study of 18,551 women Danish military employees.Women with the highest tertile of cumulative night shift military work exposure had an increased odds of BC (OR=2.3 [1.2, 4.6]) and women with a morning chronotype preference (inclination to be more active during the morning) and intense night shifts had the largest risk (OR=3.9[1.6, 9.5]) (26).The Armed Forces Health Surveillance Center (2013) published a cohort study and found that the IR of BC among active-duty service women was 40.6 per 100,000 from 2000-2012 (48).Non-Hispanic Black women, older women, senior officers, and women serving in healthcare or administrative roles had an increased risk of BC (48), and women with combat-specific duties had a mildly increased risk.Conversely, women who served in the U.S. Marine Corps, those who identified as Hispanic, younger women, junior enlistees, and women with "other" duties had a decreased risk of BC (48). Lee and colleagues (2016) examined all U.S. active-duty personnel from 2005-2014, of which 652 cases of BC were observed (IR=31.8 per 100,000).Active-duty women who were non-Hispanic Black, officers, healthcare workers, or older had an increased risk of BC (31).Zullig and colleagues (2012) found no major differences in BC incident diagnoses among Black, "Other," and White Veterans from the 2007 Veterans Affairs Central Cancer Registry (45).Zullig and colleagues (2017) updated their 2007 analysis (Zullig et al. (45),) with data from 2010, but results did not change appreciably (46).Zullig and colleagues (2019) conducted a cross-sectional study of 1,330 incident invasive cancer cases among women Veterans in 2010; BC was the most common invasive cancer (30.23%, n=402), but it did not differ by race (47).

Deployment characteristics
Four papers compared deployment characteristics.Kang and colleagues (2000) followed 3,392 Vietnam-era deployed women Veterans (n=170 cases) and 3,038 Vietnam-era women Veterans who never deployed to Vietnam (n=126).Both the crude (odds ratio  1.For papers that provided outcomes for both men and women, the full sample size was included.For papers that examined breast cancer exclusively in either men or women, only the corresponding sample size was included (vs the entire sample). [OR]=1.

General incidence rates
Four studies provided general incidence rates without a comparison group.Hoiberg & Ernst (27) conducted a cohort study of 364 active-duty Navy women (n=47 cases) and found an IR by age (overall 34.1 per 100,000) (27).Ajene and colleagues (22) conducted a cohort study of 78 women Naval personnel and found a BC IR of 8.53 per 100,000 persons (22).Strand and colleagues (2014) analyzed 268 Norwegian women military peacekeepers deployed to Kosovo, and one incident case of BC was observed (38).Strand and colleagues (2020) reanalyzed their 2014 cohort of Norwegian military peacekeepers deployed to Kosovo: 275 women were included and three BC cases were found (40).

Volatile organic compounds & endocrinedisrupting chemicals
Two papers measured exposure to VOCs.Rennix and colleagues (2005) conducted a cohort study of 274,596 enlisted Army women.Exposure to VOCs was tied to job titles, categorized as none, low, medium, or high (35).None of the top military occupational specialties were associated with BC risk.However, history of moderate or high exposure to VOCs was associated with a 48% higher incidence (IRR=1.48[1.03, 2.12]) (35).Ruckart and colleagues (2015) led a case-control study of Marines stationed at the Camp Lejeune garrison who were exposed to contaminated groundwater (71 cases of male BC, 373 controls).Adjusted ORs for high residential cumulative exposures to tetrachloroethylene, t-1,2 dichloroethylene, and vinyl chloride were 1.20 [0.16, 5.89], 1.50 [0.30, 6.11], 1.19 [0.16, 5.89], respectively (36).Ever being stationed at Camp Lejeune and high cumulative exposures to VOCs were associated with an earlier age at onset for male BC, though it was not statistically significant (36).
Carran and Shaw (24) conducted a cohort study of 71 Veterans and 76 female adult children of Veterans.They found an increased risk of BC among female adult children of New Zealand Veterans deployed to Malaya who were exposed to the EDC dibutylphthalate (24).

Tactile herbicides (agent orange)
One paper directly measured exposure to the tactile herbicide, Agent Orange.Yi & Ohrr (43) mapped each unit's post location and tactile area of responsibility to known geographic regions of chemically treated areas.Korean Vietnam-era Veterans (n=180,251 with follow-up from 1992-2003) were given an Exposure Opportunity Index and stratified into categories (high vs. low) in a cohort study (43).Exposure to high levels of Agent Orange was not associated with an increased risk of BC (adjusted HR=0.53 [0.12, 2.26]) (43).

Discussion
This scoping review covered 28 papers on the relationships between military service, MEE, and BC among active-duty personnel and Veterans.Unfortunately, evidence is still needed before conclusive remarks can be made.Most papers on military service or deployment reported a decreased or statistically equivalent risk of BC, while a few larger surveillance studies found an increased risk.When considering the effects of military service and deployment on risk of BC, individual and environmental risk factors should be considered (49).If risk factors are not controlled for, findings may be biased by the Healthy Soldier Paradox (25).
The Healthy Soldier Paradox occurs when healthier personnel are deployed in support of military operations and sicker personnel are not deployed or are given different military occupations (50).This form of sampling bias can lead to inaccurate associations between deployment and health outcomes, where deployed individuals appear to have a lower risk of adverse outcomes than non-deployed personnel.Recent work has called the Healthy Soldier Paradox into question for OEF/ OIF/OND-era Veterans (51), as OEF/OIF/OND-era Veterans have a higher risk of mortality when compared to the general U.S. population.However, it may be that healthy solider effects vary by outcome, such that OEF/OIF/OND-era Veterans may have a higher risk of mortality but a decreased risk of BC (25).In a study of 31,548 military healthcare system users with BC and 63,096 controls with BC, the military healthcare system users had a significantly lower risk of mortality (24% [20%, 29%]) (52).This lower mortality risk was found across all ages, Stages II, III, and IV tumors, and for Black and White patients (52), suggesting that military personnel may also benefit from an equalaccess healthcare system (18,19).
Many studies have few years of follow-up, small sample sizes, lack accurate measurement, and suffer from misclassification bias.Follow-Up: Epidemiological analyses of cancer require a sufficiently long follow-up so that cases can occur.Most cases of BC occur in women aged 50+ (6), and many personnel from post-9/11 conflicts may not be in this age group yet.Small Sample Size: Women personnel and Veterans with MEE concerns are still a relatively small group from an epidemiological perspective.Accurate Measurement of Exposure: Investigators often rely on participant recall for MEE (leading to recall bias) and do not consider that most MEE are transient, may occur more than once, and may occur with varying severity.Accurate Measurement of Outcome: Measurement of BC has improved in recent decades due to advances in mammography screening (53) and modern classification codes (e.g., ICD-10).However, characteristics of the breast tumor (e.g., T-stage, N-stage, M-stage) and histological and molecular subtyping are not often explored.Misclassification Bias: Misclassification bias reflects an issue with categorizing participants by exposure/outcome status.Without accurately measuring MEE, participants may be misclassified, leading to null results (54).For many articles, exposures are generalized to entire groups, but individual-level data are needed.
Most papers in this review considered military service or military deployment as an exposure when measuring BC risk in personnel.Unlike most social and environmental exposures, military service (yes/no) is not well operationalized and leaves a lot to be desired in terms of specificity.Characteristics of military service (e.g., military occupation and job duties, deployment location, rank, military branch, number of years served, and MEE) should be measured in future studies, as these factors may improve our detection of the Healthy Soldier Paradox.Additionally, tying specific military occupations and job duties to MEE will be crucial for determining causality, and will inform policy and practice regarding the expansion of personal protective equipment and environmental toxin passive monitoring devices in the field.
With these limitations stated, several conclusions may still be found.Exposure to VOCs appears to impact the downstream risk of BC among military personnel (35,36).Specific VOCs' effects are largely unknown, but the risk of BC appears stronger among women than among men, and the mechanism of action may be through oxidative damage, cytotoxicity, and genotoxicity (55)(56)(57).The effects of EDCs have not been sufficiently studied in military samples, but they are known to impact risk of BC in civilians (58).Vitamin D (34,59,60) may not be considered an environmental exposure, but ionizing radiation from the sun would be an important MEE.Future work should measure UV exposure, heat, and drought directly.No papers directly assessed the effects of AHOBP or depleted uranium, but many Veterans who deployed in support of Gulf War and OEF/OIF/OND with MEE to AHOBP and depleted uranium are now at an age where BC is a salient concern (61).One large study looked at Agent Orange and BC risk and did not find an association.Unfortunately, no studies were found on other tactile herbicides (e.g., Agents White, Blue, Purple, Pink, Green) or pesticides.Many MEE included in this review were not specific to military populations.VOCs, EDCs, and carcinogenic airborne hazards are well-known occupational exposures in the civilian sector.Too little work has been done to understand if the effects of these generic exposures are moderated by military service.Finally, it is important to recognize that a greater number of highquality articles will be needed to draw significant conclusions that link MEE and BC, as surveillance cohort studies are insufficient to draw causal links.

Recommendations
Future studies should: 1) Measure MEE in real time (e.g., dose, duration, source, route of entry) using ecological momentary assessment or passive monitoring; 2) Study specific VOCs, EDCs, and AHOBP; 3) Compare deployed to non-deployed military personnel and include a group of civilian controls when possible; 4) Recruit a diverse group of women and gender-diverse personnel, including all military branches, races/ethnicities, ranks, occupations, and deployment locations; 5) Determine warfare theater/era effects; 6) Measure BC histological/molecular subtypes; 7) Expand years of follow-up and increase recruitment; and 8) Explore biological plausibility by tying MEE to specific carcinogenic pathways.
Findings on MEE and BC are varied, in part due to the Healthy Soldier Paradox, potential misclassification of exposure(s), and modest sample sizes.The strongest evidence with reproducible findings appears to be Veterans' increased risk of BC after being exposed to VOCs.

Funding
The author(s) declare that financial support was received for the research, authorship, and/or publication of this article.This work was supported by the Women's Operational Military Exposure Network Center of Excellence (WOMEN CoE).The contents do not represent the views of the U.S. Department of Veterans Affairs or the United States Government.

Serum 25 -
hydroxyvitamin D/ultraviolet B sunlight exposureMohr and colleagues (2013)  examined the relationship between pre-diagnostic serum 25-hydroxyvitamin D (Vitamin D) and risk of BC among active-duty personnel using a nested case-control study with 600 incident cases and 600 controls.No statistically significant relationship was found between serum Vitamin D levels and odds of BC.

TABLE 1
Characteristics of the included studies.Those who deployed in support of OEF/OIF were 23% [14%, 27%] less likely to receive a breast cancer diagnosis than women who did not deploy after adjusting for age, race and ethnicity, marital status, Active-duty women who served in the Air Force (IRR=2.4),identifiedasnon-HispanicBlack (IRR=2.2),wereolder(40+)(IRR=27.1),seniorofficers(IRR=4.1),andwomenserving in healthcare roles (IRR=2.1)oradministrative/supplyroles(IRR = 1.6) had an increased risk of breast cancer.Women who served in the Marine Corps (reference group), identified as Hispanic (reference group), were younger women (<25) (reference group), were junior enlisted (reference group), and women with "other" duties (reference group) had a decreased risk of breast cancer.Women with combat-specific duties had a marginally increased risk of breast cancer (IRR = 1.1) when compared to women with "other" duties.No significant increased risk of breast cancer was found among active-duty personnel compared to the general population, and this did not differ by race: White IRR=1.06[0.98,1.13];Black IRR=1.06 [0.96, 1.16] women service members.When stratified by age, Black (IRR=0.98[0.86,1.12])and White (IRR=0.96[0.85,1.07])women military service members aged 20-39 had no significant increased risk of breast cancer when compared to the general population.However, Black (IRR=1.17[1.01,1.34])and White (IRR=1.15[1.04,1.26]) women military service members aged 40-59 had a statistically increased risk of breast cancer when compared to the general population.Further stratification by cancer stage (local, regional, and distant) showed only a significant age effect for local breast cancers (confined to the breast), but not for regional (extends to the surrounding lymph nodes, organs, or tissues) or distant cancers (extends to distant organs or lymph nodes).

TABLE 1 Continued
The incidence rate of breast cancer in women Veterans was 30.01[18.91,47.63])per100,000person-years, compared to the general population(25.06 [19.46,  32.27]per 100,000 person-years (matched on age, sex, residential geography, and community socioeconomic status).Women Veterans had no statistically significant increased risk of breast cancer when compared to the general population, before or after adjusting for the matching variables (Crude HR=1.20 [0.71, 2.03]; Adjusted HR=1.19 [0.70, 2.02]).

TABLE 1 Continued
APR, Adjusted Proportional Incidence Ratio; HR, Hazard Ratio; IRR, Incidence Rate Ratio; PIR, Proportional Incidence Rate; SIR, Standardized Incidence Ratio; OR, Odds Ratio; OEF, Operation Enduring Freedom; OIF, Operation Iraqi Freedom; U.S., United States of America; N/A, Not Available.The sample for which breast cancer outcomes are reported are included in the sample size column of Table (32)0.96, 1.55]) and adjusted ORs (OR=1.18[0.91,1.51]suggestedthat the odds of developing BC were statistically equivalent between groups(28).Macfarlane and colleagues (2003) followed a cohort of 51,721 Gulf War Veterans and 50,755 matched service personnel and found no increased risk among Veterans that deployed in support of the Gulf War (IRR=0.59[0.21,1.62])(32).