Construction and manufacturing are two of the largest industries in the United States, and both have a substantial environmental impact in the form of harmful aerosols and gases, noise, dust, and solid and liquid wastes.1 Construction activities worsen air quality by emitting air pollutants such as particulate matter (PM), volatile organic compounds (VOCs), nitrogen oxides (NOx), and sulfur dioxide (SO2),2 which may affect respiratory health. Similarly, chemical manufacturing companies produce air toxins such as metals (including cadmium, chromium, and lead compounds),3,4 hydrochloric acid,5 and hydrogen fluoride6,7 fumes, which are associated with increased risk of acute respiratory failure. Although industrial workers have the highest chance of exposure to these toxins, airborne toxins are released from industrial sites and affect nearby residents. Construction sites pose serious health risks not only to workers but also to local pedestrians and drivers.
Residents living near heavy industrial sites have an increased incidence of lung and respiratory symptoms.8–10 Older adults are particularly susceptible to both acute and chronic respiratory disease, and coexisting chronic lung, heart, and circulatory conditions may worsen following exposure to environmental pollutants.11 Indeed, older individuals who live close to industrial plants and waste sites such as mine dumps12 and coke plants13 have increased risks of asthma, pneumonia, and chronic bronchitis.
Acute respiratory distress syndrome (ARDS) is a rapidly progressive form of acute respiratory failure,14 with noncardiogenic pulmonary edema and acute ventilatory failure due to pulmonary and nonpulmonary insults.15 ARDS develops in patients with predisposing conditions such as sepsis, pneumonia, traumatic injury, or aspiration.16–18 Nearly 10% of all patients admitted to the intensive care unit (ICU) and 23% of mechanically ventilated patients develop ARDS.19 The incidence of ARDS is reported to be between 64.2 and 78.9 cases/100,000 person-years in the United States,16,20 which is comparable to the incidence of lung cancer (62 cases/100,000 person-years).21 ARDS is often lethal, and mortality in the subgroup of patients with severe ARDS ranges from 40% to 60%.22 Further, the average age of developing ARDS is 62,23 and among older patients, the mortality for ARDS has been between 69% and 80%.24
Despite substantial evidence that living in areas with dense industrial sites is associated with respiratory illness,8–10 no study has investigated a link between industrial activities and ARDS, especially among older adults. We conducted a nationwide ecologic study to estimate the association between year-to-year changes in numbers of construction and manufacturing industries and year-to-year changes in hospital admission rates for ARDS. We particularly focused on older adults who are more susceptible to developing ARDS with exposure to environmental pollutants and other hazards.
We conducted an ecologic study to investigate national associations between year-to-year changes in numbers of construction and manufacturing sites and year-to-year changes in hospital admission rates for ARDS. For each ZIP code, we obtained (1) annual counts of hospital admissions for ARDS, (2) annual numbers of older adults enrolled in Medicare, (3) annual counts of construction and manufacturing companies, and (4) yearly average values of air pollution (particulate matter with an aerodynamic diameter <2.5 mm [PM2.5] and ozone), as well as (5) information on relevant confounders (race, median household income, and smoking status). This study was approved by the institutional review board at the Harvard T.H. Chan School of Public Health. As a study of previously collected administrative data, it was exempt from informed consent requirements.
Hospital admissions data describing Medicare beneficiaries 65 years of age and older were drawn from the Medicare Provider Analysis and Review (MEDPAR) inpatient data for the years 2006 through 2012. For each admission, the ZIP code of residence, the date of admission/discharge, and the primary and secondary discharge diagnosis codes for ARDS defined by International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) were extracted (eTable 1; http://links.lww.com/EDE/B663). Annual number of hospital admissions by ZIP code provided to us as deidentified data.
We used ICD-9-CM codes 518.51, 518.52, 518.53, and 518.82 as ARDS criteria.25–28 We then calculated annual counts of hospital admissions for ARDS at the ZIP code level and defined our outcome of interest.
We extracted the industry’s Standard Industrial Classification (SIC) code, business status, number of employees, and location from ArcGIS Business Analyst data provided by the Center for Geographic Analysis, Harvard University. We classified companies by SIC codes and selected construction (SIC codes 15–17) and manufacturing (SIC codes 20–39) industries from 2006 through 2012. Among selected construction and manufacturing companies, we excluded companies classified as “headquarters” or “subsidiary headquarters” in the business status. We also created a variable “business size” using number of employees and classified sites into five groups with 0, 1–19, 20–99, 100–499, and 500+ employees. We then computed annual counts of construction and manufacturing sites at the ZIP code level by the business size. Last, we grouped ZIP codes into regions of low, medium, and high numbers of companies using a cutoff of 10 and 50 sites a year, based on distribution of construction sites, rounded to the nearest 10 of the lowest quantile (three companies a year) and the highest quantile (44 companies a year).
Potential Confounders: Regional Characteristics and Air Pollution
We extracted demographic and regional characteristics for each ZIP code, including racial distribution, smoking status, access to trauma center, and air pollution. We determined the proportions of different racial groups (white, black, Hispanic, Asian, Native American) enrolled per year in Medicare per each ZIP code. We also calculated the proportion of females in each ZIP code and extracted ZIP code-level median household income from ArcGIS Business Analyst data. We obtained the county-level proportion of ever smokers from the Behavioral Risk Factor Surveillance System (BRFSS) and assigned the same values of county-level variables to all ZIP codes within the county boundary.29 The location of verified levels I and II trauma centers (excluding Pediatric center) were obtained from the search engine provided by the American College of Surgeons Committee on Trauma (https://www.facs.org/search/trauma-centers). Finally, we used predicted ZIP code-level PM2.5 and ozone concentrations from our published spatiotemporal models.30 We computed yearly average values at each ZIP code for potential confounders, including sex, race, median household income, smoking, and air pollution. As a proxy of access to trauma center, we counted number of trauma centers per ZIP code.
Race/ethnicity and neighborhood socioeconomic status (SES) are important risk factors of ARDS incidence. African Americans31,32 have a higher incidence of ARDS and related mortality than Caucasians. One potential reason for this is that African Americans are more likely to be diagnosed with an illness associated with the development of ARDS, such as sepsis or trauma. The incidence of sepsis among blacks is twice than among whites.33,34 Once ARDS is present, Hispanics have the greatest risk of death.34 In addition, active or passive cigarette smoke exposure is independently associated with the development of ARDS after severe blunt trauma.35 Poverty has been associated with lung diseases and can influence access to and quality of health care.36 Lower SES is associated with poorer respiratory health, less access to healthy foods, and less access to timely health care.36 Additionally, recent findings suggest that long-term exposures to ozone and PM2.5 pollution are associated with ARDS and ARDS-related in-hospital mortality.37–39
We examined the characteristics of exposures and outcomes of interest, along with potential confounders at the ZIP code level and performed Pearson correlation tests between them. To investigate the associations between year-to-year changes in numbers of construction/manufacturing sites and year-to-year changes in hospital admission rates for ARDS at the ZIP code level, we applied generalized linear mixed models with a random intercept for ZIP code assuming a Poisson distribution and used quasi-likelihood methods to allow for overdispersion. We performed separate analyses for the construction and manufacturing industries because these two industries were highly correlated (correlation coefficient: 0.73, eTable 2; http://links.lww.com/EDE/B663). We adjusted for the business size and ZIP code-level socioeconomic status values including sex (proportion of females), race (proportion of whites, blacks, Hispanics, and Native Americans), median household income, and proportion of ever smokers. To control for time trends in ARDS and construction or manufacturing activity, we included a dummy variable for each year. The model is described in eAppendix 1; http://links.lww.com/EDE/B663. To deliver results in a more meaningful way, the effect estimate was converted to percent change (%) in annual hospital admission rates by an increase of 10 construction (or manufacturing) companies per year across all ZIP codes. We rounded an interquartile range of manufacturing sites down to the nearest 10 then used an increase of 10 industry sites for effect estimate. Lastly, we conducted a subgroup analysis by ARDS patients’ primary diagnosis (predisposing conditions): sepsis, pneumonia, and traumatic injury.
First, we fit the models with further adjustments for air pollution (yearly average concentrations for PM2.5 and ozone) and number of trauma centers across all analyses. Second, we restricted ZIP code areas with more than one construction (or manufacturing) site and repeated the analyses as 10% of ZIP code areas had no construction sites and 20% of ZIP code areas did not have manufacturing sites. Third, as chemical manufacturers produce air toxins most relevant to ARDS, we extracted chemical product manufacturing (SIC code 28) in manufacturing sectors (SIC codes 20–39) then conducted the same analysis. We presented the percent change (%) in annual hospital admission rates by an increase of one chemical manufacturing company per year across all ZIP codes. Fourth, to evaluate the robustness of our estimates, we repeated the analysis using propensity score modeling. We created a binary exposure for each ZIP code using a cutoff of 50 companies a year for construction and 20 companies a year for manufacturing sectors, which was the nearest 10 of third quantile for each industry. We fit two logistic regressions to create propensity scores for each binary exposure using above cutoffs. Then we fit and compared logistic regression models with a binary exposure, one adjusting for same covariates in our main model and the other adjusting for the deciles of estimated propensity scores. To check the balance of each covariate before and after matching data with propensity score, we calculated the absolute standardized difference (ASD).40 The ASD was calculated as the mean in the high construction (or manufacturing) group minus the mean in the low construction (or manufacturing) group divided by the standard deviation. Covariates having highest ASD in unmatched data are unbalanced the most and can lead to confounding bias. All data were analyzed in RStudio 1.2.1335 (http://www.rstudio.com/).
Among the more than 30 million Medicare enrollees per year, we found on average 92,363 hospital admissions for ARDS per year and 646,542 admissions over the course of 7 years. The demographic information of ARDS patients in the Medicare cohort is displayed in Table 1. The average age of ARDS patients is about 78 years old, and there were slightly more female (51.8%) than male (48.2%) patients. On average, they stayed 13 days in the hospital and spent about 7 days in ICU. The majority of patients were white. Geographic distributions of construction and manufacturing sites are shown in Figures 1 and 2. Because the distribution from 2006 through 2012 was relatively constant for both industries, we displayed distribution of industries by county in 2009 as an example. Both industries were clustered in the Northeast region, Florida, Arizona, and California, all of which have high population densities.
In our analyses, we included 28,783 ZIP codes across the United States (Figure 3). There was a median value of one hospital admission for ARDS, 12 construction sites, and four manufacturing sites (one or less than one chemical product manufacturing) a year across all ZIP codes (Table 2). Median values for annual PM2.5 and ozone concentrations were about 10 μg/m3 and 39 ppb, respectively. Hospital admissions for ARDS were highly correlated with construction (correlation coefficient: 0.63, eTable 2; http://links.lww.com/EDE/B663). In addition, correlations between ARDS admissions with a primary diagnosis of sepsis and numbers of companies were higher than those with other predisposing conditions. The proportions of black, Hispanic, and Asian residents were positively correlated with construction and manufacturing sites; however, the proportions of white and Native American residents were negatively correlated with industry sites.
The adjusted associations of construction and manufacturing industry sites with ARDS are summarized in Table 3. We found that an increase of 10 construction sites per year was associated with a 0.75% (95% confidence interval [CI] = 0.69, 0.82) increase in annual hospital admission rates for ARDS on average across all ZIP codes. Similarly, each increase of 10 manufacturing sites per year was associated with a 1.05% (95% CI = 0.94, 1.17) increase in annual hospital admission rates for ARDS across all ZIP codes. The average annual admission rate across all ZIP codes for 7 years was 0.3%. Adjusting for long-term exposure to air pollution (PM2.5 and ozone) yielded similar results for construction (0.77%, 95% CI = 0.71, 0.84) and a slightly higher percent increase in hospital admission rates for manufacturing sites (1.21%, 95% CI = 1.09, 1.33). Result of further adjusting for trauma center also did not yield substantial differences (construction: 0.75%, 95% CI = 0.69, 0.82; manufacturing: 1.18%, 95% CI = 1.06, 1.30, eTable 3; http://links.lww.com/EDE/B663). In conducting the same analyses for restricted areas with more than one construction (or manufacturing) site, we also found consistent results with the main analyses in all ZIP codes (Table 3). From the subgroup analysis, we found that estimated effects of construction and manufacturing activities were higher among ARDS patients diagnosed with sepsis (construction: 0.87%, 95% CI = 0.79, 0.95; manufacturing 1.27%, 95% CI = 1.12, 1.41) (Table 4). The estimated effects of manufacturing activities were similar among ARDS patients with pneumonia (0.82%, 95% CI = 0.30, 1.35) or traumatic injury (0.80%, 95% CI = 0.48, 1.13).
Over 7 years, we found that older adults residing in ZIP codes with a high number of construction sites (≥50 companies/year) had 9.55% (95% CI = 8.61, 10.51) increased risk of developing ARDS compared with residents living in ZIP codes with a low number of construction sites (<10 companies/year) (Table 5). Living in ZIP codes with a high number of manufacturing sites was associated with a 4.68% (95% CI = 4.07, 5.31) increased risk of ARDS compared with living in ZIP codes with a low number of manufacturing sites. Compared with total manufacturing sites, the effect of chemical product manufacturing on ARDS was more than 10 times higher (eTable 4; http://links.lww.com/EDE/B663). We found that an increase of one chemical product manufacturing site per year was associated with a 1.50% (95% CI = 1.25, 1.75) increase in annual hospital admission rates for ARDS on average across all ZIP codes.
Figure 4 shows the absolute standardized mean difference for all covariates. Business size, proportion of females, and proportion of Asians were the most unbalanced covariates when we compared high versus low number construction and manufacturing industry sites (eTable 5; http://links.lww.com/EDE/B663). After matching with propensity scores, percent increase in annual hospital admission rates for ARDS was similar to the main analysis with binary exposures (eTable 6; http://links.lww.com/EDE/B663).
We conducted an ecologic study with over 30 million Medicare enrollees per year to investigate the impact of the construction and manufacturing sites on hospital admissions for acute respiratory distress syndrome (ARDS). We found national associations between year-to-year changes in construction (and manufacturing) industries and annual changes in hospital admission rates for ARDS. In particular, the number of chemical product manufacturing sites was highly associated with increased hospital admissions for ARDS. Living in areas with a high number of construction or manufacturing sites was related to increased ARDS hospital admissions. Construction or manufacturing activities seem to be negatively associated with local environmental health, in particular with the respiratory health of older residents, even after accounting for long-term exposure to PM2.5 and ozone.
We introduced our conceptual causal model of construction and manufacturing industrial activities leading to ARDS in Figure 5. Some of the largest manufacturing facilities in the United States, such as those in chemicals and in aerospace products, as well as petroleum refineries release inhalable air toxins associated with increased risk of ARDS (P1-P8 in Figure 5). Residents in neighborhoods with heavy manufacturing industries would be more likely to inhale those air pollutants. For example, hydrochloric acid5 and hydrogen fluoride6,7 fumes from chemical manufacturers and mixed metals dust and fumes3,4,41,42 from manufacturing sites involved with welding are reported to increase the risk of ARDS. Unintentional acute releases of ARDS-causing chemicals to communities could be another hazard.43 Construction activities generate inorganic dust and metal fumes,44–46 which have been associated with increased incidence of and mortality from pneumonia among construction workers47,48 and could potentially affect respiratory health among community residents (P3-P8 in Figure 5). In our result, we observed a stronger association in manufacturing industry for the pathway that inhaling air toxins is associated with increased risk of pneumonia then further develops ARDS (P1-P8 in Figure 5) compared with construction industry (P3-P8). Considering pneumonia is a leading cause of sepsis,49 the pathway to develop ARDS after exposure to air toxins could be P1-P5-P7. In addition, because traumatic injury is one of the most common clinical risk factors for developing ARDS,50 physical hazards, e.g., construction material falloffs51 or vehicle accidents with frequent truck transportation from both industries,43 may place community residents at risk of serious traumatic injury (P4-P9 or P2-P9). As chest or head trauma is associated with increased risk of pneumonia,52,53 another pathway could be P2/P4-P6-P8 or P2/P4-P6-P5-P7.
Although several observational studies have reported that older adults residing near industrial sites have increased risk of respiratory illness such as asthma and pneumonia,12,13 no study has investigated the risk of ARDS among this population. In a cross-sectional study in South Africa, Nkosi et al12 found that people 55 years of age and older living 1–2 km from mine dumps had increased risk of asthma, chronic bronchitis, and pneumonia compared with those living 5 km from mine dumps. Aylin et al13 used distance from coke works in England and Wales with outer radii of 1, 2, 3, 4.6, 5.7, 6.7, and 7.5 km to determine exposure to particulate and sulfurous pollution. They found that people 65 years of age or over had slightly higher risk for coronary heart disease with proximity to one of the study sites. In terms of ARDS risk related to industry, 19 case reports18–22,34–47 have been published on either hospital admission for ARDS or death from ARDS in occupational settings. These cases occurred among workers after they inhaled or ingested high concentrations of phosgene,54 ethenone,55 crotonaldehyde,55 mixed metal fumes,3 hydrochloric acid fumes,5 or nickel particles.41,42 Industries that reported cases included chemical manufacturing,54,55 goldsmithing,4,5 and electric welding.3,41,42
Aside from reported ARDS cases in workplaces, there is limited information on ARDS risk for community residents living near construction or manufacturing industrial sites. Thus, we conducted to our knowledge the first ecologic study in the field to examine whether community residents have increased ARDS risk by inhalable toxicants and physical hazards related to industrial activities. Our study applied an ecologic view of analyzing an association between industry and ARDS risk, which is a common method in air pollution studies.56,57 Dominici et al56 used daily counts of county-wide hospital admissions and county-specific estimated air pollution concentrations in their analysis to examine how air pollution is associated with cardiovascular and respiratory diseases among Medicare patients.57 Similarly, in our analysis, we used yearly counts of ZIP code-level hospital admissions and yearly numbers of construction (or manufacturing) companies to investigate the impact of these industries on ARDS risk. Our study explored overall impact of construction and manufacturing industry on ARDS risk among nearby older adult residents by conducting an ecological study.
Despite our important findings, this study has limitations. First, using ICD-9-CM codes without physiologic data (e.g., positive end-expiratory pressure level or chest radiograph) to define ARDS might result in under-capturing counts of hospital admissions for ARDS. We used the ICD-9 coding system under the assumption that clinicians applied similar criteria to diagnose ARDS and coders interpreted medical records properly and coded correctly.25 However, misclassification of health outcome would have occurred randomly with respect to number of industries (exposure) and bias toward the null. We further adjusted for trauma centers to consider one possibility that there were maybe more specialties to better diagnose ARDS in areas with high number of trauma centers, which are mainly located in urban areas. However, the results were consistent (eTable 3; http://links.lww.com/EDE/B663), and it supports our explanation on nondifferential outcome misclassification. In addition, we did not include code 518.81 (i.e., respiratory failure not otherwise specified, acute, acute and chronic, or chronic, and excluding acute respiratory distress) because its descriptors do not include ARDS and inclusion could result in inflated incidence of ARDS.25 Second, we assumed that industry activities are happening in the same ZIP codes, which can lead to misclassification of exposures. We tried to reduce measurement errors on industrial activities by excluding construction or manufacturing companies classified as headquarters and subsidiary headquarters. However, air toxins could travel to neighboring ZIP codes that were not reported to have any industrial activity. This misclassification would happen randomly, which is nondifferential, and bias toward the null. Third, we assumed no unmeasured confounding variables in our study, which implies that covariates such as sex, race, median household income, smoking, and business size were adequate to adjust for residual confounding.
We conducted sensitivity analysis measuring absolute standardized difference (ASD) on every covariate we adjusted for and redid the main analysis using propensity score modeling. Figure 4 and eTable 5; http://links.lww.com/EDE/B663, show that business size, the proportion of females, and the proportion of Asians were the most unbalanced in both industries, and our results could have been affected by confounding bias. However, matching data with propensity score improved the balance of covariates. In addition, results from propensity score modeling and our main models (with binary exposures) were similar (eTable 6; http://links.lww.com/EDE/B663), indicating that adjustment for this set of covariates was adequate to account for residual confounding and did not lead to confounding bias.
Our study also has a number of strengths. To begin, our large study comprised 646,542 ARDS hospital admissions from 2006 through 2012. We used a cohort of nearly 30 million Medicare beneficiaries per year, which covers more than 97% of the older adult population in the United States. In particular, because the average age of developing ARDS is around 62 years old,23 we captured most ARDS occurrences in the United States using a Medicare cohort. Our study population was not from restricted regions but was nationwide, which makes our results more generalizable. In addition, we targeted the vulnerable elderly population for their risk of ARDS. The number of older individuals 65 years of age and over is rapidly growing globally and in the United States. Considering that older adults are more susceptible to acute respiratory disease, it is important to study environmental risk factors to prevent respiratory illness. Last, we leveraged various data sources to include ZIP code-level information such as median household income, smoking, and air pollution, which could be significant confounders.
Our ecologic study, to our knowledge the first and largest, examined the relationship between yearly changes in numbers of construction and manufacturing sites and yearly changes in hospital admission rates for ARDS among older adults and found substantial evidence of associations. Our results suggest that residence in areas with more construction and manufacturing facilities is associated with increased ARDS risk. Additionally, we found that chemical product manufacturing industry activity was strongly associated with ARDS.
We would like to thank Stacy Bogan and Jeffery C. Blossom at the Center for Geographic Analysis, Harvard University; all staff at the Faculty of Art and Science Research Computing, Harvard University; Andrea Bellavia in the Department of Environmental Health, Harvard T.H. Chan School of Public Health; and Taylor Thompson at the Massachusetts General Hospital.
1. Araújo IP, Costa DB, de Moraes RJ. Identification and characterization of particulate matter concentrations at construction jobsites. Sustainability. 2014;6:7666–7688.
2. Cheminfo Services Inc. Best Practices for the Reduction of Air Emissions from Construction and Demolition Activities. Available at: http://www.bv.transports.gouv.qc.ca/mono/1173259.pdf
. 2005. Accessed 6 April 2020.
3. Barbee JY Jr, Prince TS. Acute respiratory distress syndrome
in a welder exposed to metal fumes. South Med J. 1999;92:510–512.
4. Parikh JM, Dhareshwar S, Sharma A, Karanth R, Ramkumar VS, Ramaiah I. Acute respiratory distress in a silversmith. Indian J Occup Environ Med. 2014;18:27–28.
5. Bansal DP, Ambegaonkar R, Radhika P, Sharma M. ARDS following inhalation of hydrochloric acid. J Assoc Physicians India. 2011;59:115–117.
6. Zierold D, Chauviere M. Hydrogen fluoride inhalation injury because of a fire suppression system. Mil Med. 2012;177:108–112.
7. Christia-Lotter A, Bartoli C, Piercecchi-Marti MD, et al. Fatal occupational inhalation of hydrogen sulfide. Forensic Sci Int. 2007;169:206–209.
8. Porta D, Milani S, Lazzarino AI, Perucci CA, Forastiere F. Systematic review of epidemiological studies on health effects associated with management of solid waste. Environ Health. 2009;8:60.
9. Vrijheid M. Health effects of residence near hazardous waste landfill sites: a review of epidemiologic literature. Environ Health Perspect. 2000;108(1 suppl):101–112.
10. Wilson D, Takahashi K, Pan G, et al. Respiratory symptoms among residents of a heavy-industry province in China: prevalence and risk factors. Respir Med. 2008;102:1536–1544.
11. United Nations, Department of Economic and Social Affairs, Population Division. World population ageing 2015 report. ST/ESA/SER. A/390. 2015.
12. Nkosi V, Wichmann J, Voyi K. Chronic respiratory disease among the elderly in South Africa: any association with proximity to mine dumps? Environ Health. 2015;14:33.
13. Aylin P, Bottle A, Wakefield J, Jarup L, Elliott P. Proximity to coke works and hospital admissions for respiratory and cardiovascular disease in England and Wales. Thorax. 2001;56:228–233.
14. Donahoe M. Acute respiratory distress syndrome
: a clinical review. Pulm Circ. 2011;1:192–211.
15. Confalonieri M, Salton F, Fabiano F. Acute respiratory distress syndrome
. Euro Respir Rev. 2017;26:160116.
16. Frutos-Vivar F, Nin N, Esteban A. Epidemiology of acute lung injury and acute respiratory distress syndrome
. Curr Opin Crit Care. 2004;10:1–6.
17. Brun-Buisson C, Minelli C, Bertolini G, et al.; ALIVE Study Group. Epidemiology and outcome of acute lung injury in European intensive care units. Results from the ALIVE study. Intensive Care Med. 2004;30:51–61.
18. Irish Critical Care Trials Group. Acute lung injury and the acute respiratory distress syndrome
in Ireland: a prospective audit of epidemiology and management. Crit Care. 2008;12:R30.
19. Thompson BT, Chambers RC, Liu KD. Acute respiratory distress syndrome
. N Engl J Med. 2017;377:562–572.
20. Hudson LD, Steinberg KP. Epidemiology of acute lung injury and ARDS. Chest. 1999;116(1 suppl):74S–82S.
21. Howlader N, Noone AM, Krapcho M, et al. SEER Cancer Statistics Review, 1975–2008. 2010.Bethesda, MD: National Cancer Institute.
22. Ware LB, Matthay MA. The acute respiratory distress syndrome
. N Engl J Med. 2000;342:1334–1349.
23. Bellani G, Laffey JG, Pham T, et al.; LUNG SAFE Investigators; ESICM Trials Group. Epidemiology, patterns of care, and mortality for patients with acute respiratory distress syndrome
in intensive care units in 50 countries. JAMA. 2016;315:788–800.
24. Eachempati SR, Hydo LJ, Shou J, Barie PS. Outcomes of acute respiratory distress syndrome
(ARDS) in elderly patients. J Trauma. 2007;63:344–350.
25. Reynolds HN, McCunn M, Borg U, Habashi N, Cottingham C, Bar-Lavi Y. Acute respiratory distress syndrome
: estimated incidence and mortality rate in a 5 million-person population base. Crit Care. 1998;2:29–34.
26. Chen W, Chen YY, Tsai CF, et al. Incidence and outcomes of acute respiratory distress syndrome
: a nationwide registry-based study in Taiwan, 1997 to 2011. Medicine (Baltimore). 2015;94:e1849.
27. Knaus WA, Sun X, Hakim RB, Wagner DP. Evaluation of definitions for adult respiratory distress syndrome. Am J Respir Crit Care Med. 1994;150:311–317.
28. Ike JD, Kempker JA, Kramer MR, Martin GS. The association between acute respiratory distress syndrome
mortality and annual hospital ards case volumes from 2002-2011. In: The Nationwide Inpatient Sample. A51. Critical care: risk stratification and prognostication-from bedside to big data
. 2017:American Thoracic Society; A1819–A1819.
29. Di Q, Wang Y, Zanobetti A, et al. Air pollution and mortality in the Medicare
population. N Engl J Med. 2017;376:2513–2522.
30. Di Q, Rowland S, Koutrakis P, Schwartz J. A hybrid model for spatially and temporally resolved ozone exposures in the continental United States. J Air Waste Manag Assoc. 2016;67:39–52.
31. Moss M, Mannino DM. Race and gender differences in acute respiratory distress syndrome
deaths in the United States: an analysis of multiple-cause mortality data (1979- 1996). Crit Care Med. 2002;30:1679–1685.
32. Erickson SE, Shlipak MG, Martin GS, et al.; National Institutes of Health National Heart, Lung, and Blood Institute Acute Respiratory Distress Syndrome
Network. Racial and ethnic disparities in mortality from acute lung injury. Crit Care Med. 2009;37:1–6.
33. Martin GS, Mannino DM, Eaton S, Moss M. The epidemiology of sepsis in the United States from 1979 through 2000. N Engl J Med. 2003;348:1546–1554.
34. Barnato AE, Alexander SL, Linde-Zwirble WT, Angus DC. Racial variation in the incidence, care, and outcomes of severe sepsis: analysis of population, patient, and hospital characteristics. Am J Respir Crit Care Med. 2008;177:279–284.
35. Calfee CS, Matthay MA, Eisner MD, et al. Active and passive cigarette smoking and acute lung injury after severe blunt trauma. Am J Respir Crit Care Med. 2011;183:1660–1665.
36. Garcia JG, Sznajder JI. Healthcare disparities in patients with acute respiratory distress syndrome
. Toward equity. Am J Respir Crit Care Med. 2013;188:631–632.
37. Ware LB, Zhao Z, Koyama T, et al. Long-term ozone exposure increases the risk of developing the acute respiratory distress syndrome
. Am J Respir Crit Care Med. 2016;193:1143–1150.
38. Rush B, McDermid RC, Celi LA, Walley KR, Russell JA, Boyd JH. Association between chronic exposure to air pollution and mortality in the acute respiratory distress syndrome
. Environ Pollut. 2017;224:352–356.
39. Rhee J, Dominici F, Zanobetti A, Schwartz J, Wang Y, Di Q, et al. Impact of long-term exposures to ambient PM2. 5 and ozone on acute respiratory distress syndrome
(ARDS) risk for older adults
in the United States. Chest. 2019;156:71–79.
40. Makar M, Antonelli J, Di Q, Cutler D, Schwartz J, Dominici F. Estimating the causal effect of low levels of fine particulate matter on hospitalization. Epidemiology. 2017;28:627–634.
41. Phillips JI, Green FY, Davies JC, Murray J. Pulmonary and systemic toxicity following exposure to nickel nanoparticles. Am J Ind Med. 2010;53:763–767.
42. Rendall RE, Phillips JI, Renton KA. Death following exposure to fine particulate nickel from a metal arc process. Ann Occup Hyg. 1994;38:921–930.
43. Anderson AR, Wu J. Top five industries resulting in injuries from acute chemical incidents—hazardous substance emergency events surveillance, nine states, 1999–2008. Morb Mortal Wkly Rep Surveill Summ. 2015;64:47–53.
44. Bergdahl IA, Torén K, Eriksson K, et al. Increased mortality in COPD among construction workers exposed to inorganic dust. Eur Respir J. 2004;23:402–406.
45. Blanc PD, Iribarren C, Trupin L, et al. Occupational exposures and the risk of COPD: dusty trades revisited. Thorax. 2009;64:6–12.
46. Purdue MP, Gold L, Järvholm B, Alavanja MC, Ward MH, Vermeulen R. Impaired lung function and lung cancer incidence in a cohort of Swedish construction workers. Thorax. 2007;62:51–56.
47. Almirall J, Serra-Prat M, Bolíbar I, et al. Professions and working conditions associated with community-acquired pneumonia. Arch Bronconeumol. 2015;51:627–631.
48. Torén K, Qvarfordt I, Bergdahl IA, Järvholm B. Increased mortality from infectious pneumonia after occupational exposure to inorganic dust, metal fumes and chemicals. Thorax. 2011;66:992–996.
49. Schaaf B, Kruse J, Rupp J, et al. Sepsis severity predicts outcome in community-acquired pneumococcal pneumonia. Eur Respir J. 2007;30:517–524.
50. Matthay MA, Zemans RL. The acute respiratory distress syndrome
: pathogenesis and treatment. Annu Rev Pathol. 2011;6:147–163.
51. Haslam RA, Hide SA, Gibb GAF, et al. Causal factors in construction accidents. 2003.Health and Safety Executive 156.
52. Bronchard R, Albaladejo P, Brezac G, et al. Early onset pneumonia: risk factors and consequences in head trauma patients. Anesthesiology. 2004;100:234–239.
53. Wang KW, Chen HJ, Lu K, et al. Pneumonia in patients with severe head injury: incidence, risk factors, and outcomes. J Neurosurg. 2013;118:358–363.
54. Hardison LS Jr, Wright E, Pizon AF. Phosgene exposure: a case of accidental industrial exposure. J Med Toxicol. 2014;10:51–56.
55. Huang JF, Zhu DM, Ma JF, Zhong M. Acute respiratory distress syndrome
due to exposure to high-concentration mixture of ethenone and crotonaldehyde. Toxicol Ind Health. 2015;31:585–587.
56. Dominici F, Peng RD, Bell ML, et al. Fine particulate air pollution and hospital admission for cardiovascular and respiratory diseases. JAMA. 2006;295:1127–1134.
57. Peng RD, Chang HH, Bell ML, et al. Coarse particulate matter air pollution and hospital admissions for cardiovascular and respiratory diseases among Medicare
patients. JAMA. 2008;299:2172–2179.