Hurricane Sandy (Sandy) was the largest Atlantic hurricane on record and the second most costly hurricane in US history after Hurricane Katrina.Reference Landsea 1 When Sandy hit New York City on October 29, 2012, the East River overflowed its banks, flooding large areas of lower Manhattan. The Atlantic Ocean storm surge also caused considerable flood damage in Queens, Brooklyn, and Staten Island, displacing thousands of people.Reference Abramson and Redlener 2
The extensive flooding created environmental conditions that could potentially lead to an increased risk of lower respiratory symptoms (LRS). For instance, sediment left behind after floodwaters receded could have been disturbed during recovery efforts, thus posing a risk of inhalational injury. Such dust could potentially contain irritants and toxins such as volatile pollutants, solvent emissions, or asbestos.Reference Freund, Zuckerman, Luo, Hsu and Lucchini 3 - Reference Mandigo, DiScenza, Keimowitz and Fitzgerald 5
Flooding also creates ideal conditions for the growth of mold, which has been reported as a risk factor for LRS.Reference Johanning, Auger, Morey, Yang and Olmsted 6 , 7 In addition to mold growth, excessive indoor dampness is a source of several potentially problematic exposures such as dust mites and microbial growth. Furthermore, standing water supports cockroach and rodent infestations, and excessive moisture may initiate chemical emissions from building materials and furnishings. 8
Beyond the flooding and potential for mold growth, power outages were extensive, requiring the use of numerous generators, which created odorous exhaust fumes.Reference Freund, Zuckerman, Luo, Hsu and Lucchini 3 , 9 Relief centers were set up in parking lots and consisted of heated trailers that produced combustion fumes.Reference Freund, Zuckerman, Luo, Hsu and Lucchini 3 Construction vehicles and trucks emitted diesel exhaust and, along with automobiles, created clouds of suspended sand, debris, and dust as they moved along the roads.Reference Freund, Zuckerman, Luo, Hsu and Lucchini 3 Additionally, exposure to dust containing microbial and other agents could have occurred during demolition, removal, and repair of flood-damaged and contaminated infrastructure and building materials.Reference Barbeau, Grimsley, White, El-Dahr and Lichtveld 10 - Reference Rando, Lefante, Freyder and Jones 12 Each of these exposures has been associated with respiratory symptoms.Reference Freund, Zuckerman, Luo, Hsu and Lucchini 3 , Reference Rando, Lefante, Freyder and Jones 12 , 13
There is some evidence that Sandy-related exposures led to an increase in LRS and associated hospitalizations. For instance, a study of Long Island emergency departments found a 4% increase in visits for diseases of the respiratory system in the month following Sandy, compared with the month before.Reference Kim, Schwartz, Hirsch, Silverman, Liu and Taioli 14 However, this and other published studies of post-Sandy health effects focused on emergency department visits or hospitalizations, did not assess the relationship between specific, individual-level exposures and subsequent LRS.Reference Freund, Zuckerman, Luo, Hsu and Lucchini 3 , Reference Kim, Takematsu, Biary, Williams, Hoffman and Smith 4 , Reference Kim, Schwartz, Hirsch, Silverman, Liu and Taioli 14 , Reference Gotanda, Fogel and Husk 15 More information is needed to better understand the scope of post-Sandy respiratory issues, as well as how pre-existing respiratory symptoms and conditions contributed to the respiratory effects of Sandy-related exposures. In the current study, we sought to (1) describe exposures to respiratory hazards following Sandy, (2) examine the prevalence of post-Sandy LRS, (3) assess whether post-Sandy exposures were associated with subsequent LRS, and (4) determine whether the relationship between Sandy-related exposures and subsequent LRS differed according to the presence/absence of pre-existing respiratory problems.
Methods
Study Design and Sample Selection
We used data on a sample of World Trade Center (WTC) Health Registry (Registry) participants who were potentially affected by Hurricane Sandy. The Registry is a longitudinal cohort study designed to monitor the health of persons exposed to the 9/11/2001 WTC terrorist attacks. In 2003-2004, the Registry enrolled and surveyed 71,430 individuals exposed to the 9/11 attacks (Wave 1). Follow-up surveys were conducted in 2006-2007 (Wave 2) and 2011-2012 (Wave 3). Wave 3 data collection ended in March 2012, 7 months before Hurricane Sandy. The 43,134 enrollees who completed Wave 3 and lived in the tri-state metropolitan area (New York, New Jersey, and Connecticut) constituted the sample pool for the Sandy sub-study. Based on the Federal Emergency Management Agency (FEMA) Modeling Task Force estimated hurricane inundation zones in the NYC tri-state metropolitan area, 9 the sample pool was divided into 2 groups according to their residential location. The first group consisted of enrollees who lived in a Sandy inundation zone as defined by FEMA (n=4435), while a comparison group of the same size was selected randomly from the subset of enrollees with a current tri-state area address outside the inundation zone (n=4435). Beginning March 28, 2013, the 8870 enrollees were invited via mail or e-mail to complete a survey by paper or web regarding Sandy-related exposures and physical and mental health. After multiple reminders, 4558 (51.4%) completed surveys were received (2443 from the inundation zone; 2115 from the non-inundation zone) by November 7, 2013. Additional details on Registry recruitment and data collection and sample selection for the Hurricane Sandy sub-study are available elsewhere.Reference Brackbill, Caramanica and Maliniak 16 , Reference Caramanica, Brackbill, Stellman and Farfel 17 The Registry and Hurricane Sandy studies were approved by the New York City Department of Health and Mental Hygiene’s Institutional Review Board.
The current study included enrollees who completed the Hurricane Sandy survey as well as Registry Waves 1, 2, and 3. Enrollees aged <18 years on 9/11 were excluded (n=494), resulting in a final analytic sample of 3835.
Lower Respiratory Symptoms (LRS)
On both the Wave 3 (2010-2011) and Hurricane Sandy (2013) questionnaires, LRS were defined as wheezing, persistent cough, or shortness of breath reported on at least 1 of the 30 days preceding survey completion, excluding times when a cold, the flu, or seasonal allergies were present.Reference Friedman, Farfel, Maslow, Cone, Brackbill and Stellman 18 LRS reported prior to Sandy were used as predictors of future symptoms, and LRS reported post-Sandy were the outcome, or dependent variable.
Hurricane Sandy Exposure Variables
We examined 3 types of Hurricane Sandy-related inhalational exposures: (1) reconstruction exposure (removing water, removing mud/debris or “muck,” “tear out” work, or major repair at the respondent’s or someone else’s homes that were damaged by Sandy); (2) exposure to mold or damp environments (for those who returned home after the hurricane, having a flooded home, seeing mold, or smelling mold at home; or being exposed to sewage, dirty or contaminated flood water, or visible mold); and (3) exposure to other respiratory irritants (ie, debris, exhaust fumes from generators, diesel fuel, or heating oil leaks or spills). Hurricane Sandy exposures were analyzed separately, in a single model, and as a count variable (i.e., no exposures, 1 exposure, 2 exposures, or all 3 types of exposures).
Demographic and Physical Health Variables
Socio-demographic information (gender, age, race/ethnicity, household income in 2010, and education), smoking history, body mass index, and self-reported history of clinician-diagnosed asthma were obtained from the Wave 1, 2, or 3 surveys.
Post-Traumatic Stress Disorder (PTSD)
PTSD was included because previous work among 9/11-exposed populations showed an association between LRS and PTSD.Reference Friedman, Farfel, Maslow, Cone, Brackbill and Stellman 18 - Reference Nair, Ekenga, Cone, Brackbill, Farfel and Stellman 20 Probable PTSD was assessed at Wave 3 with the 17-item 9/11-specific PTSD Checklist (PCL-17), which reflects DSM-IV criteria for PTSD (re-experiencing, avoiding, hyper-arousal).Reference Ruggiero, Del Ben, Scotti and Rabalais 21 Respondents rated symptom severity on a 5-level Likert scale from not at all (1) to extremely (5). 9/11-related PTSD was defined as a PCL score of ≥44 on the Wave 3 survey.Reference Caramanica, Brackbill, Stellman and Farfel 17 , Reference Brackbill, Hadler and DiGrande 22
Data Analysis
χ2 tests were used to test the statistical significance of associations between LRS and socio-demographic characteristics, symptoms, PTSD, and Hurricane Sandy exposures. Multivariable logistic regression models adjusted for variables significant at the bivariate level were used to assess associations between Hurricane Sandy exposures and LRS. To examine whether factors associated with post-Sandy LRS were different among persons with and without pre-existing LRS, we also conducted a multivariable analysis stratified on the presence or absence of LRS at Wave 3. To assess for dose-response for the number of exposures reported, a Cochran-Armitage Trend test was performed. All analyses were conducted using SAS Version 9.4.
Results
Characteristics of the Study Population
Participants were predominantly male (56.8%), 45-64 years of age at the time Sandy occurred (64.7%), non-Hispanic white (73.2%), and had at least some college education (41.9%) (Table 1). Over half had never smoked (52.7%) and 31.4% were obese. At Wave 3 (2011-2012), 17.8% had probable PTSD and 45.4% reported at least 1 LRS. Over a quarter (26.8%) of participants reported a history of clinician-diagnosed asthma on the Wave 1, 2, or 3 surveys.
Table 1 Characteristics of Survey Participants and Prevalence of Lower Respiratory Symptoms (LRS)Footnote a After Hurricane Sandy (2013)
Totals might not equal 3835 due to missing data.
a LRS defined as wheezing, persistent cough, or shortness of breath reported on at least 1 of the 30 days preceding survey completion, excluding times when a cold, the flu, or seasonal allergies were present.
b Reconstruction exposure: removing water, removing mud/debris or “muck,” “tear out” work, or major repair at their or someone else’s homes that were damaged by Sandy.
c Mold or damp environments: for those who returned home after the hurricane, having a flooded home, seeing mold, or smelling mold at home; or being exposed to sewage, dirty or contaminated flood water, or visible mold.
d Other irritants: debris, exhaust fumes from generators, diesel fuel or healing oil leaks or spills.
Lower Respiratory Symptoms After Hurricane Sandy
Over one-third (34.4%) reported at least 1 LRS on their Hurricane Sandy Survey (Table 1). Of those who reported post-Sandy LRS, 41.5% reported 1 symptom, 30.1% reported 2, and 28.4% reported all 3. Older age, male sex, lower educational attainment, current smoking, and obesity were associated with post-Sandy LRS in bivariate analysis (Table 1). Compared to participants without PTSD, a higher proportion of those with PTSD reported LRS on their Hurricane Sandy survey (60.0% vs 28.2%). Moreover, a higher proportion of those who reported LRS at Wave 3 also reported LRS on their Hurricane Sandy survey (57.2%) compared with those who did not report LRS at Wave 3 (15.5%). Enrollees who reported a history of clinician-diagnosed asthma also had a higher prevalence of post-Sandy LRS (54.6% vs 26.5% among those without asthma).
Hurricane Sandy Exposures and Post-Hurricane Sandy Lower Respiratory Symptoms
One-third (33.1%) of the study participants reported reconstruction exposure. Similar proportions reported having mold/damp environment and other respiratory irritant exposures (36.8% and 37.9%, respectively; Table 1). More than half of participants (52.1%) reported having experienced at least one type of Sandy-related inhalational exposure.
When the association of Hurricane Sandy exposures and post-Sandy LRS was assessed in separate multivariable models (one for each inhalational exposure type), participants with reconstruction exposure had an adjusted odds ratio (AOR) for post-Sandy LRS of 1.8 compared with those without such exposure (95% confidence interval [CI]: 1.5-2.1) (Table 2). Those who reported mold/damp environment exposure had twice the odds of having post-Sandy LRS than those without such exposure (AOR: 2.0, 95% CI: 1.7-2.4). Those who reported exposure to other irritants also had more than twice the odds of having post-Sandy LRS compared with those who did not report exposure to other irritants (AOR: 2.2, 95% CI: 1.8-2.6) (Table 2). When all 3 exposures were put in a single model, the associations with post-Sandy LRS were attenuated but remained statistically significant for mold/damp environment exposure (AOR: 1.3, 95% CI: 1.1-1.7) and other irritant exposure (AOR: 1.7, 95% CI: 1.3-2.1; data not shown).
Table 2 Adjusted Odds Ratios (AOR) for Associations Between Lower Respiratory Symptoms (LRS) And Hurricane Sandy Exposures Overall and Stratified on Presence/Absence of LRS in 2011-2012
a Adjusted for gender, age, race/ethnicity, income (2010), post-traumatic stress disorder (2011-12), clinician-diagnosed asthma, smoking history, obesity, education, and LRS in 2011-12 in non-stratified model.
We found a significant dose-response relationship (P<0.0001) between the number of exposure types reported and the odds of post-Sandy LRS (Figure 1). Compared with participants who reported no exposures, those who reported 1 type of exposure had twice the odds of having post-Sandy LRS (AOR: 2.1, 95% CI: 1.6-2.6) whereas those who reported all 3 types of exposures had nearly 3 times the odds of having post-Sandy LRS (AOR: 2.7, 95% CI: 2.1-3.4) (Figure 1).
Figure 1 Adjusted Odds Ratios for Association of Lower Respiratory Symptoms and Number of Types of Hurricane Sandy-Related Exposures.
Stratification by LRS at Wave 3
Associations between Hurricane Sandy exposures and post-Sandy LRS were similar for enrollees who reported LRS at Wave 3 and those who did not (Table 2).
Discussion
We found that exposure to mold/damp environments, reconstruction work, and other potential respiratory irritants were each independently associated with post-Sandy LRS. In addition, there was a dose-response relationship between the number of exposure types reported and the odds of post-Sandy LRS. The magnitude of the association between these exposures and post-Sandy LRS was similar for participants with and without pre-existing LRS, suggesting that the Sandy-related exposures were risk factors for both new-onset LRS and the persistence or exacerbation of pre-existing LRS.
Exposure to damp or moldy indoor space has been found to be associated with respiratory symptoms, including wheezing and cough, 8 as well as with more severe symptoms among persons with asthma.Reference Landsea 1 However, a review of studies of the health effects of exposure to mold in the months following Hurricanes Katrina and Rita found no increase in the occurrence of adverse health outcomes, including respiratory symptoms.Reference Barbeau, Grimsley, White, El-Dahr and Lichtveld 10 Our current study thus adds new evidence that exposure to moldy or damp conditions in the home after hurricane-related flooding may cause or worsen LRS, underscoring the importance of prompt remediation of such conditions.
Less is known about reconstruction exposures and their associations with LRS. These exposures include inhalation of dust; removal of building materials; and other repair work of infrastructure post-hurricane, all of which could potentially introduce environmental triggers for respiratory symptoms. Our results align with those of Rando et al, who found that the prevalence rate ratios for respiratory issues such as sinusitis, fever, and cough were significantly elevated among restoration workers after Hurricane Katrina.Reference Rando, Lefante, Freyder and Jones 12 Prevalence rate ratios were also elevated for new-onset asthma, although this did not reach the level of statistical significance.Reference Rando, Lefante, Freyder and Jones 12
We also found that exposure to irritants such as exhaust fumes from generators or diesel fuel was associated with increased odds of post-Sandy LRS. There has been one report of diesel and silica monitoring following Sandy which found that the daily average levels of elemental carbon and airborne particulates were in the ranges that have been found to cause respiratory effects in sensitive sub-populations, such as asthmatic patients, after 2 hours of exposure.Reference Freund, Zuckerman, Luo, Hsu and Lucchini 3 An analysis of poison control center data found a significant increase in gasoline exposures after Sandy, with pulmonary symptoms as one of the top toxic effects.Reference Kim, Takematsu, Biary, Williams, Hoffman and Smith 4 This suggests that disaster plans should incorporate public health messaging regarding potentially toxic exposure to diesel and gasoline.
Additional research is needed on the relationship between post-hurricane exposures and respiratory effects. Questions of exposure and dose have not, by and large, been resolved. Beyond this study, few studies to date have considered whether there are additive or synergistic interactions among different types of exposures. Longer-term studies are ongoing to further evaluate the impact of mold on the health of children and workers.Reference Barbeau, Grimsley, White, El-Dahr and Lichtveld 10 After Sandy, government agencies provided guidance on avoiding mold and other exposures, including wearing an N95 dust mask and other personal protective equipment, having appropriate ventilation in work areas, and proper disposal procedures. 23 , 24 Studies are needed to evaluate the effectiveness of various means of protection during remediation activities.
Strengths and Limitations
The availability of pre-Sandy information on this cohort, including data on LRS, asthma, PTSD, and other comorbidities, enabled us to account for prior history when assessing the relationship between Sandy-related exposures and subsequent LRS. We were also able to describe a broad range of post-hurricane exposures that resulted from circumstances and activities related to reconstruction work or returning home after a disaster, and to assess the relationship between various types of post-Sandy exposures and LRS. Nonetheless, this study also has several limitations. We did not collect data on the use of personal protective gear, such as respirators, which may mitigate post-exposure LRS. In addition, exposure data for Sandy were self-reported and were collected at the same time when post-Sandy LRS were assessed, which could potentially bias reporting. Another possible limitation is selection bias; survey respondents may have chosen to participate because they perceived that they had greater exposure to Sandy and/or experienced more prolonged or severe symptoms.
Conclusions
Exposure to multiple sources of potential respiratory irritants after Hurricane Sandy, including mold/damp environments, irritants from reconstruction, and diesel exhaust, was associated with an increased risk of subsequent LRS. Our findings highlight the importance of safe and timely remediation of damp or moldy indoor environments, and suggest that respiratory protection is important for persons conducting reconstruction and clean-up work after a hurricane.
Acknowledgments
This publication was supported by Cooperative Agreement Numbers 2U50/OH009739 and 5U50/OH009739 from the National Institute for Occupational Safety and Health (NIOSH) of the Centers for Disease Control and Prevention (CDC); U50/ATU272750 from the Agency for Toxic Substances and Disease Registry (ATSDR), CDC, which included support from the National Center for Environmental Health, CDC; and by the New York City Department of Health and Mental Hygiene (NYC DOHMH). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of NIOSH, CDC, or the Department of Health and Human Services.
Authors’ Contribution
Dr Gargano designed the study, performed the statistical analyses, and wrote the first draft of the manuscript. Drs Brackbill and Jordan and Mr Locke contributed to the study design and analyses. All authors contributed to data interpretation and to the revision of the manuscript. All authors have approved the final manuscript.
