Estimation of Cumulative Asbestos Exposure Associated With 9/11
A modern risk assessment for asbestos-related cancer uses knowledge of the type of asbestos and the cumulative exposure, which represents the intensity and duration of exposure usually given in fibers per milliliter multiplied by years (f/mL × years). Outdoor air samples in NYC pre-9/11 were consistently less than 0.0008 f/mL for all asbestos fiber types having lengths ≥5 μm (Figs. 5 and 6).5 Because the settled dust gave no indication of any amphibole asbestos being present, we only considered chrysotile asbestos in our discussion of the upper limit (similar results regarding the asbestos fiber type present have been reported by others).9
To our knowledge, no air sampling data have been reported for the initial dust cloud on 9/11, and it is doubtful whether such a particulate dense aerosol could have been meaningfully analyzed for the type and concentration of respirable asbestos. Considering the trace amount of chrysotile asbestos in the settled dust, we estimate the maximum concentration of airborne asbestos at 50 f/mL with a length ≥5 μm. It is problematic to use the analysis of settled dust to determine the extent to which these asbestos fibers, when airborne, were respirable and at what concentration. The high exposure assumed is similar to the exposures measured historically in uncontrolled chrysotile asbestos mines and mills where the ore contains a minimum of 2% to 4% asbestos −100 times more than in the dust from the Twin Towers. This is therefore likely to be a pessimistic assumption and we assume it as an upper limit (Fig. 5).10
Photographs taken approximately 5 hours after the collapse of the first tower indicate that the suspended dust settled rapidly (Fig. 2). We assumed the airborne concentration of chrysotile asbestos to have decreased during that initial 5 hours by 50-fold to no more than 1 f/mL ≥5 μm in length. Then, we assumed that the concentration further decreased linearly to background by the time we collected our first air sample 26.4 days later on October 8th (Fig. 1, Fig. 6). Because the decrease was more likely to have been exponential, this linear assumption is conservative.
Air samples were collected by EPA starting on September 15th and continued through October 8th. The EPA collected 8870 air samples in Lower Manhattan after 9/11 for analysis by analytical transmission electron microscopy.11 Twenty-two air samples (0.24%) exceeded the AHERA standard of 70 structures per square millimeter (S/mm2), having a length greater than or equal to 0.5 μm. The samples that exceeded the AHERA standard we mainly collected at the perimeter of Ground Zero and the landfill on Staten Island. The AHERA standard is not a health benchmark but rather reflects the upper limit of possible asbestos contamination of the collection filter. Based on the area of the filter examined for fibers and the volume of air sampled required in the AHERA protocol, the 70 S/mm2 corresponds to approximately 0.02 S/mL ≥0.5 μm in length. Not all structures are fibers and the number of S/mL will always be equal to or greater than the number of f/mL. Therefore when S/mm2 are converted to f/mL, the exposure values are upper limits.
The first air samples collected after 9/11 (by the EPA on September 15th) correspond to 0.038 f/mL and 0.048 f/mL. An additional 10 air samples (range 0.021–0.164 f/mL, mean 0.04 f/mL ≥0.5 μm in length for all 12 air samples) were above 0.02 f/mL before collecting our first air sample on October 8th. Although these air samples are of limited use for risk assessment that requires that the number and type of airborne asbestos fiber ≥5 μm in length be determined, it is interesting to know that the numbers for total fibers present are below what we have assumed for the ≥5 μm in length fraction.
In our exposure estimate, we assume a mean exposure to 0.5 f/mL of chrysotile asbestos ≥5 μm in length for the 26.4-day period from the initial clearing of the heavy airborne dust after 5 hours until the background level is re-established on October 8th. The mean of the 12 highest concentrations reported by EPA during that time period, 0.04 f/mL ≥0.5 μm in length, is an order of magnitude lower than our assumed value and would be even lower if fiber length were considered. As with the initial 5-hour exposure period, we have assumed what is likely to be a worst-case estimate of exposure prior to establishing background on October 8th.
On the basis of the analysis of settled dust and air sampling, we estimated the 9/11-based incremental increase in the ambient asbestos exposure for a typical resident of Lower Manhattan. Our objective was to assess the incremental cancer risk associated with this exposure by applying established quantitative risk assessment models. We calculated two exposure indexes: cumulative lifetime exposure for use with separate risk models for lung cancer and mesothelioma, which was developed by Hodgson and Darnton,12 and lifetime average daily exposure (LADE) for use with EPA's aggregate risk model for lung cancer plus mesothelioma.13 The assumed exposure levels from 9/11 until our first air sample was collected on October 8th is shown graphically in Fig. 6. The cumulative exposure for this time period is calculated below:
Although the initial level of 50 f/mL fell to approximately 1 f/mL during the first 5 hours after the first tower collapsed on 9/11, our estimate is an upper bound on exposures by assuming 50 f/mL throughout the 5-hour interval:
for the next 26.4 days, our estimate is as follows:
The total cumulative exposure is the sum of the exposures for these two time-periods is as follows:
Anyone not exposed to the initial 5-hour dust cloud on 9/11 had less than half the cumulative environmental chrysotile asbestos exposure given. Because no asbestos fibers were detected in any of the air samples, the upper 95% confidence limit for the combined samples, 0.00004 f/mL, was used as the background concentration of asbestos fibers.
The risk models we used were derived from occupational exposure data. Therefore, we must restate our continuous environmental exposure estimates as equivalent occupational exposures. Occupational exposures occur over the course of 250 days per year for 8 hours per day. Continuous environmental exposure occurs over 365 days per year 24 hours per day. Therefore, multiplying continuous exposure by the ratio (365 days/yr × 24 hours/d)/(250 working days /yr × 8 hours/d) = 4.38 produces equivalent occupational exposure. The equivalent occupational exposure associated with the events of 9/11 is 4.38-fold larger than the environmental exposure, or 0.28 f/mL − years.
The Lifetime Average Daily Exposure (LADE), the exposure index used with EPA's aggregate cancer risk model for asbestos is a measure of exposure for 24 hours per day every day of the year. LADE, therefore, is the environmental exposure calculated above, 0.065f/mL − years, divided 70 years, the lifetime duration EPA uses for risk assessment. LADE for the 9/11 exposure is 0.0009 f/mL (= 0.065/70).
Risk Assessment for Asbestos-Related Cancer
The number of asbestos-related mesothelioma (OM) depends on the type of asbestos one is exposed to, the cumulative exposure and the age at which exposure first occurs12 and can be calculated by the following:
Where RM is the risk of mesothelioma as a percentage of the total expected mortality. The RM used, 0.001, is obtained from Table 1 of Hodgson and Darnton12 (adjusted to 30 years of age at first exposure) and over estimates the chrysotile asbestos risk as some exposure to amphibole asbestos occurred in the cohorts used to determine the value of RM. This is derived from occupational exposure, assumed to be 8 hours/d for 250 days per year. ECA is the cumulative chrysotile asbestos environmental exposure (assumed to be continuous) 0.065f/mL × years is multiplied by 4.38 to the equivalent occupational exposure 0.28 f/mL × years (Fig. 6). Tpop is the adjusted total exposed population for Lower Manhattan. The total population is 57,514 residents estimated from United States Census 2000 (see Fig. 4 for area included). Multiplying the Tpop by 0.47 adjusts the age at first exposure to the average age of Lower Manhattan residents of 38.12 Tpop used in the calculation is 57,514 × 0.47 = 27,302.
Solving for OM:
OM = 0.08 mesothelioma cases due to 9/11 exposure to chrysotile asbestos and the lifetime risk of mesothelioma is OM/Opop = 1.39 × 10−6.
For a given cumulative asbestos exposure, the risk of developing lung cancer will increase as a percentage of the existing lung cancer risk in the population. We will assume that on average 8% of cigarette smokers develop lung cancer, 90% of the lung cancers are found in smokers, and 25% of the residents of Lower Manhattan smoke. The risk of lung cancer increases linearly with cumulative asbestos exposure following the relationship:
We wish to calculate the increase in the observed number of lung cancers (ObsL) caused by exposure to chrysotile asbestos. ExpL is the expected background of lung cancer deaths, 1,278, among the 57,514 residents of Lower Manhattan. This background rate is determined by solving equations that reflect the relationship between the percentage of smokers who get lung cancer and the percentage of lung cancers that occur in smokers. Specifically, 0.9 × (no. lung cancers) = 0.08 × (no. smokers) = 0.08 × 0.25 × 57,514/0.9 = 1,278.
RL is the risk of lung cancer expressed as a percentage of lung cancer deaths per f/mL × years of asbestos exposure. The RL used is 0.062 obtained from Table 2 of Hodgson and Darnton12 and is specific for chrysotile asbestos. ECA is the cumulative chrysotile asbestos environmental exposure (assumed to be continuous) 0.065 f/mL × years is converted to the equivalent occupational exposure of 0.28 f/mL × years (Fig. 6). Using these values ObsL = 0.22 and the relative risk of lung cancer associated with the events of 9/11 is ObsL/ExpL = 1.7 × 10−4.
The US EPA's aggregate asbestos cancer risk model does not differentiate asbestos fiber types. The risk for the sum of lung cancer and mesothelioma is calculated as 0.23 × LADE, where the increment to LADE (lifetime average daily exposure) for the events of 9/11 is 0.0009 f/mL. The risk of cancer equals 2.1 × 10−4, which is equivalent to 12 excess cancers, for the population of Lower Manhattan.
The attacks on NYC's WTC and the collapse of both towers created a pressure wave, which dispersed an enormous amount of dust containing asbestos into the outside air of Lower Manhattan (Figs. 1 and 2). Our analysis of representative settled dust samples by XRD, PLM, and ATEM indicates that of the six regulated asbestos fiber types, only chrysotile asbestos was present. The chrysotile asbestos concentration was less than 0.01% by volume. Although estimating the airborne concentration of asbestos on and shortly after 9/11 has limitations, it undoubtedly was above the background in the air for some period of time.9,11 The potential for an increased incidence of asbestos-related cancer from 9/11-related exposure depends principally on two factors: asbestos fiber type(s) and the cumulative asbestos exposure. For mesothelioma age at first exposure is an additional important factor. For lung cancer the synergy between asbestos and cigarette smoking can be important, although only at higher cumulative asbestos exposures than those associated with 9/11 (Fig. 7).
This risk assessment makes two fundamental assumptions about the carcinogenicity of chrysotile asbestos. First, it is assumed, following Hodgson and Darnton,12 that chrysotile is a less potent inducer of mesothelioma and lung cancer than amosite or crocidolite. Second, it is assumed that at low doses there is a linear dose–response. Our approach was to interpolate linearly the increased risk from high cumulative exposures, for which there is a known risk for the asbestos-related cancer, to very low exposure. Epidemiology studies of some workers with low chrysotile exposures found no increased risk of lung cancer even though the workers smoked.14 Indicating the dose–response may be sub linear and this is an additional reason why our estimates are only an upper limit. The average age of Lower Manhattan residents at the time of the exposure was 38 years. Applying the adjustment indicated;12 we calculated the risk of mesothelioma associated with the incremental ambient asbestos exposure as the result of the events of 9/11 to be 1.39 × 10−6.
For a cumulative asbestos exposure of 0.28 f/mL × years, the increment in mesothelioma for the 57,514 residents of Lower Manhattan would be less than 1 case (expected number of cases equals 0.08). The probability of more than one case occurring is less than 0.01. Mesothelioma is a very rare tumor with a lifetime background rate estimated to be 3.6 × 10–4; therefore, in a population of 57,514, the expected number of background mesothelioma cases is 21.15 The 9/11 related increase is less than 1% of the background and cannot be observed using epidemiological methods. If 9/11 caused even one asbestos-related cancer case, it would be indistinguishable among the background cases by any pathologist.
If the exposure was to crocidolite asbestos, the mesothelioma risk would be almost 500-fold higher.12 This justifies our claim that determining asbestos fiber type(s) is important. Environmental exposure to airborne crocidolite and tremolite asbestos have been shown to increase the risk of mesothelioma where mine tailings or local outcrops have been used in the construction of unpaved roads or building materials,16 whereas there is a paucity of epidemiological evidence demonstrating similar occurrences of nonoccupational mesotheliomas in chrysotile mining communities.6 The latter have experienced for the last 100 years much greater cumulative exposures to chrysotile asbestos than in Lower Manhattan after 9/11 without convincing evidence of chrysotile related environmental mesotheliomas occurring.6
Lung cancer risk resulting from asbestos exposure is modeled as an increment relative to the background risk of lung cancer. If the entire population smoked cigarettes, approximately one asbestos-related lung cancer case would be expected. If no one smoked, the risk of lung cancer would be 10-fold lower. Given the smoking rates for the residents of Lower Manhattan of approximately 25%, 1278 background lung cancers cases would be expected (1150 in smokers).17–19 The model projects a relative risk of lung cancer associated with incremental exposure to asbestos from the events of 9/11 equal to 1.7 × 10−4.12 The expected number of lung cancers is 0.22 and the probability of more than one incremental case occurring is approximately 0.02. If an additional 9/11 related case were to occur, it would be indistinguishable among the 1278 background lung cancer cases. At this very low cumulative asbestos exposure, the synergy with smoking is expressed solely as a difference between smokers and non-smokers in the assumed linear risk coefficient (Fig. 7).
On the basis of the results presented above, we conclude that the exposure to asbestos in ambient air after the collapse of the WTC towers has resulted in no more than a negligible increase in the risk of cancer for the residents of Lower Manhattan. The critical underpinnings of this conclusion are (1) assuming that the dust particles sampled were representative, both in space and time of the dust from the collapse; (2) identifying the asbestos fiber-type as chrysotile; (3) expending sufficient resources on air sampling and analysis to produce accurate estimates of airborne asbestos concentrations and establish the return to background following 9/11; and (4) assessing the risks of mesothelioma and lung cancer separately rather than as an aggregate of asbestos-related cancers. Differentiating mesothelioma from lung cancer and chrysotile asbestos from other asbestos fiber-types are both essential for meaningful risk calculations.
EPA's aggregate risk model does not differentiate fiber types and combines mesothelioma and lung cancer. The EPA aggregate model indicates a risk of cancer equal to 2.1 × 10−4, which is equivalent to 12 excess cancers, for the incremental ambient asbestos exposure during and after 9/11 in Lower Manhattan. Camus and coworkers evaluated the two component parts of EPA's aggregate risk model, the model for lung cancer and the model for mesothelioma.20,21 They found that both the lung cancer model and the mesothelioma model substantially overstated risk when compared to actual cases in areas of environmental chrysotile exposure in Canada.
To further demonstrate the importance of differentiating among fiber-types, we applied the mesothelioma model used by EPA to exposures in Lower Manhattan but incorporated a potency factor specific for chrysotile rather than EPA's potency factor that treat all fiber-types alike. The chrysotile potency (Km) factor for mesothelioma developed in research conducted for EPA22 is 4 × 10−10, EPA's all-inclusive fiber-type potency factor (Km) for mesothelioma is 1 × 10−8.23 We projected the number of expected mesothelioma cases using EPA's mesothelioma risk model (their Table 6–323), adjusted to reflect an average continuous exposure of 0.065 f/mL for 1 year, and the population age distribution of Lower Manhattan. With EPA's all-inclusive fiber-type potency factor for mesothelioma, 9.5 mesothelioma cases are expected corresponding to the asbestos exposures resulting from the events of 9/11. With the chrysotile potency factor, 0.4 mesothelioma cases are expected (Table 2
The studies by Camus and coworkers and our analysis of EPA's mesothelioma model described above further support our estimates of less than one expected mesothelioma and less than one expected lung cancer. The difference between EPA estimates and our estimates is the consideration of asbestos fiber-type, which clearly is an important risk factor for mesothelioma. Recent estimates of the relative mesothelioma potencies are 500:100:1 for crocidolite, amosite and chrysotile respectively,12 and 750:1 for amphibole fibers (amosite and crocidolite) versus chrysotile fibers.22 By averaging the mesothelioma risk for the three different asbestos fiber-types the EPA substantially overstates the mesothelioma risk for chrysotile, which is the most common and least potent of the three fiber- types.
This report shows that the risk of developing cancer from asbestos exposures during, and subsequent to, the collapse of the World Trade Center towers is negligible; we make no estimate of the risk from inhaling fine particulate matter.
We thank Drs. Arthur M. Langer, J.B.L. Gee, and E.E. McConnell for very helpful suggestions that led to improvements of this report. The authors would also like to thank Jesse W. Liss for his valuable help in collecting the air samples and Barbara G. Mensch, Robert Essel NYC/Corbis, and John Ricasoli for the photographs and Angelina Nikolova for the graphics.
1. Office of Inspector General. Evaluation Report on EPA's Response to the World Trade Center Collapse: Challenges, Successes and Areas of Improvement. Report No. 2003-P-00012, August 21, 2003. Available from: URL: http://www.epa.gov/oigearth/index.htm
; Internet; accessed April 22, 2005.
2. Reitz WB, Nicholson WJ, Holaday DA, Selikoff IJ. Application of sprayed inorganic fiber containing asbestos: occupational health hazard. Am Indust Hyg Assn J
3. Federal Emergency Management Agency (FEMA) 2002 World Trade Center Building Performance Study Data Collection, Preliminary Observations and Recommendations, FEMA 402/May 2002. Federal Insurance and Mitigation Administration, Washington, DC and FEMA Region II, New York Available from: URL: http://www.fema.gov/library/wtcstudy.shtm
; Internet; accessed April 22, 2005.
4. ISO 10312 International Standard Ambient Air— Determination of Asbestos Fibers—Direct Transfer. Transmission Electron Microscopy Method, 1995.
5. Nolan RP, Langer AM. Concentration and type of asbestos fibers in air inside buildings In: Nolan RP, Langer AM, Ross M, Wicks FJ & Martin RF, eds. The Health Effects of Chrysotile Asbestos: Contribution of Science to Risk-Management Decisions.
Ottawa, Canada: Canadian Mineralogist, Spec Publ 5; 2001:39–51.
6. Shcherbakov SV, Kashansky S, Domnin SG, et al. The health effects of mining and milling chrysotile: The Russian experience. In: Nolan RP, Langer AM, Ross M, Wicks FJ, Martin RF, eds. The Health Effects of Chrysotile Asbestos: Contribution of Science to Risk-Management Decisions.
Ottawa, Canada: Canadian Mineralogist, Spec Publ 5; 2001:187–198.
7. Langer AM, Morse RG. The World Trade Center catastrophe: was the type of spray proofing a factor in the collapse of the towers? Indoor Built Environ
8. World Health Organization (1986) International Programme on Chemical Safety (IPCS), Environmental Health Criteria 203, Chrysotile Asbestos. Available from URL: http://www.who.int/docs/ehc_203.htm
9. NYC Department of Health and US Agency for Toxic Substances and Disease Registry, Final Report of the Public Health Investigations to Assess Potential Exposure to Airborne and Settled Surface Dust in Residential Areas of Lower Manhattan. September, 2002:20.
10. Gibbs GW, DuToit RS. Environmental data in mining. In: Biological Effects of Asbestos.
Geneva: IARC Publication No. 8; 1973:311–320.
11. US-Environmental Protection Agency. Exposure and Human Health Evaluation of Airborne Pollution from the World Trade Center Disaster. External Review Draft, October, 2002.
12. Hodgson JT, Darnton A. The quantitative risk of mesothelioma and lung cancer in relationship to asbestos exposure. Ann Occup Hyg
14. Gibbs GW. Health effect associated with mining and milling chrysotile asbestos in Quebec and the role of tremolite. In: Nolan RP, Langer AM, Ross M, Wicks FJ, Martin RF, eds. The Health Effects of Chrysotile Asbestos: Contribution of Science to Risk-Management Decisions.
Ottawa, Canada: Canadian Mineralogist, Spec Publ 5; 2001:165–175.
15. Price B, Ware A. Mesothelioma trends in the United States: an update based on SEER data for 1973 through 2003. Am J Epidemiol
16. Browne K, Wagner JC. Environmental exposure to amphibole-asbestos and mesothelioma. In: Nolan RP, Langer AM, Ross M, Wicks FJ, Martin RF, eds. The Health Effects of Chrysotile Asbestos: Contribution of Science to Risk-Management Decisions.
Ottawa, Canada: Canadian Mineralogist, Spec Publ 5; 2001: 21–28.
17. Doll R, Peto R. The causes of cancer: quantitative estimates of avoidable risks of cancer in the United States today. J Natl Cancer Inst
18. Zang EA, Wynder EL. Smoking trends in the United States between 1969 and 1995 based on patients hospitalized with non-smoking related disease. Prev Med
19. Wilson R, Crouch EAC. Risk-benefit analysis. Cambridge, MA: Harvard University Press; 2001:230–231.
20. Camus M, Siemtiatycki J, Case BW, Desy M, Richardson L, Campbell S. Risk of mesothelioma among women living near chrysotile mines versus US EPA Asbestos Risk Model: Preliminary Findings. Ann Occup Hyg
. 2002;46: 95–98.
21. Camus M, Siemiatycki J, Meek B. Nonoccupational exposure to chrysotile asbestos and the risk of lung cancer. N Engl J Med
22. Final Draft: Technical Support Document for a Protocol to Assess Asbestos-Related Risk. EPA 9345.4-06, October 2003.
Copyright © 2005 by the American College of Occupational and Environmental Medicine
23. Airborne Asbestos Health Assessment Update 1986. EPA/600/8-84/003F