Burden of mortality due to ambient fine particulate air pollution ([PM.sub.2.5]) in Interior and Northern BC.

Elliott, Catherine T. ; Copes, Ray

It is widely accepted that air pollution is detrimental to health. The relationship between mortality and short-term exposure to high concentrations of fine particulate matter and sulphur oxides was demonstrated in the Great London Smog in 1952, (1) and subsequently replicated in several natural experiments. (2,3) Recent epidemiologic research suggests that long-term exposure to low-level air pollution may be associated with a greater population health burden than short-term exposure. Prospective longitudinal cohort studies have demonstrated that the risk of mortality from long-term exposure to levels of ambient fine particulate matter ([PM.sub.2.5]) found in Western cities is up to ten times greater than the risk from short-term exposure to the same concentration increment, (4-14) and that life expectancy improves as [PM.sub.2.5] concentrations are reduced. (15)

Nevertheless, measures used to inform public health action and policies in Canada, such as the air quality health index (16) and burden of disease estimates, (17) rely on effect measures for short-term exposure. There is no information about burden of disease associated with long-term exposure to air pollution in rural British Columbia, despite major industrial sources in some communities corresponding to higher pollutant levels than those found in urban areas. Our analysis uses existing data to estimate the burden of mortality from long-term exposure to fine particulate air pollution in rural BC.

METHODS

Setting, population and exposure assessment

Interior Health Authority and Northern Health Authority are rural health authorities which make up the majority of the geographic area of British Columbia, but only contain 16% (n=733,285) and 6% (285,493) of the population, respectively. (18) The major contributors to fine particulate air pollution vary by community and include traffic, industries, forest fires and wood-burning stoves.

Mortality data for 2005 to 2009 for each local health area (LHA) were obtained from BC Vital Statistics Agency. Outdoor ambient [PM.sup.2.5] concentrations ([[PM.sub.2.5]]) were measured by the BC Ministry of Environment using tapered element oscillating microbalance (TEOM) continuous monitors, corrected to compensate for losses of volatile material. Monitors closest to the main populations within each monitored LHA were applied to the population of that LHA. We estimated the [[PM.sub.2.5]] for unmonitored LHAs as the median of monitored communities in the corresponding health authority (HA) and conducted sensitivity analyses for this estimation.

Calculation of burden of mortality

Since the relationship between [[PM.sub.2.5]] and mortality is widely accepted to be log-linear over the range of concentrations observed in BC, (19) the mortality attributable to each increment of [[PM.sub.2.5]] can be estimated using concentration response functions derived from longitudinal cohort studies. The largest such study using data collected by the American Cancer Society (ACS) followed 300,000 participants in 51 cities over 25 years (13) and modeled relative risk estimates that were robust to analytic scrutiny. (10)

Biological plausibility for these effects has been demonstrated through pathophysiologic mechanisms including oxidative stress and inflammation in the lungs and leading to accelerated atherosclerosis and altered autonomic activity in the heart. (19) Other air pollutants have been associated with mortality, however none have the weight of evidence that exists for [PM.sub.2.5].

We used the relative risk for all-cause mortality from the ACS study (1.06 per 10 ug/[m.sup.3]) and locally measured 5-year mean [[PM.sup.2.5]] to calculate a relative risk for each local health area in the Interior and Northern Health Authorities (equation 1). The attributable fraction--the proportion of mortality from all causes that is due to air pollution--is derived from this relative risk (equation 2). For ambient air pollution, the attributable fraction and the population attributable fraction are identical since the entire population is exposed. The burden of mortality attributable to long-term exposure to [PM.sub.2.5] is the product of the attributable fraction and measured all-cause mortality.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] [equation 1]

Where [beta] is the regression coefficient, derived from RR ([beta]= [(log RR)/10]). C is the five-year mean observed [[PM.sub.2.5]] and Co is the counterfactual concentration, the concentration below which no mortality effects are assumed.

In order to calculate the attributable fraction (AF), RR' was used in the classic attributable risk calculation: (20)

AF = [1-(1/RR')] [equation 2]

Attributable mortality (AM) is the product of the attributable fraction and the five-year mean all-cause mortality.

[FIGURE 1 OMITTED]

AM = AF x (all-cause mortality) [equation 3]

We used mortality for adults 30 and older, since this corresponds to the population in the ACS study. Confidence intervals were calculated using the method of Greenland et al. (1987). (21)

Using this method, mortality is estimated only for [PM.sub.2.5] concentrations above a reference concentration that is set a priori (counterfactual concentration, [C.sub.o]). We set [C.sub.o] to the lowest concentration measured in the ACS (5 ug/[m.sup.3]) for the base case scenario, since the effects of [PM.sub.2.5] have not been studied below this concentration.

Anthropogenic Burden of Mortality

In order to estimate mortality burden due to pollution (i.e., [[PM.sub.2.5]] above natural levels), we set the counterfactual concentration to that found in the absence of emissions from human sources, by using the lowest observed concentration in BC (3.1 ug/[m.sup.3], in Terrace). This is similar to the minimum [[PM.sub.2.5]] used in the WHO burden of disease estimates (3 ug/[m.sup.3]) and the minimum background [[PM.sub.2.5]] observed in the United States. (22, p.1401)

We conducted a sensitivity analysis for the concentration estimated for unmonitored communities, by setting their concentration to background levels (3.1 ug/[m.sup.3]). This results in no mortality contribution from these communities.

RESULTS

The mean annual [[PM.sub.2.5]] in Interior and Northern BC LHAs ranged from 3.1 to 7.4 ug/[m.sup.3], with higher concentrations in cities with forest processing, particularly those in valleys where winter inversions trap pollutants (e.g., Prince George, Quesnel; Table 1). Most of the adult population (60%) lived in monitored LHAs.

In our base case, we estimate that fine particulate air pollution causes 0.1% (8 deaths/year) and 0.4% (6 deaths/year) of all-cause mortality among adults in Interior and Northern Health, respectively. For the entire region, an estimated 0.20% (16 deaths/year) of mortality was caused by fine particulate air pollution. The anthropogenic burden of mortality was much higher at 0.93% (74 deaths/year) for the region (Table 2). This demonstrates the sensitivity of estimates to the counterfactual concentration, which was set to 5.0 ug/[m.sup.3] in the base case and 3.1 ug/[m.sup.3] for the anthropogenic burden of mortality.

DISCUSSION

We estimate that 0.20% of mortality among adults in Interior and Northern Health Authorities is due to fine particulate air pollution, corresponding to 8 and 6 deaths each year in each health authority respectively.

The attributable fraction for [PM.sub.2.5] ranged from 0.20 to 0.93% depending on the case scenario. Our base case is lower than the World Health Organization global burden of disease estimates for mortality fraction attributable to urban [PM.sub.2.5] pollution for this region of North America (0.42%), (22) due to low pollution levels in BC relative to other regions of North America.

Our estimates are very sensitive to the counterfactual concentration--the [PM.sub.2.5] concentration above which we attribute mortality. This method assumes a threshold concentration below which no mortality effects occur, set to 5 ug/[m.sup.3] in the base case. The threshold model is debated among scientists. Mortality effects of ambient [PM.sub.2.5] have been demonstrated down to the lowest measured levels (~5 ug/[m.sup.3]) and there is biological plausibility that they occur at even lower levels. (19) Furthermore, the relationship between concentration and mortality is non-linear, with incremental effects greater at lower concentrations. (23) Therefore the true fraction of mortality due to air pollution is likely closer to our estimates for all anthropogenic [PM.sub.2.5], which is 0.93%.

Our estimates fall within a credible range relative to other attributable causes of mortality. The estimates of [PM.sub.2.5] attributable mortality (0.20 to 0.93%) are less than global estimates for population attributable fraction (PAF) for mortality due to second-hand smoke exposure (1%)24 and are a fraction of the population attributable fraction for mortality due to smoking estimated for Northern and Interior BC (19%). (25)

This type of analysis has inherent uncertainties. The major limitations of the methodology relate to: the exposure assessment and the appropriateness of the effect estimate for the target population.

Exposure assessment is restricted because 1) [[PM.sub.2.5]] from one central monitor was applied to the population of each LHA and 2) [[PM.sub.2.5]] was not measured in all LHAs. This method does not account for individual variation in exposure due, for example, to spatial variations in [PM.sub.2.5], penetration indoors, and individual time spent outdoors. It is, however, comparable to the exposure assessment used in the cohort studies where the effect estimates were derived. (26) Since monitors were selected for proximity to major population centres, it would underestimate exposures for those living near pollution hotspots and overestimate exposures for those living outside the city. The overall effect on the estimates is uncertain. [[PM.sub.2.5]] estimates for unmonitored LHAs have a large influence on attributable fraction, and since 40% of the population lives in unmonitored regions, this is a major limitation. [[PM.sub.2.5]] in monitored sites may not be representative of those in unmonitored ones since, for example, monitors may be deliberately sited near sources of emissions. [PM.sub.2.5] measurement by TEOM underestimates [[PM.sub.2.5]] at lower temperatures, causing a conservative bias in our estimates. Our exposure estimates are for current exposure, however, the current burden is due to past exposure which was higher in this BC region. However, a method to incorporate past exposures into burden of disease estimations has not been established. (22)

Effect estimates are never completely transferable from one population to another; however, suitability can be assessed based on 1) [[PM.sub.2.5]] and 2) population. Differences between the physicochemical properties of [PM.sub.2.5] between the 51 cities in the ACS and these BC sites are uncertain. Furthermore, even if these differences were quantified, a method to apply this information to burden of disease estimates has not been developed. Our sites included small cities and rural towns, where the main pollution sources include forest-related industry, forest fires, wood-burning stoves, and traffic. Although ACS included smaller cities with similar pollution profiles, the ideal effect estimate for our estimations would be derived from a cohort study with emissions profiles and population characteristics better matched to BC; however, these studies have not occurred. The age range of the BC population was matched to that of the ACS study and the populations had similar smoking rates (21-26% vs. 22%); however, population susceptibility to air pollution may differ between these groups. Therefore the ACS effect estimate is the most suitable, albeit imperfect, effect estimate available. It is important to note that effect estimates are remarkably similar across cohort studies across different populations in North America, (11,13,27,28) therefore the contribution of these uncertainties is likely to be small.

In conclusion, the findings of this study demonstrate that fine particulate air pollution does have an important mortality burden in a region of Canada with relatively low pollution levels. The magnitude of health impact is a critical piece of information that can be considered along with other factors (e.g., preventability, public acceptability of interventions) when assessing public health priorities.

Conflict of Interest: None to declare.

Received: April 9, 2010

Accepted: May 11, 2011

REFERENCES

(1.) Bell ML, Davis DL. Reassessment of the lethal London fog of 1952: Novel indicators of acute and chronic consequences of acute exposure to air pollution. Environ Health Perspect 2001;109(Suppl 3):389-94.

(2.) Pope CA, III, Muhlestein JB, May HT, Renlund DG, Anderson JL, Horne BD. Ischemic heart disease events triggered by short-term exposure to fine particulate air pollution. Circulation 2006;114(23):2443-48.

(3.) Kwon HJ, Cho SH, Chun Y, Lagarde F, Pershagen G. Effects of the Asian dust events on daily mortality in Seoul, Korea. Environ Res 2002;90(1):1-5.

(4.) Committee on the Medical Effects of Air Pollutants. Cardiovascular disease and air pollution. A report by the Committee on the Medical Effects of Air Pollutants. London, UK: Department of Health, 2006. Available at: http://www.advisorybodies.doh.gov.uk/comeap/statementsreports/ CardioDisease.pdf (Accessed July 22, 2011).

(5.) Klemm RJ, Mason R. Replication of reanalysis of Harvard Six-City mortality study. Revised analyses of time-series studies of air pollution and health, Part II. Boston, MA: Health Effects Institute, 2003;165-72.

(6.) Dominici F, Daniels M, McDermott A, Zeger SL, Samet J. Shape of the exposure-response relation and mortality displacement in the NMAPS database. Revised analyses of time-series studies of air pollution and health, Part II. Boston: Health Effects Institute, 2003;91-96.

(7.) Katsouyanni K, Touloumi G, Samolu E, Petasakis Y, Analitis A, Le Tertre A, et al. Sensitivity analysis of various models of short-term effects of ambient particles on total mortality in 29 cities in APHEA2. Revised analyses of time-series studies of air pollution and health, Part II. Boston: Health Effects Institute, 2003;157-64.

(8.) Ostro B, Broadwin R, Green S, Feng WY, Lipsett M. Fine particulate air pollution and mortality in nine California counties: Results from CALFINE. Environ Health Perspect 2006;114(1):29-33.

(9.) Dockery D, Pope C, Xu X, Spengler J, Ware J, Fay M, et al. An association between air pollution and mortality in six U.S. cities. N Engl J Med 1993;329:1753-59.

(10.) Krewski D, Burnett RT, Goldberg MS, Hoover K, Siemiatycki J, Jarret M, et al. Reanalysis of the Harvard Six Cities Study and the American Cancer Society Study of Particulate Air Pollution and Mortality. Boston: Health Effects Institute, 2000. Available at: http://pubs.healtheffects.org/view.php?id=6 (Accessed August 16, 2011).

(11.) Laden F, Schwartz J, Speizer F, Dockery D. Reduction in fine particulate air pollution and mortality: Extended follow-up of the Harvard Six Cities study. Am J Respir Crit Care Med 2006;173:667-72.

(12.) Pope C, Thun M, Namboodiri M, Dockery D, Evans J, Speizer F, et al. Particulate air pollution as a predictor of mortality in a prospective study of US adults. Am J Respir Crit Care Med 1995;151:669-74.

(13.) Pope C, Burnett R, Thun M, Calle E, Krewski D, Ito K, et al. Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. JAMA 2002;287:1132-41.

(14.) Pope C, Burnett R, Thurston G, Thun M, Calle E, Krewski D, et al. Cardiovascular mortality and long-term exposure to particulate air pollution: Epidemiological evidence of general pathophysiological pathways of disease. Circulation 2004;109:71-77.

(15.) Pope CA, III, Ezzati M, Dockery DW. Fine-particulate air pollution and life expectancy in the United States. N Engl J Med 2009;360(4):376-86.

(16.) Stieb DM, Burnett RT, Smith-Doiron M, Brion O, Shin HH, Economou V. A new multipollutant, no-threshold air quality health index based on short-term associations observed in daily time-series analyses. J Air Waste Manag Assoc 2008;58(3):435-50.

(17.) British Columbia Provincial Health Officer. Every breath you take. Provincial Health Officer's annual report, 2003. Air quality in British Columbia, a public health perspective. Victoria, BC: Ministry of Health Services, 2004. Available at: http://www.health.gov.bc.ca/pho/pdf/phoannual2003pres.pdf (Accessed July 22, 2011).

(18.) British Columbia Vital Statistics Agency. Regional population estimates and projections. Victoria, BC: Ministry of Health Services, 2009. Available at: http://www.bcstats.gov.bc.ca/data/pop/pop/dynamic/PopulationStatistics/ Query.asp?category=Health&type=HY&topic=Estimate s&agegrouptype=Custom (Accessed February 11, 2011).

(19.) Pope C, Dockery D. Health effects of fine particulate air pollution: Lines that connect. J Air Waste Manage 2006;56(6):709-42.

(20.) Steenland K, Armstrong B. An overview of methods for calculating the burden of disease due to specific risk factors. Epidemiol 2006;17(5):512-19.

(21.) Greenland S. Variance estimators for attributable fraction estimates consistent in both large strata and sparse data. Stat Med 1987;6(6):701-8.

(22.) Cohen AJ, Ross Anderson H, Ostro B, Pandey KD, Krzyzanowski M, Kunzli N, et al. Chapter 17. Urban air pollution. In: Ezzati M, Lopez AD, Rodgers A, Murray CJL (Eds.), Comparative Quantification of Health Risks: Global and Regional Burden of Disease Attribution to Selected Major Risk Factors, Volume 2, Part 2: Environmental and occupational risk factors. Copenhagen, Denmark: World Health Organization Regional Office for Europe, 2004;1353-433. Available at: http://www.who.int/publications/cra/chapters/volume2/part2/en/index.html (Accessed July 22, 2011).

(23.) Smith KR, Peel JL. Mind the gap. Environ Health Perspect 2010;118(12):1643-45.

(24.) Oberg M, Jaakkola MS, Woodward A, Peruga A, Pruss-Ustun A. Worldwide burden of disease from exposure to second-hand smoke: A retrospective analysis of data from 192 countries. Lancet 2011;377(9760):139-46.

(25.) British Columbia Vital Statistics Agency. Appendix 3: Selected health status indicators by local health area, health service delivery area and health authority. British Columbia, 2005-2009. Selected vital statistics and health status indicators (One hundred thirty-eighth annual report). Victoria: Ministry of Health Services, 2009. Available at: http://www.vs.gov.bc.ca/stats/annual/2009/pdf/app309.pdf (Accessed July 22, 2011).

(26.) Krzyzanowski M, Cohen A. Quantification of the health effects of exposure to air pollution. Report of a WHO Working Group. Bilthoven, Netherlands: World Health Organisation European Centre for Environment and Health, 2001. Available at: http://www.euro.who.int/_data/assets/pdf_file/0011/112160/E74256.pdf (Accessed July 22, 2011).

(27.) McDonnell WF, Nishino-Ishikawa NAOMI, Petersen FF, Chen LH, Abbey DE. Relationships of mortality with the fine and coarse fractions of long-term ambient PM10 concentrations in nonsmokers. J Expo Anal Environ Epidemiol 2000;10(5):427.

(28.) Lipfert F, Baty J, Miller J, Wyzga R. [PM.sub.2.5] constituents and related air quality variables as predictors of survival in a cohort of U.S. military veterans. Inhal Toxicol 2006;18(9):645-57.

Catherine T. Elliott, MD, MHSc, FRCPC, [1] Ray Copes, MD, MSc [2]

Author Affiliations

[1.] Environmental Health Services, BC Centre for Disease Control, Vancouver, BC

[2.] Environmental and Occupational Health, Ontario Agency for Health Protection and Promotion, Toronto, ON; Dalla Lana School of Public Health, University of Toronto, Toronto, ON; School of Population and Public Health, University of British Columbia, Vancouver, BC

Correspondence: Dr. Catherine Elliott, Environmental Health Services, BC Centre for Disease Control, Main Floor, 655 12th Ave W, Vancouver, BC V5Z 4R4, E-mail: doctor.elliott@gmail.com

Table 1. Population and Annual Mean [PM.sub.2.5] Concentration
in Interior and Northern BC, 2005-2009

                              Population [greater
                              than or equal to] 30    Annual Mean
                                   Years Old         [[PM.sub.2.5]]
Local Health Area                   Mean (%)

Golden                              4448 (0.94)            6.7
Grand Forks                         6335 (1.3)             6.9
Kamloops                          68,818 (15)              5.1
Central Okanagan                 113,888 (24)              4.9
Nelson                            16,115 (3.4)             4.4
Southern Okanagan                 14,447 (3.1)             4.5
Vernon                            42,611 (9.0)             5.7
Cariboo-Chilcotin                 16,368 (3.5)             6.3
Unmonitored Interior Health      189,361 (40)              n/a
Burns Lake                          4637 (2.8)             4.6
Fort Nelson                         3309 (2.0)             3.4
Kitimat                             6570 (3.9)             3.6
Prince George                     57,311 (34)              7.2
Quesnel                           14,832 (8.9)             7.0
Terrace                           12,067 (7.2)             3.1
Unmonitored Northern Health       68,845 (41)              n/a

Table 2. Estimates of Attributable Fraction and Annual Mortality
Due to Air Pollution in Interior and Northern BC Under Different
Scenarios (2005-2009)

                          All Local Health Authorities

                                                Annual
                          Attributable      Mortality Count
Case                      Fraction (%)         (95% CI)

Base case              0.20 (0.19, 0.21)      16 (15, 16)
Anthropogenic burden   0.93 (0.89, 0.97)      74 (71, 77)
  of mortality

                       Monitored Loca1 Health Authorities

                                                Annual
                          Attributable      Mortality Count
Case                      Fraction (%)         (95% CI)

Base case              0.20 (0.20, 0.21)      10 (15, 17)
Anthropogenic burden   0.84 (0.80, 0.87)      43 (41, 44)
  of mortality