Identifying subgroups of the general population that may be susceptible to short-term increases in particulate air pollution: a time-series study in Montreal, Quebec

Abstract

This study was undertaken in order to shed light on which groups of the general population may be susceptible to the effects of ambient particles. The objectives of the study were (1) to determine whether concentrations of particles in the ambient air of Montreal, Quebec, were associated with daily all-cause and cause-specific mortality in the period 1984 to 1993, and (2) to determine whether groups of the population had higher than average risks of death from exposure to particles. From the network of fixed-site air pollution monitors in Montreal we obtained daily mean levels of various measures of particles, gaseous pollutants, and weather variables measured at Dorval International Airport. We also used measurements of sulfate from an acid rain monitoring station 150 km southeast of the city (Sutton, Quebec). We estimated associations for particulate matter (PM) with an aerodynamic diameter of 10 microns or smaller (PM10), or 2.5 microns or smaller (PM2.5), total suspended particles (TSP), coefficient of haze (COH), an extinction coefficient, and sulfate. Because substantial data for fine particles were missing, we developed a regression model to predict PM2.5 and to predict sulfate from PM2.5. In the main body of the report, we present results for COH, predicted PM2.5, and sulfate. Detailed results for all pollutants are included in Appendices H through O, which are available on request from Health Effects Institute and from the HEI web site at www.healtheffects.org. To address the first objective, we made use of the underlying causes of death among all 140,939 residents of Montreal who died between 1984 and 1993. We regressed the logarithm of daily counts of cause-specific mortality on the daily mean levels for a variety of measures of particles, accounting for seasonal and subseasonal fluctuations in the mortality time series, overdispersion, and weather factors. To address the second objective, we developed algorithms to define conditions that subjects had prior to death, with the focus on cardiopulmonary diseases. These algorithms were based on information retained on the databases of the universal Quebec Health Insurance Plan (QHIP). The databases include records of all procedures (e.g., type of surgery), physician visits, and consultations carried out by all physicians in Quebec. For persons > or = 65 years and for all recipients of social assistance the prescription database contains records of all pharmaceuticals dispensed (type of medication, dose, quantity). For each group of conditions defined, we used the same statistical model that was used in the analyses of all nonaccidental causes of death. In the analyses of cause-specific mortality, we found evidence of associations for all nonaccidental causes of death and specific causes of death--cancer, coronary artery disease, respiratory diseases, and diabetes--that were consistent across most metrics of ambient air particle concentrations, evaluated as the 3-day mean of particle concentrations measured on the day of death (lag 0) and on each of the two days before death (lag 1, lag 2). Associations for all cardiovascular diseases combined were found only with sulfate. As well, we generally found increased daily mortality for persons 65 years of age and over. The results for all nonaccidental causes of death are similar to findings from other studies; the mean percent increase in mortality for a 100 micrograms/m3 increase in daily TSP at lag 0 was 6.7%. In the analyses of the groups defined from the QHIP data, there was little evidence of associations with air pollutants among persons who before death were classified as having acute or chronic upper respiratory diseases, airways diseases, hypertension, acute coronary artery diseases, and cerebrovascular diseases. On the other hand, we found consistent increases across most types of ambient particles for persons who had cancer, acute lower respiratory diseases, any form of cardiovascular disease, chronic coronary artery diseases, and congestive heart failure. As well, we found an association for individuals who did not have any cardiovascular disease, lower respiratory diseases, and cancer. This latter group consisted of persons who had no interactions with the health care system one year before death (12%) and individuals with a wide variety of potentially fatal diseases (52%), including neurological conditions (12%), diabetes (8%), cardiac dysrhythmias (8%), dementia (6%), organic psychotic disorders (6%), and anemias (4%). As statistical power was reduced in the analyses presented above, differences between groups (e.g., < 65 and > or = 65 year age groups) were not usually statistically significant. The association with diabetes has not been reported previously, and this needs to be replicated in other studies. (ABSTRACT TRUNCATED)