Validation of the Harvard Six Cities Study of Particulate Air Pollution and Mortality
http://www.100md.com
《新英格兰医药杂志》
To the Editor: In a 16-year prospective cohort study of six cities in the northeastern and midwestern United States beginning in the 1970s, Dockery et al.1 reported that long-term exposure to ambient fine particles (with an aerodynamic diameter of 2.5 μm) was positively associated with both overall mortality and mortality from cardiopulmonary causes. Because the Harvard Six Cities Study played a pivotal and controversial role2,3,4,5,6 in the establishment of the current U.S. objective for ambient air quality with respect to fine particles,7 the Environmental Protection Agency, industry, and nongovernmental organizations called for an independent reanalysis of this study to validate the original findings. This reanalysis involved a detailed statistical audit of a random sample of 250 questionnaires and 250 death certificates and the replication of the original results with the use of the same data and analytic methods used by the original investigators.8
The audit identified very few discrepancies between the original questionnaires and death certificates in the audit sample and the analytic file used by the original investigators; the identified discrepancies included differences in the occupational exposure to dust (14 discrepancies among 249 questionnaires that were located for evaluation) and fumes (15 discrepancies among the 249 questionnaires), the date of death (2 discrepancies among 250 questionnaires), and the cause of death (2 discrepancies among 250 questionnaires). The discrepancies in the date of death involved a maximal period of one year; discrepancies in the cause of death did not result in misclassification within the three main causes of death considered (any cause, cardiopulmonary causes, and lung cancer). Although the audit uncovered a programming problem that resulted in early censorship in five cities and the loss of 1 percent of the reported person-years of follow-up, the inclusion of these missing person-years did not appreciably alter the original estimates of risk. Otherwise, we were able to reproduce virtually all of the original numerical results, including the 26 percent difference in overall mortality between the most polluted city (Stubenville, Ohio) and the least polluted city (Portage, Wis.). A less conservative method of calculating confidence intervals for mortality-rate ratios associated with tobacco consumption resulted in narrower 95 percent confidence limits: for example, we obtained a 95 percent confidence interval of 1.20 to 1.61 for overall mortality among former smokers as compared with the interval of 1.10 to 1.75 given in the original analysis.
Attempts to reproduce the original air-pollution data, for which intact records no longer existed, resulted in some notable discrepancies in the total levels of suspended particulates and the sulfur dioxide levels. However, the discrepancies noted during the audit were not of major epidemiologic importance and did not substantively alter the original risk estimates associated with particulate air pollution nor the main conclusions that were reached.
Daniel Krewski, Ph.D., M.H.A.
University of Ottawa
Ottawa, ON K1N 6N5, Canada
Richard T. Burnett, Ph.D.
Health Canada
Ottawa, ON K1A 0K9, Canada
Mark S. Goldberg, Ph.D.
McGill University
Montreal, QC H3A 2T5, Canada
Kristin Hoover, M.A.
Hoover Consultants
West Chester, PA 19382-6816
Jack Siemiatycki, Ph.D.
Michael Abrahamowicz, Ph.D.
McGill University
Montreal, QC H3A 2T5, Canada
Warren H. White, Ph.D.
Washington University
St. Louis, MO 63110
References
Dockery DW, Pope CA III, Xu X, et al. An association between air pollution and mortality in six U.S. cities. N Engl J Med 1993;329:1753-1759.
Greenbaum DS, Bachmann JD, Krewski D, Samet JM, White R, Wyzga RE. Particulate air pollution standards and morbidity and mortality: case study. Am J Epidemiol 2001;154:Suppl:S78-S90.
National Research Council. Research priorities for airborne particulate matter. Vol. 1. Immediate priorities and a long-term research portfolio. Washington, D.C.: National Academy Press, 1998.
Gamble JF. PM2.5 and mortality in long-term prospective cohort studies: cause-effect or statistical associations? Environ Health Perspect 1998;106:535-549.
Vedal S. Ambient particles and health: lines that divide. J Air Waste Manag Assoc 1997;47:551-581.
Moolgavkar SH. Air pollution and mortality. N Engl J Med 1994;330:1237-1238.
Environmental Protection Agency. Air quality criteria for particulate matter. Washington, D.C.: Government Printing Office, 1996. (EPA/600/P-95/001cF.)
Reanalysis of the Harvard Six Cities Study and the American Cancer Society Study of Particulate Air Pollution and Mortality: a special report of the institute's Particle Epidemiology Reanalysis Project. Cambridge, Mass.: Health Effects Institute, 2000.
The audit identified very few discrepancies between the original questionnaires and death certificates in the audit sample and the analytic file used by the original investigators; the identified discrepancies included differences in the occupational exposure to dust (14 discrepancies among 249 questionnaires that were located for evaluation) and fumes (15 discrepancies among the 249 questionnaires), the date of death (2 discrepancies among 250 questionnaires), and the cause of death (2 discrepancies among 250 questionnaires). The discrepancies in the date of death involved a maximal period of one year; discrepancies in the cause of death did not result in misclassification within the three main causes of death considered (any cause, cardiopulmonary causes, and lung cancer). Although the audit uncovered a programming problem that resulted in early censorship in five cities and the loss of 1 percent of the reported person-years of follow-up, the inclusion of these missing person-years did not appreciably alter the original estimates of risk. Otherwise, we were able to reproduce virtually all of the original numerical results, including the 26 percent difference in overall mortality between the most polluted city (Stubenville, Ohio) and the least polluted city (Portage, Wis.). A less conservative method of calculating confidence intervals for mortality-rate ratios associated with tobacco consumption resulted in narrower 95 percent confidence limits: for example, we obtained a 95 percent confidence interval of 1.20 to 1.61 for overall mortality among former smokers as compared with the interval of 1.10 to 1.75 given in the original analysis.
Attempts to reproduce the original air-pollution data, for which intact records no longer existed, resulted in some notable discrepancies in the total levels of suspended particulates and the sulfur dioxide levels. However, the discrepancies noted during the audit were not of major epidemiologic importance and did not substantively alter the original risk estimates associated with particulate air pollution nor the main conclusions that were reached.
Daniel Krewski, Ph.D., M.H.A.
University of Ottawa
Ottawa, ON K1N 6N5, Canada
Richard T. Burnett, Ph.D.
Health Canada
Ottawa, ON K1A 0K9, Canada
Mark S. Goldberg, Ph.D.
McGill University
Montreal, QC H3A 2T5, Canada
Kristin Hoover, M.A.
Hoover Consultants
West Chester, PA 19382-6816
Jack Siemiatycki, Ph.D.
Michael Abrahamowicz, Ph.D.
McGill University
Montreal, QC H3A 2T5, Canada
Warren H. White, Ph.D.
Washington University
St. Louis, MO 63110
References
Dockery DW, Pope CA III, Xu X, et al. An association between air pollution and mortality in six U.S. cities. N Engl J Med 1993;329:1753-1759.
Greenbaum DS, Bachmann JD, Krewski D, Samet JM, White R, Wyzga RE. Particulate air pollution standards and morbidity and mortality: case study. Am J Epidemiol 2001;154:Suppl:S78-S90.
National Research Council. Research priorities for airborne particulate matter. Vol. 1. Immediate priorities and a long-term research portfolio. Washington, D.C.: National Academy Press, 1998.
Gamble JF. PM2.5 and mortality in long-term prospective cohort studies: cause-effect or statistical associations? Environ Health Perspect 1998;106:535-549.
Vedal S. Ambient particles and health: lines that divide. J Air Waste Manag Assoc 1997;47:551-581.
Moolgavkar SH. Air pollution and mortality. N Engl J Med 1994;330:1237-1238.
Environmental Protection Agency. Air quality criteria for particulate matter. Washington, D.C.: Government Printing Office, 1996. (EPA/600/P-95/001cF.)
Reanalysis of the Harvard Six Cities Study and the American Cancer Society Study of Particulate Air Pollution and Mortality: a special report of the institute's Particle Epidemiology Reanalysis Project. Cambridge, Mass.: Health Effects Institute, 2000.