© 2006 American Thoracic Society doi: 10.1164/rccm.2601001
A Tale of Six CitiesInstitute for Risk Assessment Sciences, Universiteit Utrecht and Julius Center for Health Sciences and Primary Care, University Medical Center, Utrecht, The Netherlands In this issue of the Journal (pp. 667–672), Laden and colleagues report on the extended follow-up of the Harvard Six Cities cohort study on air pollution and mortality (1). Few environmental issues have been as hotly debated in the last decade as the long-term effects of low concentrations of ambient fine particles on survival. This debate was triggered by the original Harvard cohort study (2). It showed that during a 14-year follow-up, survival was reduced in well-characterized cohorts of adults randomly selected from cities across the eastern part of the United States, in close association with concentrations of ambient particles smaller than 2.5 µm in aerodynamic diameter (PM2.5). Remarkably, this association was observed at annual average PM2.5 concentrations ranging from 10 to not more than 30 µg/m3. I'm sure in due time this episode in air pollution epidemiology will be remembered in Charles Dickens' words: "It was the epoch of belief, it was the epoch of incredulity" (3). As always in epidemiology, important findings require replication, and support from other lines of scientific enquiry. In the years since 1993, a few more cohort studies have addressed the issue of long-term effects of contemporary air pollution on survival—just a few, because long-term cohort studies are difficult to organize and conduct. One of these is the large American Cancer Society II (ACS-II) study. An initial 1995 report showed a significant association between ambient PM2.5 and sulfate, on the one hand, and survival, on the other hand, in subjects living in more than 100 different cities all over the United States (4). These two studies became an important basis for the U.S. Environmental Protection Agency to establish an annual average air quality standard for PM2.5 of just 15 µg/m3 in 1997. Understandably, the studies were strongly attacked by scientists associated with industrial stakeholder organizations, and this prompted a unique and independent reanalysis and validation of the original findings of both studies (5). Since then, an extended follow-up of the ACS-II study has been published, which confirmed the original findings and extended them to a significant association between ambient PM2.5 and lung cancer mortality (6). The extended follow-up of the Harvard Six Cities study published in this issue of the Journal in a sense completes the sequence of publications started in 1993. Some important features of the new study are worth highlighting. One is that it has a much stronger air pollution exposure assessment in the early phase of the study than the ACS-II study ever had. Dedicated measurements of a range of pollutants were conducted over several years. In addition, the populations studied, being from relatively small towns, were living much closer to the monitoring stations used for assessing their exposure. Commentators have always speculated that this is one reason why the effect estimates from the Harvard study are higher than those from the ACS-II study. Interestingly, a recent analysis of exposure contrasts within the Los Angeles ACS-II cohort (7) has demonstrated a much higher effect estimate as well, providing empiric support for the notion that reduced exposure misclassification reduces bias in effect estimates. Another, and clearly the most important, feature of the new study is that it specifically looked at effects of pollution reduction. Over the years of the study, air pollution levels gradually decreased, but they decreased more in the cities that were more heavily polluted. The investigators show that the city-specific reduction of PM2.5 was associated with a reduction in mortality rates. The reason why this is so important is that, until now, it was not clear whether the cohort studies were showing effects that resulted from lifetime cumulative exposures. If so, late changes in exposure would have little, if any, effect on survival. These new findings suggest another dynamic—namely, that recent exposures do matter. This would be consistent with pollution affecting primarily a dynamic "pool" of susceptible individuals whose susceptibility itself may to some extent have been increased by lifelong, cumulative pollution exposure. We do know that smoking cessation leads to reductions in respiratory, cardiovascular, and lung cancer risks, with different lags (8); the findings in this study, which show no effect of pollution reduction on lung cancer and the strongest effects on respiratory and cardiovascular mortality reduction, seem to show a similar pattern. The practical implication is that pollution reduction, even beyond the already relatively low levels that have been achieved in the past half-century, will lead to public health benefits. There are limitations to this study, as is inevitable in these difficult analyses. The study population was relatively small in view of the effect sizes to be detected. Hence, confidence intervals are wide, and effects of clear medical importance (> 50% respiratory mortality reduction) did not reach statistical significance. The investigators were not able to continue their intensive air pollution monitoring efforts beyond the mid 1980s, so that PM2.5 concentrations for the second phase of the study needed to be estimated. Also, subjects were no longer interviewed regularly in the last phase of the study, so that unmeasured changes in variables such as smoking habits may have confounded the findings. Clearly, additional studies are needed. Four studies from Europe have shown that intraurban exposure contrasts to traffic-related pollution mixtures (characterized by nitrogen oxides and soot measurements) are associated with cardiovascular and lung cancer incidence and mortality. Three of these were cohort studies from Oslo (9, 10), the Netherlands (11), and France (12); the fourth was a lung cancer case-control study from Stockholm (13). Whereas these studies have not used direct PM2.5 measurements, there is good evidence that intraurban contrasts in nitrogen oxides and soot in European urban study areas are highly correlated with intraurban contrasts in PM2.5 (14, 15). It therefore seems reasonable to suggest that the European studies provide additional support for mortality effects of urban air pollution mixtures, including PM2.5. Experimental support for these types of epidemiologic findings has been lacking until very recently; but now, the animal experimental equivalent of a cohort study has shown a clear increase in cardiovascular lesions in mice exposed for 6 months to concentrated ambient PM2.5 at a 6-month average concentration of only 15 µg/m3 (16). These are very low concentrations, and much financial and political effort is likely needed to bring down pollution levels even more. Weighing the costs and benefits of pollution control properly is the next challenge. Increasingly, pollution control will have to compete with other expenditures for which our tax euros and dollars are required. Excellent analyses such as the Harvard Six Cities study improve the much-needed knowledge for this, but, to quote Dickens once more, we will also need to enter "the age of wisdom" to build consensus in society on the proper course of action to take. FOOTNOTES Conflict of Interest Statement: B.B. does not have a financial relationship with commercial entity that has an interest in the subject of this manuscript. REFERENCES
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
