Skip Navigation LinksHome > November 2010 - Volume 21 - Issue 6 > The Color of Smoke
Epidemiology:
doi: 10.1097/EDE.0b013e3181f4e1e6
Air Pollution: Commentary

The Color of Smoke

Brunekreef, Bert

Free Access
Article Outline
Collapse Box

Author Information

From the University Medical Center, University of Utrecht, Utrecht, The Netherlands.

Correspondence: Bert Brunekreef, Institute for Risk Assessment Sciences and Julius Center for Health Sciences and Primary Care, University Medical Center, University of Utrecht, PO Box 80178, 3508 TD, Utrecht, The Netherlands. E-mail: B.BRUNEKREEF@UU.NL.

Need smoke be “black” to constitute an offence?

This question was raised in The Lancet the other day—on 10 October 1903 to be precise.1 The editorial lamented that the Public Health Act of London specified that smoke needed to be “black” before measures could be taken pointing out that there were many examples of brown, yellow, blue, or even white smoke being offensive. The attributes that determine the adverse effects of smoke (particulate air pollution) has been a question for at least a century—and is still unresolved.

Over the years, we have busied ourselves with many different characteristics of airborne particulate matter (PM) in an attempt to find out which one matters most. After coal burning ceased to be a major source, we came to realize that the “blackness” of smoke could no longer be translated into the mass concentration of particles in the air we breathe. We have specified size ranges such as particles smaller than 10 or 2.5 μm to define those that could enter the respiratory tract. We have become interested in particle numbers and surface areas, which might be more important for the biologically-relevant dose of harmful constituents carried on the particles. Above all, we have tried to analyze PM composition to find out which components determine the harmful effects of PM seen in many epidemiologic studies.

At first glance, it seems rather straightforward to make progress in this regard: measure PM composition, then analyze which components have the strongest or least effect on some health outcome, and declare these components to be responsible for the health effect (or for the lack thereof). Alternatively, PM composition data could be subjected to some form of factor analysis to group them into classes that then can be attributed to specific sources.

In the real world of epidemiology, things are not that simple, however. This issue of Epidemiology contains 2 papers that illustrate the complexities involved. One paper examines the association between exposure to PM constituents and birth weight.2 The other assesses airway and systemic inflammation in a panel of elderly subjects in relation to primary and secondary aerosols.3

Bell et al2 used PM filters from 5 monitoring sites in Connecticut and Massachusetts collected over a period of 3.5 years. Filter availability ranged from only 30% up to 92% of days. Filters were analyzed for 51 elements by x-ray fluorescence, and elemental carbon was estimated by light reflectance—essentially the method used in the old days to measure “black” smoke. Elements were attributed to 5 sources (traffic, oil combustion, road dust, salt, and regional sources)—with most constituents originating from more than 1 source. The study found minute (5-7 g) changes in birth weight associated with interquartile-range changes in exposure to zinc and elemental carbon (as markers of traffic), silicon and aluminum (as markers of road dust), and vanadium and nickel (as markers of oil combustion). The change with an interquartile difference of 3.6 μg/m3 for PM2.5 was 3 g and not statistically significant. This suggests that something was gained by looking at particle composition, and not just particle mass concentration. The correlations among these markers were high to very high (0.53-0.98), effectively ruling out a multivariate approach to shed further light on the relative merits of the individual elements or source categories.

Delfino et al3 measured outdoor particle number concentrations, organic, elemental and black carbon, and a range of 92 organics in the context of a panel study of 60 elderly subjects who contributed up to 12 measurements of fractional exhaled nitric oxide (NO) as a marker of airway inflammation, and plasma IL-6 as a marker of systemic inflammation. The investigators also measured the potential of semi-ultrafine PM (PM0.25) to produce reactive oxygen species, using an in vitro test system. Results were rather different for the 2 markers of inflammation: eNO was found to be positively associated with PM2.5 mass and secondary organics carbon, but negatively with the particle number count; plasma IL-6 was associated with elemental carbon and primary organic carbon. This article is the latest in a series of multiple, partially overlapping articles from the same study.4-6

Comparison of these 2 contributions shows that the suite of PM attributes studied was almost completely different between these studies. Little if any justification was given why elements as measured by x-ray fluorescence and elemental carbon would be of special interest in birth outcome studies, and primary and secondary organic components in studies of biomarkers of inflammation. One wonders what would have been found if the suite of PM attributes had been exchanged between the studies?

There is no lack of recent articles examining associations between PM attributes and one health outcome or another. Examples include Atkinson et al7 who related fine and coarse PM, particle number count, and sulfates, nitrates, carbon and chloride to daily mortality in a time-series study from London. These authors found positive associations of particle number count with cardiovascular deaths, and of sulfates and nitrates with respiratory deaths. A study by Hirshon et al,8 on pediatric asthma emergency-room admissions, found an association of asthma with zinc only, treating other metals as potential confounders. These are just 2 examples of studies looking at yet other combinations of PM attributes.

This literature has not yet been systematically reviewed, and this brief editorial should not be seen as an attempt to do so. A few observations might be useful, though:

1. Studies of the health effects of PM beyond the regulated metrics such as PM2.5 invariably suffer from lack of statistical power. This is because the number of data points for other PM attributes is usually much smaller than for the regulated metrics.

2. The measured attributes of PM vary considerably from study to study, as do the health endpoints. As a result, one may find many unique associations between one or another attribute and one or another health end point—but precious little replication or even harmonization of study designs and analyses.

3. Some studies have tested fairly large numbers of associations with little if any discussion of the perils of multiple testing.

Is this, then, a problem too large for epidemiology to tackle? I don't think so. None of the methodological challenges seem unique to the study of differential health effects of PM metrics. We do seem to be in a situation, however, in which investigators are following many different leads, with many different approaches, which makes a systematic review difficult. Perhaps it is time for those interested in the study of PM attributes (or multipollutant studies more generally) to convene a workshop on methodological principles?

Back to Top | Article Outline

ABOUT THE AUTHOR

BERT BRUNEKREEF is professor of environmental epidemiology and director of the Institute for Risk Assessment Sciences (IRAS) at the Utrecht University in The Netherlands. He has been studying the health effects of pollutants, notably indoor and outdoor air pollution, for the last 30 years. He is currently coordinating a Europe-wide effort to study the long-term effects of outdoor air pollution in some 30 ongoing cohort studies.

Back to Top | Article Outline

REFERENCES

1. Need smoke be “black” to constitute an offence? Lancet. 1903;162:1029.

2. Bell M, Belanger K, Ebisu K, et al. Prenatal exposure to fine particulate matter and birth weight: variations by particulate constituents and sources. Epidemiology. 2010;21:884–891

3. Delfino RJ, Staimer N, Tjoa T, et al. Association of primary and secondary organic aerosols with airway and systemic inflammation in an elderly panel cohort. Epidemiology. 2010;21:892–902

4. Delfino RJ, Staimer N, Tjoa T, et al. Association of biomarkers of systemic inflammation with organic components and source tracers in quasi-ultrafine particles. Environ Health Perspect. 2010;118:756–762.

5. Delfino RJ, Staimer N, Tjoa T, et al. Air pollution exposures and circulating biomarkers of effect in a susceptible population: clues to potential causal component mixtures and mechanisms. Environ Health Perspect. 2009;117:1232–1238.

6. Delfino RJ, Staimer N, Tjoa T, et al. Circulating biomarkers of inflammation, antioxidant activity, and platelet activation are associated with primary combustion aerosols in subjects with coronary artery disease. Environ Health Perspect. 2008;116:898–906.

7. Atkinson RW, Fuller GW, Anderson HR, Harrison RM, Armstrong B. Urban ambient particle metrics and health: a time-series analysis. Epidemiology. 2010;21:501–511.

8. Hirshon JM, Shardell M, Alles S, et al. Elevated ambient air zinc increases paediatric asthma morbidity. Environ Health Perspect. 2008; 116:826–831.

Cited By:

This article has been cited 2 time(s).

Plos One
Composition of PM Affects Acute Vascular Inflammatory and Coagulative Markers - The RAPTES Project
Strak, M; Hoek, G; Godri, KJ; Gosens, I; Mudway, IS; van Oerle, R; Spronk, HMH; Cassee, FR; Lebret, E; Kelly, FJ; Harrison, RM; Brunekreef, B; Steenhof, M; Janssen, NAH
Plos One, 8(3): -.
ARTN e58944
CrossRef
Occupational and Environmental Medicine
Acute nasal pro-inflammatory response to air pollution depends on characteristics other than particle mass concentration or oxidative potential: the RAPTES project
Steenhof, M; Mudway, IS; Gosens, I; Hoek, G; Godri, KJ; Kelly, FJ; Harrison, RM; Pieters, RHH; Cassee, FR; Lebret, E; Brunekreef, BA; Strak, M; Janssen, NAH
Occupational and Environmental Medicine, 70(5): 341-348.
10.1136/oemed-2012-100993
CrossRef
Back to Top | Article Outline

© 2010 Lippincott Williams & Wilkins, Inc.

Twitter  Facebook

Login

Article Tools

Share