Respiratory and Cardiovascular Effects of Metals in Ambient Particulate Matter: A Critical Review - Reviews of Environmental Contamination and Toxicology - page 151

New York City (Bell et al. 2009 ; Patel et al. 2009 ; Lippmann et al. 2006 ; and
Dominici et al. 2007 ) were infl uenced by the historical use of residual oil (which is
high in these metals) to heat residential and commercial buildings in those neigh-
borhoods during the time periods covered by the studies. Thus, conclusions regard-
ing associations between Ni and respiratory and cardiovascular health effects in
those studies should be viewed within the unique context of air quality in residential
neighborhoods of Manhattan and the Bronx.
Ambient PM is a complex mixture, and thus, exposure is to multiple agents that
may interact in undefi ned ways. For any study discussed in this review article, the
associations found for any given PM 2.5 -associated metal component and the result-
ing health effect may not wholly refl ect the effects of the single metal in question,
but may represent the effect of another co-pollutant, particularly those that are
unmeasured, that may be part of that particular source (Ostro et al. 2009 ).
The majority of the studies are population-based. Therefore, despite the (generally)
large number of subjects, the infl uence of individual risk factors and exposures that
are related to the same health outcomes, and are sources of many of the same transi-
tion metals, such as tobacco smoke, could not be completely accounted for in the
study designs.
Observations from even a relatively large number of subjects in a specifi c geo-
graphic location may or may not be relevant to other geographic locations or to
other demographic groups. Different mixtures of PM components, weather condi-
tions, and even general population health in different areas may contribute to the
results in unknown ways. For example, the group of studies published by Ostro
et al. focused on a select set of California counties, and the Ostro et al. ( 2008 ) study
looked specifi cally at differences between whites and Hispanics. The study by Patel
focused on children of Dominican Republic and African American heritage living in
Northern Manhattan and the South Bronx. The studies were designed to address
specifi c populations in specifi c geographic locations, and thus caution must be used
in extrapolating the fi ndings from them to the general US population.
Even when appropriate statistical methods were used to control bias introduced
by chance, the very large number of associations tested in some of the studies may
have produced fi ndings that were signifi cant by chance alone. For example, Patel
et al. ( 2009 ) tested over 400 potential associations.
The limitations of the epidemiological studies to defi ne the role of transition met-
als in health effects are primarily related to characterization of exposure on an indi-
vidual and population level and the presence of co-pollutants that may contribute to
the same effects. For the majority of the epidemiological studies that we reviewed,
the levels of exposure to PM and its components were estimated from centrally-
located monitors that may or may not have been representative for the populations
of interest. The indoor and personal environments of the study subjects may have
been additional sources of exposure to particulate and some transition metals.
Furthermore, most study designs were limited in the ability to account for other
personal risk factors that could be associated with morbidity and mortality from
respiratory and cardiovascular causes.