Authors' response: Traffic-related air pollution and acute myocardial infarction
Dear Editor,
We are grateful for the thoughtful comments from Mills and Newby on
our study investigating the short-term associations between pollution
levels and risk of myocardial infarction (MI).[1] We agree that our data
were consistent with some prothrombotic effect of air pollution, but we
concluded from the lack of any apparent net increase in short-term MI risk
associated with pollution increases that other, perhaps non-thrombotic,
mechanisms may be more important drivers of the pollution-associated net
mortality increases that have been observed previously.[2]
The various valid points that Mills and Newby raise do not seem to
challenge this basic interpretation. They point out that the harvesting
phenomenon that we observed is consistent with an effect of traffic-
related pollution on coronary thrombosis, and we agree: were there no such
effect, we would not have expected to see any association between
pollution levels and MI risk at all. A number of experimental studies also
point to plausible pathways through which pollution might trigger MI.[3]
Our findings were consistent with such a triggering effect, as we observed
a transiently increased risk of MI for up to 6 hours after exposure. This
is also consistent with the very short-term and transient observed effects
of diesel exposure on ST-segment changes in the correspondents' own
experimental work.[4] However our data did not support an effect of
pollution on the overall MI risk over a 72-hour lag period; any pollution-
associated triggering of MIs appeared to result only in short-term
displacement (the bringing forward of MI events by a few hours or days).
In contrast with our observations for MI, short-term increases in broader
mortality outcomes after pollution peaks do not appear to be explained by
short-term displacement alone,[5] hence our conclusion that other
mechanisms may be more important drivers of net mortality increases.
A number of important limitations of our data are raised by Mills and
Newby. As we discuss at some length in the manuscript, we used data from
central pollution monitors to capture pollution levels across
conurbations. Unpublished data from our group suggest that correlations
between pollution levels measured at different sites within a conurbation
remain high at several kilometres distance; nevertheless, this strategy is
likely to have led to some degree of measurement error, and may have
resulted in a bias towards the null in single pollutant models. Hales and
Edwards point out in their editorial accompanying our article that this
factor could explain a missed or underestimated effect, but not the
temporal (harvesting) pattern that we observed.[6] In large-scale studies
of this type using routinely collected data, there is inevitably a trade-
off between study size and proximity of monitoring sites; restricting to
individuals living very close to a monitor would result in drastically
reduced study power. Predicted small-area pollution levels based on
spatial models may offer a solution to this dilemma in future studies, by
allowing a more personalised exposure without restrictive geographical
inclusion criteria; however, such modelled data would require careful
validation.
The MINAP dataset, which holds detailed timing data on acute coronary
events, offered an unprecedented opportunity to study associations at an
hourly temporal resolution in a large scale study. As Mills and Newby
point out, one disadvantage is that MIs resulting in death before hospital
admission are not recorded; MIs resulting in death at any time following
hospital admission would of course have been included. We find it
intuitively unlikely that pollution would be associated with a net
increase only in those MIs that resulted in death before hospital
admission, and yet no net increase at all in the remainder of MIs that
were admitted to hospital. However, we cannot exclude the possibility that
high levels of pollutants could exacerbate the severity of an MI and thus
the probability of early death by amplifying the thrombotic process. We
would therefore welcome future research clarifying this issue, using
alternative data sources which incorporate out-of-hospital deaths.
Finally, our study did not attempt to assess the impact of chronic
exposure to air pollution on MI risk. Whilst it is valid to consider acute
effects as a separate phenomenon, pollution policy must also be based on
an assessment of long-term risks. Unfortunately, the evidence base
regarding the long-term effects of air pollution exposure on MI risk
specifically is limited; only a handful of studies have been carried out,
and potential confounding by factors such as socioeconomic status and
occupation hampers many study designs.[7] The development of methods and
strategies to clarify the long-term effects of pollution on specific
diseases should be a priority for future research in this area.
References
1. Bhaskaran K, Hajat S, Armstrong B, Haines A, Herrett E, Wilkinson
P, et al. The effects of hourly differences in air pollution on the risk
of myocardial infarction: case crossover analysis of the MINAP database.
BMJ. 2011; 343: d5531.
2. Samet JM, Dominici F, Curriero FC, Coursac I, Zeger SL. Fine
particulate air pollution and mortality in 20 U.S. cities, 1987-1994. N
Engl J Med. 2000; 343(24): 1742-9.
3. Bhaskaran K, Wilkinson P, Smeeth L. Cardiovascular consequences of
air pollution: what are the mechanisms? Heart. 2011; 97(7): 519-20.
4. Mills NL, ouml, rnqvist H, Gonzalez MC, Vink E, Robinson SD, et
al. Ischemic and thrombotic effects of dilute diesel-exhaust inhalation in
men with coronary heart disease. N Engl J Med 2007, Sep 13; 357(11):1075-
82 [The New England journal of medicine]; 2007.
5. Zeger SL, Dominici F, Samet J. Harvesting-resistant estimates of
air pollution effects on mortality. Epidemiology. 1999; 10(2): 171-5.
6. Hales S, Edwards R. Cardiovascular effects of exposure to air
pollution. Bmj. 2011; 343: d5814.
7. Bhaskaran K, Hajat S, Haines A, Herrett E, Wilkinson P, Smeeth L.
Effects of ambient temperature on the incidence of myocardial infarction.
Heart. 2009; 95(21): 1760-9.
Competing interests:
No competing interests
11 October 2011
Krishnan Bhaskaran
lecturer in statistical epidemiology
Professor Sir Andy Haines (professor of public health and primary care), Dr Paul Wilkinson (reader in environmental epidemiology), Professor Liam Smeeth (professor of clinical epidemiology)
Rapid Response:
Authors' response: Traffic-related air pollution and acute myocardial infarction
Dear Editor,
We are grateful for the thoughtful comments from Mills and Newby on
our study investigating the short-term associations between pollution
levels and risk of myocardial infarction (MI).[1] We agree that our data
were consistent with some prothrombotic effect of air pollution, but we
concluded from the lack of any apparent net increase in short-term MI risk
associated with pollution increases that other, perhaps non-thrombotic,
mechanisms may be more important drivers of the pollution-associated net
mortality increases that have been observed previously.[2]
The various valid points that Mills and Newby raise do not seem to
challenge this basic interpretation. They point out that the harvesting
phenomenon that we observed is consistent with an effect of traffic-
related pollution on coronary thrombosis, and we agree: were there no such
effect, we would not have expected to see any association between
pollution levels and MI risk at all. A number of experimental studies also
point to plausible pathways through which pollution might trigger MI.[3]
Our findings were consistent with such a triggering effect, as we observed
a transiently increased risk of MI for up to 6 hours after exposure. This
is also consistent with the very short-term and transient observed effects
of diesel exposure on ST-segment changes in the correspondents' own
experimental work.[4] However our data did not support an effect of
pollution on the overall MI risk over a 72-hour lag period; any pollution-
associated triggering of MIs appeared to result only in short-term
displacement (the bringing forward of MI events by a few hours or days).
In contrast with our observations for MI, short-term increases in broader
mortality outcomes after pollution peaks do not appear to be explained by
short-term displacement alone,[5] hence our conclusion that other
mechanisms may be more important drivers of net mortality increases.
A number of important limitations of our data are raised by Mills and
Newby. As we discuss at some length in the manuscript, we used data from
central pollution monitors to capture pollution levels across
conurbations. Unpublished data from our group suggest that correlations
between pollution levels measured at different sites within a conurbation
remain high at several kilometres distance; nevertheless, this strategy is
likely to have led to some degree of measurement error, and may have
resulted in a bias towards the null in single pollutant models. Hales and
Edwards point out in their editorial accompanying our article that this
factor could explain a missed or underestimated effect, but not the
temporal (harvesting) pattern that we observed.[6] In large-scale studies
of this type using routinely collected data, there is inevitably a trade-
off between study size and proximity of monitoring sites; restricting to
individuals living very close to a monitor would result in drastically
reduced study power. Predicted small-area pollution levels based on
spatial models may offer a solution to this dilemma in future studies, by
allowing a more personalised exposure without restrictive geographical
inclusion criteria; however, such modelled data would require careful
validation.
The MINAP dataset, which holds detailed timing data on acute coronary
events, offered an unprecedented opportunity to study associations at an
hourly temporal resolution in a large scale study. As Mills and Newby
point out, one disadvantage is that MIs resulting in death before hospital
admission are not recorded; MIs resulting in death at any time following
hospital admission would of course have been included. We find it
intuitively unlikely that pollution would be associated with a net
increase only in those MIs that resulted in death before hospital
admission, and yet no net increase at all in the remainder of MIs that
were admitted to hospital. However, we cannot exclude the possibility that
high levels of pollutants could exacerbate the severity of an MI and thus
the probability of early death by amplifying the thrombotic process. We
would therefore welcome future research clarifying this issue, using
alternative data sources which incorporate out-of-hospital deaths.
Finally, our study did not attempt to assess the impact of chronic
exposure to air pollution on MI risk. Whilst it is valid to consider acute
effects as a separate phenomenon, pollution policy must also be based on
an assessment of long-term risks. Unfortunately, the evidence base
regarding the long-term effects of air pollution exposure on MI risk
specifically is limited; only a handful of studies have been carried out,
and potential confounding by factors such as socioeconomic status and
occupation hampers many study designs.[7] The development of methods and
strategies to clarify the long-term effects of pollution on specific
diseases should be a priority for future research in this area.
References
1. Bhaskaran K, Hajat S, Armstrong B, Haines A, Herrett E, Wilkinson
P, et al. The effects of hourly differences in air pollution on the risk
of myocardial infarction: case crossover analysis of the MINAP database.
BMJ. 2011; 343: d5531.
2. Samet JM, Dominici F, Curriero FC, Coursac I, Zeger SL. Fine
particulate air pollution and mortality in 20 U.S. cities, 1987-1994. N
Engl J Med. 2000; 343(24): 1742-9.
3. Bhaskaran K, Wilkinson P, Smeeth L. Cardiovascular consequences of
air pollution: what are the mechanisms? Heart. 2011; 97(7): 519-20.
4. Mills NL, ouml, rnqvist H, Gonzalez MC, Vink E, Robinson SD, et
al. Ischemic and thrombotic effects of dilute diesel-exhaust inhalation in
men with coronary heart disease. N Engl J Med 2007, Sep 13; 357(11):1075-
82 [The New England journal of medicine]; 2007.
5. Zeger SL, Dominici F, Samet J. Harvesting-resistant estimates of
air pollution effects on mortality. Epidemiology. 1999; 10(2): 171-5.
6. Hales S, Edwards R. Cardiovascular effects of exposure to air
pollution. Bmj. 2011; 343: d5814.
7. Bhaskaran K, Hajat S, Haines A, Herrett E, Wilkinson P, Smeeth L.
Effects of ambient temperature on the incidence of myocardial infarction.
Heart. 2009; 95(21): 1760-9.
Competing interests: No competing interests