Using data from the URMC Cardiac Catheterization Laboratory with information on the time of symptom onset, acute coronary syndrome subtype (STEMI or NSTEMI), and demographic and clinical variables, we found that increased ambient PM2.5 concentrations were associated with immediate (within 1 hour) increases in the risk of STEMI, but not NSTEMI. Effect estimates were largest for those with prior cardiovascular disease/events, specifically those with pre-existing hypertension, and non-smokers, those 65 years and older, Caucasians, and women. This finding of fine particle triggering of STEMI, but not NSTEMI, within 1 hour, suggests potential mechanisms by which this response could occur, and also that these mechanisms must act on this rapid time scale. Such mechanisms might result in more extensive plaque rupture or the promotion of thrombus formation in the subsequent interplay between prothrombotic and antithrombotic vascular processes.
Previous studies have reported an increased risk of myocardial infarction associated with increased PM2.5 and other pollutants in the preceding day or two . Only our previous study  investigated PM associations by type of myocardial infarction, finding that increased PM2.5 concentration in the 24 hours before emergency room arrival for the myocardial infarction was associated with an increased risk of those infarctions progressing to a transmural infarction but not those progressing only as far as a partial wall infarction . Only a few studies had symptom onset time data and thus were able to examine whether increased pollutant concentrations in the previous few hours, and not just the same day or just within 24 hours, triggered the myocardial infarction [1, 4, 5, 7]. Only our current study and that of Peters et al.  report such an immediate response, while other studies do not [1, 5, 7]. This inconsistency may be due, at least in part, to the distribution of STEMI/NSTEMI within the sample of acute coronary syndrome included in each study. Our findings of greater risk of STEMI associated with increased PM2.5 concentrations among those subjects with pre-existing cardiovascular disease and/or prior acute cardiovascular events are consistent with findings from previous studies, [6, 8, 9] and affirm a hypothesis of increased susceptibility for those with pre-existing coronary artery disease.
There are multiple mechanisms for acute coronary syndrome and specifically STEMI and NSTEMI, including coagulation, inflammation, vascular dysfunction, and autonomic dysfunction. However, when a thrombus forms, circulating platelets are tethered via intermediate filaments to one another and to the injured vascular wall. Mechanistically, STEMI differ from NSTEMI in that STEMIs progress acutely to complete arterial occlusion following plaque rupture, whereas plaque-rupture mediated NSTEMIs do not. It is important to note that plaque rupture itself is necessary, but in isolation is insufficient to promote full arterial occlusion, and a cascade of coordinated cellular events must occur prior to thrombus formation. Furthermore, endogenous thrombolysis is an important protective process which can abrogate complete arterial occlusion (i.e. STEMI). Due to fundamental mechanistic differences between STEMI and NSTEMI, morbidity and mortality are higher with STEMI [18, 19]. Thus, one may propose that the intracellular signal transduction mechanisms committing a patient to a STEMI distinguish it from NSTEMI by a different balance between thrombosis and thrombolysis. PM may be more likely to cause a STEMI than a NSTEMI if PM enhances the rate or extent of thrombus formation following plaque rupture and/or if PM exposure impairs thrombolysis, increasing the chance for complete vessel occlusion.
Plaque rupture and subsequent acute thrombosis of previously narrowed or unstable plaque-lined arteries is recognized as the proximal event in the evolution of myocardial infarction . Previously, we reported large changes in circulating markers of platelet and endothelial cell activation (p-selectin, CD40L, von Willebrand factor [vWF]), in healthy young medical residents, associated with increased PM2.5 and other pollutants during the 2008 Beijing Summer Olympics [21, 22]. We have also found increased platelet activation following controlled ultrafine particle exposure in diabetics . We and others have found increased fibrinogen and C-reactive protein levels (inflammatory markers) associated with increased ambient PM2.5 concentrations, and accumulation mode and ultrafine particle number concentrations in the preceding few days,  especially in subjects with underlying cardiovascular risk factors [24, 25]. However, some controlled human exposure studies have not reproduced these data [26, 27]. Furthermore, changes in vWF, p-selectin, CD40L, and platelet aggregation have also been repeatedly associated with acute increases in PM concentration in both humans and animals, [25, 27–30] suggesting these interacting coagulation/inflammation pathways may be acutely affected by PM air pollution exposure. Consistent with this, controlled exposure studies have reported impaired vascular function and increased thrombogenesis following diesel exhaust exposure [31–35]. Only a few studies have contrasted the circulating concentration of inflammatory and thrombotic biomarker levels in STEMI and NSTEMI from blood samples drawn upon arrival in the emergency room, and reported increases in thrombotic/inflammatory markers (e.g. WBC, C-reactive protein, ferritin, serum amyloid A) [36, 37]. Other studies found no difference in these and other platelet markers (e.g. mean platelet volume, platelet count) [38, 39]. However, of the biochemical and physiological mechanisms associated with ambient PM exposure, only thrombus formation and vasoconstriction occur on such a rapid time scale. Vascular constriction is an instantaneous response to offending stimuli mediated by endothelial and non-endothelial factors, whereas thrombus formation is a continuum of events that is initiated by platelet activation, aggregation, and adhesion to an injured vessel wall. This time continuum makes thrombus formation more variable, and may explain fundamental differences between STEMI and NSTEMI with respect to PM exposure.
Although our study had several strengths including well characterized acute coronary syndrome events in the Cardiac Catheterization Laboratory, there are several important limitations that should be considered when making inference. First, we had limited sample size due to our use of existing data, the requirement of symptom onset time data, and a residence within 15 miles of the monitoring site. This limited sample size resulted in reduced statistical power and limited precision in our risk estimates. Second, we used ambient PM2.5 concentrations, and ultrafine particle and accumulation mode particle number concentrations, measured at one central site monitoring location, to represent each myocardial infarction patient’s exposure to PM of outdoor origin, regardless of how close subjects lived, worked, or spent time to the monitoring site. This exposure error, however, is not likely different for time periods immediately preceding the myocardial infarction and other time periods in the weeks before or after the STEMI/NSTEMI, likely resulting in biases toward the null and underestimates of the risk. Further, when we included only those STEMI patients living within 5 miles of the monitoring site, our estimates of the risk of STEMI associated with increased PM2.5 concentration, ultrafine particle number concentration and accumulation mode particle number concentration in the previous hour were similar. Since UFP and AMP number concentrations are more spatially variable across space than PM2.5 concentrations, in part due to differences in sources (AMP formed by atmospheric chemistry oxidizing precursor gases like SO2 and NO2; UFP locally emitted and having a shorter atmospheric lifetime), it is not surprising they are not well correlated with PM2.5 concentrations . Thus it is also not surprising we did not find increased risks of STEMI associated with increased UFP and AMP number concentrations.
Third, symptom onset time was determined by patient self-report upon arrival to the Cardiac Catheterization Laboratory. This is a reasonable surrogate marker for ischemia onset, but cannot be extrapolated to infarction onset with confidence. The temporal relationship between plaque rupture, ischemia and infarction may differ between individuals as much as it differs between STEMI and NSTEMI. If STEMI and NSTEMI have the same true OR associated with increased PM2.5 concentration (e.g. 1.50), but there is a greater degree of error in estimating NSTEMI onset time than STEMI onset time, the risk of NSTEMI associated with increased PM2.5 concentration could be underestimated to a greater degree (i.e. OR = 1.00) than the risk of STEMI associated with increased PM2.5 concentration (i.e. OR = 1.18). Further work is therefore needed to confirm that our findings were not due to this exposure error.
It is also possible that our finding is due to residual confounding by an unmeasured confounder. However, for such a factor to exist, it would have to vary temporally between weeks within each subject, be correlated with ambient air pollution levels, and also be a risk factor for myocardial infarction independent of air pollution. Outside of temperature and relative humidity, numerous studies examining the acute risk of cardiovascular events including myocardial infarction associated with air pollution exposure have not identified such a factor.
As shown in Table 1, STEMI patients were slightly younger, less likely to be obese, more likely to be smokers, and were less likely to have several cardiovascular co-morbidities or procedures than NSTEMI patients. Although not significantly different for most factors, STEMI patients with many of these factors had an even larger relative risk of STEMI associated with increased PM2.5 in the previous hour than STEMI patient without them. If these factors put one at increased risk for an ACS, then we would have expected to see an increased risk of NSTEMI associated with increased PM2.5 concentrations, rather than STEMI. Thus, although it is possible that our findings of PM2.5 triggering of STEMI but non NSTEMI might simply reflect differences in the population presenting rather than a different biological mechanism, our effect modification findings do not support this.
We only studied patients receiving care in the Cardiac Catheterization Laboratory of a tertiary care hospital. Thus our subjects may not be representative of the broader population of acute coronary syndrome patients that do not have access to cardiac catheterization, or who are medically managed. However, if our findings can be confirmed in a prospective study addressing many of the weaknesses described above, the findings can be generalized to all US adults who would be treated at a hospital’s cardiac catheterization laboratory, since plaque rupture with thrombus formation generally is the most prevalent cause of STEMI and NSTEMI .