Prognostic Implications of Procedural vs Spontaneous MI
Prognostic Implications of Procedural vs Spontaneous MI
Among 7,380 patients included in this analysis, 4,568 (62%) patients had no MI, 580 (8%) patients had a procedural MI, and 2,232 (30%) patients presented with a spontaneous MI (595 with STEMI and 1,495 with NSTEMI) at baseline. The baseline characteristics of the 3 groups are summarized in Table I. As expected, there were considerable differences between the 3 groups. When compared with the no-MI and procedural MI groups, patients with spontaneous MI were younger and were less likely to have hypertension, diabetes, hyperlipidemia, prior PCI, prior coronary artery bypass graft surgery, prior stroke, or peripheral arterial disease but were more likely to be smokers, have heart failure, prior MI, ejection fraction <25%, platelet count >300,000, creatinine >2 mg/dL, New York Heart Association (NYHA) class III or IV heart failure, or Canadian Cardiovascular Society (CCS) class IV angina. In general, patients with procedural MI had more complex coronary artery disease than did the other 2 groups, including more diseased coronary arteries and more frequent bifurcation lesions, and they required more stents and longer total stent lengths. As expected, patients with no MI were more likely to have Thrombolysis In Myocardial Infarction (TIMI) 3 flow at baseline and at the end of the procedure and less likely to have experienced angiographic complications (site reported) compared with the other 2 groups (Table II). After multiple propensity score adjustment, there were no significant differences in the baseline characteristics of the 3 groups (Table I).
The baseline and peak cardiac biomarker levels were considerably different between the 3 groups (Table II). Compared with patients with procedural MI, those presenting with spontaneous MI had higher peak CK-MB (4.08 ± 6.64 vs 8.82 ± 19.19; P < .0001) or cTn (24.05 ± 71.40 vs 133.53 ± 783.51; P < .0001). In addition, when compared with patients with procedural MI, those presenting with spontaneous MI were more likely to have peak cTn >10× ULN (Figure 2).
(Enlarge Image)
Figure 2.
Distribution of peak troponin levels among patients with procedural and spontaneous MI.
There was a graded increase in 1-year all-cause mortality across the 3 patient subgroups (no MI: 1.9%, procedural MI: 3.1%, spontaneous MI: 3.9%; P < .001) (Figure 3). In the unadjusted analysis, both spontaneous MI (hazard ratio [HR] 2.15, 95% CI 1.16-2.90, P < .001) and procedural MI (HR 1.65, 95% CI 0.99-2.74, P = .053) were associated with an increased risk for all-cause mortality.
(Enlarge Image)
Figure 3.
All-cause mortality among patients with no MI, procedural MI, and spontaneous MI.
In a proportional hazards regression model that incorporated the multiple propensity scores, spontaneous MI (adjusted HR 1.62, 95% CI 1.11-2.37; P = .01) remained independently associated with all-cause mortality when compared with the no-MI group. Although there was a trend toward a similar association for procedural MI, after adjustment for multiple propensity scores, the association was not statistically significant (adjusted HR 1.51, 95% CI 0.89-2.54, P = .12). Similar results were seen in a traditional proportional hazards regression model adjusted to baseline covariates (data not shown).
When the analysis was restricted to cardiovascular mortality, there was a graded increase in risk over the 12-month follow-up period (no MI: 0.5%, procedural MI: 1.0%, spontaneous MI: 1.7%; P < .001) (Figure 4). In a univariate model with the no-MI group as reference, spontaneous MI (HR 3.56, 95% CI 2.13-5.96, P < .001) was a significant predictor of cardiovascular death but not procedural MI (HR 2.06, 95% CI 0.84-5.05, P = .12). When we adjusted for the multiple propensity scores, the independent association with cardiovascular mortality was significant for spontaneous (adjusted HR 3.15, 95% CI 1.68-5.90, P < .0001) but not procedural MI (adjusted HR 1.74, 95% CI 0.69-4.40, P = .24). Similar results were seen in a traditional proportional hazards regression model adjusted to baseline covariates (data not shown).
(Enlarge Image)
Figure 4.
Cardiovascular mortality among patients with no MI, procedural MI, and spontaneous MI.
In a sensitivity analysis, excluding patients presenting with STEMI yielded largely similar results, with spontaneous MI being a predictor of all-cause mortality (adjusted HR 1.69, 95% CI 1.15-2.50, P = .008) and cardiovascular mortality (adjusted HR 3.26, 95% CI 1.71-6.22, P < .001) but not procedural MI (adjusted HR for all-cause mortality 1.50, 95% CI 0.89-2.52, P = .13; adjusted HR for cardiovascular mortality 1.75, 95% CI 0.69-4.42, P = .24).
Analysis was then restricted to troponin levels >10× ULN for both procedural and spontaneous MIs. Both procedural and spontaneous MIs were associated with similar risk of all-cause mortality (3.1% vs 3.9%; P = .59) and cardiovascular mortality (0.4% vs 2.0%; P = .11), although the event rates were lower for procedural MI.
Results
Patient Population
Among 7,380 patients included in this analysis, 4,568 (62%) patients had no MI, 580 (8%) patients had a procedural MI, and 2,232 (30%) patients presented with a spontaneous MI (595 with STEMI and 1,495 with NSTEMI) at baseline. The baseline characteristics of the 3 groups are summarized in Table I. As expected, there were considerable differences between the 3 groups. When compared with the no-MI and procedural MI groups, patients with spontaneous MI were younger and were less likely to have hypertension, diabetes, hyperlipidemia, prior PCI, prior coronary artery bypass graft surgery, prior stroke, or peripheral arterial disease but were more likely to be smokers, have heart failure, prior MI, ejection fraction <25%, platelet count >300,000, creatinine >2 mg/dL, New York Heart Association (NYHA) class III or IV heart failure, or Canadian Cardiovascular Society (CCS) class IV angina. In general, patients with procedural MI had more complex coronary artery disease than did the other 2 groups, including more diseased coronary arteries and more frequent bifurcation lesions, and they required more stents and longer total stent lengths. As expected, patients with no MI were more likely to have Thrombolysis In Myocardial Infarction (TIMI) 3 flow at baseline and at the end of the procedure and less likely to have experienced angiographic complications (site reported) compared with the other 2 groups (Table II). After multiple propensity score adjustment, there were no significant differences in the baseline characteristics of the 3 groups (Table I).
Cardiac Biomarkers and Short-term Outcomes
The baseline and peak cardiac biomarker levels were considerably different between the 3 groups (Table II). Compared with patients with procedural MI, those presenting with spontaneous MI had higher peak CK-MB (4.08 ± 6.64 vs 8.82 ± 19.19; P < .0001) or cTn (24.05 ± 71.40 vs 133.53 ± 783.51; P < .0001). In addition, when compared with patients with procedural MI, those presenting with spontaneous MI were more likely to have peak cTn >10× ULN (Figure 2).
(Enlarge Image)
Figure 2.
Distribution of peak troponin levels among patients with procedural and spontaneous MI.
Mortality Outcomes
There was a graded increase in 1-year all-cause mortality across the 3 patient subgroups (no MI: 1.9%, procedural MI: 3.1%, spontaneous MI: 3.9%; P < .001) (Figure 3). In the unadjusted analysis, both spontaneous MI (hazard ratio [HR] 2.15, 95% CI 1.16-2.90, P < .001) and procedural MI (HR 1.65, 95% CI 0.99-2.74, P = .053) were associated with an increased risk for all-cause mortality.
(Enlarge Image)
Figure 3.
All-cause mortality among patients with no MI, procedural MI, and spontaneous MI.
In a proportional hazards regression model that incorporated the multiple propensity scores, spontaneous MI (adjusted HR 1.62, 95% CI 1.11-2.37; P = .01) remained independently associated with all-cause mortality when compared with the no-MI group. Although there was a trend toward a similar association for procedural MI, after adjustment for multiple propensity scores, the association was not statistically significant (adjusted HR 1.51, 95% CI 0.89-2.54, P = .12). Similar results were seen in a traditional proportional hazards regression model adjusted to baseline covariates (data not shown).
When the analysis was restricted to cardiovascular mortality, there was a graded increase in risk over the 12-month follow-up period (no MI: 0.5%, procedural MI: 1.0%, spontaneous MI: 1.7%; P < .001) (Figure 4). In a univariate model with the no-MI group as reference, spontaneous MI (HR 3.56, 95% CI 2.13-5.96, P < .001) was a significant predictor of cardiovascular death but not procedural MI (HR 2.06, 95% CI 0.84-5.05, P = .12). When we adjusted for the multiple propensity scores, the independent association with cardiovascular mortality was significant for spontaneous (adjusted HR 3.15, 95% CI 1.68-5.90, P < .0001) but not procedural MI (adjusted HR 1.74, 95% CI 0.69-4.40, P = .24). Similar results were seen in a traditional proportional hazards regression model adjusted to baseline covariates (data not shown).
(Enlarge Image)
Figure 4.
Cardiovascular mortality among patients with no MI, procedural MI, and spontaneous MI.
Sensitivity Analysis
In a sensitivity analysis, excluding patients presenting with STEMI yielded largely similar results, with spontaneous MI being a predictor of all-cause mortality (adjusted HR 1.69, 95% CI 1.15-2.50, P = .008) and cardiovascular mortality (adjusted HR 3.26, 95% CI 1.71-6.22, P < .001) but not procedural MI (adjusted HR for all-cause mortality 1.50, 95% CI 0.89-2.52, P = .13; adjusted HR for cardiovascular mortality 1.75, 95% CI 0.69-4.42, P = .24).
Analysis was then restricted to troponin levels >10× ULN for both procedural and spontaneous MIs. Both procedural and spontaneous MIs were associated with similar risk of all-cause mortality (3.1% vs 3.9%; P = .59) and cardiovascular mortality (0.4% vs 2.0%; P = .11), although the event rates were lower for procedural MI.
Source...