Author + information
- Received October 31, 2001
- Revision received February 27, 2002
- Accepted March 13, 2002
- Published online June 5, 2002.
- Warren J Cantor, MD*,* (, )
- L.Kristin Newby, MD, FACC†,
- Robert H Christenson, PhD‡,
- Robert H Tuttle, MSPH†,
- Vic Hasselblad, PhD†,
- Paul W Armstrong, MD, FACC§,
- David J Moliterno, MD, FACC∥,
- Robert M Califf, MD, FACC†,
- Eric J Topol, MD, FACC∥,
- E.Magnus Ohman, MD, FACC†,
- the SYMPHONY and 2nd SYMPHONY Cardiac Markers Substudy Investigators
- ↵*Reprint request and correspondence:
Dr. Warren J. Cantor, St. Michael’s Hospital, Division of Cardiology, 30 Bond St., Toronto, Ontario, Canada M5B-1W8.
Objectives We sought to assess the incidence and clinical significance of elevated cardiac troponin I (cTnI) after percutaneous coronary intervention (PCI).
Background Elevated creatine kinase-MB (CK-MB) is prognostically important after PCI, but the prognostic significance of elevated cTnI after PCI is uncertain.
Methods In a prospective substudy of the Sibrafiban Versus Aspirin to Yield Maximum Protection From Ischemic Heart Events Post-acute Coronary Syndromes (SYMPHONY) trials, which randomized patients with acute coronary syndromes (ACS) to receive aspirin or sibrafiban, we measured cTnI (positive, ≥1.5 ng/ml) and CK-MB (positive, ≥7 ng/ml) in 481 patients with PCI. Samples were collected immediately before and at 0, 8 and 16 h after PCI and analyzed by a core laboratory. The primary end point was the Kaplan-Meier estimate of death, myocardial infarction or severe, recurrent ischemia at 90 days.
Results Overall, 230 patients (48%) had elevated cTnI after PCI. Such patients underwent PCI sooner and were more likely to have coronary stenting. Elevated cTnI was associated with nonsignificantly higher risks of the primary end point (11.5% vs. 8.7%; p = 0.15) and of death (1.8% vs. 0.4%; p = 0.4) and a significantly higher risk of death or infarction (10.6% vs. 4.2%; p = 0.005). This pattern was more pronounced for patients who became positive only after PCI: primary end point, 20.7% vs. 10.1% for patients who remained negative after PCI (p = 0.05); death, 5.2% vs. 0% (p = 0.02); death or infarction, 18.1% vs. 4.1% (p = 0.007).
Conclusions Elevated cTnI, often observed after PCI in patients with ACS, is associated with worse 90-day clinical outcomes. This marker, therefore, is a useful prognostic indicator in such patients.
After percutaneous coronary intervention (PCI), about 20% of patients develop elevated creatine kinase (CK) or its MB isoform (CK-MB) (1). Such elevations have been associated with increased risks for death, myocardial infarction (MI) and repeat revascularization (1). A recent study by Stone et al. (2)of 7,147 patients with systematically collected CK-MB after PCI suggested that the relationship with in-hospital death may be manifest only at higher CK-MB levels. Compared with CK or CK-MB, the cardiac troponins are more specific biomarkers of myocardial necrosis and may be more sensitive in detecting myonecrosis after PCI. However, the prognostic significance of elevated troponin levels after PCI has not been prospectively evaluated (3).
We report the results of a prospective substudy within the Sibrafiban Versus Aspirin to Yield Maximum Protection From Ischemic Heart Events Post-acute Coronary Syndromes (SYMPHONY) and 2nd SYMPHONY protocols, designed to assess the incidence and clinical significance of elevated cardiac troponin I (cTnI) after PCI.
The SYMPHONY and 2nd SYMPHONY trials (4–6)randomized patients within seven days after presentation with unstable angina or MI to receive either an oral glycoprotein IIb/IIIa inhibitor, sibrafiban or placebo. In SYMPHONY, 9,233 patients were randomly allocated to receive either aspirin 80 mg or low-dose or high-dose sibrafiban twice daily for 90 days. The 2nd SYMPHONY trial randomized 6,671 patients to receive aspirin 80 mg, low-dose sibrafiban plus aspirin 80 mg, or high-dose sibrafiban twice daily for a final median treatment duration of 90 days.
All North American patients enrolled in either trial who underwent PCI after randomization were required by protocol to have periprocedural serum samples drawn for CK-MB analysis. In a prospective substudy, these samples were also analyzed for cTnI. The troponin substudy was approved by the institutional review boards or ethics committees of the participating hospitals.
In total, 15,904 patients were enrolled in the SYMPHONY trials, 8,056 in North America. Of North American patients, 1,547 (19%) underwent PCI, and 657 had serum samples sent to the core biochemistry laboratory. A total of 176 patients were excluded—17 had only preprocedural samples, 6 had only an immediately postprocedural sample, 24 had samples drawn outside the 24-h window surrounding PCI, and 129 underwent PCI before randomization—leaving 481 patients who underwent PCI and had samples drawn during the appropriate time window.
Sample collection and analysis
Samples were collected within 1 h before PCI, immediately after PCI and at 8 h (6 to 10 h) and 16 h (12 to 20 h) after PCI. All samples were prepared and stored locally at −70°C until shipped to the core laboratory (University of Maryland), where they were analyzed for CK-MB and cTnI using the Dimension RxL/HM analyzer (Dade Behring, Glasgow, Delaware). The analytical sensitivities for CK-MB and cTnI assays were 0.5 ng/ml and 0.04 ng/ml, respectively. The manufacturer’s suggested upper cutoff levels for the diagnosis of acute MI were considered positive for the primary analysis (cTnI level >1.5 ng/ml and CK-MB level >7 ng/ml, respectively). A secondary analysis was performed using 0.8 ng/ml as the cutoff level for positive cTnI.
Procedures and adjunctive medications
The decision to perform angiography or revascularization was left to treating physicians. All devices approved for PCI, including stents, atherectomy and lasers, were permitted. All patients received open-label aspirin the day of and day after PCI. After stenting, patients in the aspirin groups received ticlopidine 250 mg twice daily for two to four weeks. In SYMPHONY, patients in the sibrafiban groups who underwent stent implantation initially had received placebo for two to four weeks. After an interim safety analysis detected a higher-than-expected rate of stent thrombosis in the low-dose sibrafiban group, the protocol was amended to treat patients in this group with ticlopidine 250 mg twice daily for two to four weeks after stenting. In 2nd SYMPHONY, patients in both sibrafiban groups received placebo for two to four weeks. In all groups, study drug was continued without interruption through the periprocedural period, unless abciximab or another intravenous glycoprotein IIb/IIIa inhibitor was used during PCI. In these cases, study drug was held for 48 h after abciximab infusion (or for 8 h after eptifibatide or tirofiban infusion).
The primary end point for SYMPHONY was the 90-day composite incidence of death (all-cause), myocardial (re-)infarction or severe, recurrent ischemia (SRI) requiring urgent revascularization. The time to this composite end point was the primary end point in 2nd SYMPHONY. For this substudy, the primary assessment was the relationship between postprocedural cTnI status (positive or negative) and time to the 90-day composite of death, MI or SRI after PCI. Secondary analyses included the relationship of postprocedural cTnI status with the end points of death, MI and their composite. The MI and ischemia end points were used as defined and adjudicated by blinded central events classification committees for the main trials (4–6). The criteria for MI included: CK-MB above the upper limit of normal after an ischemic event, >3 times the upper limit of normal after PCI, >5 times the upper limit of normal after bypass surgery or new, significant Q waves on electrocardiogram. In 2nd SYMPHONY, only the first MI and first SRI episode after randomization were adjudicated. For this analysis, adjudicated events were used unless a patient in 2nd SYMPHONY had suffered an event before PCI, in which case investigator-reported events were used.
To eliminate the influence of positive cardiac markers before PCI, we also assessed the impact of postprocedural cTnI status on clinical outcomes after excluding patients with positive or missing preprocedural CK-MB or cTnI values, leaving only patients known to be preprocedural cTnI-negative. Comparisons between preprocedural cTnI-positive and preprocedural negative/postprocedural negative and preprocedural negative/postprocedural positive groups were also explored.
Continuous variables are summarized as medians (25th, 75th percentiles), discrete variables as percentages. Groups were compared using the likelihood-ratio chi-square test for discrete variables and the Wilcoxon rank-sum test for continuous variables. P values are presented for descriptive purposes; no adjustments were made for multiple comparisons.
The occurrence of postprocedural end points is presented in tabular form as Kaplan-Meier (KM) estimates (95% confidence interval [CI]). Multivariable Cox proportional-hazards modeling was performed with time to death or MI as the dependent variable. Candidate variables used in the models included: cTnI, age, gender, weight, creatinine clearance, days from qualifying event to PCI, stent use, prior MI, prior angina, elevated enzymes during qualifying event and multivessel intervention. Two models were developed incorporating cTnI as a binary variable and cTnI as a continuous variable. Kaplan-Meier survival curves were constructed to compare event-free survival after PCI by postprocedural cTnI status and were compared using the log-rank chi-square test. A p value of <0.05 was considered statistically significant.
Among the 481 substudy patients, 230 (48%) had elevated cTnI, and 137 (29%) had elevated CK-MB after PCI. A strong correlation was present between the peak postprocedural CK-MB and cTnI levels (r= 0.85, p = 0.0001). Postprocedural cTnI-positive patients less often had prior angina and more often hypertension and positive cardiac markers during the qualifying event (Table 1).
About 75% of interventions were elective, regardless of cTnI status (Table 2). Postprocedural cTnI-negative patients more often had PCI for recurrent or refractory ischemia, whereas postprocedural cTnI-positive patients more often had PCI for (re)MI. Patients who were cTnI-positive after PCI had intervention much sooner than patients negative after PCI (median 3 vs. 11 days; p < 0.001). They were also significantly more likely to receive a stent. Abciximab was used during PCI in 21% of cases; patients treated with abciximab were more likely to have elevated postprocedural cTnI (58% vs. 45%; p = 0.02).
A nonsignificant trend was present for greater KM estimates of the primary end point among patients with positive postprocedural cTnI compared with patients who were negative after PCI (11.5% vs. 8.7%; p = 0.15) (Table 3). The hazard ratio (HR) (95% CI) for the primary end point among patients with positive postprocedural cTnI was 1.49 (0.86 to 2.57). The KM estimates of death and SRI were low and not significantly different, but the KM estimate of MI alone was significantly higher among postprocedural cTnI-positive patients (p = 0.007). The KM estimate of death or MI was significantly greater in patients with elevated postprocedural cTnI (10.6% vs. 4.2%; p = 0.005). The KM estimates for each quartile of postprocedural cTnI elevation are listed in Table 4. With increasing postprocedural cTnI levels, there were trends toward increasing 90-day death and death or MI. Figure 1shows the timing of death or MI after PCI for patients with positive and negative postprocedural cTnI.
When patients who had an end point MI <48 h after PCI were removed from the analysis, those with elevated postprocedural cTnI still had a significantly higher KM estimate of 90-day death or MI (7.8% vs. 3.7%; p = 0.04). Assessing clinical outcomes using a cTnI cutoff of 0.8 ng/ml, more patients were cTnI-positive after PCI (57%), but otherwise the results were similar. Likewise, in an analysis stratified by type of PCI (elective vs. nonelective), the troponin effect remained statistically significant (p = 0.003), and, in a Breslow-Day test, the p value was 0.571, suggesting that the association of cTnI with outcome is not significantly different for elective and nonelective patients.
After excluding patients with positive or missing preprocedural cTnI values, postprocedural cTnI was elevated in 39 (26%) of 151 preprocedural cTnI-negative patients. The pattern of increased risk among patients with new positive cTnI after PCI was similar to that in the primary analysis (Fig. 2). Among patients with new elevations of cTnI after PCI, that is, those who were known to be cTnI-negative before the procedure, the HR for 90-day death or MI was 4.3 (1.4 to 13.5) compared with 2.7 (1.3 to 5.4) for the entire substudy cohort. Further, the HR for patients known to be positive before the procedure was similar (2.4 [0.8 to 7.0]). Estimated 90-day death or MI rates were 6.1% for patients with negative preprocedural cTnI who remained negative after PCI, 18.1% for patients with negative pre-PCI cTnI who became positive after PCI and 9.0% for pre-PCI cTnI-positive patients.
In multivariable models including cTnI as either a continuous or binary variable, cTnI independently predicted time to death or MI: HR 1.02 (1.01 to 1.03), p = 0.0001 when cTnI was continuous and 3.0 (1.5 to 6.1), p = 0.003 when cTnI was binary. The only other variables independently associated with death or MI were previous MI (HR 2.1 [1.0 to 4.2], p = 0.045) and history of angina (HR 2.3 [1.1 to 4.9], p = 0.03).
This prospective substudy of the SYMPHONY trials confirms that cTnI is often elevated after PCI and that this is an important predictor of later cardiac events. Nearly half of all substudy patients had elevated cTnI after intervention. These patients were at significantly increased risk for death or MI at 90 days. The increase in risk is of similar magnitude to that observed previously for postprocedural CK-MB elevations.
Troponins and risk for cardiac events
The cardiac troponins are highly sensitive and specific in detecting myonecrosis. In patients presenting with acute coronary syndromes (ACS), elevated troponin is associated with significantly higher mortality (7,8). However, the prognostic value of troponins after PCI has not been well established. Prior studies have reported elevated post-PCI troponin in up to 45% of patients (9–17). Bertichant et al. (16)found that troponin was elevated in 18% to 21% but found no association between cTnI status and clinical outcomes at a median 16 months after PCI. Fuchs et al. (17)recently analyzed 1,129 patients undergoing PCI, of which 31% had elevated cTnI after intervention. These patients had higher rates of death, MI and target-lesion revascularization during initial hospitalization, but there was no association between cTnI positivity and late (up to eight months) clinical events.
To our knowledge, our report represents the first large, prospective study to measure troponin levels before and after PCI using a standardized assay in a core biochemical laboratory and to systematically follow patients for clinical events up to 90 days after PCI. Overall, postprocedural cTnI was elevated in 48% of patients. After excluding those with positive or unknown preprocedural cTnI, 26% had an elevated postprocedural cTnI, consistent with prior studies (12,16,17). We found that patients with positive postprocedural cTnI had significantly higher 90-day rates of MI and the composite of death or MI. This difference persisted even after excluding patients with end point MI within 48 h after the procedure from the analysis. There also was a trend toward greater mortality in postprocedural cTnI-positive patients, especially among patients with new postprocedural cardiac marker elevations.
Potential mechanisms for post-PCI troponin elevation.
As with CK-MB, the pathophysiologic mechanism for elevated cTnI after PCI is unclear. Potential mechanisms include necrosis or ischemia from prolonged balloon inflations, transient abrupt closure, distal embolization and side-branch occlusion. Bertichant et al. (16)showed that elevated postprocedural cTnI correlated most closely with long balloon-inflation times and was not associated with side-branch occlusion or other procedural complications. Talasz et al. (9)found that, of 16 patients without side-branch occlusion, none had elevated cTnT after intervention, but three of five patients with side-branch occlusion did show such elevations.
Coronary stents may increase the frequency of minor myonecrosis by occlusion of small side-branches, increased distal embolization or both. La Vecchia et al. (12)found higher peak cTnI levels after stenting compared with a matched cohort undergoing angioplasty; the difference mainly reflected the occurrence of minor side-branch occlusion. Two other studies have found no difference in the incidence of cTnI elevation after stenting (16,17). In our analysis, patients who underwent coronary stenting were more likely to have elevated cTnI afterward than were patients who did not (51% vs. 33%; p = 0.002). These findings must be interpreted cautiously because patients with high-risk angiographic characteristics, dissection or suboptimal results would more likely undergo stenting. However, because about 80% of the patients in the SYMPHONY trials received a stent, this is less likely to be a major explanation in our study.
Even small CK-MB elevations (one to three times the upper limit of normal) after PCI are associated with higher six-month death or MI (18–20). Our findings indicate that elevations in cardiac troponins, the most sensitive markers of cardiac myonecrosis, also are significantly associated with the risk of death or MI. Periprocedural myonecrosis appears to represent a spectrum of risk. Large elevations in CK-MB after PCI are associated with the worst clinical outcomes, but even small CK-MB or cTnI elevations are associated with adverse outcomes.
The sample size of the study was small with low event rates, particularly for death and SRI. Therefore, the estimates of association may be somewhat unstable. However, the consistency of the point estimates in both the overall analysis and in patients with newly positive troponin after PCI favoring an increased risk for each end point in troponin-positive patients is reassuring.
Although we report an association between postprocedural cTnI elevation and 90-day outcomes, a causal relationship cannot be inferred from these data alone. Discussion continues about whether elevated postprocedural CK-MB is merely a marker of vessel or plaque characteristics predisposing to recurrent plaque rupture or whether the myonecrosis that occurs during PCI relates directly to long-term outcomes. Detailed angiographic data such as lesion morphology were not collected in the SYMPHONY trials and, therefore, could not be included in the multivariable model. Glycoprotein IIb/IIIa inhibitors have been shown to reduce both the incidence of elevated CK-MB after PCI and the risk of death or MI, suggesting a causal relationship between periprocedural cardiac markers and clinical outcomes (21). The impact of glycoprotein IIb/IIIa inhibitors on postprocedural troponin release has not been studied. In our study, abciximab was used more often in postprocedural cTnI-positive patients. This difference likely reflects the more frequent use of abciximab in patients with high-risk clinical and angiographic characteristics, such as intracoronary thrombus, or in “bail-out” situations after procedural complications have developed.
Troponin can remain elevated for 7 to 10 days (7). As a result, a substantial proportion of patients with ACS in this analysis had elevated preprocedural cTnI. In such cases, adverse long-term prognosis may relate more to the recent ACS than to elevated cTnI after PCI. However, the association between elevated cTnI after PCI and the risk of death or MI was even stronger when considering only patients with documented negative preprocedural cardiac markers. This suggests that the prognostic association of postprocedural cTnI elevation with 90-day outcomes is not driven by risk related to preprocedural ischemic events with resultant cTnI elevation. Furthermore, the comparison of patients known to be preprocedural cTnI-positive, reflecting elevations from prior ischemic events, with those known to be preprocedural cTnI-negative who then had new postprocedural cTnI elevation suggests that the risk for subsequent events is at least as great, if not greater, among patients with new post-PCI cTnI elevations as it is among those with elevations due to ischemic events who then undergo PCI.
In this study, follow-up was available only up to 90 days after PCI. Further studies are required to determine the relationship between postprocedural troponin positivity and long-term clinical outcomes. This analysis was limited to patients undergoing PCI within 90 days of ACS. Such patients are at higher risk for ischemic events during and after PCI. Whether these findings can be extrapolated to patients with stable coronary disease undergoing PCI requires further study.
Cardiac troponin I often is elevated after PCI, with positive results in 25% to 50% of patients. Elevated cTnI after PCI is strongly associated with increased 90-day risks of MI and the composite of death or MI, even after adjustment for baseline characteristics, procedural variables and preprocedural marker status. The prognostic value of cTnI after PCI is consistent with and similar to that of CK-MB. Further study of the pathophysiology of troponin elevation after PCI, its relationship to adverse outcomes and measures to alter the pathophysiology is warranted.
The authors thank the following colleagues for their efforts to ensure the quality of the substudy and its presentation: Lindsay Lambe, project leader; Linda Zillman and Kathy Galan, overall SYMPHONY and 2nd SYMPHONY project coordinators; Kristina Sigmon, biostatistician; Tonya Miller, trial assistant; and the research staff and principal investigators at the North American SYMPHONY and 2nd SYMPHONY hospitals. A complete list of investigators is provided in references 6and 8.
☆ Supported by Dade Behring, Inc., Glasgow, Delaware.
- acute coronary syndromes
- confidence interval
- creatine kinase-MB
- cardiac troponin I
- hazard ratio
- myocardial infarction
- percutaneous coronary intervention
- severe recurrent ischemia
- Sibrafiban Versus Aspirin to Yield Maximum Protection From Ischemic Heart Events Post-acute Coronary Syndromes trial
- Received October 31, 2001.
- Revision received February 27, 2002.
- Accepted March 13, 2002.
- American College of Cardiology Foundation
- Califf R.M.,
- Abdelmeguid A.E.,
- Kuntz R.E.,
- et al.
- Stone G.W.,
- Mehran R.,
- Dangas G.,
- Lansky A.J.,
- Kornowski R.,
- Leon M.B.
- Holmes D.R. Jr..,
- Berger P.B.
- Newby L.K.
- The 2nd SYMPHONY Investigators
- Rupprecht H.J.,
- Terres W.,
- Ozbek C.,
- et al.
- Johansen O.,
- Brekke M.,
- Stromme J.H.,
- et al.
- Abdelmeguid A.E.,
- Topol E.J.,
- Whitlow P.L.,
- Sapp S.K.,
- Ellis S.G.
- Abdelmeguid A.E.,
- Whitlow P.L.,
- Sapp S.K.,
- et al.
- Tardiff B.E.,
- Califf R.M.,
- Tcheng J.E.,
- et al.