Prognostic Significance of Multiple Previous Percutaneous Coronary Interventions in Patients Undergoing Elective Coronary Artery Bypass Surgery
Background— A possible relationship between increased perioperative risk during coronary artery bypass grafting (CABG) and previous percutaneous coronary intervention (PCI) is debatable. We sought to determine the impact of previous PCI on patient outcome after elective CABG.
Methods and Results— Between January 2000 and January 2005, 2626 consecutive patients undergoing first-time isolated elective CABG as the primary revascularization procedure (group 1) were evaluated for in-hospital mortality and major adverse cardiac events (MACEs) and were compared with 360 patients after single PCI (group 2) and with 289 patients after multiple PCI sessions (group 3) before elective CABG. Unadjusted univariate and risk-adjusted multivariate logistic-regression analysis revealed previous multiple PCIs to be strongly associated with in-hospital mortality (odds ratio [OR], 2.24; 95% confidence interval [CI], 1.52 to 3.21; P<0.001) and MACEs (OR, 2.28; 95% CI, 1.38 to 3.59; P<0.001). To control for selection bias, a computed propensity-score matching based on 13 patient characteristics and preoperative risk factors was performed separately comparing group 1 versus 2 and group 1 versus 3. After propensity matching, conditional logistic-regression analysis confirmed previous multiple PCIs to be strongly associated with in-hospital mortality (OR, 3.01; 95% CI, 1.51 to 5.98; P<0.0017) and MACEs (OR, 2.31; 95% CI, 1.45 to 3.67; P<0.0004).
Conclusions— In patients with a history of multiple PCI sessions, perioperative risk for in-hospital mortality and MACEs during subsequent elective CABG is increased.
Since the advent of percutaneous coronary intervention (PCI) for treatment of coronary artery disease (CAD), there has been a dramatic increase in stenting procedures. With the increased stent experience, aggressive repeated PCI therapy with multiple stent-graft placement has become more common.1–3
In this current “stent era,” coronary artery bypass graft (CABG) surgery is faced with a rapidly increasing number of patients in whom previous PCI procedures have been performed before they are finally referred to CABG surgery. Whether there is a relationship between increased perioperative risk during CABG and previous PCI is debatable, and the extent of previous stenting procedures has not been well studied so far as a prognostic factor of CABG.4 We sought to determine the impact of previous single or multiple PCI sessions on patient outcome finally referred to primary elective CABG therapy.
This study was a retrospective, single-center, cohort study including 3275 consecutive patients who underwent first-time isolated CABG at the West German Heart Center Essen between January 2000 and August 2005. Patients were classified to 1 of 3 groups, depending on whether they had no previous PCI procedures (group 1), a successful single previous PCI procedure (group 2), or multiple repeated ≥2) PCI procedures (group 3) before CABG. We evaluated the association between previous PCI procedures before CABG and the postoperative in-hospital patient outcome.
Emergency surgery and previous myocardial infarction (<4weeks) before CABG, reoperative procedures, or concomitant surgery were exclusion criteria. The study was approved by the institutional review board. All of the patients had previously granted permission for use of their medical records for research purposes.
Data used in this analysis were retrieved from the West German Heart Center cardiovascular surgical database. This database prospectively collects a comprehensive list of prespecified data points, with ≥1800 data items per patient for all consecutive patients undergoing CABG surgery at our institution, including demographic, clinical, and outcome data. Within the database, patients were coded as having had previous stent-graft placement procedures, with the number of implanted stent-grafts and the time point of the last PCI procedure also noted.
All of the outcome measures used in this analysis were prespecified. Given the subjective nature of many clinical outcomes, we only prespecified all-cause in-hospital mortality after CABG as the primary study end point. The prespecified secondary end point was the rate of major adverse cardiac events (MACEs), including perioperative myocardial infarction (PMI), low cardiac output syndrome (LCOS), cardiac death (CD), and sudden cardiac death (SCD) during the postoperative hospitalization period. An independent review of the medical records of the patients who died after CABG surgery was performed, and a cardiac versus a noncardiac cause of death was adjudicated.
In-hospital death was defined as death after CABG during the index hospitalization. A PMI was considered to have occurred when 1 of the following diagnostic criteria were present: (1) a cardiac troponin I level >10.5 ng/mL after CABG, as previously described,5 (2) a creatine kinase MB level 3 times above the upper normal level, (3) new persistent ST-segment or T-wave changes (Minnesota codes 4-1, 4-2, 5-1, 5-2, or 9-2), or (4) the development of new Q waves (Minnesota code 1-1-1 to 1-2-7). LCOS was presumed in patients with a cardiac index <2.0 L·min−1·m−2 or a systolic arterial pressure <90 mm Hg, despite high-dose inotropic support (IV dopamine ≥8 μg·kg−1·min−1, or dobutamine ≥6 μg·kg−1·min−1, epinephrine >0.1 μg·kg−1·min−1, or norepinephrine >0.1 μg·kg−1·min−1). Death was considered cardiac in origin if it was caused by PMI, significant cardiac arrhythmias, or refractory LCOS. Sudden unexpected death occurring without another explanation was defined as SCD.
Surgical revascularization was performed as previously described5 in all patients via median sternotomy, standard cardiopulmonary bypass technique with ascending aortic and 2-stage venous cannulation, mild hypothermia (>32°C), and cold crystalloid cardioplegic arrest. Heparin was administered to achieve an activated coagulation time >400 seconds. Bypass graft flow was assessed for each graft by Doppler transit-time flowmetry. Protamine was administered to reverse heparin according to standard practice. Aprotinin was given before and during surgery. A medication of 500 mg aspirin was routinely administered within the first 6 hours after surgery followed by a daily dose of 100 mg. Preoperative medication with clopidogrel was routinely discontinued at least 24 hours before surgery and restarted within the first 48 hours after CABG.
Descriptive statistics are summarized for categorical variables as frequencies (percentages) and compared between groups with Pearson’s χ2 exact test or a 1-sided Cochran-Armitage trend test. Continuous variables, expressed as mean±SD or as median (interquartile range), were compared between groups with the Kruskal-Wallis test or Jonckheere-Terpstra’s trend test. When a significant overall effect was detected, 2-group comparisons were performed with Fisher’s exact test or the Mann-Whitney U test. Univariate and multivariate logistic regressions were performed to identify preoperative independent predictors for in-hospital mortality and MACEs. Those variables identified by univariate regression analysis with a probability value ≤0.1 for at least 1 study end point were added to the multivariate logistic-regression model. To control for selection bias as a result of nonrandom assignment to the 3 groups, propensity score (n:m) matching was performed. The propensity scores were calculated separately by comparing between group 1 versus 2 and group 1 versus 3. The following patient characteristics and preoperative risk factors were used to calculate propensity scores: age, sex, diabetes, hypertension, hyperlipidemia, left ventricular ejection fraction, renal disease, previous MI, left mainstem disease, chronic obstructive pulmonary disease, peripheral vascular disease, angina class III–IV, and the number of bypassed vessels. For both comparisons, patients with similar propensity scores were matched into 20 matched sets of equal size. Once patients were matched, conditional logistic regression was used.6 A probability value <0.05 was considered to indicate statistical significance. All statistical analyses were performed with StatXact 6.0 software (Cytel Software Corp, Cambridge, Mass) and the SAS System, version 8 (SAS Institute Inc, Cary, NC).
The authors had full access to the data and take responsibility for its integrity. All authors have read and agree to the manuscript as written.
Of the 3275 patients included in the present analysis, 2626 patients underwent CABG surgery without having had a previous PCI procedure in group 1, 360 patients with a single previous PCI procedure in group 2, and 289 patients having with multiple repeated PCI procedures in group 3. Preoperative characteristics of the patients are presented in Table 1. Patients did not differ according to their demographics, risk factors, and comorbidities, except that a significantly higher number of patients with hypertension and hyperlipidemia were in group 3. More group 3 patients received aspirin and statins, they had suffered more often from MI, and they had more often been treated with thrombolysis. The average time between the last PCI and CABG was 12±22 and 8±15 months in groups 2 and 3, respectively. Angiographic characteristics revealed no difference in the extent of CAD between the groups. When comparing the coronary “stent load,” group 2 patients had more single and multiple stents in a single vessel, for a total of 451 stent-grafts, whereas group 3 patients had more multiple stents in multiple vessels, for a total of 1040 stent-grafts (Table 2; P<0.0001). The majority of patients had bare-metal stent-grafts, and only in 9% of group 2 patients and 22% of group 3 patients were drug-eluting stents implanted (Table 2; P<0.0001). As a result of CAD progression, CABG was predominantly indicated for de novo stenosis in groups 1 and 2, but most of the group 3 patients had in-stent restenosis or combined de novo and in-stent restenosis (Table 2; P<0.0001).
Intraoperative results did not differ between the groups. The postoperative outcome data, eg, ventilation time and length of intensive care unit and hospital stays, did not differ between the groups, despite a higher incidence for postoperative use of an intra-aortic balloon pump (P<0.04). Major thoracic bleeding (>600 mL within the first 12 hours postoperatively) occurred more often in group 3 patients (P<0.001), and cardiopulmonary resuscitation was performed significantly more often in group 3 compared with group 1 patients (Table 3; P=0.04). The postoperative incidence of primary and secondary study end points are presented in Figures 1 and 2⇓. A 1-sided Cochran-Armitage trend test revealed a significant difference of all-cause in-hospital mortality between the groups (P<0.001). This was accompanied by a difference in the postoperative MACE rate (P<0.0001), with a higher incidence of all secondary end points, including SCD (P<0.02), CD (P<0.001), LCOS (P<0.04), and PMI (P<0.007) in group 3 compared with groups 1 and 2 (Figure 2). The MACE rate seemed to be related to the number of previous PCIs. The more previous PCI sessions, the higher the MACE rate (P<0.006; Figure 3); but no relationship could be found between the time from last PCI to CABG and MACEs.
To evaluate independent preoperative predictors of death, a logistic-regression analysis model was constructed, and several univariate factors were found to be predictive of death (Table 4). A multivariate logistic-regression analysis revealed that age, chronic obstructive pulmonary disease, and multiple previous PCI procedures were independent predictors of death (Table 4). As independent preoperative predictors for overall MACEs, previous MI and previous PCI procedures were identified (Table 5). Finally, after propensity-score matching, the conditional logistic-regression model did not change but rather confirmed the results indicating previous multiple PCIs as binge strongly associated with in-hospital mortality (odds ratio [OR], 3.01; 95% confidence interval [CI], 1.51 to 5.98; P<0.0017) and MACEs ([OR], 2.31; 95% CI, 1.45 to 3.67; P<0.0004).
The present study is the first to clearly demonstrate a significantly increased risk for in-hospital mortality and MACEs in patients with prior multiple PCI procedures subsequently undergoing isolated first-time elective CABG compared with patients without or with only a single previous PCI procedure.
PCI procedures are emerging as the main treatment option for CAD, and the number and frequency of PCI procedures are rapidly increasing worldwide. Because of this overwhelming evolution of PCI treatment and due to the fact that 15% to 30% of these patients require “poststent” coronary revascularization1,7 and nearly 20% are referred to CABG surgery after stenting,7–9 the number of “stent-loaded” patients requiring CABG is likewise rapidly increasing and will probably further increase. Because previous PCI has been reported in several recent clinical studies to have a negative impact on patient outcome after noncardiac surgery,10,11 the increasing number of multiple stent-loaded CAD patients may also have an impact on CABG surgery outcomes. Numerous preoperative predictors of patient outcome after CABG have been identified and are reliably used to stratify risk and predict morbidity and mortality, but previous PCI procedures before elective CABG have not been well studied so far as a prognostic factor after CABG.4,12 In the present study, however, by retrospectively analyzing a complete consecutive first-time CABG patient cohort of a single institution during a 5 year period, multiple previous PCI procedures before elective CABG could be clearly identified as an additional, major, independent predictor of in-hospital death and MACEs.
Although almost all outcome measures were significantly increased in this subset of CABG patients, our study could not show direct evidence of the underlying pathomechanisms leading to the increased surgical risk. However, several possible perioperative coronary pathomechanisms should be addressed. The significantly worse postoperative outcome after surgery may be a consequence of the increased stent load leading to coronary endothelial injury with intimal hyperplasia due to repeated stent lesions.13,14 In addition, the vascular wall, with its dysfunctional and denuded coronary endothelium, activates cytokines and the contact system, subsequently leading to an inflammatory response with an accumulation of platelets and neutrophils causing microvascular thrombotic obstruction and/or distal microembolization owing to platelet microaggregates.15,16 Moreover, coronary side-branch obstruction or occlusion due to multiple contiguous and overlapping stents (“stent jail”) may lead to a compromised collateral blood flow,17 affecting coronary runoff and the patency rate of the bypass grafts. Finally, owing to the fact that there is no possibility of removing the stent-graft intraoperatively, the graft anastomosis has to be inserted more distally where target vessels are smaller in diameter, possibly leading again to a compromised runoff and/or impaired patency rate.
The present study was retrospective in design, and the preoperative stent load of the patients was not blinded; thus, our results may have been influenced by treatment bias. However, statistical adjustment was performed by using propensity-score matching and logistic-regression analysis. Our study encompasses the experience at a single tertiary medical center; therefore, the generalizability of our findings may not extend to all clinical centers performing CABG. Furthermore, our study has no direct evidence of postoperative stent thrombosis, except for 3 patients in whom stent thrombosis was evidenced at autopsy. Very limited knowledge exists about postoperative coronary pathomechanisms associated with multiple stents. Moreover, a systematic evaluation of the psychological burden of undergoing multiple stent procedures over time is lacking, and only limited information exists about the quality of life of these patients.18 Finally, the cost-effectiveness of multiple PCI procedures compared with CABG has to be questioned.
Patients with prior multiple PCI procedures undergoing isolated first-time elective CABG have a significantly increased risk for in-hospital mortality and MACEs compared with patients without or with only a single previous PCI. There is no doubt that PCI may provide successful myocardial revascularization. However, in an increasing subset of patients, PCI procedures have to be repeated over time, may delay CABG, and may additionally increase the surgical risk. Thus, patients who potentially undergo multiple PCI procedures should be encouraged to undergo CABG instead of more PCI procedures. To what extent the increasing number of PCIs with drug-eluting stent-grafts will have a different influence on surgical outcome after CABG has to be investigated in the near future.
Presented at the American Heart Association Scientific Sessions, Dallas, Tex, November 13–16, 2005.
Mercado N, Wijns W, Serruys PW, Sigwart U, Flather MD, Stables RH, O’Neill WW, Rodriguez A, Lemos PA, Hueb WA. One-year outcomes of coronary artery bypass graft surgery versus percutaneous coronary intervention with multiple stenting for multisystem disease: a meta-analysis of individual patient data from randomized clinical trials. J Thorac Cardiovasc Surg. 2005; 130: 512–519.
Thielmann M, Massoudy P, Schmermund A, Neuhauser M, Marggraf G, Kamler K, Herold U, Aleksic I, Mann K, Haude M, Heusch G, Erbel R, Jakob H. Diagnostic discrimination between graft-related and non-graft-related perioperative myocardial infarction with cardiac troponin I after coronary artery bypass surgery. Eur Heart J. 2005; 26: 2440–2447.
Cutlip DE, Chhabra AG, Baim DS, Chauhan MS, Marulkar S, Massaro J, Bakhai A, Cohen DJ, Kuntz RE, Ho KKL. Beyond restenosis: five-year clinical outcomes from second-generation coronary stent trials. Circulation. 2004; 110: 1226–1230.
Hofma SH, van der Giessen WJ, van Dalen BM, Lemos PA, McFadden EP, Sianos G, Ligthart JMR, van Essen D, De Feyter PJ, Serruys PW. Indication of long-term endothelial dysfunction after sirolimus-eluting stent implantation. Eur Heart J. 2006; 27: 166–170.
Herrmann J. Peri-procedural myocardial injury: 2005 update. Eur Heart J. 2005; 26: 2493–2519.
Toutouzas K, Colombo A, Stefanadis C. Inflammation and restenosis after percutaneous coronary interventions. Eur Heart J. 2004; 25: 1679–1687.
Alfonso F, Hernandez C, Perez-Vizcayno MJ, Hernandez R, Fernandez-Ortiz A, Escaned J, Banuelos C, Sabate M, Sanmartin M, Fernandez C, Macaya C. Fate of stent-related side branches after coronary intervention in patients with in-stent restenosis. J Am Coll Cardiol. 2000; 36: 1549–1556.
Spertus JA, Nerella R, Kettlekamp R, House J, Marso S, Borkon AM, Rumsfeld JS. Risk of restenosis and health status outcomes for patients undergoing percutaneous coronary intervention versus coronary artery bypass graft surgery. Circulation. 2005; 111: 768–773.