- Email: [email protected]
Baseline clinical and angiographic variables associated with long-term outcome after successful intracoronary stent implantation
Baseline Clinical and Angiographic Variables Associated With Long-Term Outcome After Successful Intracoronary Stent Implantation Verghese Mathew, MD, Diane E. Grill, MS, Christopher G. Scott, Kirk N. Garratt, MD, and David R. Holmes, Jr., MD
BS,
Although randomized studies have demonstrated improved outcomes with stents over balloon angioplasty in straightforward coronary narrowings in low-risk patients, this advantage is less clear for complex lesions and high-risk patients. This study was designed to identify clinical and angiographic variables that are associated with long-term outcome after stent implantation. We identified 1,709 patients undergoing successful stent placement without in-hospital major adverse events. We analyzed clinical, lesional, and procedural variables to determine their correlation with outcome. Mean duration of follow-up was 1.6 ⴞ 1.4 years. Cox proportional-hazards models and stepwise methods were used to assess which covariates were potentially related to each end point. The occurrence of death/myocardial infarction (MI) was associated with any history of congestive heart failure (relative risk [RR] 3.3, 95% confidence interval [CI] 2.3 to 4.7, p <0.0001), procedure within 24 hours of MI (RR 2.3, CI 1.3 to 4.1, p ⴝ 0.0048), vein graft intervention (RR 1.8, CI 1.3 to 2.6, p ⴝ 0.0007), and prior MI (RR 1.8, CI 1.2 to 2.6, p ⴝ 0.004). Repeat
revascularization was associated with multivessel stent placement (RR 1.8, CI 1.2 to 2.8, p ⴝ 0.006) and stent for abrupt closure (RR 1.7, CF 1.1 to 2.7, p ⴝ 0.03), but was less frequent with de novo lesions and right coronary artery lesions (RR 0.6, CI 0.5 to 0.8, p ⴝ 0.0007, and RR 0.8, CI 0.6 to 1.0, p ⴝ 0.05, respectively). The cumulative end point of death/MI/repeat revascularization was associated with congestive heart failure (RR 1.7, CI 1.3 to 2.2, p <0.0001), multivessel stent placement (RR 1.6, CI 1.1 to 2.3, p ⴝ 0.03), warfarin therapy (RR 1.4, CI 1.2 to 1.8, p ⴝ 0.001), and procedure within 24 hours of MI (RR 1.5, CI 1.1 to 2.1, p ⴝ 0.02), but was less frequent with complete revascularization and right coronary artery intervention (RR 0.8, CI 0.7 to 0.99, p ⴝ 0.04, and RR 0.7, CI 0.6 to 0.9, p ⴝ 0.009, respectively). Thus, this study demonstrates that there are readily identifiable characteristics in patients treated successfully with stents that are associated with long-term outcome. 䊚1999 by Excerpta Medica, Inc. (Am J Cardiol 1999;84:789 –794)
oronary stent implantation is increasingly used to improve procedural outcome and reduce the likeC lihood of repeat revascularization procedures after
METHODS
percutaneous coronary intervention. Although the early randomized prospective studies demonstrating the advantages of stents over balloon angioplasty included relatively straightforward coronary lesions in low-risk patients,1,2 the use of stents in current clinical practice has evolved to include complex lesions and high-risk patients. However, the superiority of stent implantation over angioplasty in such cases has not been clearly demonstrated, and extrapolation of data from the initial randomized trials to such cases may not be appropriate. The present study was designed to analyze a broad range of clinical, angiographic, and lesional variables and identify those that are associated with long-term clinical outcome after successful percutaneous coronary interventions using stents. From the Division of Cardiovascular Diseases and Internal Medicine, Mayo Clinic, Rochester, Minnesota. Manuscript received February 24, 1999; revised manuscript received and accepted May 11, 1999. Address for reprints: Verghese Mathew, MD, Division of Cardiovascular Diseases, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905. E-mail: [email protected]. ©1999 by Excerpta Medica, Inc. All rights reserved. The American Journal of Cardiology Vol. 84 October 1, 1999
We performed a retrospective analysis of the Mayo Clinic cardiac catheterization database from October 1989 through December 1996 and identified all patients undergoing coronary revascularization in which stents were placed in at least 1 treated segment. From this group, the study cohort was defined as all angiographically successful procedures (reduction in luminal diameter stenosis of ⱖ20% with a final residual diameter stenosis of ⬍50%) without the occurrence of death, coronary artery bypass graft (CABG) surgery, or Q-wave myocardial infarction (MI) during initial hospitalization. We analyzed clinical, lesional, and procedural variables to identify factors associated with better or worse long-term outcome. Clinical characteristics included patient age, gender, MI within 24 hours of procedure, angina of Canadian Cardiovascular Society Class ⱖ3, prior CABG, prior MI (not within 24 hours of procedure), current smoker, prior history of congestive heart failure (CHF) or CHF at presentation of procedural admission, ejection fraction ⬍50%, diabetes, hypertension, multivessel coronary disease, and complete revascularization at the time of stent procedure. Angiographic and procedural variables included vessel treated (right, circumflex, left 0002-9149/99/$–see front matter PII S0002-9149(99)00454-3
789
anterior descending artery, or vein graft), proximal versus distal segment, discrete (ⱕ10 mm) versus diffuse (⬎10 mm) stenosis, ostial lesion, eccentric lesion, calcium in stenosis, calcium in vessel, bifurcation lesion, American College of Cardiology/American Heart Association type B lesion, American College of Cardiology/American Heart Association type C lesion, multiple (ⱖ2) stents in a vessel, thrombus-containing lesion, nominal stent size 2.5 mm, unplanned stent (stent for abrupt closure, dissection, or suboptimal angioplasty result), de novo lesion, total occlusion, abciximab use, warfarin use, low molecular weight heparin use, and multivessel intervention. Definitions: Acute MI was considered to have occurred when at least 2 of the following 3 criteria were met: (1) chest pain lasting ⬎30 minutes, (2) persistent electrocardiographic changes suggestive of ischemia, or (3) greater than or equal to twofold elevations in serum creatine kinase levels with a corresponding increase in the MB isoform. Multivessel disease was defined as the presence of ⱖ70% stenosis of the luminal diameter in a major epicardial coronary artery with ⱖ50% stenosis in a second major epicardial vessel (2-vessel disease) or both of the other epicardial vessels (3-vessel disease). Complete revascularization was defined as successful percutaneous treatment with no residual stenosis of ⱖ70% in any coronary artery. Follow-up of all patients was obtained in a prospective fashion. The study protocol was approved by the Mayo Clinic Institutional Review Board. Angioplasty technique for stent implantation: All angioplasty procedures were performed using standard techniques as described previously.3 Intracoronary stent implantation techniques evolved over the study period. Gianturco-Roubin (Cook Inc., Bloomington, Indiana) and Wiktor (Medtronic Inc., Minneapolis, Minnesota) stents were generally placed across target lesions using 0.014-inch extra support guidewires for delivery, as described previously.4,5 Coronary PalmazSchatz stents (SDS, Johnson & Johnson Interventional Systems, Warren, New Jersey) were positioned within the delivery sheath over 0.014-inch extra support guidewires according to standard techniques.6 Delivery balloons were inflated to nominal pressure. As of 1994, stent delivery has been followed routinely by high pressure (ⱖ14 atm) balloon inflations using noncompliant or minimally compliant balloons at sizes equivalent to or slightly larger than nominal stent size, with the goal of achieving a residual diameter stenosis of 0%. Intravascular ultrasound was used to guide stent deployment at the discretion of the operator. Anticoagulation: All patients received preprocedural oral aspirin (325 mg) and received intraprocedural heparin to achieve an activated clotting time ⬎300 seconds. Until 1994, patients received dextran immediately before the procedure in accordance with the stent manufacturer’s instructions. Our protocol for stent implantation until late 1994 included warfarin (target International Normalized Ratio of 2.0 to 4.0), dipyridamole 75 mg 3 times daily, and 81 to 325 mg of aspirin daily indefinitely. Therapy with heparin was discontinued after the procedure, and sheaths were 790 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 84
removed when the activated clotting time was ⬍160 seconds. Administration of heparin was resumed after sheath removal and continued until the target International Normalized Ratio was achieved. Between late 1994 and early 1995, warfarin was administered only to patients with suboptimal stent deployment as assessed by coronary angiography or intravascular ultrasonography, with standard poststent therapy consisting of ticlopidine 250 mg twice daily for 4 to 6 weeks and aspirin 81 to 325 mg/day. In patients receiving only antiplatelet therapy after the procedure, no further heparin was administered after the procedure. Later, warfarin was eliminated completely and patients believed to be at increased risk for subacute stent occlusion received subcutaneous injections of low molecular weight heparin 30 mg or 60 mg twice daily for 10 to 14 days in addition to aspirin and ticlopidine. Dipyridamole use was discontinued in late 1994. Adverse events: Adverse events of death, nonfatal MI, repeat percutaneous intervention of the target lesion, and CABG surgery were recorded for each patient. Adverse event rates were reported as follows: death/nonfatal MI; repeat percutaneous target lesion revascularization/CABG; and death/nonfatal MI/ repeat percutaneous target lesion revascularization/ CABG. The cumulative end points were calculated on a per patient basis. Follow-up: The mean duration of follow-up was 1.6 ⫾ 1.4 years (range 1 day to 7.1 years); 49 patients did not have complete follow-up and were excluded from multivariate analysis. Patients were contacted by telephone by the nurse stent coordinator weekly for the first 8 weeks after the procedure. Patients were also contacted 6 and 12 months after the procedure, and annually thereafter. The case records of all patients followed at this institution were reviewed. Documentation of adverse events that occurred at other institutions during the follow-up period were obtained from the local physicians and hospital records. Of 263 patients who underwent repeat percutaneous intervention, records of 181 patients were known to be target lesion revascularization procedures, 66 were nontarget lesions, and in 16 patients, the record was incomplete as to whether the procedure involved a previously treated segment or a new lesion/segment. These 16 patients were handled in 2 ways: (1) included as target lesion procedures for multivariate analysis of the end point of repeat revascularization and the composite end point of death/nonfatal MI/repeat revascularization (presented data); and (2) excluded from analysis. Exclusion of these patients did not alter the results. Statistical analysis: Cox proportional-hazards models and stepwise methods were used to assess which covariates were potentially related to each end point.
RESULTS
Clinical/angiographic characteristics: One thousand seven hundred nine patients were identified who had undergone successful stent implantation without adverse events during initial hospitalization. Table I lists baseline demographics of the study population. A OCTOBER 1, 1999
TABLE I Clinical and Angiographic Characteristics of the Study Population Number of patients Number of lesions Age (yrs) Men Acute myocardial infarction Canadian heart class ⱖ3 Prior coronary artery bypass grafting Prior myocardial infarction Congestive heart failure on presentation History of congestive heart failure Ejection fraction (%) Active smoker Diabetes mellitus Systemic hypertension Total cholesterol ⬎250 mg/dl Multivessel CAD Complete revascularization Multivessel intervention Vessel treated Right coronary artery Left anterior descending Left circumflex Saphenous vein graft
1,709 2,669 63.6 ⫾ 11.3 1,254 (73.4%) 150 (8.8%) 1,005 (58.8%) 468 (27.4%) 745 (44.2%) 123 (7.2%) 215 (12.7%) 63.6 ⫾ 15.2 318 (18.7%) 333 (19.6%) 977 (57.6%) 906 (53%) 1,083 (66.3%) 969 (56.7%) 270 (15.8%) 561 665 338 273
(32.8%) (38.9%) (19.8%) (16.0%)
CAD ⫽ coronary artery disease.
TABLE II Lesion Characteristics Percent of Lesions Discrete lesions Ostial Proximal Eccentric Moderate and severe calcium lesion Moderate and severe calcium vessel Bifurcation lesion ⱖ2 stents/vessel Angiographic evidence of thrombus ACC/AHA B lesion ACC/AHA C lesion Nominal stent size ⫽ 2.5 mm Unplanned stenting De novo lesion Total occlusion
40.7% 10.0% 62.8% 67.3% 32.1% 40.6% 6.5% 14.4% 18.1% 45.1% 34.9% 2.7% 42.8% 80.5% 13.5%
ACC/AHA ⫽ American College of Cardiology/American Heart Association.
significant proportion of patients had a prior history of coronary disease as manifest by the frequency of prior MI and prior coronary artery bypass grafting. Although multivessel disease was common in the series, multivessel intervention was infrequent. Lesion characteristics are listed in Table II. Complex lesion characteristics were common in this series, although most lesions were de novo. Abciximab was given to 15.4% of patients, and warfarin to 36.9% of patients after the procedure. Clinical events: During follow-up, death occurred in 86 patients, MI in 93. Target lesion revascularization (percutaneous transluminal coronary angioplasty or CABG) was performed in 280 patients. Kaplan-Meier
TABLE III Twelve-Month Event-Free Survival Rates
Freedom from death/MI Freedom from repeat percutaneous revascularization of the target lesion or CABG Freedom from death/MI/repeat percutaneous intervention of the target lesion or CABG
Survival
95% CI
92.5% 84.6%
91.2%–93.8% 82.8%–86.5%
80.2%
78.2%–82.2%
estimates of event-free survival at 12 months are listed in Table III. Univariate and multivariate analyses: Univariate correlates of the cumulative end points of death/MI, percutaneous target vessel revascularization/CABG, and death/MI/percutaneous target vessel revascularization/CABG are listed in Tables IV, V, and VI6, respectively. Stepwise regression analysis of clinical, angiographic, and lesional variables was performed for the cumulative end points of death/nonfatal MI, repeat percutaneous target lesion revascularization/CABG, and death/nonfatal MI/repeat percutaneous target lesion revascularization/CABG. Of the clinical, lesion, and procedural variables examined, the occurrence of death/nonfatal MI (Figure 1) was associated with any history of CHF (previous or on presentation) (RR 3.3, 95% confidence interval [CI] 2.3 to 4.7, p ⬍0.0001), MI within 24 hours of procedure (RR 2.3, CI 1.3 to 4.1, p ⫽ 0.0048), vein graft intervention (RR 1.8, CI 1.3 to 2.6, p ⫽ 0.0007), previous MI (RR 1.8, CI 1.2 to 2.6, p ⫽ 0.004). There was a nonsignificant trend toward increased events with warfarin use after the procedure (RR 1.4, CI 0.97 to 1.94, p ⫽ 0.07). Percutaneous target lesion revascularization or CABG (Figure 2) was associated with multivessel stent placement (RR 1.8, CI 1.2 to 2.8, p ⫽ 0.006); this end point was less frequent in de novo lesions and in right coronary artery interventions (RR 0.6, CI 0.5 to 0.8, p ⫽ 0.0007, and RR 0.8, CI 0.6 to 1.0, p ⫽ 0.05, respectively). A nonsignificant trend toward a need for greater repeat revascularization was noted when MI had occurred within 24 hours of the procedure (RR 1.4, CI 0.98 to 2.1, p ⫽ 0.07). Although unplanned stent placement generally was not correlated with repeat revascularization, the subset of stent procedures for abrupt vessel closure after percutaneous transluminal coronary angioplasty was associated with an increased risk of repeat revascularization (RR 1.7, CI 1.1 to 2.7, p ⫽ 0.03). The cumulative end point of death/nonfatal MI/ repeat percutaneous target lesion intervention/CABG (Figure 3) was associated with CHF (RR 1.7, CI 1.3 to 2.2, p ⬍0.0001), multivessel stent placement (RR 1.6, CI 1.1 to 2.3, p ⫽ 0.03), warfarin after the procedure (RR 1.4, CI 1.2 to 1.8, p ⫽ 0.001), or MI within 24 hours of the procedure (RR 1.5, CI 1.1 to 2.1, p ⫽ 0.02); the frequency of this cumulative end point was reduced in complete revascularization or right coro-
CORONARY ARTERY DISEASE/CLINICAL OUTCOMES WITH INTRACORONARY STENTS
791
TABLE IV Univariate Correlates of Death/MI 95% Confidence Interval Effect Age Prior CABG graft Prior MI Prior history of CHF Diabetes Multivessel disease Ejection fraction ⬍50% Proximal lesion Right coronary artery Circumflex lesion Graft lesion ACC/AHA type C lesion Warfarin
Risk Ratio
Lower Limit
Upper Limit
p Value
1.022 2.099 2.194 4.223 1.782 1.692 2.825 0.502 0.553 0.669 2.462 1.790 1.549
1.007 1.538 1.587 3.046 1.264 1.169 1.738 0.367 0.379 0.427 1.771 1.279 1.105
1.037 2.863 3.032 5.855 2.513 2.448 4.590 0.686 0.808 1.050 3.422 2.504 2.171
0.005 0.0001 0.0001 0.0001 0.001 0.005 0.0001 0.0001 0.002 0.08 0.0001 0.0007 0.01
Abbreviation as in Table II.
target vessel) were included, diabetes mellitus was associated with an increased risk of the cumulative end point of death/MI/CABG/repeat percutaneous intervention (RR 1.3, CI 1.0 to 1.6, p ⫽ 0.04) and percutaneous intervention/CABG (RR 1.4, CI 1.08 to 1.81, p ⫽ 0.012). Additionally, multiple stent implantation within a single vessel, which has been associated with higher rates of target lesion revascularization in other series, was not found to be predictive of repeat target lesion percutaneous intervention in this series (RR 1.4, CI 0.9 to 2.0, p ⫽ 0.11).
DISCUSSION The current study demonstrates that several easily identifiable clinical, procedural, and lesion character95% Confidence Interval istics, which in large part are known Effect Risk Ratio Lower Limit Upper Limit p Value before performing a percutaneous revascularization procedure, are assoMen 1.290 0.973 1.711 0.08 Angina ⱖclass III 1.301 1.015 1.668 0.04 ciated with long-term clinical events Stent for abrupt vessel closure 1.7 1.1 2.7 0.03 after coronary stent placement. Thrombus 0.739 0.550 0.994 0.046 Death/myocardial infarction: The Bifurcation 1.421 0.967 2.088 0.07 presence of CHF and/or left ventricLeft anterior descending 1.284 1.014 1.626 0.04 lesion ular dysfunction has been clearly asRight coronary artery lesion 0.809 0.624 1.049 0.11 sociated with increased morbidity Warfarin use 1.262 0.987 1.613 0.06 and mortality in patients with coronary artery disease.7,8 The current findings are in accord with pooled TABLE VI Univariate Correlates of Death/MI/Repeat Percutaneous Intervention data from 4 interventional trials and Target Lesion/CABG a large single center database where it has been demonstrated that a his95% Confidence Interval tory of CHF is associated with early Effect Risk Ratio Lower Limit Upper Limit p Value and intermediate-term mortality in Prior CABG 1.376 1.115 1.699 0.003 patients undergoing percutaneous Prior history of CHF 1.811 1.404 2.336 0.0001 coronary intervention.9 Similarly, a CHF on presentation 1.726 1.247 2.387 0.001 history of coronary artery disease Diabetes 1.375 1.084 1.743 0.009 manifesting as an acute MI, or a prior Multivessel disease 1.388 1.100 1.750 0.006 Complete revascularization 0.735 0.601 0.899 0.003 history of MI, has been demonstrated Ejection fraction ⬍50% 1.438 1.035 1.998 0.03 to impact survival.10,11 However, the Ostial lesion 1.402 1.060 1.855 0.02 impact on mortality in these settings Right coronary artery lesion 0.707 0.562 0.889 0.003 is more likely a reflection of the presLeft anterior descending lesion 1.193 0.973 1.464 0.09 Graft lesion 1.419 1.113 1.810 0.005 ence of these characteristics rather than a consequence of the stent procedure. The finding that saphenous vein graft interventions are also asnary artery intervention (RR 0.8, CI 0.7 to 0.99, p ⫽ sociated with an increased rate of adverse events has 0.04, and RR 0.7, CI 0.6 to 0.9, p ⫽ 0.009, respectively). previously been reported,12,13 although a portion of Diabetes mellitus, which has been associated with these events has been shown to occur in association increased rates of adverse events after percutaneous with previously mild untreated stenoses within the intervention, was not found to be independently asso- vein graft.14 Repeat revascularization: Multivessel intervention ciated with the occurrence of death/MI in this population (RR 1.2, CI 0.85 to 1.8, p ⫽ 0.28), target lesion has clearly been established to increase the likelihood revascularization (RR 1.1, CI 0.7 to 1.6, p ⫽ 0.7), or of repeat revascularization procedures after percutadeath/MI/CABG/repeat target lesion percutaneous in- neous transluminal coronary angioplasty.15 Although tervention (RR 1.2, CI 0.9 to 1.5, p ⫽ 0.16). However, multivessel stent placement can be performed with a when all revascularization procedures (target and non- high degree of procedural success and a low rate of TABLE V Univariate Correlates of Repeat Percutaneous Intervention of Target Lesion/CABG
792 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 84
OCTOBER 1, 1999
FIGURE 1. Variables associated with death/nonfatal MI during the course of long-term follow-up after successful stent implantation.
FIGURE 2. Variables associated with repeat revascularization (percutaneous revascularization of the target lesion or coronary artery bypass grafting) during the course of long-term follow-up after successful stent implantation. RCA ⴝ right coronary artery.
intermediate-term events,16 the finding that the need for repeat revascularization is increased is likely a combination of restenosis and disease progression in previously untreated coronary segments.13 Additionally, the finding that de novo lesions are associated with a lower likelihood of repeat revascularization is in keeping with previous studies.17 Interestingly, right coronary artery interventions were associated with a decreased likelihood of repeat revascularization, which was apparent in the Stent Restenosis Study1 as well. Notably, the presence of diabetes mellitus did not portend an increased risk of target lesion revascularization. Studies have demonstrated an increased rate of angiographic restenosis after stent placement in diabetics versus nondiabetics18,19 due to exaggerated neointimal hyperplasia,18 although it has also been reported that restenosis rates are similar in diabetics versus nondiabetics after stent placement due to reduction in negative arterial remodeling after stenting.20 Importantly, the current series evaluates clinically driven target lesion revascularization and may potentially have missed cases of angiographic restenosis that were not apparent clinically. In addition, the
finding in this series that diabetes was associated with increased need for repeat revascularization rates when all repeat revascularization procedures (target lesion and nontarget lesion) were included supports the observation from the Bypass Angioplasty Revascularization Investigation that diabetes was associated with a higher likelihood of new lesions during follow-up in treated and untreated vessels than in nondiabetics.21 Additionally, the current series did not identify multiple stent implantation in a single vessel as a correlate of repeat revascularization. Numerous studies have demonstrated that multiple contiguous stent implantation (ⱖ2 or 3 stents) is associated with higher rates of target lesion revascularization, although these reports included a large proportion of patients with severe dissections,22 or did not attempt to establish the independent effect of multiple stents on target lesion revascularization. Current data are in accord with recently published data that report similar target lesion revascularization rates in patients with ⱖ3 contiguous stents compared with patients with 1 or 2 stents.23 Abrupt vessel closure but not unplanned stent placement was generally associated with greater need for repeat vascularization, which is in keeping with recent data.24 Death, MI, repeat target lesion percutaneous intervention, or CABG: The defi-
nition of completeness of revascularization varies among published studies on angioplasty as well as surgery, although there is ample evidence to suggest that complete revascularization is associated with improved long-term outcome,25 in keeping with the current study. Interestingly, the finding that warfarin use was associated with a greater frequency of the composite end point is in accord with previous data that demonstrated an increase in death, MI, and repeat revascularization at 30 days in patients treated with warfarin after the procedure compared with those treated with only antiplatelet agents.26 By design, the current study excluded patients who encountered adverse events during the initial hospital stay; the long-term effect of warfarin on clinical outcome has not been reported and may warrant further investigation. Although our statistical modeling was designed to eliminate the possibility of confounding variables, the possibility that warfarin use correlated with adverse variables cannot be ruled out with certainty in this retrospective analysis. Impact of lesion characteristics on outcome: Notably, of all the lesion characteristics assessed, only vein graft interventions, multivessel stenting, and restenosis lesions were associated with adverse events. However, among univariate correlates of adverse events,
CORONARY ARTERY DISEASE/CLINICAL OUTCOMES WITH INTRACORONARY STENTS
793
7. Emond M, Mock MB, Davis KB, Fisher LD, Holmes DR Jr, Chaitman BR, Kaiser GC, Alderman E, Killip T III. Long-term survival of medically treated patients in coronary artery surgery study (CASS) registry. Circulation 1994;90:2645–2657. 8. Judge KW, Pawitan Y, Caldwell J, Gersh BJ, Kennedy JW. Congestive heart failure symptoms in patients with preserved left ventricular systolic function: analysis of the CASS registry. J Am Coll Cardiol 1991;18:377–382. 9. Anderson RD, Ohman EM, Holmes DR, Harrington RA, Barsness GW, Wildermann NM, Phillips HR, Topol EJ, Califf RM. Prognostic value of congestive heart failure history in patients undergoing percutaneous coronary interventions. J Am Coll Cardiol 1998; 32:936 –941. 10. Maseri A. Mode of first presentation: acute, chronic, and quiescent phases of IHD. In: Ischemic Heart Disease. New York: Churchill Livingstone, 1995;323–335. 11. Goldstein S, Friedman L, Hutchinson R, Canner P, Romhilt D, Schlant R, Sobrino R, Verter J, Wasserman A. Timing, mechanism and clinical setting of witnessed FIGURE 3. Variables associated with a cumulative end point of death/nonfatal MI/ deaths in postmyocardial infarction patients. J Am Coll repeat revascularization during the course of long-term follow-up after successful Cardiol 1984;3:1111–1117. 12. Waksman R, Weintraub WS, Ghazzal Z, Scott NA, stent implantation. RCA ⴝ right coronary artery. Shen Y, King SB III, Douglas JS Jr. Short- and long-term outcome of narrowed saphenous vein bypass graft: a comof Palmaz-Schatz stent, directional coronary atherectomy, and balloon several lesion characteristics were found to influence parison angioplasty. Am Heart J 1997;134:274 –281. outcome, which is in accord with prior studies.24,25 13. Laham RJ, Ho KK, Baim DS, Kuntz RE, Cohen DJ, Carrozza JP Jr. The fact that many of these variables were no longer Multivessel Palmaz-Schatz stenting: early results and one-year outcome. J Am Coll Cardiol 1997;30:180 –185. significant in multivariate analysis may in part be 14. Ellis SG, Brener SJ, DeLuca S, Tuzcu EM, Raymond RE, Whitlow PL, Topol related to the exclusion of patients with short-term EJ. Late myocardial ischemic events after saphenous vein graft intervention— of initially “nonsignificant” vein graft lesions. Am J Cardiol 1997; adverse events in this study. Such variables identified importance 79:1460 –1464. with univariate analysis may, however, shed light on 15. Lambert M, Bonan R, Cote G, Crepeau J, deGuise P, Lesperance J, David PR, Waters DD. Early results, complications and restenosis rates after multilesion and important variables to evaluate prospectively. percutaneous transluminal coronary angioplasty. Am J Cardiol 1987; In cases in which these adverse clinical or lesion multivessel 60:788 –791. characteristics are present, clinical follow-up should 16. Mathew V, Rihal CS, Berger PB, Bell MR, Garratt KN, Holmes DR Jr. outcome of patients undergoing multivessel coronary stent implantation. be increased. Adjunctive therapies to reduce recurrent Clinical Int J Cardiol 1998;64:1–7. ischemic events may also be considered; for example, 17. Black AJR, Anderson V, Rabin GS, Powelson SW, Douglas JS, Kinky SB. intracoronary brachytherapy appears to be a promising Repeat coronary angioplasty: correlates of a second restenosis. J Am Coll Cardiol 1988;11:714 –718. modality to reduce restenosis,27 and may thereby re- 18. Kornowski R, Mintz GS, Kent KM, Pichard AD, Satler LF, Bucher TA, Hong MK, Popma JJ, Leon MB. Increased restenosis in diabetes mellitus after coronary duce adverse clinical events. is due to exaggerated intimal hyperplasia: a serial intravascular Study limitations: The current study is a retrospec- interventions ultrasound study. Circulation 1997;95:1366 –1369. tive analysis of a single center experience, and is 19. Kastrati A, Schomig A, Elezi S, Schuhlen H, Dirschinger J, Hadamitzky M, therefore subject to the bias innate to such studies. Wehinger A, Hausleiter J, Walter H, Neumann F-J. Predictive factors of resteafter coronary stent placement. J Am Coll Cardiol 1997;30:1428 –1436. Because this includes our earlier stent experience, nosis 20. Van Belle E, Bauters C, Hubert E, Bodart J-C, Abolmaali K, Meurice T, warfarin was used in a significant portion of patients, McFadden EP, Lablanche J-M, Bertrand ME. Restenosis rates in diabetic paa comparison of coronary stenting and balloon angioplasty in native which no longer is the standard of postprocedure tients; coronary vessels. Circulation 1997;96:1454 –1460. anticoagulation. Follow-up angiographic data were 21. Rozenman Y, Sapoznikov D, Mosseri M, Gilon D, Lotan C, Nassar H, Weiss not obtained in most patients; however, this study AT, Hasin Y, Gotsman MS. Long-term angiographic follow-up of coronary angioplasty in patients with diabetes mellitus: a clue to the explanation of reflects current practice in which patients are usually balloon the results of the BARI study. J Am Coll Cardiol 1997;30:1420 –1425. evaluated clinically, and subject to repeat angiography 22. Mathew V, Hasdai D, Holmes DR, Garratt KN, Bell MR, Lerman A, Melby S, Grill DE, Berger PB. Clinical outcome of patients undergoing endoluminal only if there is evidence of recurrent ischemia. coronary artery reconstruction with three or more stents. J Am Coll Cardiol 1997;30:676 – 681. 23. Kornowski R, Mehran R, Hong MK, Satler LF, Pichard AD, Kent KM, Mintz GS, Waksman R, Laird JR, Lansky AJ, Bucher TA, Popma JJ, Leon MB. 1. Fischman DL, Leon MB, Baim DS, Schatz RA, Savage MP, Penn I, Detre K, Procedural results and late clinical outcomes after placement of three or more Veltri L, Ricci D, Nobuyoshi M. A randomized comparison of coronary-stent stents in single coronary lesions. Circulation 1998;97:1355–1361. placement in the treatment of coronary artery disease. Stent Restenosis Study 24. Piana RN, Ahmed WH, Chaitman B, Ganz P, Kinlay S, Strony J, Adelman B, Investigators. N Engl J Med 1994;331:496 –501. Bittl JA, on Behalf of the Hirulog Angioplasty Study Investigators. Effect of transient 2. Serruys PW, de Jaegere P, Kiemeneij F, Macaya C, Rutsch W, Heyndrickx G, abrupt vessel closure during otherwise successful angioplasty for unstable angina on Emanuelsson H, Marco J, Legrand V, Materne P. A comparison of balloonclinical outcome at six months. J Am Coll Cardiol 1999;33:73–78. expandable-stent implantation with balloon angioplasty in patients with coronary 25. Bell MR, Gersh BJ, Schaff HV, Holmes DR Jr, Fisher LD, Alderman EL, artery disease. Benestent Study Group. N Engl J Med 1994;331:489 – 495. Meyers WO, Parsons LS, Reeder GS. Effect of completeness of revascularization 3. Holmes DR Jr. Technical aspects. In: Vliesttra RE, Holmes DR Jr, eds. PTCA: on long-term outcome of patients with three-vessel disease undergoing coronary Percutaneous Transluminal Coronary Angioplasty. Philadelphia, PA: FA Davis, artery bypass surgery. A report from the Coronary Artery Surgery Study (CASS) 1987:35– 48. Registry. Circulation 1992;86:446 – 457. 4. Cannon AD, Roubin GS. The Gianturco-Roubin stent. In: Topol EJ, ed. 26. Schomig A, Neumann FJ, Kastrati A, Schuhlen H, Blasini R, Hadamitzky M, Textbook of Interventional Cardiology. 2nd ed. Philadelphia, PA. WB Saunders, Walter H, Zitzmann-Roth EM, Richardt G, Alt E, Schmitt C, Ulm K. A random1994:712–726. ized comparison of antiplatelet and anticoagulant therapy after placement of coronary-artery stents. N Engl J Med 1996;334:1084 –1089. 5. Whitlow PL, de Jaegere PP, Serruys PW. The Wiktor stent. In: Topol EJ, ed. 27. Teirstein PS, Massullo V, Jani S, Popma JJ, Mintz GS, Russo RJ, Schatz RA, Textbook of Interventional Cardiology. 2nd ed. Philadelphia, PA. WB Saunders, Guarneri EM, Steuterman S, Morris NB, Leon MB, Tripuraneni P. Catheter1994:727–741. based radiotherapy to inhibit restenosis after coronary stenting. N Engl J Med 6. Ellis SG. The Palmaz-Schatz stent. In: Topol EJ, ed. Textbook of Interven1997;336:1697–1703. tional Cardiology. 2nd ed. Philadelphia, PA. WB Saunders, 1994:702–711.
794 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 84
OCTOBER 1, 1999
BS,
Although randomized studies have demonstrated improved outcomes with stents over balloon angioplasty in straightforward coronary narrowings in low-risk patients, this advantage is less clear for complex lesions and high-risk patients. This study was designed to identify clinical and angiographic variables that are associated with long-term outcome after stent implantation. We identified 1,709 patients undergoing successful stent placement without in-hospital major adverse events. We analyzed clinical, lesional, and procedural variables to determine their correlation with outcome. Mean duration of follow-up was 1.6 ⴞ 1.4 years. Cox proportional-hazards models and stepwise methods were used to assess which covariates were potentially related to each end point. The occurrence of death/myocardial infarction (MI) was associated with any history of congestive heart failure (relative risk [RR] 3.3, 95% confidence interval [CI] 2.3 to 4.7, p <0.0001), procedure within 24 hours of MI (RR 2.3, CI 1.3 to 4.1, p ⴝ 0.0048), vein graft intervention (RR 1.8, CI 1.3 to 2.6, p ⴝ 0.0007), and prior MI (RR 1.8, CI 1.2 to 2.6, p ⴝ 0.004). Repeat
revascularization was associated with multivessel stent placement (RR 1.8, CI 1.2 to 2.8, p ⴝ 0.006) and stent for abrupt closure (RR 1.7, CF 1.1 to 2.7, p ⴝ 0.03), but was less frequent with de novo lesions and right coronary artery lesions (RR 0.6, CI 0.5 to 0.8, p ⴝ 0.0007, and RR 0.8, CI 0.6 to 1.0, p ⴝ 0.05, respectively). The cumulative end point of death/MI/repeat revascularization was associated with congestive heart failure (RR 1.7, CI 1.3 to 2.2, p <0.0001), multivessel stent placement (RR 1.6, CI 1.1 to 2.3, p ⴝ 0.03), warfarin therapy (RR 1.4, CI 1.2 to 1.8, p ⴝ 0.001), and procedure within 24 hours of MI (RR 1.5, CI 1.1 to 2.1, p ⴝ 0.02), but was less frequent with complete revascularization and right coronary artery intervention (RR 0.8, CI 0.7 to 0.99, p ⴝ 0.04, and RR 0.7, CI 0.6 to 0.9, p ⴝ 0.009, respectively). Thus, this study demonstrates that there are readily identifiable characteristics in patients treated successfully with stents that are associated with long-term outcome. 䊚1999 by Excerpta Medica, Inc. (Am J Cardiol 1999;84:789 –794)
oronary stent implantation is increasingly used to improve procedural outcome and reduce the likeC lihood of repeat revascularization procedures after
METHODS
percutaneous coronary intervention. Although the early randomized prospective studies demonstrating the advantages of stents over balloon angioplasty included relatively straightforward coronary lesions in low-risk patients,1,2 the use of stents in current clinical practice has evolved to include complex lesions and high-risk patients. However, the superiority of stent implantation over angioplasty in such cases has not been clearly demonstrated, and extrapolation of data from the initial randomized trials to such cases may not be appropriate. The present study was designed to analyze a broad range of clinical, angiographic, and lesional variables and identify those that are associated with long-term clinical outcome after successful percutaneous coronary interventions using stents. From the Division of Cardiovascular Diseases and Internal Medicine, Mayo Clinic, Rochester, Minnesota. Manuscript received February 24, 1999; revised manuscript received and accepted May 11, 1999. Address for reprints: Verghese Mathew, MD, Division of Cardiovascular Diseases, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905. E-mail: [email protected]. ©1999 by Excerpta Medica, Inc. All rights reserved. The American Journal of Cardiology Vol. 84 October 1, 1999
We performed a retrospective analysis of the Mayo Clinic cardiac catheterization database from October 1989 through December 1996 and identified all patients undergoing coronary revascularization in which stents were placed in at least 1 treated segment. From this group, the study cohort was defined as all angiographically successful procedures (reduction in luminal diameter stenosis of ⱖ20% with a final residual diameter stenosis of ⬍50%) without the occurrence of death, coronary artery bypass graft (CABG) surgery, or Q-wave myocardial infarction (MI) during initial hospitalization. We analyzed clinical, lesional, and procedural variables to identify factors associated with better or worse long-term outcome. Clinical characteristics included patient age, gender, MI within 24 hours of procedure, angina of Canadian Cardiovascular Society Class ⱖ3, prior CABG, prior MI (not within 24 hours of procedure), current smoker, prior history of congestive heart failure (CHF) or CHF at presentation of procedural admission, ejection fraction ⬍50%, diabetes, hypertension, multivessel coronary disease, and complete revascularization at the time of stent procedure. Angiographic and procedural variables included vessel treated (right, circumflex, left 0002-9149/99/$–see front matter PII S0002-9149(99)00454-3
789
anterior descending artery, or vein graft), proximal versus distal segment, discrete (ⱕ10 mm) versus diffuse (⬎10 mm) stenosis, ostial lesion, eccentric lesion, calcium in stenosis, calcium in vessel, bifurcation lesion, American College of Cardiology/American Heart Association type B lesion, American College of Cardiology/American Heart Association type C lesion, multiple (ⱖ2) stents in a vessel, thrombus-containing lesion, nominal stent size 2.5 mm, unplanned stent (stent for abrupt closure, dissection, or suboptimal angioplasty result), de novo lesion, total occlusion, abciximab use, warfarin use, low molecular weight heparin use, and multivessel intervention. Definitions: Acute MI was considered to have occurred when at least 2 of the following 3 criteria were met: (1) chest pain lasting ⬎30 minutes, (2) persistent electrocardiographic changes suggestive of ischemia, or (3) greater than or equal to twofold elevations in serum creatine kinase levels with a corresponding increase in the MB isoform. Multivessel disease was defined as the presence of ⱖ70% stenosis of the luminal diameter in a major epicardial coronary artery with ⱖ50% stenosis in a second major epicardial vessel (2-vessel disease) or both of the other epicardial vessels (3-vessel disease). Complete revascularization was defined as successful percutaneous treatment with no residual stenosis of ⱖ70% in any coronary artery. Follow-up of all patients was obtained in a prospective fashion. The study protocol was approved by the Mayo Clinic Institutional Review Board. Angioplasty technique for stent implantation: All angioplasty procedures were performed using standard techniques as described previously.3 Intracoronary stent implantation techniques evolved over the study period. Gianturco-Roubin (Cook Inc., Bloomington, Indiana) and Wiktor (Medtronic Inc., Minneapolis, Minnesota) stents were generally placed across target lesions using 0.014-inch extra support guidewires for delivery, as described previously.4,5 Coronary PalmazSchatz stents (SDS, Johnson & Johnson Interventional Systems, Warren, New Jersey) were positioned within the delivery sheath over 0.014-inch extra support guidewires according to standard techniques.6 Delivery balloons were inflated to nominal pressure. As of 1994, stent delivery has been followed routinely by high pressure (ⱖ14 atm) balloon inflations using noncompliant or minimally compliant balloons at sizes equivalent to or slightly larger than nominal stent size, with the goal of achieving a residual diameter stenosis of 0%. Intravascular ultrasound was used to guide stent deployment at the discretion of the operator. Anticoagulation: All patients received preprocedural oral aspirin (325 mg) and received intraprocedural heparin to achieve an activated clotting time ⬎300 seconds. Until 1994, patients received dextran immediately before the procedure in accordance with the stent manufacturer’s instructions. Our protocol for stent implantation until late 1994 included warfarin (target International Normalized Ratio of 2.0 to 4.0), dipyridamole 75 mg 3 times daily, and 81 to 325 mg of aspirin daily indefinitely. Therapy with heparin was discontinued after the procedure, and sheaths were 790 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 84
removed when the activated clotting time was ⬍160 seconds. Administration of heparin was resumed after sheath removal and continued until the target International Normalized Ratio was achieved. Between late 1994 and early 1995, warfarin was administered only to patients with suboptimal stent deployment as assessed by coronary angiography or intravascular ultrasonography, with standard poststent therapy consisting of ticlopidine 250 mg twice daily for 4 to 6 weeks and aspirin 81 to 325 mg/day. In patients receiving only antiplatelet therapy after the procedure, no further heparin was administered after the procedure. Later, warfarin was eliminated completely and patients believed to be at increased risk for subacute stent occlusion received subcutaneous injections of low molecular weight heparin 30 mg or 60 mg twice daily for 10 to 14 days in addition to aspirin and ticlopidine. Dipyridamole use was discontinued in late 1994. Adverse events: Adverse events of death, nonfatal MI, repeat percutaneous intervention of the target lesion, and CABG surgery were recorded for each patient. Adverse event rates were reported as follows: death/nonfatal MI; repeat percutaneous target lesion revascularization/CABG; and death/nonfatal MI/ repeat percutaneous target lesion revascularization/ CABG. The cumulative end points were calculated on a per patient basis. Follow-up: The mean duration of follow-up was 1.6 ⫾ 1.4 years (range 1 day to 7.1 years); 49 patients did not have complete follow-up and were excluded from multivariate analysis. Patients were contacted by telephone by the nurse stent coordinator weekly for the first 8 weeks after the procedure. Patients were also contacted 6 and 12 months after the procedure, and annually thereafter. The case records of all patients followed at this institution were reviewed. Documentation of adverse events that occurred at other institutions during the follow-up period were obtained from the local physicians and hospital records. Of 263 patients who underwent repeat percutaneous intervention, records of 181 patients were known to be target lesion revascularization procedures, 66 were nontarget lesions, and in 16 patients, the record was incomplete as to whether the procedure involved a previously treated segment or a new lesion/segment. These 16 patients were handled in 2 ways: (1) included as target lesion procedures for multivariate analysis of the end point of repeat revascularization and the composite end point of death/nonfatal MI/repeat revascularization (presented data); and (2) excluded from analysis. Exclusion of these patients did not alter the results. Statistical analysis: Cox proportional-hazards models and stepwise methods were used to assess which covariates were potentially related to each end point.
RESULTS
Clinical/angiographic characteristics: One thousand seven hundred nine patients were identified who had undergone successful stent implantation without adverse events during initial hospitalization. Table I lists baseline demographics of the study population. A OCTOBER 1, 1999
TABLE I Clinical and Angiographic Characteristics of the Study Population Number of patients Number of lesions Age (yrs) Men Acute myocardial infarction Canadian heart class ⱖ3 Prior coronary artery bypass grafting Prior myocardial infarction Congestive heart failure on presentation History of congestive heart failure Ejection fraction (%) Active smoker Diabetes mellitus Systemic hypertension Total cholesterol ⬎250 mg/dl Multivessel CAD Complete revascularization Multivessel intervention Vessel treated Right coronary artery Left anterior descending Left circumflex Saphenous vein graft
1,709 2,669 63.6 ⫾ 11.3 1,254 (73.4%) 150 (8.8%) 1,005 (58.8%) 468 (27.4%) 745 (44.2%) 123 (7.2%) 215 (12.7%) 63.6 ⫾ 15.2 318 (18.7%) 333 (19.6%) 977 (57.6%) 906 (53%) 1,083 (66.3%) 969 (56.7%) 270 (15.8%) 561 665 338 273
(32.8%) (38.9%) (19.8%) (16.0%)
CAD ⫽ coronary artery disease.
TABLE II Lesion Characteristics Percent of Lesions Discrete lesions Ostial Proximal Eccentric Moderate and severe calcium lesion Moderate and severe calcium vessel Bifurcation lesion ⱖ2 stents/vessel Angiographic evidence of thrombus ACC/AHA B lesion ACC/AHA C lesion Nominal stent size ⫽ 2.5 mm Unplanned stenting De novo lesion Total occlusion
40.7% 10.0% 62.8% 67.3% 32.1% 40.6% 6.5% 14.4% 18.1% 45.1% 34.9% 2.7% 42.8% 80.5% 13.5%
ACC/AHA ⫽ American College of Cardiology/American Heart Association.
significant proportion of patients had a prior history of coronary disease as manifest by the frequency of prior MI and prior coronary artery bypass grafting. Although multivessel disease was common in the series, multivessel intervention was infrequent. Lesion characteristics are listed in Table II. Complex lesion characteristics were common in this series, although most lesions were de novo. Abciximab was given to 15.4% of patients, and warfarin to 36.9% of patients after the procedure. Clinical events: During follow-up, death occurred in 86 patients, MI in 93. Target lesion revascularization (percutaneous transluminal coronary angioplasty or CABG) was performed in 280 patients. Kaplan-Meier
TABLE III Twelve-Month Event-Free Survival Rates
Freedom from death/MI Freedom from repeat percutaneous revascularization of the target lesion or CABG Freedom from death/MI/repeat percutaneous intervention of the target lesion or CABG
Survival
95% CI
92.5% 84.6%
91.2%–93.8% 82.8%–86.5%
80.2%
78.2%–82.2%
estimates of event-free survival at 12 months are listed in Table III. Univariate and multivariate analyses: Univariate correlates of the cumulative end points of death/MI, percutaneous target vessel revascularization/CABG, and death/MI/percutaneous target vessel revascularization/CABG are listed in Tables IV, V, and VI6, respectively. Stepwise regression analysis of clinical, angiographic, and lesional variables was performed for the cumulative end points of death/nonfatal MI, repeat percutaneous target lesion revascularization/CABG, and death/nonfatal MI/repeat percutaneous target lesion revascularization/CABG. Of the clinical, lesion, and procedural variables examined, the occurrence of death/nonfatal MI (Figure 1) was associated with any history of CHF (previous or on presentation) (RR 3.3, 95% confidence interval [CI] 2.3 to 4.7, p ⬍0.0001), MI within 24 hours of procedure (RR 2.3, CI 1.3 to 4.1, p ⫽ 0.0048), vein graft intervention (RR 1.8, CI 1.3 to 2.6, p ⫽ 0.0007), previous MI (RR 1.8, CI 1.2 to 2.6, p ⫽ 0.004). There was a nonsignificant trend toward increased events with warfarin use after the procedure (RR 1.4, CI 0.97 to 1.94, p ⫽ 0.07). Percutaneous target lesion revascularization or CABG (Figure 2) was associated with multivessel stent placement (RR 1.8, CI 1.2 to 2.8, p ⫽ 0.006); this end point was less frequent in de novo lesions and in right coronary artery interventions (RR 0.6, CI 0.5 to 0.8, p ⫽ 0.0007, and RR 0.8, CI 0.6 to 1.0, p ⫽ 0.05, respectively). A nonsignificant trend toward a need for greater repeat revascularization was noted when MI had occurred within 24 hours of the procedure (RR 1.4, CI 0.98 to 2.1, p ⫽ 0.07). Although unplanned stent placement generally was not correlated with repeat revascularization, the subset of stent procedures for abrupt vessel closure after percutaneous transluminal coronary angioplasty was associated with an increased risk of repeat revascularization (RR 1.7, CI 1.1 to 2.7, p ⫽ 0.03). The cumulative end point of death/nonfatal MI/ repeat percutaneous target lesion intervention/CABG (Figure 3) was associated with CHF (RR 1.7, CI 1.3 to 2.2, p ⬍0.0001), multivessel stent placement (RR 1.6, CI 1.1 to 2.3, p ⫽ 0.03), warfarin after the procedure (RR 1.4, CI 1.2 to 1.8, p ⫽ 0.001), or MI within 24 hours of the procedure (RR 1.5, CI 1.1 to 2.1, p ⫽ 0.02); the frequency of this cumulative end point was reduced in complete revascularization or right coro-
CORONARY ARTERY DISEASE/CLINICAL OUTCOMES WITH INTRACORONARY STENTS
791
TABLE IV Univariate Correlates of Death/MI 95% Confidence Interval Effect Age Prior CABG graft Prior MI Prior history of CHF Diabetes Multivessel disease Ejection fraction ⬍50% Proximal lesion Right coronary artery Circumflex lesion Graft lesion ACC/AHA type C lesion Warfarin
Risk Ratio
Lower Limit
Upper Limit
p Value
1.022 2.099 2.194 4.223 1.782 1.692 2.825 0.502 0.553 0.669 2.462 1.790 1.549
1.007 1.538 1.587 3.046 1.264 1.169 1.738 0.367 0.379 0.427 1.771 1.279 1.105
1.037 2.863 3.032 5.855 2.513 2.448 4.590 0.686 0.808 1.050 3.422 2.504 2.171
0.005 0.0001 0.0001 0.0001 0.001 0.005 0.0001 0.0001 0.002 0.08 0.0001 0.0007 0.01
Abbreviation as in Table II.
target vessel) were included, diabetes mellitus was associated with an increased risk of the cumulative end point of death/MI/CABG/repeat percutaneous intervention (RR 1.3, CI 1.0 to 1.6, p ⫽ 0.04) and percutaneous intervention/CABG (RR 1.4, CI 1.08 to 1.81, p ⫽ 0.012). Additionally, multiple stent implantation within a single vessel, which has been associated with higher rates of target lesion revascularization in other series, was not found to be predictive of repeat target lesion percutaneous intervention in this series (RR 1.4, CI 0.9 to 2.0, p ⫽ 0.11).
DISCUSSION The current study demonstrates that several easily identifiable clinical, procedural, and lesion character95% Confidence Interval istics, which in large part are known Effect Risk Ratio Lower Limit Upper Limit p Value before performing a percutaneous revascularization procedure, are assoMen 1.290 0.973 1.711 0.08 Angina ⱖclass III 1.301 1.015 1.668 0.04 ciated with long-term clinical events Stent for abrupt vessel closure 1.7 1.1 2.7 0.03 after coronary stent placement. Thrombus 0.739 0.550 0.994 0.046 Death/myocardial infarction: The Bifurcation 1.421 0.967 2.088 0.07 presence of CHF and/or left ventricLeft anterior descending 1.284 1.014 1.626 0.04 lesion ular dysfunction has been clearly asRight coronary artery lesion 0.809 0.624 1.049 0.11 sociated with increased morbidity Warfarin use 1.262 0.987 1.613 0.06 and mortality in patients with coronary artery disease.7,8 The current findings are in accord with pooled TABLE VI Univariate Correlates of Death/MI/Repeat Percutaneous Intervention data from 4 interventional trials and Target Lesion/CABG a large single center database where it has been demonstrated that a his95% Confidence Interval tory of CHF is associated with early Effect Risk Ratio Lower Limit Upper Limit p Value and intermediate-term mortality in Prior CABG 1.376 1.115 1.699 0.003 patients undergoing percutaneous Prior history of CHF 1.811 1.404 2.336 0.0001 coronary intervention.9 Similarly, a CHF on presentation 1.726 1.247 2.387 0.001 history of coronary artery disease Diabetes 1.375 1.084 1.743 0.009 manifesting as an acute MI, or a prior Multivessel disease 1.388 1.100 1.750 0.006 Complete revascularization 0.735 0.601 0.899 0.003 history of MI, has been demonstrated Ejection fraction ⬍50% 1.438 1.035 1.998 0.03 to impact survival.10,11 However, the Ostial lesion 1.402 1.060 1.855 0.02 impact on mortality in these settings Right coronary artery lesion 0.707 0.562 0.889 0.003 is more likely a reflection of the presLeft anterior descending lesion 1.193 0.973 1.464 0.09 Graft lesion 1.419 1.113 1.810 0.005 ence of these characteristics rather than a consequence of the stent procedure. The finding that saphenous vein graft interventions are also asnary artery intervention (RR 0.8, CI 0.7 to 0.99, p ⫽ sociated with an increased rate of adverse events has 0.04, and RR 0.7, CI 0.6 to 0.9, p ⫽ 0.009, respectively). previously been reported,12,13 although a portion of Diabetes mellitus, which has been associated with these events has been shown to occur in association increased rates of adverse events after percutaneous with previously mild untreated stenoses within the intervention, was not found to be independently asso- vein graft.14 Repeat revascularization: Multivessel intervention ciated with the occurrence of death/MI in this population (RR 1.2, CI 0.85 to 1.8, p ⫽ 0.28), target lesion has clearly been established to increase the likelihood revascularization (RR 1.1, CI 0.7 to 1.6, p ⫽ 0.7), or of repeat revascularization procedures after percutadeath/MI/CABG/repeat target lesion percutaneous in- neous transluminal coronary angioplasty.15 Although tervention (RR 1.2, CI 0.9 to 1.5, p ⫽ 0.16). However, multivessel stent placement can be performed with a when all revascularization procedures (target and non- high degree of procedural success and a low rate of TABLE V Univariate Correlates of Repeat Percutaneous Intervention of Target Lesion/CABG
792 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 84
OCTOBER 1, 1999
FIGURE 1. Variables associated with death/nonfatal MI during the course of long-term follow-up after successful stent implantation.
FIGURE 2. Variables associated with repeat revascularization (percutaneous revascularization of the target lesion or coronary artery bypass grafting) during the course of long-term follow-up after successful stent implantation. RCA ⴝ right coronary artery.
intermediate-term events,16 the finding that the need for repeat revascularization is increased is likely a combination of restenosis and disease progression in previously untreated coronary segments.13 Additionally, the finding that de novo lesions are associated with a lower likelihood of repeat revascularization is in keeping with previous studies.17 Interestingly, right coronary artery interventions were associated with a decreased likelihood of repeat revascularization, which was apparent in the Stent Restenosis Study1 as well. Notably, the presence of diabetes mellitus did not portend an increased risk of target lesion revascularization. Studies have demonstrated an increased rate of angiographic restenosis after stent placement in diabetics versus nondiabetics18,19 due to exaggerated neointimal hyperplasia,18 although it has also been reported that restenosis rates are similar in diabetics versus nondiabetics after stent placement due to reduction in negative arterial remodeling after stenting.20 Importantly, the current series evaluates clinically driven target lesion revascularization and may potentially have missed cases of angiographic restenosis that were not apparent clinically. In addition, the
finding in this series that diabetes was associated with increased need for repeat revascularization rates when all repeat revascularization procedures (target lesion and nontarget lesion) were included supports the observation from the Bypass Angioplasty Revascularization Investigation that diabetes was associated with a higher likelihood of new lesions during follow-up in treated and untreated vessels than in nondiabetics.21 Additionally, the current series did not identify multiple stent implantation in a single vessel as a correlate of repeat revascularization. Numerous studies have demonstrated that multiple contiguous stent implantation (ⱖ2 or 3 stents) is associated with higher rates of target lesion revascularization, although these reports included a large proportion of patients with severe dissections,22 or did not attempt to establish the independent effect of multiple stents on target lesion revascularization. Current data are in accord with recently published data that report similar target lesion revascularization rates in patients with ⱖ3 contiguous stents compared with patients with 1 or 2 stents.23 Abrupt vessel closure but not unplanned stent placement was generally associated with greater need for repeat vascularization, which is in keeping with recent data.24 Death, MI, repeat target lesion percutaneous intervention, or CABG: The defi-
nition of completeness of revascularization varies among published studies on angioplasty as well as surgery, although there is ample evidence to suggest that complete revascularization is associated with improved long-term outcome,25 in keeping with the current study. Interestingly, the finding that warfarin use was associated with a greater frequency of the composite end point is in accord with previous data that demonstrated an increase in death, MI, and repeat revascularization at 30 days in patients treated with warfarin after the procedure compared with those treated with only antiplatelet agents.26 By design, the current study excluded patients who encountered adverse events during the initial hospital stay; the long-term effect of warfarin on clinical outcome has not been reported and may warrant further investigation. Although our statistical modeling was designed to eliminate the possibility of confounding variables, the possibility that warfarin use correlated with adverse variables cannot be ruled out with certainty in this retrospective analysis. Impact of lesion characteristics on outcome: Notably, of all the lesion characteristics assessed, only vein graft interventions, multivessel stenting, and restenosis lesions were associated with adverse events. However, among univariate correlates of adverse events,
CORONARY ARTERY DISEASE/CLINICAL OUTCOMES WITH INTRACORONARY STENTS
793
7. Emond M, Mock MB, Davis KB, Fisher LD, Holmes DR Jr, Chaitman BR, Kaiser GC, Alderman E, Killip T III. Long-term survival of medically treated patients in coronary artery surgery study (CASS) registry. Circulation 1994;90:2645–2657. 8. Judge KW, Pawitan Y, Caldwell J, Gersh BJ, Kennedy JW. Congestive heart failure symptoms in patients with preserved left ventricular systolic function: analysis of the CASS registry. J Am Coll Cardiol 1991;18:377–382. 9. Anderson RD, Ohman EM, Holmes DR, Harrington RA, Barsness GW, Wildermann NM, Phillips HR, Topol EJ, Califf RM. Prognostic value of congestive heart failure history in patients undergoing percutaneous coronary interventions. J Am Coll Cardiol 1998; 32:936 –941. 10. Maseri A. Mode of first presentation: acute, chronic, and quiescent phases of IHD. In: Ischemic Heart Disease. New York: Churchill Livingstone, 1995;323–335. 11. Goldstein S, Friedman L, Hutchinson R, Canner P, Romhilt D, Schlant R, Sobrino R, Verter J, Wasserman A. Timing, mechanism and clinical setting of witnessed FIGURE 3. Variables associated with a cumulative end point of death/nonfatal MI/ deaths in postmyocardial infarction patients. J Am Coll repeat revascularization during the course of long-term follow-up after successful Cardiol 1984;3:1111–1117. 12. Waksman R, Weintraub WS, Ghazzal Z, Scott NA, stent implantation. RCA ⴝ right coronary artery. Shen Y, King SB III, Douglas JS Jr. Short- and long-term outcome of narrowed saphenous vein bypass graft: a comof Palmaz-Schatz stent, directional coronary atherectomy, and balloon several lesion characteristics were found to influence parison angioplasty. Am Heart J 1997;134:274 –281. outcome, which is in accord with prior studies.24,25 13. Laham RJ, Ho KK, Baim DS, Kuntz RE, Cohen DJ, Carrozza JP Jr. The fact that many of these variables were no longer Multivessel Palmaz-Schatz stenting: early results and one-year outcome. J Am Coll Cardiol 1997;30:180 –185. significant in multivariate analysis may in part be 14. Ellis SG, Brener SJ, DeLuca S, Tuzcu EM, Raymond RE, Whitlow PL, Topol related to the exclusion of patients with short-term EJ. Late myocardial ischemic events after saphenous vein graft intervention— of initially “nonsignificant” vein graft lesions. Am J Cardiol 1997; adverse events in this study. Such variables identified importance 79:1460 –1464. with univariate analysis may, however, shed light on 15. Lambert M, Bonan R, Cote G, Crepeau J, deGuise P, Lesperance J, David PR, Waters DD. Early results, complications and restenosis rates after multilesion and important variables to evaluate prospectively. percutaneous transluminal coronary angioplasty. Am J Cardiol 1987; In cases in which these adverse clinical or lesion multivessel 60:788 –791. characteristics are present, clinical follow-up should 16. Mathew V, Rihal CS, Berger PB, Bell MR, Garratt KN, Holmes DR Jr. outcome of patients undergoing multivessel coronary stent implantation. be increased. Adjunctive therapies to reduce recurrent Clinical Int J Cardiol 1998;64:1–7. ischemic events may also be considered; for example, 17. Black AJR, Anderson V, Rabin GS, Powelson SW, Douglas JS, Kinky SB. intracoronary brachytherapy appears to be a promising Repeat coronary angioplasty: correlates of a second restenosis. J Am Coll Cardiol 1988;11:714 –718. modality to reduce restenosis,27 and may thereby re- 18. Kornowski R, Mintz GS, Kent KM, Pichard AD, Satler LF, Bucher TA, Hong MK, Popma JJ, Leon MB. Increased restenosis in diabetes mellitus after coronary duce adverse clinical events. is due to exaggerated intimal hyperplasia: a serial intravascular Study limitations: The current study is a retrospec- interventions ultrasound study. Circulation 1997;95:1366 –1369. tive analysis of a single center experience, and is 19. Kastrati A, Schomig A, Elezi S, Schuhlen H, Dirschinger J, Hadamitzky M, therefore subject to the bias innate to such studies. Wehinger A, Hausleiter J, Walter H, Neumann F-J. Predictive factors of resteafter coronary stent placement. J Am Coll Cardiol 1997;30:1428 –1436. Because this includes our earlier stent experience, nosis 20. Van Belle E, Bauters C, Hubert E, Bodart J-C, Abolmaali K, Meurice T, warfarin was used in a significant portion of patients, McFadden EP, Lablanche J-M, Bertrand ME. Restenosis rates in diabetic paa comparison of coronary stenting and balloon angioplasty in native which no longer is the standard of postprocedure tients; coronary vessels. Circulation 1997;96:1454 –1460. anticoagulation. Follow-up angiographic data were 21. Rozenman Y, Sapoznikov D, Mosseri M, Gilon D, Lotan C, Nassar H, Weiss not obtained in most patients; however, this study AT, Hasin Y, Gotsman MS. Long-term angiographic follow-up of coronary angioplasty in patients with diabetes mellitus: a clue to the explanation of reflects current practice in which patients are usually balloon the results of the BARI study. J Am Coll Cardiol 1997;30:1420 –1425. evaluated clinically, and subject to repeat angiography 22. Mathew V, Hasdai D, Holmes DR, Garratt KN, Bell MR, Lerman A, Melby S, Grill DE, Berger PB. Clinical outcome of patients undergoing endoluminal only if there is evidence of recurrent ischemia. coronary artery reconstruction with three or more stents. J Am Coll Cardiol 1997;30:676 – 681. 23. Kornowski R, Mehran R, Hong MK, Satler LF, Pichard AD, Kent KM, Mintz GS, Waksman R, Laird JR, Lansky AJ, Bucher TA, Popma JJ, Leon MB. 1. Fischman DL, Leon MB, Baim DS, Schatz RA, Savage MP, Penn I, Detre K, Procedural results and late clinical outcomes after placement of three or more Veltri L, Ricci D, Nobuyoshi M. A randomized comparison of coronary-stent stents in single coronary lesions. Circulation 1998;97:1355–1361. placement in the treatment of coronary artery disease. Stent Restenosis Study 24. Piana RN, Ahmed WH, Chaitman B, Ganz P, Kinlay S, Strony J, Adelman B, Investigators. N Engl J Med 1994;331:496 –501. Bittl JA, on Behalf of the Hirulog Angioplasty Study Investigators. Effect of transient 2. Serruys PW, de Jaegere P, Kiemeneij F, Macaya C, Rutsch W, Heyndrickx G, abrupt vessel closure during otherwise successful angioplasty for unstable angina on Emanuelsson H, Marco J, Legrand V, Materne P. A comparison of balloonclinical outcome at six months. J Am Coll Cardiol 1999;33:73–78. expandable-stent implantation with balloon angioplasty in patients with coronary 25. Bell MR, Gersh BJ, Schaff HV, Holmes DR Jr, Fisher LD, Alderman EL, artery disease. Benestent Study Group. N Engl J Med 1994;331:489 – 495. Meyers WO, Parsons LS, Reeder GS. Effect of completeness of revascularization 3. Holmes DR Jr. Technical aspects. In: Vliesttra RE, Holmes DR Jr, eds. PTCA: on long-term outcome of patients with three-vessel disease undergoing coronary Percutaneous Transluminal Coronary Angioplasty. Philadelphia, PA: FA Davis, artery bypass surgery. A report from the Coronary Artery Surgery Study (CASS) 1987:35– 48. Registry. Circulation 1992;86:446 – 457. 4. Cannon AD, Roubin GS. The Gianturco-Roubin stent. In: Topol EJ, ed. 26. Schomig A, Neumann FJ, Kastrati A, Schuhlen H, Blasini R, Hadamitzky M, Textbook of Interventional Cardiology. 2nd ed. Philadelphia, PA. WB Saunders, Walter H, Zitzmann-Roth EM, Richardt G, Alt E, Schmitt C, Ulm K. A random1994:712–726. ized comparison of antiplatelet and anticoagulant therapy after placement of coronary-artery stents. N Engl J Med 1996;334:1084 –1089. 5. Whitlow PL, de Jaegere PP, Serruys PW. The Wiktor stent. In: Topol EJ, ed. 27. Teirstein PS, Massullo V, Jani S, Popma JJ, Mintz GS, Russo RJ, Schatz RA, Textbook of Interventional Cardiology. 2nd ed. Philadelphia, PA. WB Saunders, Guarneri EM, Steuterman S, Morris NB, Leon MB, Tripuraneni P. Catheter1994:727–741. based radiotherapy to inhibit restenosis after coronary stenting. N Engl J Med 6. Ellis SG. The Palmaz-Schatz stent. In: Topol EJ, ed. Textbook of Interven1997;336:1697–1703. tional Cardiology. 2nd ed. Philadelphia, PA. WB Saunders, 1994:702–711.
794 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 84
OCTOBER 1, 1999