- Email: [email protected]
Comparison of medicine alone, coronary angioplasty, and left internal mammary artery–coronary artery bypass for one-vessel proximal left anterior descending coronary artery disease
Comparison of Medicine Alone, Coronary Angioplasty, and Left Internal Mammary Artery–Coronary Artery Bypass for OneVessel Proximal Left Anterior Descending Coronary Artery Disease Adam B. Greenbaum, MD, Robert M. Califf, MD, Robert H. Jones, MD, Laura H. Gardner, BSPH, Harry R. Phillips, MD, Michael H. Sketch, Jr., MD, Richard S. Stack, MD, and Joseph A. Puma, DO Despite the deleterious and sometimes catastrophic consequences of proximal left anterior descending (LAD) artery occlusion, there is a paucity of data to guide the treatment of patients with such disease. Our aim was to describe outcomes with medical therapy, angioplasty, or left internal mammary artery (LIMA) bypass grafting in patients with 1-vessel, proximal LAD disease. We retrospectively analyzed prospectively collected data from 1,188 patients first presenting only with proximal LAD disease at 1 center over 9 years. We assessed the rates of death, acute myocardial infarction, and repeat intervention by initial treatment over a median 5.7 years of follow-up. Patients undergoing angioplasty or LIMA bypass were more often men and had progressive or unstable angina; those receiving medical therapy had a lower median ejection fraction. Both revascularization procedures offered slightly better adjusted survival versus medicine (hazard ratio for angioplasty, 0.82; 95%
confidence interval, 0.60 to 1.11; hazard ratio for bypass, 0.74; 95% confidence interval, 0.44 to 1.23). Bypass, but not angioplasty, was associated with significantly fewer composite end point events (death, infarction, or reintervention, p <0.0001), and angioplasty was associated with a higher composite event rate than bypass or medical therapy (p <0.0001 and p ⴝ 0.0003, respectively). The initial advantages of bypass and medicine over angioplasty diminished over time; angioplasty became more advantageous than medicine after 1 year (p ⴝ 0.05) and not significantly different from bypass. Treatment of 1-vessel, proximal LAD disease with medicine, angioplasty, or LIMA bypass resulted in comparable adjusted survival. However, LIMA bypass alone reduced the long-term incidence of infarctions and repeat procedures. 䊚2000 by Excerpta Medica, Inc. (Am J Cardiol 2000;86:1322–1326)
espite generally worse outcomes for patients with 1-vessel proximal left anterior descending (LAD) D than for those whose disease is located elsewhere,
with an abnormal ejection fraction, prior AMI, or unstable coronary syndromes.4 The purpose of this investigation was to describe the long-term experience of patients with 1-vessel, proximal LAD stenoses (defined as ⱖ75% stenosis before the first septal perforator with ⱕ50% lesions in any other major epicardial segments)10 treated with medical therapy, angioplasty, or LIMA bypass grafting at Duke University Medical Center between 1986 and 1994.
treatment decisions often are guided by anecdotal experience, physician bias, small retrospective studies, or extrapolation of data comparing 2 of the 3 possible therapies—medicine, angioplasty, or left internal mammary artery (LIMA) bypass grafting.1– 8 Only 1 randomized trial has compared all 3 treatments for 1-vessel, proximal LAD disease, the Medicine, Angioplasty or Surgery Study.9 In this comparison, 214 patients were randomized to receive medicine, angioplasty, or LIMA bypass and were followed for 3 years. No treatment difference was observed for the end point of death or acute myocardial infarction (AMI). However, this study had serious drawbacks, including a small sample size, short follow-up (benefits of revascularization generally emerge only after 3 to 9 years after bypass), and the exclusion of patients From the Duke Clinical Research Institute, Durham, North Carolina. Manuscript received February 24, 2000; revised manuscript received and accepted June 30, 2000. Address for reprints: Adam B. Greenbaum, MD, K-2 Cardiac Cath Lab, Henry Ford Hospital, 2799 West Grand Boulevard, Detroit, Michigan 48202. E-mail: [email protected].
1322
©2000 by Excerpta Medica, Inc. All rights reserved. The American Journal of Cardiology Vol. 86 December 15, 2000
METHODS
Patients: Between April 1, 1986, and February 28, 1994, 1,218 patients with proximal LAD disease undergoing a first diagnostic angiogram were identified. Of these, 754 underwent angioplasty, 149 underwent LIMA bypass (another 21 had vein grafts, which were excluded from analysis), and the remainder (n ⫽ 294) underwent no procedures within the first 30 days after angiography. These 1,197 eligible patients had coronary syndromes ranging from stable angina to evolving AMI, and reflected left ventricular function of all degrees. Data collection system: Pertinent baseline variables from the history, physical examination, laboratory tests, chest x-ray, and 12-lead electrocardiogram, as 0002-9149/00/$–see front matter PII S0002-9149(00)01235-2
study was 2%. Treatment selection (angioplasty, bypass, or medicine) Medicine Angioplasty LIMA Bypass was at the discretion of the attending (n ⫽ 285) (n ⫽ 754) (n ⫽ 149) cardiologist. For those treated medically, the cardiologist and primary Male sex 71% 70% 78% Age (yrs; median [IQR]) 60 (50–67) 59 (50–67) 60 (51–66) care physician determined therapy. Anginal symptoms Follow-up procedures: Patients Chest pain within 6 weeks 63% 73% 80% were contacted at 6 months and 1 Typical angina 67% 81% 73% year after presentation, then annualStable angina 28% 8% 14% Progressive angina 15% 20% 21% ly.10 The anniversary was the date of Unstable angina 6% 21% 23% angiography for medical patients, Recent infarction, stable 22% 18% 22% and the date of the first revascularRecent infarction, unstable 5% 8% 5% ization for those who underwent anAcute myocardial infarction 23% 24% 14% gioplasty or LIMA bypass. Patients Nocturnal angina 5% 8% 11% Pain frequency (median per 1 (0–4) 3 (0–5) 3 (1–7) were contacted by mailed questionweek [IQR]) naires; with telephone follow-up, Duration of ischemic symptoms 2 (0–57) 1 (0–8) 2 (0–13) National Death Index search, and vi(mo; median [IQR]) tal record review if the questionnaire Coexisting illness Diabetes 20% 16% 11% was not returned. The primary end Smoker 63% 64% 68% point for this study was all-cause Hypertension 42% 44% 46% mortality. The secondary end point Hyperlipidemia 32% 36% 36% was freedom from death, AMI, or Peripheral vascular disease 9% 5% 5% (repeat) revascularization. Cerebrovascular disease 8% 4% 7% Clinical measures of myocardial Data analysis: Baseline characterdamage istics and outcomes were summaPrior infarction 72% 56% 47% rized as medians and interquartile Prior congestive heart failure 17% 6% 8% ranges (for continuous measures) or New York Heart Association 13% 5% 7% class I or II percentages (for discrete measures). New York Heart Association 4% 2% ⬍1% The treatment group was defined by class III or IV the first revascularization within 30 IQR ⫽ interquartile range. days of angiography; we assumed an intent to treat with angioplasty or LIMA bypass based on the angiogram. If patients did not undergo revascularization within 30 days, we TABLE 2 Baseline Angiographic Characteristics assumed an intent to treat the patient LIMA medically. Once patients were asMedicine Angioplasty Bypass signed to a revascularization group (n ⫽ 285) (n ⫽ 754) (n ⫽ 149) (medicine, angioplasty, or LIMA byDisease severity pass), they remained in that group Proximal LAD occlusion 75% to 94% only 19% 16% 13% until death or the end of follow-up, Proximal LAD occlusion 75% to 94% and 10% 6% 15% ⱖ95% occlusion elsewhere in LAD regardless of later treatment. With Proximal LAD occlusion ⱖ95% 71% 78% 72% this definition, any patient who died Abnormal left ventricular contraction 83% 81% 58% within a few days of angiography Ejection fraction (%, median [IQR]) 44 (34–56) 52 (44–62) 56 (47–63) would have been ineligible to unMitral insufficiency dergo revascularization. Thus, all 1⫹ or 2⫹ 16% 12% 12% 3⫹ or 4⫹ 3% 2% 2% early deaths would be included in the medical therapy arm, unfairly penalAbbreviation as in Table 1. izing that group. To address this, we excluded from analysis all medical well as angiographic and procedural details, were col- patients who died or were lost to follow-up before the lected prospectively on standard forms during patient mean time to revascularization (1.5 days; n ⫽ 9). To control for baseline differences between care, and stored at the Duke Clinical Research Instigroups, we used an established model of long-term tute.10 Revascularization and medical therapy: Standard survival developed from a larger population of Duke angioplasty and cardiac surgical and anesthetic tech- patients, all treated for coronary disease between 1984 niques at Duke University Medical Center during this and 1990.13 The significant predictors of mortality in period have been described.11,12 During this study, the this model include age, sex, presentation, congestive overall procedural success rate for angioplasty (ⱕ50% heart failure, chest pain, comorbidity, ejection fracresidual stenosis) was 93%, and the emergency bypass tion, vascular disease, coronary artery disease index, rate for this cohort was 4.5%. The unadjusted 30-day and mitral insufficiency. We estimated disease severmortality rate for surgical patients included in this ity for patients in our study with the baseline risk TABLE 1 Baseline Clinical Characteristics
CORONARY ARTERY DISEASE/OPTIMAL TREATMENT OF PROXIMAL LAD DISEASE
1323
FIGURE 1. Observed (A) and adjusted (B) Kaplan-Meier survival estimates with LIMA bypass (thick solid line), angioplasty (thin solid line), or medical therapy (dashed line) of 1-vessel, proximal LAD disease. Med ⴝ medicine.
Index of Comorbidity.14 These variables, including treatment (medicine, angioplasty, or LIMA bypass) were added to the models. To avoid weakening the precision of estimated hazard ratios, inconsequential variables were removed from the models. Treatment comparisons are presented as Kaplan-Meier survival curves15 (unadjusted and adjusted) to contrast absolute survival differences, and as hazard ratios from Cox models (to compare relative survival differences of the 3 treatments, taken 2 at a time). For revascularized patients, days to an event were calculated from the date of procedure. For patients in the medical arm, timing began at the date of angiography. Because early procedural mortality creates a pattern of crossing survival curves, the proportional-hazards assumptions of the Cox model are violated. To account for this when comparing treatments, 3 Cox models were used to also calculate hazard ratios for 3 time intervals: baseline (date of treatment) to 30 days (with survival truncated at 30 days), 30 days to 1 year from baseline (conditional upon surviving 30 days and truncating survival at 1 year), and ⬎1 year from baseline (conditional upon surviving 1 year). Hazard ratios are shown with 95% confidence intervals.
RESULTS
Fewer women underwent surgery than did those taking other treatments (Table 1). Patients who had TABLE 3 Observed and Adjusted Survival Estimates by Treatment undergone angioplasty or LIMA LIMA Bypass grafting more often had either proMedicine (n ⫽ 285) Angioplasty (n ⫽ 754) (n ⫽ 149) gressive or unstable angina than did Interval Observed Adjusted Observed Adjusted Observed Adjusted medically treated patients. Risk factor profiles were similar except for 30 d 0.95 0.98 0.98 0.99 0.98 0.98 1 yr 0.90 0.95 0.95 0.96 0.96 0.96 more diabetics in the medical and 3 yrs 0.85 0.92 0.91 0.93 0.93 0.93 angioplasty groups. The greatest dif5 yrs 0.80 0.88 0.87 0.90 0.91 0.91 ferences among the groups occurred 7 yrs 0.73 0.82 0.82 0.86 0.88 0.88 in clinical measures of myocardial 9 yrs 0.67 0.78 0.76 0.80 0.80 0.79 damage. On average, medically treated patients had a longer duration (hazard) score from this model, calculated by first of symptoms, more prior AMI, and more prior conmultiplying each factor by its corresponding Cox re- gestive heart failure. Patients who received medicine or underwent angioplasty also had an abnormal ejecgression coefficient and then summing the products. A potentially important characteristic, presence of tion fraction more often (lowest in the medicine mitral insufficiency of grade 3 or 4, was an exclusion group), and grade 1⫹ and 2⫹ mitral insufficiency was criterion for the larger population. We included this most prevalent in medically treated patients (Table 2). The follow-up interval (time to death or last convariable in adjusted survival modeling (with the baseline hazard score), as well as year of angiography and tact) for the 1,188 patients after assignment to a treata weighted index of comorbidity. The index, modified ment group was a median 5.66 years (interquartile due to data availability, was based on the Charlson range, 3.66 to 8.01 years). Figure 1A and 1B show 1324 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 86
DECEMBER 15, 2000
FIGURE 2. Hazard ratios (boxes) and 95% confidence intervals (lines) for mortality differences between treatments. Vertical line at 1.0 indicates equivalence between treatments; values to the left of 1.0 indicate benefit from the first treatment listed. Abbreviations as in Figure 1.
FIGURE 3. Observed estimates of event-free survival with LIMA bypass (thick solid line), angioplasty (thin solid line), or medical therapy (dashed line) of 1-vessel, proximal LAD disease. Abbreviation as in Figure 1.
survival curves by treatment before and after adjustment for baseline differences. Observed and adjusted survival estimates by treatment over time are summarized in Table 3. A significant survival benefit was seen with revascularization compared with medical therapy (p ⫽ 0.001 for either LIMA bypass or angioplasty vs medical therapy). After adjusting for baseline clinical and angiographic differences, a trend toward improved survival with revascularization versus medicine persisted (hazard ratio for LIMA bypass, 0.74; 95% confidence interval, 0.44 to 1.23; hazard ratio for angioplasty, 0.82; 95% confidence interval, 0.60 to 1.11). Figure 2 shows the adjusted hazard ratios for the relative treatment effects on survival during follow-up and at 30 days, 1 year, and ⬎1 year (conditional upon surviving the previous period). After adjustment, survival did not differ significantly between revascularization approaches (hazard ratio, 0.90; 95% confidence interval, 0.55 to 1.48 for bypass vs angioplasty). Observed Kaplan-Meier curves and survival estimates at 30 days, and at 1, 3, 5, and 7 years for freedom from death, AMI, or (repeat) revasculariza-
tion by treatment are shown in Figure 3 and Table 4. Hazard ratios for the relative effect of the treatments on event-free survival throughout follow-up and at 30 days, 1 year, and ⬎1 year (conditional upon surviving the previous interval) are shown in Figure 4. Overall, LIMA bypass offered significant benefits over both medical therapy and angioplasty (hazard ratio, 0.43; 95% confidence interval, 0.28 to 0.64; p ⬍0.0001; hazard ratio, 0.29, 95% confidence interval, 0.20 to 0.42; p ⬍0.0001, respectively). Angioplasty also was associated with significantly more composite events than medical therapy (hazard ratio, 1.47; 95% confidence interval, 1.20 to 1.81; p ⫽ 0.0003). The event-free survival advantage of LIMA bypass over medical therapy began after the first 30 days (p ⫽ 0.005) and persisted throughout follow-up (p ⫽ 0.004). The advantage of LIMA bypass over angioplasty also persisted throughout follow-up but was no longer significant after the first year (conditional on surviving event free for the first year). The disadvantage of angioplasty versus medical therapy also diminished over time with an event-free survival advantage to angioplasty versus medicine evident after the first year, conditional on surviving event free until then (hazard ratio, 0.69; 95% confidence interval, 0.48 to 0.99; p ⫽ 0.05).
DISCUSSION
In this observational comparison of patients with 1-vessel, proximal LAD disease, surgical and percutaneous revascularizations were associated with nonsignificant survival advantages compared with medical therapy after adjustment for baseline clinical and angiographic differences. Further, LIMA bypass (but not angioplasty) was associated with a significant, long-term reduction in a composite of adverse cardiac events, largely driven by fewer repeat interventions. Angioplasty resulted in significantly more composite events than medical therapy overall. However, the disadvantages to angioplasty appeared to diminish over time; after the first year, angioplasty became more advantageous than medical therapy, and no longer differed significantly from surgical revascularization. Although no significant mortality benefit was seen with revascularization in this study, the results were directionally appropriate and consistent with a small survival benefit implied for this cohort. When grouped together, patients with more severe forms of 1-vessel disease (ⱖ95% proximal LAD occlusion) and those with 2-vessel disease (with a ⱖ95% LAD occlusion) have a significant 5-year survival advantage of 5 to 6 lives saved per 100 treated with revascularization versus medicine (a 30% to 35% relative reduction).5 In our study of patients
CORONARY ARTERY DISEASE/OPTIMAL TREATMENT OF PROXIMAL LAD DISEASE
1325
TABLE 4 Observed, Event-Free Survival Estimates by Treatment Interval 30 d 1 yr 3 yrs 5 yrs 7 yrs
Medicine (n ⫽ 285)
Angioplasty (n ⫽ 754)
LIMA Bypass (n ⫽ 149)
0.93 0.78 0.66 0.57 0.46
0.86 0.60 0.53 0.48 0.42
0.97 0.91 0.86 0.80 0.68
mortality.16 –18 Thus, the event-free survival curves for angioplasty-treated patients likely would have improved, but not the survival analysis in our study.
Acknowledgment: We thank Pat French and A. J. Mayhew for editorial assistance, and Anthony Doll for graphic assistance.
1. CASS Principal Investigators and their associates. Coronary artery surgery study (CASS): a randomized trial of coronary artery bypass surgery: survival data. Circulation 1983:68:939 – 950. 2. Varnauskas E. Twelve-year follow-up of survival in the randomized European Coronary Surgery Study. N Engl J Med 1988;319:332–337. 3. Veterans Administration Coronary Artery Bypass Surgery Cooperative Study Group. Eleven-year survival in the Veterans Administration randomized trial of coronary bypass surgery for stable angina. N Engl J Med 1984;311:1333–1339. 4. Yusuf S, Zucker D, Peduzzi P, Fisher LD, Takaro T, Kennedy JW, Davis K, Killip T, Passamani E, Norris R, Morris C, Mathur V, Varnauskas E, Chalmers TC. Effect of coronary artery bypass FIGURE 4. Hazard ratios (boxes) and 95% confidence intervals (lines) for differgraft surgery on survival: overview of 10-year results from ranences in event-free survival between treatments. Vertical line at 1.0 indicates domised trials by the Coronary Artery Bypass Graft Surgery equivalence between treatments; values to the left of 1.0 indicate benefit from Trialists Collaboration. Lancet 1994;344:563–570. the first treatment listed. Abbreviations as in Figure 1. 5. Jones RH, Kesler K, Phillips HR III, Mark DB, Smith PK, Nelson CL, Newman MF, Reves JG, Anderson RW, Califf RM. Long-term survival benefits of coronary artery bypass grafting and percutaneous transluminal angioplasty in patients with coronary artery disease. J Thorac Cardiovasc Surg 1996;111:1013–1025. with 1-vessel, proximal LAD disease only, evaluated 6. Kramer JR, Proudfit WL, Loop FD, Goormastic M, Zimmerman K, Simpfenseparately from those with 2-vessel disease involving the dorfer C, Horner G. Late follow-up of 781 patients undergoing percutaneous LAD, revascularization was associated with a nonsignif- transluminal coronary angioplasty or coronary artery bypass grafting for an isolated obstruction in the left anterior descending coronary artery. Am Heart J icant 5-year survival advantage over medical therapy of 1989;118:1144 –1153. about 2 to 3 lives saved per 100 thus treated (a 17% to 7. Cameron J, Mahanonda N, Aroney C, Hayes J, McEniery P, Gardner M, Bett 25% relative reduction). This subtler revascularization N. Outcome five years after percutaneous transluminal coronary angioplasty or artery bypass grafting for significant narrowing limited to the left benefit for 1-vessel, proximal LAD disease might have coronary anterior descending coronary artery. Am J Cardiol 1994;74:544 –549. been predicted, but our study had limited power to detect 8. Goy JJ, Eeckhout E, Burnand B, Vogt P, Stauffer J-C, Hurni M, Stumpe F, P, Sadeghi H, Kappenberger L. Coronary angioplasty versus left internal small differences. Nonetheless, a trend toward improved Ruchat mammary artery grafting for isolated proximal left anterior descending artery survival with revascularization was present and was stenosis. Lancet 1994;343:1449 –1453. noted within a population, with an overall 5-year mor- 9. Hueb WA, Bellotti G, de Oliveira SA, Arie S, de Albuquerque CP, Jatene AD, F. The Medicine, Angioplasty or Surgery Study (MASS): a prospective, tality rate similar to that of previous studies of this cohort Pileggi randomized trial of medical therapy, balloon angioplasty or bypass surgery for 5 (13%). single proximal left anterior descending artery stenoses. J Am Coll Cardiol Other differences between this and previous studies 1995;26:1600 –1605. 10. Rosati RA, McNeer JF, Starmer CF, Mittler BS, Morris JJ, Wallace AG. A new include our definition of significant LAD disease and information system for medical practice. Arch Intern Med 1975;135:1017–1024. how treatment groups were assigned. We defined signif- 11. Buller C, Mark DB, Phillips HR III, Stack RS. Cardiac catheterization and coronary angioplasty. In: Sabiston DC Jr, ed. The Davis-Christoicant LAD disease as ⱖ75% diameter stenosis. Although percutaneous pher Textbook of Surgery. Philadelphia: WB Saunders, 1991;1826 –1843. most patients (76%) also qualified as having ⱖ95% 12. Rankin JS, Sabiston DC Jr. The coronary circulation. In: Sabiston DC Jr, ed. occlusive lesions, previous definitions have ranged any- Textbook of Surgery. Philadelphia: WB Saunders, 1991;1957–1972. 13. Mark DB, Nelson CL, Califf RM, Harrell FE Jr, Lee KL, Jones RH, Fortin where from 50% to 95% stenosis.5,6 We also used a more DF, Stack RS, Glower DD, Smith LR, DeLong ER, Smith PK, Reves JG, Jollis traditional “intent-to-treat” approach than the Coronary JG, Tcheng JE, Muhlbaier LH, Lowe JE, Phillips HR, Pryor DB. Continuing of therapy for coronary artery disease: initial results from the era of Artery Surgery Study “method A” of assigning treatment evolution coronary angioplasty. Circulation 1994;89:2015–2025. 1 groups. Method A considered all patients to be medi- 14. Charlson ME, Pompei P, Alex KL, MacKenzie CR. A new method of cally treated until they underwent a first procedure, then classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–383. reassigned patients to the respective revascularization 15. Kaplan EL, Meier P. Nonparametric estimation from incomplete observagroup after intervention. We assigned treatment groups tions. J Am Stat Assoc 1958;53:457– 481. Serruys PW, de Jaegere P, Kiemeneij F, Macaya C, Rutsch W, Heyndrickx G, at first angiography, and eliminated from analysis all 16. Emanuelsson H, Marco J, Legrand V, Materne P, Belardi J, Sigwart U, Colombo A, patients (n ⫽ 9) who died before the median time to Goy JJ, Van den Heuvel P, Delcan J, Morel M, for the BENESTENT Study Group. A comparison of balloon-expandable-stent implantation with balloon angioplasty in revascularization (1.5 days). with coronary artery disease. N Engl J Med 1994;331:489-495. Although the study ended before the widespread patients 17. Versaci F, Gaspardone A, Tomai F, Crea F, Chiariello L, Gioffre PA. A use of intracoronary stents or platelet glycoprotein comparison of coronary-artery stenting with angioplasty for isolated stenosis of the IIb/IIIa inhibitors, we believe the results to remain proximal left anterior descending coronary artery. N Engl J Med 1997;336:817– 822. Kong DF, Califf RM, Miller DP, Moliterno DJ, White HD, Harrington RA, relevant to clinical patient care in 2000. Improved 18. Tcheng JE, Lincoff AM, Hasselblad V, Topol EJ. Clinical outcomes of theraoutcomes with these agents have resulted from re- peutic agents that block the platelet glycoprotein IIb/IIIa integrin in ischemic duced target vessel revascularization and AMI, not heart disease. Circulation 1998;98:2829 –2835.
1326 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 86
DECEMBER 15, 2000
confidence interval, 0.60 to 1.11; hazard ratio for bypass, 0.74; 95% confidence interval, 0.44 to 1.23). Bypass, but not angioplasty, was associated with significantly fewer composite end point events (death, infarction, or reintervention, p <0.0001), and angioplasty was associated with a higher composite event rate than bypass or medical therapy (p <0.0001 and p ⴝ 0.0003, respectively). The initial advantages of bypass and medicine over angioplasty diminished over time; angioplasty became more advantageous than medicine after 1 year (p ⴝ 0.05) and not significantly different from bypass. Treatment of 1-vessel, proximal LAD disease with medicine, angioplasty, or LIMA bypass resulted in comparable adjusted survival. However, LIMA bypass alone reduced the long-term incidence of infarctions and repeat procedures. 䊚2000 by Excerpta Medica, Inc. (Am J Cardiol 2000;86:1322–1326)
espite generally worse outcomes for patients with 1-vessel proximal left anterior descending (LAD) D than for those whose disease is located elsewhere,
with an abnormal ejection fraction, prior AMI, or unstable coronary syndromes.4 The purpose of this investigation was to describe the long-term experience of patients with 1-vessel, proximal LAD stenoses (defined as ⱖ75% stenosis before the first septal perforator with ⱕ50% lesions in any other major epicardial segments)10 treated with medical therapy, angioplasty, or LIMA bypass grafting at Duke University Medical Center between 1986 and 1994.
treatment decisions often are guided by anecdotal experience, physician bias, small retrospective studies, or extrapolation of data comparing 2 of the 3 possible therapies—medicine, angioplasty, or left internal mammary artery (LIMA) bypass grafting.1– 8 Only 1 randomized trial has compared all 3 treatments for 1-vessel, proximal LAD disease, the Medicine, Angioplasty or Surgery Study.9 In this comparison, 214 patients were randomized to receive medicine, angioplasty, or LIMA bypass and were followed for 3 years. No treatment difference was observed for the end point of death or acute myocardial infarction (AMI). However, this study had serious drawbacks, including a small sample size, short follow-up (benefits of revascularization generally emerge only after 3 to 9 years after bypass), and the exclusion of patients From the Duke Clinical Research Institute, Durham, North Carolina. Manuscript received February 24, 2000; revised manuscript received and accepted June 30, 2000. Address for reprints: Adam B. Greenbaum, MD, K-2 Cardiac Cath Lab, Henry Ford Hospital, 2799 West Grand Boulevard, Detroit, Michigan 48202. E-mail: [email protected].
1322
©2000 by Excerpta Medica, Inc. All rights reserved. The American Journal of Cardiology Vol. 86 December 15, 2000
METHODS
Patients: Between April 1, 1986, and February 28, 1994, 1,218 patients with proximal LAD disease undergoing a first diagnostic angiogram were identified. Of these, 754 underwent angioplasty, 149 underwent LIMA bypass (another 21 had vein grafts, which were excluded from analysis), and the remainder (n ⫽ 294) underwent no procedures within the first 30 days after angiography. These 1,197 eligible patients had coronary syndromes ranging from stable angina to evolving AMI, and reflected left ventricular function of all degrees. Data collection system: Pertinent baseline variables from the history, physical examination, laboratory tests, chest x-ray, and 12-lead electrocardiogram, as 0002-9149/00/$–see front matter PII S0002-9149(00)01235-2
study was 2%. Treatment selection (angioplasty, bypass, or medicine) Medicine Angioplasty LIMA Bypass was at the discretion of the attending (n ⫽ 285) (n ⫽ 754) (n ⫽ 149) cardiologist. For those treated medically, the cardiologist and primary Male sex 71% 70% 78% Age (yrs; median [IQR]) 60 (50–67) 59 (50–67) 60 (51–66) care physician determined therapy. Anginal symptoms Follow-up procedures: Patients Chest pain within 6 weeks 63% 73% 80% were contacted at 6 months and 1 Typical angina 67% 81% 73% year after presentation, then annualStable angina 28% 8% 14% Progressive angina 15% 20% 21% ly.10 The anniversary was the date of Unstable angina 6% 21% 23% angiography for medical patients, Recent infarction, stable 22% 18% 22% and the date of the first revascularRecent infarction, unstable 5% 8% 5% ization for those who underwent anAcute myocardial infarction 23% 24% 14% gioplasty or LIMA bypass. Patients Nocturnal angina 5% 8% 11% Pain frequency (median per 1 (0–4) 3 (0–5) 3 (1–7) were contacted by mailed questionweek [IQR]) naires; with telephone follow-up, Duration of ischemic symptoms 2 (0–57) 1 (0–8) 2 (0–13) National Death Index search, and vi(mo; median [IQR]) tal record review if the questionnaire Coexisting illness Diabetes 20% 16% 11% was not returned. The primary end Smoker 63% 64% 68% point for this study was all-cause Hypertension 42% 44% 46% mortality. The secondary end point Hyperlipidemia 32% 36% 36% was freedom from death, AMI, or Peripheral vascular disease 9% 5% 5% (repeat) revascularization. Cerebrovascular disease 8% 4% 7% Clinical measures of myocardial Data analysis: Baseline characterdamage istics and outcomes were summaPrior infarction 72% 56% 47% rized as medians and interquartile Prior congestive heart failure 17% 6% 8% ranges (for continuous measures) or New York Heart Association 13% 5% 7% class I or II percentages (for discrete measures). New York Heart Association 4% 2% ⬍1% The treatment group was defined by class III or IV the first revascularization within 30 IQR ⫽ interquartile range. days of angiography; we assumed an intent to treat with angioplasty or LIMA bypass based on the angiogram. If patients did not undergo revascularization within 30 days, we TABLE 2 Baseline Angiographic Characteristics assumed an intent to treat the patient LIMA medically. Once patients were asMedicine Angioplasty Bypass signed to a revascularization group (n ⫽ 285) (n ⫽ 754) (n ⫽ 149) (medicine, angioplasty, or LIMA byDisease severity pass), they remained in that group Proximal LAD occlusion 75% to 94% only 19% 16% 13% until death or the end of follow-up, Proximal LAD occlusion 75% to 94% and 10% 6% 15% ⱖ95% occlusion elsewhere in LAD regardless of later treatment. With Proximal LAD occlusion ⱖ95% 71% 78% 72% this definition, any patient who died Abnormal left ventricular contraction 83% 81% 58% within a few days of angiography Ejection fraction (%, median [IQR]) 44 (34–56) 52 (44–62) 56 (47–63) would have been ineligible to unMitral insufficiency dergo revascularization. Thus, all 1⫹ or 2⫹ 16% 12% 12% 3⫹ or 4⫹ 3% 2% 2% early deaths would be included in the medical therapy arm, unfairly penalAbbreviation as in Table 1. izing that group. To address this, we excluded from analysis all medical well as angiographic and procedural details, were col- patients who died or were lost to follow-up before the lected prospectively on standard forms during patient mean time to revascularization (1.5 days; n ⫽ 9). To control for baseline differences between care, and stored at the Duke Clinical Research Instigroups, we used an established model of long-term tute.10 Revascularization and medical therapy: Standard survival developed from a larger population of Duke angioplasty and cardiac surgical and anesthetic tech- patients, all treated for coronary disease between 1984 niques at Duke University Medical Center during this and 1990.13 The significant predictors of mortality in period have been described.11,12 During this study, the this model include age, sex, presentation, congestive overall procedural success rate for angioplasty (ⱕ50% heart failure, chest pain, comorbidity, ejection fracresidual stenosis) was 93%, and the emergency bypass tion, vascular disease, coronary artery disease index, rate for this cohort was 4.5%. The unadjusted 30-day and mitral insufficiency. We estimated disease severmortality rate for surgical patients included in this ity for patients in our study with the baseline risk TABLE 1 Baseline Clinical Characteristics
CORONARY ARTERY DISEASE/OPTIMAL TREATMENT OF PROXIMAL LAD DISEASE
1323
FIGURE 1. Observed (A) and adjusted (B) Kaplan-Meier survival estimates with LIMA bypass (thick solid line), angioplasty (thin solid line), or medical therapy (dashed line) of 1-vessel, proximal LAD disease. Med ⴝ medicine.
Index of Comorbidity.14 These variables, including treatment (medicine, angioplasty, or LIMA bypass) were added to the models. To avoid weakening the precision of estimated hazard ratios, inconsequential variables were removed from the models. Treatment comparisons are presented as Kaplan-Meier survival curves15 (unadjusted and adjusted) to contrast absolute survival differences, and as hazard ratios from Cox models (to compare relative survival differences of the 3 treatments, taken 2 at a time). For revascularized patients, days to an event were calculated from the date of procedure. For patients in the medical arm, timing began at the date of angiography. Because early procedural mortality creates a pattern of crossing survival curves, the proportional-hazards assumptions of the Cox model are violated. To account for this when comparing treatments, 3 Cox models were used to also calculate hazard ratios for 3 time intervals: baseline (date of treatment) to 30 days (with survival truncated at 30 days), 30 days to 1 year from baseline (conditional upon surviving 30 days and truncating survival at 1 year), and ⬎1 year from baseline (conditional upon surviving 1 year). Hazard ratios are shown with 95% confidence intervals.
RESULTS
Fewer women underwent surgery than did those taking other treatments (Table 1). Patients who had TABLE 3 Observed and Adjusted Survival Estimates by Treatment undergone angioplasty or LIMA LIMA Bypass grafting more often had either proMedicine (n ⫽ 285) Angioplasty (n ⫽ 754) (n ⫽ 149) gressive or unstable angina than did Interval Observed Adjusted Observed Adjusted Observed Adjusted medically treated patients. Risk factor profiles were similar except for 30 d 0.95 0.98 0.98 0.99 0.98 0.98 1 yr 0.90 0.95 0.95 0.96 0.96 0.96 more diabetics in the medical and 3 yrs 0.85 0.92 0.91 0.93 0.93 0.93 angioplasty groups. The greatest dif5 yrs 0.80 0.88 0.87 0.90 0.91 0.91 ferences among the groups occurred 7 yrs 0.73 0.82 0.82 0.86 0.88 0.88 in clinical measures of myocardial 9 yrs 0.67 0.78 0.76 0.80 0.80 0.79 damage. On average, medically treated patients had a longer duration (hazard) score from this model, calculated by first of symptoms, more prior AMI, and more prior conmultiplying each factor by its corresponding Cox re- gestive heart failure. Patients who received medicine or underwent angioplasty also had an abnormal ejecgression coefficient and then summing the products. A potentially important characteristic, presence of tion fraction more often (lowest in the medicine mitral insufficiency of grade 3 or 4, was an exclusion group), and grade 1⫹ and 2⫹ mitral insufficiency was criterion for the larger population. We included this most prevalent in medically treated patients (Table 2). The follow-up interval (time to death or last convariable in adjusted survival modeling (with the baseline hazard score), as well as year of angiography and tact) for the 1,188 patients after assignment to a treata weighted index of comorbidity. The index, modified ment group was a median 5.66 years (interquartile due to data availability, was based on the Charlson range, 3.66 to 8.01 years). Figure 1A and 1B show 1324 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 86
DECEMBER 15, 2000
FIGURE 2. Hazard ratios (boxes) and 95% confidence intervals (lines) for mortality differences between treatments. Vertical line at 1.0 indicates equivalence between treatments; values to the left of 1.0 indicate benefit from the first treatment listed. Abbreviations as in Figure 1.
FIGURE 3. Observed estimates of event-free survival with LIMA bypass (thick solid line), angioplasty (thin solid line), or medical therapy (dashed line) of 1-vessel, proximal LAD disease. Abbreviation as in Figure 1.
survival curves by treatment before and after adjustment for baseline differences. Observed and adjusted survival estimates by treatment over time are summarized in Table 3. A significant survival benefit was seen with revascularization compared with medical therapy (p ⫽ 0.001 for either LIMA bypass or angioplasty vs medical therapy). After adjusting for baseline clinical and angiographic differences, a trend toward improved survival with revascularization versus medicine persisted (hazard ratio for LIMA bypass, 0.74; 95% confidence interval, 0.44 to 1.23; hazard ratio for angioplasty, 0.82; 95% confidence interval, 0.60 to 1.11). Figure 2 shows the adjusted hazard ratios for the relative treatment effects on survival during follow-up and at 30 days, 1 year, and ⬎1 year (conditional upon surviving the previous period). After adjustment, survival did not differ significantly between revascularization approaches (hazard ratio, 0.90; 95% confidence interval, 0.55 to 1.48 for bypass vs angioplasty). Observed Kaplan-Meier curves and survival estimates at 30 days, and at 1, 3, 5, and 7 years for freedom from death, AMI, or (repeat) revasculariza-
tion by treatment are shown in Figure 3 and Table 4. Hazard ratios for the relative effect of the treatments on event-free survival throughout follow-up and at 30 days, 1 year, and ⬎1 year (conditional upon surviving the previous interval) are shown in Figure 4. Overall, LIMA bypass offered significant benefits over both medical therapy and angioplasty (hazard ratio, 0.43; 95% confidence interval, 0.28 to 0.64; p ⬍0.0001; hazard ratio, 0.29, 95% confidence interval, 0.20 to 0.42; p ⬍0.0001, respectively). Angioplasty also was associated with significantly more composite events than medical therapy (hazard ratio, 1.47; 95% confidence interval, 1.20 to 1.81; p ⫽ 0.0003). The event-free survival advantage of LIMA bypass over medical therapy began after the first 30 days (p ⫽ 0.005) and persisted throughout follow-up (p ⫽ 0.004). The advantage of LIMA bypass over angioplasty also persisted throughout follow-up but was no longer significant after the first year (conditional on surviving event free for the first year). The disadvantage of angioplasty versus medical therapy also diminished over time with an event-free survival advantage to angioplasty versus medicine evident after the first year, conditional on surviving event free until then (hazard ratio, 0.69; 95% confidence interval, 0.48 to 0.99; p ⫽ 0.05).
DISCUSSION
In this observational comparison of patients with 1-vessel, proximal LAD disease, surgical and percutaneous revascularizations were associated with nonsignificant survival advantages compared with medical therapy after adjustment for baseline clinical and angiographic differences. Further, LIMA bypass (but not angioplasty) was associated with a significant, long-term reduction in a composite of adverse cardiac events, largely driven by fewer repeat interventions. Angioplasty resulted in significantly more composite events than medical therapy overall. However, the disadvantages to angioplasty appeared to diminish over time; after the first year, angioplasty became more advantageous than medical therapy, and no longer differed significantly from surgical revascularization. Although no significant mortality benefit was seen with revascularization in this study, the results were directionally appropriate and consistent with a small survival benefit implied for this cohort. When grouped together, patients with more severe forms of 1-vessel disease (ⱖ95% proximal LAD occlusion) and those with 2-vessel disease (with a ⱖ95% LAD occlusion) have a significant 5-year survival advantage of 5 to 6 lives saved per 100 treated with revascularization versus medicine (a 30% to 35% relative reduction).5 In our study of patients
CORONARY ARTERY DISEASE/OPTIMAL TREATMENT OF PROXIMAL LAD DISEASE
1325
TABLE 4 Observed, Event-Free Survival Estimates by Treatment Interval 30 d 1 yr 3 yrs 5 yrs 7 yrs
Medicine (n ⫽ 285)
Angioplasty (n ⫽ 754)
LIMA Bypass (n ⫽ 149)
0.93 0.78 0.66 0.57 0.46
0.86 0.60 0.53 0.48 0.42
0.97 0.91 0.86 0.80 0.68
mortality.16 –18 Thus, the event-free survival curves for angioplasty-treated patients likely would have improved, but not the survival analysis in our study.
Acknowledgment: We thank Pat French and A. J. Mayhew for editorial assistance, and Anthony Doll for graphic assistance.
1. CASS Principal Investigators and their associates. Coronary artery surgery study (CASS): a randomized trial of coronary artery bypass surgery: survival data. Circulation 1983:68:939 – 950. 2. Varnauskas E. Twelve-year follow-up of survival in the randomized European Coronary Surgery Study. N Engl J Med 1988;319:332–337. 3. Veterans Administration Coronary Artery Bypass Surgery Cooperative Study Group. Eleven-year survival in the Veterans Administration randomized trial of coronary bypass surgery for stable angina. N Engl J Med 1984;311:1333–1339. 4. Yusuf S, Zucker D, Peduzzi P, Fisher LD, Takaro T, Kennedy JW, Davis K, Killip T, Passamani E, Norris R, Morris C, Mathur V, Varnauskas E, Chalmers TC. Effect of coronary artery bypass FIGURE 4. Hazard ratios (boxes) and 95% confidence intervals (lines) for differgraft surgery on survival: overview of 10-year results from ranences in event-free survival between treatments. Vertical line at 1.0 indicates domised trials by the Coronary Artery Bypass Graft Surgery equivalence between treatments; values to the left of 1.0 indicate benefit from Trialists Collaboration. Lancet 1994;344:563–570. the first treatment listed. Abbreviations as in Figure 1. 5. Jones RH, Kesler K, Phillips HR III, Mark DB, Smith PK, Nelson CL, Newman MF, Reves JG, Anderson RW, Califf RM. Long-term survival benefits of coronary artery bypass grafting and percutaneous transluminal angioplasty in patients with coronary artery disease. J Thorac Cardiovasc Surg 1996;111:1013–1025. with 1-vessel, proximal LAD disease only, evaluated 6. Kramer JR, Proudfit WL, Loop FD, Goormastic M, Zimmerman K, Simpfenseparately from those with 2-vessel disease involving the dorfer C, Horner G. Late follow-up of 781 patients undergoing percutaneous LAD, revascularization was associated with a nonsignif- transluminal coronary angioplasty or coronary artery bypass grafting for an isolated obstruction in the left anterior descending coronary artery. Am Heart J icant 5-year survival advantage over medical therapy of 1989;118:1144 –1153. about 2 to 3 lives saved per 100 thus treated (a 17% to 7. Cameron J, Mahanonda N, Aroney C, Hayes J, McEniery P, Gardner M, Bett 25% relative reduction). This subtler revascularization N. Outcome five years after percutaneous transluminal coronary angioplasty or artery bypass grafting for significant narrowing limited to the left benefit for 1-vessel, proximal LAD disease might have coronary anterior descending coronary artery. Am J Cardiol 1994;74:544 –549. been predicted, but our study had limited power to detect 8. Goy JJ, Eeckhout E, Burnand B, Vogt P, Stauffer J-C, Hurni M, Stumpe F, P, Sadeghi H, Kappenberger L. Coronary angioplasty versus left internal small differences. Nonetheless, a trend toward improved Ruchat mammary artery grafting for isolated proximal left anterior descending artery survival with revascularization was present and was stenosis. Lancet 1994;343:1449 –1453. noted within a population, with an overall 5-year mor- 9. Hueb WA, Bellotti G, de Oliveira SA, Arie S, de Albuquerque CP, Jatene AD, F. The Medicine, Angioplasty or Surgery Study (MASS): a prospective, tality rate similar to that of previous studies of this cohort Pileggi randomized trial of medical therapy, balloon angioplasty or bypass surgery for 5 (13%). single proximal left anterior descending artery stenoses. J Am Coll Cardiol Other differences between this and previous studies 1995;26:1600 –1605. 10. Rosati RA, McNeer JF, Starmer CF, Mittler BS, Morris JJ, Wallace AG. A new include our definition of significant LAD disease and information system for medical practice. Arch Intern Med 1975;135:1017–1024. how treatment groups were assigned. We defined signif- 11. Buller C, Mark DB, Phillips HR III, Stack RS. Cardiac catheterization and coronary angioplasty. In: Sabiston DC Jr, ed. The Davis-Christoicant LAD disease as ⱖ75% diameter stenosis. Although percutaneous pher Textbook of Surgery. Philadelphia: WB Saunders, 1991;1826 –1843. most patients (76%) also qualified as having ⱖ95% 12. Rankin JS, Sabiston DC Jr. The coronary circulation. In: Sabiston DC Jr, ed. occlusive lesions, previous definitions have ranged any- Textbook of Surgery. Philadelphia: WB Saunders, 1991;1957–1972. 13. Mark DB, Nelson CL, Califf RM, Harrell FE Jr, Lee KL, Jones RH, Fortin where from 50% to 95% stenosis.5,6 We also used a more DF, Stack RS, Glower DD, Smith LR, DeLong ER, Smith PK, Reves JG, Jollis traditional “intent-to-treat” approach than the Coronary JG, Tcheng JE, Muhlbaier LH, Lowe JE, Phillips HR, Pryor DB. Continuing of therapy for coronary artery disease: initial results from the era of Artery Surgery Study “method A” of assigning treatment evolution coronary angioplasty. Circulation 1994;89:2015–2025. 1 groups. Method A considered all patients to be medi- 14. Charlson ME, Pompei P, Alex KL, MacKenzie CR. A new method of cally treated until they underwent a first procedure, then classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–383. reassigned patients to the respective revascularization 15. Kaplan EL, Meier P. Nonparametric estimation from incomplete observagroup after intervention. We assigned treatment groups tions. J Am Stat Assoc 1958;53:457– 481. Serruys PW, de Jaegere P, Kiemeneij F, Macaya C, Rutsch W, Heyndrickx G, at first angiography, and eliminated from analysis all 16. Emanuelsson H, Marco J, Legrand V, Materne P, Belardi J, Sigwart U, Colombo A, patients (n ⫽ 9) who died before the median time to Goy JJ, Van den Heuvel P, Delcan J, Morel M, for the BENESTENT Study Group. A comparison of balloon-expandable-stent implantation with balloon angioplasty in revascularization (1.5 days). with coronary artery disease. N Engl J Med 1994;331:489-495. Although the study ended before the widespread patients 17. Versaci F, Gaspardone A, Tomai F, Crea F, Chiariello L, Gioffre PA. A use of intracoronary stents or platelet glycoprotein comparison of coronary-artery stenting with angioplasty for isolated stenosis of the IIb/IIIa inhibitors, we believe the results to remain proximal left anterior descending coronary artery. N Engl J Med 1997;336:817– 822. Kong DF, Califf RM, Miller DP, Moliterno DJ, White HD, Harrington RA, relevant to clinical patient care in 2000. Improved 18. Tcheng JE, Lincoff AM, Hasselblad V, Topol EJ. Clinical outcomes of theraoutcomes with these agents have resulted from re- peutic agents that block the platelet glycoprotein IIb/IIIa integrin in ischemic duced target vessel revascularization and AMI, not heart disease. Circulation 1998;98:2829 –2835.
1326 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 86
DECEMBER 15, 2000