The clinical outcome after coronary bypass surgery: a 30-year follow-up study

Transcription

1 European Heart Journal (2009) 30, doi: /eurheartj/ehn530 CLINICAL RESEARCH Coronary heart disease The clinical outcome after coronary bypass surgery: a 30-year follow-up study Ron T. van Domburg*, Arie Pieter Kappetein, and Ad J.J.C. Bogers Thoraxcenter, Erasmus Medical Centre, Room Ba559, Dr Molewaterplein 40, Rotterdam, GD 3015, The Netherlands Received 25 February 2008; revised 5 November 2008; accepted 18 November 2008; online publish-ahead-of-print 9 December 2008 Aims To investigate the long-term clinical outcome (up to 30 years) after coronary artery bypass graft (CABG) surgery and to assess the life expectancy (LE) among subgroups of patients.... Methods We analysed the 30-year outcome of the first 1041 consecutive patients in our institution (age at operation 53 years, and results 88% male) who underwent venous CABG between 1971 and During follow-up, every 5 7 years follow-up status was obtained by reviewing the hospital records and from general practitioners and civil registries. Data were collected on death and repeat coronary revascularization procedures. Follow-up was complete in 98%. Median follow-up was 29 years (26 36 years). The cumulative 10-, 20-, and 30-year survival rates were 77%, 40%, and 15%, respectively. Overall, 623 coronary re-interventions were performed in 373 patients (36%). The cumulative 10-, 20-, and 30-year freedom from death and coronary re-intervention rates were 60%, 20%, and 6%, respectively. Age [hazard ratio (HR) 1.04/year], extent of vessel disease (VD) (two-vd HR 1.4; three-vd HR 1.9), left main disease (HR 1.6) and impaired left ventricular ejection fraction (LVEF) (HR 1.8) were independent predictors of mortality. We were able to assess the exact LE by calculating the area under the Kaplan Meier curves. Overall LE after first CABG was 17.6 years. LE in patients with one-, two-, and three-vd was 21.4, 18.8, and 15.4 years, respectively (P, ). Patients with impaired LVEF had a significant shorter LE than patients with normal LVEF (13.9% vs. 19.3%; P, ).... Conclusion This 30-year follow-up study comprises the almost complete life cycle after CABG surgery. Overall median LE was 17.6 years. As the majority of the patients (94%) needed a repeat intervention, we conclude that the classic venous bypass technique is a useful but palliative treatment of a progressive disease Keywords Life expectancy Surgery Prognsis Survival Mortality Epidemiology Introduction Coronary artery bypass graft (CABG) surgery is a well-accepted treatment in patients with severe angina since the 1970s. 1 Relief of anginal symptoms is still the major goal of CABG to obtain better quality of life, as well as increase of life expectancy (LE). Today, several different grafting techniques are used, but in the beginning only venous conduits were used. Although nowadays arterial grafts are preferred, 2 venous grafts are still being used extensively. 3 As vein graft atherosclerosis has been shown to be more aggressive than native vessel disease (VD), 4,5 it remains very important to evaluate this on the long-term. 6 Furthermore, it has been shown that certain subsets of patients may have better LE. 7,8 For this reason, we updated the 30-year outcome of the first 1041 consecutive patients who underwent venous CABG between 1971 and Because that period comprises almost the whole post-procedural life cycle, we were able to calculate the exact LE of subgroups of patients and to investigate the occurrence of atherosclerosis in the vein grafts. Methods Patient population The baseline characteristics and earlier follow-up reports up to 20 years have been described in detail before In summary, all 1041 consecutive patients between 1971 and 1980 with severe ischaemia who underwent a first isolated venous CABG procedure in the Erasmus Medical Centre Rotterdam were included in this study. * Corresponding author. Tel: þ , Fax: þ , r.vandomburg@erasmusmc.nl Published on behalf of the European Society of Cardiology. All rights reserved. & The Author For permissions please journals.permissions@oxfordjournals.org.

2 454 R.T. van Domburg et al. No internal mammary artery grafts were used. All patients underwent pre-operative catherization. Baseline clinical variables, data concerning pre-operative catheterization and peri-procedural data were collected, when available and stored, first on a HP-85 micro computer, and later on a main-frame PDP computer. The first database program was written in BASIC. The indication for surgery was based on angina, stable or unstable, refractory to maximal pharmacological treatment available at that time (mainly a combination of nitrates and beta-blockers; calcium antagonists were not yet available) and the findings at catheterization. Both elective and urgent surgery procedures were included. Excluded were those patients who needed additional surgery such as valve replacement or aneurysmectomy. The study complies with the Declaration of Helsinki. Follow-up Follow-up status was obtained by reviewing the hospital records and from general practitioners and civil registries in September 2007 of all patients who were alive at the previous follow-up in Data were collected on death and repeat coronary revascularization procedures. If necessary, patients were contacted by telephone. Follow-up was complete in 98%. Survival status of 18 patients, who had moved abroad, could not be retrieved and the last available follow-up data were used. Statistical analysis Cumulative survival analyses were constructed using the Kaplan Meier method. Among patient subgroups, the Mantel and Haenszel log-rank test was used to compare survival curves. The expected survival in a reference population was calculated using age- and gender-specific mortality data from the Netherlands in 1975 ( and compared with survival in patients after CABG. Since the mean age of our study population was 53 years and 88% were male, mortality risks were weighted accordingly. The Cox proportional hazards model was used to identify independent risk factors for 30 years mortality. Preselected baseline characteristics were age, gender, diabetes, hypertension, dyslipidaemia, current smoking, extent of VD, and impaired left ventricular ejection fraction (LVEF) (,0.55). To investigate whether the appearance or need of re-intervention has changed the LE of any coronary patient in his/her life span, we added repeat revascularization as a time-dependent covariable in the Cox model. Proportionality of hazards was tested graphically based on visual inspection of log log survival curves, and by performing a formal test of proportionality based on Schoenfeld residuals for each variable in the model. Cox regression analyses showed no statistically significant interactions with time. We calculated a linear shrinkage factor for the regression coefficients with bootstrapping. A key problem of regression modelling is that the regression coefficients are overestimated for predictive purposes. We calculated a linear shrinkage factor for the regression coefficients with bootstrapping, including the confidence intervals. The difference between performance in the bootstrap samples and the original sample is an estimate of the optimism in the apparent performance (miscalibration). This difference is averaged to obtain a stable estimate of the optimism. Shrunk coefficients were calculated by multiplication of the standard coefficients with the shrinkage factor, which might take values between 0 and 1. The prognostic ability of the model, i.e. the power to discriminate between the survivors and those who died, was estimated with the c-statistic. The c-statistic provides a quantitative summary of the discriminative ability of a predictive model. A value of 0.5 indicates that the model does not have any discriminatory ability, and a value of 1.0 represents perfect discrimination. The c-index is the generalized form of the c-statistic for censored data, 12 which specifically incorporates censoring information, is reasonable and reliable. 13 Calibration refers to whether the model agrees with the observed probabilities and was assessed with the Hosmer Lemeshow statistic. These statistical analyses were performed with S-plus software (MathSoft, Inc., Seattle WA, version 2000). In the multivariable model, multiple imputation was used to handle the 167 patients in whom the LVEF was missing. Multiple imputation methods are known to be superior to complete case analyses. 14 For all tests, a P-value,0.05 (two-sided) was considered significant. Life expectancy LE after CABG was calculated from the area under the Kaplan Meier curve. 15 To calculate the exact LE, the curves were extended beyond 30 years using the age- and sex-specific mortality data from the reference population in the Netherlands, assuming that those few CABG patients who had survived 30 years would have similar further LE as their age- and sex-matched peers. Among patient subgroups, the Mantel and Haenszel log-rank test was used to compare life expectancies. Results Survival Mean age at first operation was 53 years (+8 years) and 88% were male (Table 1). Single VD was present in 19% of the patients and LVEF was impaired in 27%. Diabetes was present in 9% of the patients, dyslipidaemia in 22%, hypertension in 22%, and current smoking in 53%. Median follow-up was 29 years (range years). At follow-up, 196 patients were still alive, of whom 10 patients (1%) reached the age of 90 years. Cumulative survival rates at 10-, 20-, and 30 years were 77%, 40%, and 17%, respectively (Figure 1). The mortality rate was 3.2% in the first year (including 1.2% peri-operative mortality), which then decreased to 0.9% in the third year. Thereafter, mortality increased to 4% per year until the 15th year. Between 15 and 20 years, the yearly mortality Table 1 Baseline characteristics Number of patients 1041 Age (year) (SD) 53 (+8) Range (year) Men 915 (87.9) Diabetes 82 (8.4) Hypertension 224 (21.6) Dyslipidemia 232 (22.4) Current smoker 589 (57.8) One vessel 192 (18.5) Two vessels 320 (30.8) Three vessels 445 (42.8) Left main 83 (8.0) Impaired LV ejection fraction 274 (31.4)

3 Thirty years after CABG 455 Table 2 Multivariable predictors of 30 years mortality Figure 1 Cumulative survival and event-free survival. Cumulative survival (represented by a red line) and cumulative event-free survival (death or repeat revascularization) (in blue) after venous coronary bypass surgery of 1041 patients. The dotted line is the survival of the normal population (matched for age and gender). Figure 2 Yearly reintervention rates. Yearly incidence of repeat revascularization (repeat CABG in red, PCI in blue) as a percentage of the number of pts at risk in that year. decreased to 3.5%, followed by a further decrease towards 2.5% per year beyond 20 years. Between 10 and 25 years follow-up, survival after CABG was 10 15% lower than in the normal population (matched for age and gender). Beyond 25 years, survival after CABG converged to that of the normal population. Re-intervention Overall, 623 coronary re-interventions were performed in 373 of the 1041 patients (36%). Repeat CABG was performed in 309 patients, angioplasty in 148 patients and in 64 both a re-cabg and angioplasty were performed. At least three re-interventions were needed in 35 patients, four re-interventions were needed in 15 patients, and five re-interventions were needed in 12 patients. In the first 3 years after the index operation, only re-cabg was performed while angioplasty was not available yet. Up to the 8th year, the yearly incidence of re-intervention (adjusted for the number of patients at risk) was 1.8% (Figure 2). Between 8 and 13 years, the re-intervention rates more than doubled to 4.2% per year. Thereafter, up to Year 20, the re-intervention rates HR 95% CI Clinical Age (years) Male Diabetes Smoking Hypertension Dyslipidaemia Angiographic One vessel 1.0 Two vessels Three vessels Left main Impaired LVEF Repeat revascularization HR, hazard ratio; 95% CI, 95% confidence interval. decreased to 3% per year. Beyond 20 years, re-intervention procedures became rare and remained constant at 1%. Cumulative freedom from death and re-intervention rates was at 10-, 20-, and 30 years 60, 20, and 6%, respectively. Multivariable analysis The most important independent predictors of mortality were three-vd [hazard ratio (HR) 1.9; 95% CI ] and impaired LVEF (HR 1.8; 95% CI ) (Table 2). The other independent predictors were elderly age (HR 1.04; 95% CI ), two-vd (HR 1.4; 95% CI ), and left main disease (HR % CI ), while gender was not. Also repeat revascularization, forced in the model as a time-dependent covariable, was no predictor of lower mortality. The c-statistic was 0.76 which means a reasonable fit of the model. To test the stability of the Cox proportional hazards regression model, we performed a bootstrap investigation (500 replications) using the same variables tested in the original model. The significant variables in the bootstrap model were those selected in the original analysis. Survival in subgroups Survival rates between males and females were similar over time (Table 3). A clear survival benefit was observed in patients with two-vd as compared with those with three-vd (benefit: 11 years at 10 years follow-up and 17 years at 20 years). There was a clear survival advantage for those with a normal LVEF with almost 50% remaining alive at 20 years compared with only 20% of those with impaired LVEF. Life expectancy Overall LE was 17.6 years (Table 4). The highest LE was observed in patients with normal LVEF. These patients had a 5.4 year higher LE than in those patients with impaired LVEF (19.3 vs years; P, ). Furthermore, patients with two-vd had a significant

4 456 R.T. van Domburg et al. Table 3 Thirty-year cumulative survival overall and in specific subgroups Cumulative survival rates (%)*... P-value 10 years years years... Survival 95% CI Survival 95% CI Survival 95% CI... Normal population Overall Clinical Age 60 years ,0.0001,60 years Sex Male Female Diabetes Yes ,0.01 No Hypertension Yes No Dyslipidaemia Yes No Angiographic One vessel , Two vessels Three vessels Left main Yes No LV ejection fraction Normal , Impaired *Thirty-year cumulative mortality rates (%) are derived from Kaplan Meier lifetables; P-value, log-rank test; 95% CI, 95% confidence interval. longer LE (3.4 years; 18.8 vs years; P, 0.01) than patients with three-vd. The LE in patients with normal LVEF without three-vd was almost similar to that of the normal population (20.8 vs years; P ¼ 0.3). Discussion This report represents the longest and most complete follow-up, comprising almost the complete life cycle after CABG surgery, of the first patients who underwent CABG. The main finding is that although mortality is higher after CABG than in the normal population in the first 15 years after the index operation, survival rates decreased more slowly in the CABG group thereafter and eventually converged to the normal population. Or, in other words, in about half of the patients, CABG provides the patients a prognosis which is the same as in the normal population. This yields especially for those with a normal left ventricular function and without triple VD. We were in the unique position to calculate the LE accurately almost without making assumptions. Overall LE after venous CABG was 18 years. We observed that patients with a normal left ventricular function had a better gain LE of 5.4 years (19.3 vs years) as compared with patients with impaired left ventricular function. Also the extent of VD had a great impact on prognosis. Patients with triple VD had a 3.3 years shorter LE (15.5 vs years) than patients with two-vd. Perhaps not expected but left main disease did not have an influence on survival. These observations indicate that estimates of LE and cost-effectiveness

5 Thirty years after CABG 457 Table 4 Cumulative life expectancy overall and in specific subgroups Life expectancy*... Years 95% CI (years) Normal population 21.6 Overall Clinical Age 60 years ,60 years Sex Male Female Diabetes Yes No Hypertension Yes No Dyslipidaemia Yes No Smoking Yes No Angiographic One vessel Two vessel Three vessels Left main Yes No LV ejection fraction Impaired Normal *Thirty-year cumulative mortality rates (%) are derived from Kaplan Meier lifetables; P-value, log-rank test; 95% CI, 95% confidence interval. analyses of randomized clinical trials should be interpreted with caution as long as true long-term data are lacking. Therefore, true long-term follow-up studies of randomized trials should be conducted more frequently. A hypothesis was that repeat revascularization may have changed the LE. We entered repeat intervention as a timedependent variable in the multivariable model and it turned out that it was far from significant. Thus, a repeat intervention provides relief of symptoms, but we were not able to prove that this influenced the LE in a positive or negative way. Although no randomized trials are known, large registries have shown that arterial grafts improve survival significantly, especially if an IMA was constructed to the LAD at the time of the intervention. 2,3 At present, vein grafts are still used extensively in combination with arterial grafts. Saphenous vein grafts, the conduits exclusively used in our study, were protective in the first 7 years. Thereafter, mortality as well as repeat interventions increased significantly because of gradual loss of patency. In an earlier study, we reported that 80% of the repeat revascularizations were due to failure of the new venous bypass graft. 16 Furthermore, progression of coronary artery disease in native vessels, especially distal to the vein grafts may further increase mortality and morbidity. 6 Limitations Our study has clear limitations because treatment for coronary artery disease has changed and also the risk profiles of today s populations are different. Although our study population was a consecutive patient series, it was a very selective one. Mean age was 10 years younger than today and patients over 70 years did not undergo surgery. Furthermore, the waiting lists were very long and more importantly, aspirin, beta-blockers, lipid treatment, and life-style changes are now routinely performed which was not the case at that time. Furthermore, medical ischaemic treatments such as calcium antagonists were not available then. Operation techniques have been changed. Internal mammary artery grafts are being used as frequently as possible and this conduit remains longer patent than vein grafts with subsequently less repeat coronary interventions. Furthermore, in contrast to current knowledge, only a few risk factors were known to influence survival in the 1970s and only those were collected. Also, since the introduction of stents, especially drug-eluting stents, PCI has replaced CABG for many indications and only patients with complex coronary VD remain candidates for CABG. Conclusion This 30-year follow-up study comprises the almost complete life cycle after CABG surgery. Overall median LE was 17.6 years. As the majority of the patients (94%) needed a repeat intervention, we concluded that the classic venous bypass technique is a useful but palliative treatment of a progressive disease. Conflict of interest: none declared. References 1. ACC/AHA guidelines and indications for coronary artery bypass surgery. A report of the American College of Cardiology/American Heart Association task force on assessment of diagnostic and therapeutic cardiovascular procedures (subcommission on coronary artery bypass graft surgery). Circulation 1991;83: Cameron A, Davis KB, Green G, Schaff HV. Coronary bypass surgery with internal-thoracic-artery grafts effects on survival over a 15-year period. N Engl J Med 1996;334: Bradshaw PJ, Jamrozik K, Le M, Gilfillan I, Thompson PL. Mortality and recurrent cardiac events after coronary artery bypass graft: long term outcomes in a population study. Heart 2002;88: Angelini GD, Newby AC. The future of saphenous vein as a coronary artery bypass conduit. Eur Heart J 1989;10: Lytle BW, Loop FD, Cosgrove DM, Ratliff NB, Easley K, Taylor PC. Long-term (5 to 12 years) serial studies of internal mammary artery and saphenous vein coronary bypass grafts. J Thorac Cardiovasc Surg 1985;89:

6 458 R.T. van Domburg et al. 6. Campeau L, Enjalbert M, Lespérance J, Bourassa MG, Kwiterovich P Jr, Wacholder S, Sniderman A. The relation of risk factors to the development of atherosclerosis in saphenous-vein bypass grafts and the progression of disease in the native circulation. A study 10 years after aortocoronary bypass surgery. N Engl J Med 1984;311: Myers WO, Blackstone EH, Davis K, Foster ED, Kaiser GC. CASS Registry long term surgical survival. Coronary Artery Surgery Study. J Am Coll Cardiol 1999; 33: Califf RM, Harrell FE Jr, Lee KL, Rankin JS, Hlatky MA, Mark DB, Jones RH, Muhlbaier LH, Oldham HN Jr, Pryor DB. The evolution of medical and surgical therapy for coronary artery disease. A 15-year perspective. J Am Med Assoc 1989;261: Laird-Meeter K, Penn OC, Haalebos MM, van Domburg R, Lubsen J, Bos E, Hugenholtz PG. Survival in 1041 patients with consecutive aorto-coronary bypass operations. Eur Heart J 1984;5: Laird-Meeter K, van Domburg R, Bos E, Hugenholtz PG. Survival at 5 to 10 years after aorto-coronary bypass operations in 1041 consecutive patients. 1. Eur Heart J 1987;8: Veldkamp RF, Valk SD, van Domburg RT, van Herwerden LA, Meeter K. Mortality and repeat interventions up until 20 years after aorto-coronary bypass surgery with saphenous vein grafts. A follow-up study of 1041 patients. 1. Eur Heart J 2000;21: Harrell FE Jr, Califf RM, Pryor DB, Lee KL, Rosati RA. Evaluating the yield of medical tests. J Am Med Assoc 1982;247: May M, Royston P, Egger M, Justice AC, Sterne JA, ART Cohort Collaboration. Development and validation of a prognostic model for survival time data: application to prognosis of HIV positive patients treated with antiretroviral therapy. Stat Med 2004;23: van der Heijden GJ, Donders AR, Stijnen T, Moons KG. Imputation of missing values is superior to complete case analysis and the missing-indicator method in multivariable diagnostic research: a clinical example. J Clin Epidemiol 2006;59: van Domburg RT, Sonnenschein K, Nieuwlaat R, Kamp O, Storm CJ, Bax JJ, Simoons ML. Sustained benefit 20 years after reperfusion therapy in acute myocardial infarction. J Am Coll Cardiol 2005;46: Laird-Meeter K, van Domburg R, van den Brand MJ, Lubsen J, Bos E, Hugenholtz PG. Incidence, risk, and outcome of reintervention after aortocoronary bypass surgery. Br Heart J 57: