Author + information
- Received January 29, 2001
- Revision received May 21, 2001
- Accepted June 14, 2001
- Published online October 1, 2001.
- Jurriën M ten Berg, MD∗,* (, )
- Johannes C Kelder, MD∗,
- Maarten Jan Suttorp, MD, PhD∗,
- Freek W.A Verheugt, MD, PhD, FACC† and
- H.W Thijs Plokker, MD, PhD, FACC∗
- ↵*Reprint requests and correspondence:
Dr. J. M. ten Berg, Department of Cardiology, St. Antonius Hospital, Koekoekslaan 1, 3435 CM Nieuwegein, the Netherlands
The goal of this research was to study the effect of planned angiography on late clinical outcome after percutaneous coronary intervention.
It is still largely unknown whether planned follow-up angiography after coronary angioplasty influences late outcome.
Randomization assigned 527 patients to clinical follow-up alone and 531 to clinical and six-month angiographic follow-up. The effect of planned angiography on clinical outcome at one and three years after coronary angioplasty was studied.
The two groups were well matched. At one year, more events occurred in the angiographic group than in the clinical group: 122 (23.2%) versus 88 (16.7%) (p = 0.01). While the incidence of death or myocardial infarction (MI) was similar at one year, the revascularization rate was higher in the angiographic group: 113 (21.3%) versus 67 (12.7%) (relative risk = 1.7, 95% confidence interval: 1.3 to 2.3, p = 0.0003). At three years, still more events had occurred in the angiographic group (146 [34.5%] vs. 114 [26.3%], p = 0.03). More reinterventions did not improve late survival. However, there was a nonsignificant reduction in MI (7 [1.3%] vs. 13 [2.5%], p = NS) and a significant improvement in functional class at the end of follow-up (freedom from angina 81% vs. 74%, p = 0.03). The effect of follow-up angiography on the reintervention rate was similar for stented and nonstented patients.
Planned follow-up angiography to evaluate the late results of coronary intervention led to a 1.7 times higher reintervention rate. This effect was similar for stented and nonstented patients. More reinterventions did not improve survival but tended to reduce the incidence of MI and led to a significantly better functional class at follow-up.
In daily practice, late results of percutaneous coronary intervention (PCI) are usually evaluated on the basis of cardiac events and the recurrence of angina pectoris. Follow-up angiography is mostly restricted to a subgroup of symptomatic patients. In the setting of a clinical trial, however, planned follow-up angiography is often used to assess late angioplasty results. Yet the effect of performing routine follow-up angiography on late clinical outcome is still largely unknown. This question was addressed in the BElgium NEtherlands STENT II study (Benestent-II ). Patients were randomized to stent implantation or balloon angioplasty, and there was a subrandomization to clinical follow-up alone or to both clinical and angiographic follow-up. It was found that follow-up angiography led to more reinterventions. Particularly, patients assigned to balloon angioplasty and to follow-up angiography had a strikingly higher reintervention rate. The investigators suggested that knowledge about the received treatment, stent or balloon, probably influenced the treating physician’s post-angioplasty management (1). The Balloon Angioplasty and Anticoagulation Study (BAAS) evaluated the effect of pre-treatment with coumarins on early and one-year events in nonselected patients (2). Also in BAAS, a subrandomization to planned six-month follow-up angiography was used to study its effect on patient management and outcome. In contrast with Benestent-II, BAAS patients were not randomized to balloon angioplasty or stenting. Therefore, BAAS allows evaluation of the effect of routine follow-up angiography in a patient population closely resembling daily practice.
Patients and procedure
From March 1996 to November 1997, all patients with symptomatic coronary artery disease referred by seven participating hospitals to our department (performing 2,000 procedures/year) for PCI were candidates for enrollment. Patients were randomized to aspirin alone or to aspirin plus coumarin treatment as previously described (2). Patients with an acute myocardial infarction (MI) with current use of oral anticoagulants, contraindications to coumarins or aspirin or a bypass graft lesion were excluded. Study medication was continued for at least six months. Stents were used for bailout situations and suboptimal results (provisional stenting), based on visual assessment only. Ticlopidine became available in the Netherlands during the trial period. Since then, when a stent was placed, it was left to the discretion of the operator whether to start ticlopidine (loading dose 500 mg followed by 250 mg twice a day for four weeks) or to continue coumarins. When ticlopidine was given to patients randomized to coumarins, the oral anticoagulants were discontinued. There was a 1:1 subrandomization to clinical follow-up alone or to both clinical and angiographic follow-up. The study was carried out according to the principles of the Declaration of Helsinki and was approved by our institutional ethics committee.
All patients were seen by their referring physicians in outpatient clinics at six weeks, three months and one year. Clinical data and angina pectoris class were obtained from the patients and the referring physicians. In March 2000, all patients were additionally contacted by telephone to study the occurrence of events and the recurrence of angina pectoris between one and three years of follow-up. For those patients randomized to angiographic follow-up, angiography was performed at six months. Angiography was performed earlier only on request from the referring physician when objective evidence of ischemia was obtained. For those patients randomized to clinical follow-up alone, the decision to perform angiography was left to the discretion of the referring physicians. Reintervention was based on angiographic restenosis and recurrent chest pain with electrocardiographic or scintigraphic evidence of ischemia. Quantitative coronary analysis was performed with the Computer Measurement System (Medis, the Netherlands) by an independent core laboratory blinded to the assigned therapy (Heartcore, Leiden, the Netherlands).
This study was undertaken to evaluate the effect of undergoing planned six-month follow-up angiography and the subsequent knowledge of the atherosclerotic status on the performance of revascularization and the occurrence of late events. The analysis was performed on an intention-to-treat basis. We used a policy of provisional stenting. Thus, many stents were placed in bailout situations, which were associated with the occurrence of events. Therefore, for the comparison of the late effect of planned follow-up angiography in stented versus nonstented patients, the early procedure-related events were excluded from this analysis.
Definitions and end points
The primary end point was the composite of death, MI and revascularization at one year. The same end point was evaluated at three years of follow-up. Revascularization included coronary angioplasty or coronary artery bypass surgery (CABG) for restenosis and for new lesions. Events were reviewed by a safety committee blinded to study medication and mode of follow-up assignment. Definitions of end points were described previously (2).
The two groups were compared using the ttest for continuous variables and the chi-square test or, when appropriate, Fisher exact test for discrete variables. Discrete variables were compared in terms of relative risks (RR) with 95% confidence intervals (CI). Event-free survival was calculated by means of the Kaplan-Meier method. Differences in survival times were assessed by the log-rank test. A p value <0.05 was considered significant. Multivariate Cox proportional hazard regression was used to study predictors for undergoing reintervention during follow-up in the stented and nonstented group.
In BAAS, 1,028 patients were randomized to aspirin alone or to aspirin plus coumarins. The clinical and angiographic baseline characteristics were described previously and were well-matched (2). Statistically, the clinical (Table 1)and angiographic baseline (Table 2)characteristics of the 527 patients in the clinical follow-up group and of the 531 patients in the angiographic follow-up group did not differ significantly. Additionally, the percentages of patients with multivessel intervention (1.43 vs. 1.46 interventions per patient, p = NS) were similar in the two groups. One hundred and ninety-eight patients (35.9%) in the clinical follow-up group and 178 patients (33.5%) in the angiographic follow-up group received one or more stents (p = NS). The clinical baseline characteristics of the stented and nonstented patients did not differ significantly (data not shown). However, in the stented group, more proximal left anterior descending lesions (19.3% vs. 11.4%) and less total occlusions (4.9% vs. 10.3%) were treated than in the balloon-treated group. Patients were followed for a mean of 29.5 ± 7.0 months.
Clinical versus angiographic follow-up
Of the 531 patients randomized to follow-up angiography, 51 patients (9.6%) did not undergo angiography for various reasons: failed angioplasty (6 patients), subacute stent thrombosis (6 patients), death (6 patients), CABG based on the pre-angioplasty angiogram (3 patients), administrative error (2 patients), groin complication during index angioplasty (9 patients) and consent withdrawn (21 patients). These 51 patients without angiographic follow-up did not differ significantly from the 480 patients with angiographic follow-up with respect to baseline characteristics or event rates (data not shown). Of 528 patients randomized to clinical follow-up alone, 134 patients (25.3%) underwent angiography during follow-up. Of these, angiography was performed in 27.4% of the stented patients, compared with 24.2% of the nonstented patients (p = 0.70).
At one year, more events occurred in the angiographic group than in the clinical group: 122 (23.2%) vs. 88 (16.7%), p = 0.01 (Fig. 1). The incidence of death or MI at one year was similar. The difference was, therefore, due to a higher target lesion reintervention rate in the angiographic group: 102 (19.2%) vs. 61 (11.6%) (RR = 1.7, 95% CI: 1.2 to 2.3, p = 0.001) (Fig. 2and Table 3). The higher reintervention rate concurred with more patients being asymptomatic at six months in the angiographic group (76.9%) compared with those in the clinical group (71%, p = 0.03).
At three years, there were 146 events (34.5%) in the angiographic group and 114 (26.3%) in the clinical group (p = 0.03) (Fig. 1). Again, the difference was mainly due to more reinterventions in the angiographic group. More reinterventions did not improve survival (Fig. 3). However, there was a nonsignificant reduction in MIs after the first year in the angiographic group (7 [1.3%] vs. 13 [2.5%], p = NS). Most of these late MIs occurred in the last year of follow-up. At three years, the functional class was significantly better in the angiographic group; 81% of the patients were free from angina pectoris versus 74% in the clinical group (p = 0.03) (Table 4).
Stent versus balloon
There was no difference in event rates between the group treated with a stent and the group treated with a balloon alone, whether in the clinical (Fig. 4)or angiographic follow-up group (Fig. 5). Thus, for patients treated with a balloon alone (Fig. 6)and for stented patients (Fig. 7), planned follow-up angiography led to an increased revascularization rate. The effect of planned angiography on the performance of revascularizations was, however, more pronounced in the balloon group (Fig. 5). Because the baseline angiographic characteristics differed in the stented and nonstented group with respect to the number of proximal left anterior descending lesions and total occlusions, multivariate analysis was performed. None of the baseline characteristics was shown to be a predictor for undergoing revascularization during follow-up. In the clinical follow-up group, the reintervention rate was similar for stented and nonstented patients (11.0% vs. 10.3%, p = NS).
For patients undergoing coronary intervention, there is still controversy about the optimal means of evaluating the late success of the procedure. Follow-up angiography is the best method for detecting restenosis and is, therefore, used in most clinical trials to evaluate the effect of new drugs or devices. But is it also in the interest of the patients to perform routine follow-up angiography in all patients? Follow-up angiography may lead to unnecessary reinterventions in asymptomatic patients. On the other hand, it may detect symptomatic patients with restenosis who would otherwise have escaped reintervention. It is, therefore, essential to know whether a policy of performing routine follow-up angiography prevents deaths or MIs and whether it reduces angina pectoris.
In the Benestent-II trial, patients were randomly assigned to stenting or balloon angioplasty. Furthermore, it used the unique setting of subrandomization to angiographic follow-up to study the effect on patient management and outcome. It was shown that planned follow-up angiography led to more reinterventions. In addition, it was concluded that the initial treatment, stent or balloon, had a major influence on the reintervention rate. In the stent group, the patients with angiographic follow-up underwent 2.5 times more reinterventions than those with clinical follow-up alone. However, in the balloon group, this ratio was only 1.5. It was concluded that receiving a stent could have reassured the treating physician, resulting in not performing ischemia detection and/or follow-up angiography in case of the recurrence of symptoms. On the other hand, the known higher restenosis rate after balloon angioplasty could have lowered the threshold for follow-up angiography and subsequent stent implantation (1).
The present BAAS trial confirms that planned follow-up angiography leads to more reinterventions. However, in contrast with Benestent-II, the effect of follow-up angiography in BAAS was similar for stented and nonstented patients. In particular, the percentage of patients undergoing angiography in the clinical follow-up group of BAAS was similar for stented and nonstented patients. Thus, the BAAS trial could not confirm the conclusions of Benestent-II that receiving a stent could have reassured the treating physician or that the known higher restenosis rate after balloon angioplasty could have lowered the threshold for follow-up angiography. The BAAS data suggest that the difference between a stented and a balloon-treated lesion occurs after the follow-up angiogram is made. We could find an only slightly higher reintervention rate in the balloon-treated patients in the angiographic follow-up group (Fig. 5). We speculate that this difference is due to reluctance to redilate in-stent restenosis knowing the high recurrence rates (3). Thus, the outcome of a restenotic lesion that is subsequently stented is presumed to be better than that of an in-stent restenosis that is redilated. This may be the reason why restenosis after balloon angioplasty was more frequently treated than in-stent restenosis. We must, however, consider the fact that BAAS patients were not randomized to undergo stenting. Nevertheless, none of the differences in baseline characteristics between the stented and nonstented patients was a predictor for undergoing revascularization during follow-up. Therefore, cautious comparisons between the two groups can be made.
Planned follow-up angiography and mortality
A further question in relation to the effect of follow-up angiography is whether the knowledge on the atherosclerotic status and the following reinterventions are beneficial or harmful to the patient. Rupprecht et al. (4)studied 400 patients who underwent successful coronary angioplasty and were advised to undergo a six-month follow-up angiogram. Of these, 85 patients did not undergo follow-up angiography. The study showed a 2.5 times increased revascularization rate in the patients who underwent follow-up angiography. Coronary artery bypass surgery and MI occurred at a similar rate in the two groups. However, the absence of a six-month angiographic follow-up was an independent predictor of death associated with a 2.7 times higher mortality rate during the 10-year follow-up period (cardiac death 16/315 [5%] vs. 16/85 [16%]). It was concluded that the survival advantage of patients with follow-up angiography was related to the significantly higher rate of repeat angioplasties in this group. Therefore, their advice was that routine follow-up angiography after successful angioplasty should be strongly considered (4).
In BAAS and Benestent-II, there was no effect of follow-up angiography on mortality. It might be that the follow-up in these two studies was too short to detect an effect on mortality. On the other hand, in the study by Rupprecht et al. (4), the survival curves diverted after one year, whereas there was no divergence even after three years in BAAS. Furthermore, the study by Rupprecht et al. (4)was retrospective, which might have led to a bias, as asymptomatic patients are less likely to undergo follow-up angiography (5). More importantly, as Rupprecht et al. (4)rightfully suggested, the patients who are willing to undergo follow-up angiography might be more compliant with a healthier life style. In addition, these patients might receive better medical treatment based on knowledge about the atherosclerotic status of the patient, which could affect late outcome. Thus, in the study by Rupprecht et al. (4), follow-up angiography might have led to better secondary prevention and, therefore, less progression, which is suggested by the fact that the main difference in survival occurs only six years post-intervention, while the majority of repeat angioplasties were performed at the time of the angiographic follow-up (4).
Planned follow-up angiography and MI
In BAAS, the patients with follow-up angiography had 50% fewer late MIs. For the most part, these MIs occurred in the third year of follow-up; therefore, we do not think that the lower incidence was caused solely by the increased number of reinterventions, almost all of which were performed in the first year of follow-up. It is more probable that aggressive secondary prevention, based on knowledge of the atherosclerotic status of the patient, could have been part of the reason for fewer MIs. However, because we did not systematically study risk factor management during the study, we can only speculate on the cause of the reduction of MIs. Rupprecht et al. (4), in their explanation for the survival advantage of patients with angiographic follow-up, suggested asymptomatic restenosis or silent reocclusion to be of greater consequence in cases of disease progression in another vessel, because of the compromise of noninvolved systolic function when an acute event occurs. Surprisingly, in their study, MI was not increased in the group without angiographic follow-up.
Planned follow-up angiography and symptoms
In BAAS, more reinterventions led to fewer symptoms at six months and at three years. Coronary angioplasty is known to be more effective than medical treatment in decreasing symptoms (6). Furthermore, patients in clinical trials always do better than patients in the real world. With this in mind, one can surmise that the chances of a symptomatic patient’s not receiving the benefit from repeat revascularization would be larger than observed in this trial. As a limitation, we did not study the extra costs of routine follow-up angiography and of the increased reintervention rate. On the other hand, routine follow-up angiography could lead to an earlier relief of symptoms, an earlier discontinuation of anti-anginal medication in case there is no restenosis, better risk factor management, a lower late-event rate and, very importantly, an earlier reintegration into daily social and working life. Therefore, our observations call for a prospective randomized trial including costs and cost-effectiveness measurements.
In the nonselected patient population of BAAS, planned follow-up angiography to evaluate the late results of coronary angioplasty led to an increased reintervention rate. This effect was similar for stented and nonstented patients. More reinterventions did not improve survival but tended to reduce the incidence of MIs and significantly improved the functional class.
☆ Supported by grant 94,138 of the Netherlands Heart Foundation.
- Balloon Angioplasty and Anticoagulation Study
- BElgium NEtherlands STENT II study
- coronary artery bypass surgery
- confidence interval
- myocardial infarction
- percutaneous coronary intervention
- relative risk
- Received January 29, 2001.
- Revision received May 21, 2001.
- Accepted June 14, 2001.
- American College of Cardiology
- ten Berg J.M,
- Kelder J.C,
- Suttorp M.J,
- et al.
- Narins C.R,
- Ellis S.G
- Kuntz R.E,
- Keaney K.M,
- Senerchia C,
- Baim D.S