Three-Year Clinical and Angiographic Follow-Up After Intracoronary Radiation

Restenosis continues to be the major limitation of catheter-based vascular procedures. Early preclinical studies have demonstrated radiation therapy to be a uniquely efficacious treatment for restenosis.^12345678 Although short-term clinical results have been promising, the long-term efficacy and, most importantly, safety of this technique are not known. The possibility of late adverse angiographic findings such as aneurysm formation, perforation, or accelerated vascular disease is of significant concern.²³ The objective of this report is to document angiographic and clinical outcome 3 years after treatment of restenotic stented coronary arteries with catheter-based ¹⁹²Ir.

Methods

The Scripps Coronary Radiation to Inhibit Proliferation Post-Stenting (SCRIPPS) trial was a double-blind, randomized trial comparing ¹⁹²Ir with placebo sources. The methods have been described previously.⁹ This clinical trial was approved by the institution’s human subjects and radiation safety committees. Patient inclusion criteria required a target lesion in a restenotic coronary artery that either already contained a stent or was a candidate for

stent placement. If the lesion was not already stented, single or (if required) tandem coronary stenting (Johnson and Johnson Interventional Systems) was performed. If stents had been placed previously, redilation was undertaken, and often, additional stents were placed within the original stent to optimize the angiographic result. In all cases, high-pressure (≥18 atm) balloon dilations were performed in an attempt to achieve a 0% residual stenosis within the stented segment. Patients were then randomly assigned to receive a 0.76-mm (0.03-inch) ribbon (Best Industries) containing sealed sources of either ¹⁹²Ir or placebo. The study ribbon was left in place for 20 to 45 minutes, as required to administer the prescribed dose of 800 to 3000 cGy to the adventitial border.⁹

All patients were requested to return for repeat coronary angiography at 6 months and again at 3 years. Revascularization was repeated after follow-up angiography only if the patient had recurrent symptoms or if functional tests demonstrated the presence of coronary ischemia. At 3-year follow-up, patients were queried regarding any hospitalizations or procedures occurring since their index procedure. Medical records were obtained from each patient’s primary treating physician along with copies of hospital records from all admissions and procedures. Where necessary, the county coroner’s office was contacted to obtain data regarding the cause and date of patient deaths. Several patients who initially appeared lost to follow-up were located by a commercial service (1–800-US-SEARCH).

Core laboratory quantitative angiography was performed at the Brigham and Women’s Hospital Center by individuals blinded to the treatment allocation, as previously described.⁹¹⁰ Selected serial cine frames, obtained from 2 unforeshortened projections and matched for position within the cardiac cycle with the use of side-by-side projectors, were digitized with a cine video converter, with the contrast-filled catheter used as the calibration standard. The reference vessel was defined as the vessel segment 5 mm proximal and distal to the radiation sources. Binary restenosis was defined as stenosis ≥50% of the luminal diameter of the stent and/or stent margin 5 mm proximal and distal to the radiation sources at follow-up. The assessment of binary restenosis at 3 years included only those patients with angiographic follow-up beyond 27 months, unless a target-lesion revascularization occurred earlier. Patients with restenosis at the 6-month angiogram but no target-lesion revascularization who lacked 3-year angiography (1 in each group) were excluded from the late binary restenosis analysis because stenosis regression could not be ruled out. An analysis of serial changes in minimal luminal diameter and diameter stenosis (Figures 4 and 5) included only those patients with 3-year angiograms who had not had a target-lesion revascularization by the 6-month angiogram.

Target-lesion revascularization was defined as coronary angioplasty or surgical bypass of the target vessel due to the presence of ≥50% diameter stenosis of the target lesion as measured by the core angiographic laboratory. The target lesion was defined as the stented segment in addition to the stent margins 5 mm proximal and distal to the radioactive or placebo sources. Thus, target-lesion revascularization included revascularization of both in-stent restenosis and restenosis at the stent or source margins due to “edge effect.” Target-vessel revascularization included revascularization of the target lesion or a segment outside the target lesion but within the same vessel. Non–target-vessel revascularization was defined as revascularization of an epicardial vessel that did not contain the target lesion. Myocardial infarction was defined as an elevation of the MB fraction of creatinine kinase to a value 3 times the upper limit of the normal range.

For the analysis of continuous data, Mann-Whitney rank sum tests were used to assess differences between the 2 treatment groups, except for serial comparisons of luminal diameter and percent diameter stenosis, which were done with a 2-way ANOVA. The results are expressed as mean±SD. Categorical data were compared with the use of χ² or Fisher’s exact test except for the composite clinical end point, which was analyzed by means of Kaplan-Meier survival analysis, with differences between the 2 treatment groups compared with the use of a Mantel-Cox test of significance.

Results

Between March 24 and December 22, 1995, 55 patients were enrolled in this study; 26 were randomized to ¹⁹²Ir and 29 to placebo. Baseline clinical and angiographic factors were similar in the 2 groups. In-stent restenosis was present in 62% of both treated and placebo groups (Table 1).

Clinical follow-up was obtained on or after the 3-year anniversary following the index procedure in 100% of living patients (Table 2). The mean time from index study procedure to clinical follow-up was similar in ¹⁹²Ir and placebo groups (39.1±2.3 versus 39.6±2.8 months; P=NS). Follow-up times ranged from 36 to 44 months in ¹⁹²Ir patients and 36 to 46 months in placebo patients.

At 3-year follow-up, target-lesion revascularization occurred in 4 patients (15.4%) in the ¹⁹²Ir group compared with 14 (48.3%) in the placebo group (P<0.01). One patient in each group sustained a new target-lesion revascularization after the 6-month follow-up angiogram: at 36 months for an asymptomatic ¹⁹²Ir patient and at 40 months for a symptomatic placebo patient (Figure 1). Both restenoses were focal and occurred within the target segment (not margins) at sites with <50% diameter stenosis on the 6-month angiogram. Target-vessel revascularization was also lower in ¹⁹²Ir patients, occurring in 8 (30.8%) treated versus 17 (58.7%) placebo patients (P=0.04). Target-vessel revascularization was higher than target-lesion revascularization in both groups because 4 patients in the treated group and 3 in the placebo group had revascularization of disease that was located at a significant distance (>5 mm) from the target lesion and believed therefore to be unrelated to edge effect. Two of these revascularization procedures in the treated group and 1 in the placebo group occurred between the 6-month and 3-year follow-up periods. Non–target-vessel revascularization, with or without target-vessel revascularization, was similar in both groups, occurring in 7 (26.9%) treated and 8 (27.6%) placebo patients. In both groups, revascularization of nontarget lesions over the follow-up period was common. Eight (30.8%) treated and 7 (24.1%) placebo patients who did not need target-lesion revascularization did require revascularization of other, nontarget lesions. Thus, by 3 years, a total of 12 (46.2%) treated and 21 (72.4%) placebo patients had undergone subsequent revascularization procedures (Table 2).

There were 3 deaths in each group. Two of the deaths in the placebo group were cardiac deaths (at 8 and 11 months) associated with myocardial infarction, and the third (at 30 months) occurred in the postoperative period after bypass surgery for a target-lesion restenosis. One death in the ¹⁹²Ir group (at 23 months) occurred in the postoperative period after bypass surgery for a non–target-lesion stenosis. Another death in the ¹⁹²Ir group occurred in a patient who had self-terminated ticlopidine on day 3 and sustained a stent thrombosis that resulted in acute myocardial infarction on day 18 after the index procedure. Angiography during the acute thrombotic event and again at 6-month follow-up demonstrated 100% occlusion of the target lesion. This patient died 18 months after the study procedure of complications of abdominal surgery for diverticulitis. The third death was sudden (at 39 months) in a ¹⁹²Ir-treatment-failure patient who had a target-lesion revascularization at 8 months.

The composite end point of death, myocardial infarction, or target-lesion revascularization was significantly lower in ¹⁹²Ir versus placebo patients (23.1% versus 55.2%; P=0.01). Life-table analysis of this composite end point is displayed in Figure 2. Differences in clinical events were driven largely by differences in the need for target-lesion revascularization and become apparent at ≈3 months. The 2 curves continue to diverge until 10 months, after which clinical events are infrequent. However, owing to the large number of non–target-lesion revascularizations, differences in the composite end point of death, myocardial infarction, and target-vessel revascularization (38.5% versus 65.5%; P<0.05) and differences in death, myocardial infarction, and any revascularization (50% versus 79.3%; P=0.02) were less pronounced.

As previously reported, initial follow-up angiography was obtained at a mean of 6.7 months in 96.4% of patients.⁹ At the 6-month time point, restenosis rates were significantly reduced in the ¹⁹²Ir group (17% versus 54%; P=0.01). For the present report, all patients were asked to return for a repeat angiogram at 36 months. The mean interval from index procedure to late follow-up angiography was 36.7±4.5 months in ¹⁹²Ir patients compared with 39.1±2.6 months in placebo patients (P=0.05). This late angiogram was obtained in 19 (73.1%) of ¹⁹²Ir and 18 (62.1%) of placebo patients (P=NS). This represents 82.6% of living ¹⁹²Ir and 69.2% of living placebo patients (P=NS). None of the 4 living ¹⁹²Ir and 8 living placebo patients who refused 3-year angiography had symptoms of angina. Each of these patients had undergone numerous catheterization procedures in the past, had now broken the cycle of repeated invasive procedures, and adamantly refused catheterization.

At the 3-year follow-up, angiographic restenosis (≥50% stenosis of the luminal diameter) either within the stent or at its border (outside the stent but spanned by the study ribbon) was observed in 33.3% of ¹⁹²Ir patients compared with 63.6% of placebo patients (P<0.05). Thus, whereas restenosis rates were reduced by 69% by the 6-month angiogram, restenosis was only reduced by 48% by the 3-year angiogram. However, at 3 years, this reduction in restenosis was still significant (P<0.05) (Figure 3).

To better characterize the late natural history after radiation therapy, angiograms were analyzed in patients who were alive and had not had a target-lesion revascularization by the 6-month angiographic time point. This is an important subgroup of patients, because these lesions were not subjected to interim interventions, and therefore, they best represent the late natural history after radiation therapy. Late follow-up angiography was obtained in 17 (81%) of 21 eligible ¹⁹²Ir and 10 (71.4%) of 14 eligible placebo patients in this subgroup. The mean minimal luminal diameter between the 6-month and 3-year angiograms decreased from 2.49±0.81 to 2.12±0.73 mm (P=0.15) in ¹⁹²Ir patients but was unchanged in placebo patients in this subgroup (Figure 4). Thus, when the analysis was confined only to lesions that were “untouched by human hands” by the 6-month angiogram, a small reduction in minimal luminal diameter was observed between 6 months and 3 years in the treated group that was not found in the placebo group. Correspondingly, there was a small amount of late “catch up” in percent diameter stenosis between 6 months and 3 years in ¹⁹²Ir-treated patients that was not matched by the placebo group. The percent diameter stenosis increased from 14±28% at 6 months to 26±28% (P=0.25) at 3 years in ¹⁹²Ir-treated patients but only increased from 21±24% to 23±17% (P=0.75) in placebo patients (Figure 5).

Although our sample size was small, it is notable that only 24% of treated and 20% of placebo patients had a reduction in diameter stenosis of >15% between the 6-month and 3-year angiograms. Thus, most late serial changes in percent diameter stenosis were small. However, between the 6-month and 3-year angiograms, 4 patients in the treated group had an increase in diameter stenosis that brought the final luminal diameter to >50%, whereas only 1 patient in the placebo group crossed this 50% threshold. Of the 3 patients in the treated group who crossed over the 50% threshold but did not have a target-lesion revascularization, the absolute increases in diameter stenosis were small (14%, 15%, and 35%); however, this small amount of late catch up in the treated group did narrow the difference in dichotomous restenosis rates between the 2 groups. No aneurysms, pseudoaneurysms, or perforations were observed by the core angiographic laboratory on any follow-up angiograms of either the ¹⁹²Ir group or the placebo group.

Discussion

After 3 years of clinical and angiographic follow-up, treatment with ¹⁹²Ir continues to demonstrate improved clinical outcome compared with placebo. With 100% clinical follow-up and each living patient followed up for ≥36 months, the clinical benefits initially observed at 6 and 24 months⁹¹¹ were maintained, and no unexpected, radiation-related adverse events were identified. Interestingly, in ¹⁹²Ir-treated patients, there were no clinically apparent restenoses after the initial follow-up angiogram until protocol-mandated 3-year angiography documented a high-grade target-lesion restenosis in an asymptomatic treated patient. This lesion was located in the proximal segment of a large-diameter left anterior descending artery. Although the patient was asymptomatic, noninvasive testing (performed after the follow-up diagnostic angiogram) did demonstrate silent coronary ischemia, which prompted a return to the catheterization laboratory for repeat angioplasty. There was 1 late target-lesion revascularization after the 6-month angiogram in the placebo group. This patient, who initially resisted 3-year angiography, underwent restudy owing to recurrence of symptoms at 40 months, was found to have significant proximal right coronary artery stenosis, and underwent repeat angioplasty.

Over the 3-year follow-up period, the composite clinical event rate (death, myocardial infarction, or target-lesion revascularization) in the treated group was lower than in the placebo group (23.1% versus 55.2%; P=0.01). The 3 deaths in the ¹⁹²Ir group had clearly identified causes: 1 was a consequence of elective bypass surgery of a nontarget lesion, another was a sudden death in a treatment-failure patient, and another was due to complications following abdominal surgery 18 months after a stent thrombosis. Although it is possible that the stent thrombosis was related to radiation exposure (in the absence of ticlopidine), this vessel was documented to be 100% occluded on angiography 6 months after the index procedure and was therefore unlikely to be a site of sudden coronary closure, perforation, or other acute cardiac event.

Another important finding of the present study was the large number of non–target-lesion revascularization procedures in both the treated and placebo groups over the 3-year follow-up period. Fully 30.7% of treated and 24.1% of placebo patients who did not need target-lesion revascularization did require revascularization of other nontarget lesions (Table 2). Thus, by the end of the follow-up period, 50% of treated patients and 79.3% of placebo patients had sustained a clinical event. It should be emphasized, therefore, that although γ-radiation effectively reduced events associated with the target lesion, this local therapy did not affect the development of disease outside of the target segment. Late events were common, even in treated patients.

At 3 years, the angiographic restenosis rate in the treated group was 48% lower than in the placebo group (33% versus 64%; P<0.05). In ¹⁹²Ir patients who did not undergo target-lesion revascularization by the 6-month follow-up angiogram, the mean minimal luminal diameter decreased and the mean percent diameter stenosis increased by a small amount between the 6-month and 3-year angiograms. This reduction in diameter was not matched by the placebo group, whose mean minimal luminal diameter was unchanged over the late follow-up period. The number of patients with serial angiographic measurements was small, and standard deviations were large, and therefore, the possibility exists that the play of chance alone may account for these observations. Nevertheless, our findings may contradict the concept that radiation, in a stented artery, “freezes” the postprocedure angiographic result. The response of the vessel to radiation appeared to be somewhat dynamic over the 3-year follow-up period, and in some patients, there was continued late loss. This late loss was small, but in 4 treated patients, it was enough to cross the 50% diameter threshold, provoking an increase in the dichotomous restenosis rate. Although the increase in restenosis rate in treated patients from 17% to 33% over the 3-year period raises some concerns about the prognosis for an even longer follow-up period, it should be emphasized that all 4 patients were asymptomatic, and only 1 patient had a high-grade stenosis with evidence of ischemia that prompted revascularization. Also, this initial pilot trial intentionally prescribed a relatively low dose of radiation; higher radiation exposures may further improve long-term results. Importantly, late angiographic follow-up in the present study revealed no evidence of perforation, aneurysm, or pseudoaneurysm in ¹⁹²Ir-treated patients. Thus, no safety issues unique to radiotherapy have been identified.

Our late results differ from the reported long-term follow-up after stent implantation for de novo disease. In the only published study examining 3-year angiographic outcome after initial stent implantation (without radiation),¹² the minimal luminal diameter increased slightly but significantly between the 6-month and 3-year angiograms. Several other clinical trials using vascular radiotherapy have been published.^1314151617 In 1 study, 7-year follow-up documented high patency rates after femoropopliteal arteries undergoing angioplasty were exposed to intravascular γ-radiation.¹⁴¹⁵¹⁶ Other reports evaluating the long-term effects of coronary radiation are pending.¹³

Study Limitations

The most important limitation of this study is its small sample size. The distribution of baseline characteristics was not entirely even, with a trend toward more diabetic patients in the placebo group. Also, more asymptomatic patients in the placebo group refused 3-year angiography, which may have increased the restenosis rate documented in the placebo arm.

Conclusions

With 100% clinical follow-up 3 years after study entry, the clinical efficacy of ¹⁹²Ir appears durable. In ¹⁹²Ir-treated patients, target-lesion revascularization was reduced by 74% at 6 months and 68% at 3 years. Angiographic restenosis was reduced by 69% at 6 months but only 48% at 3 years because of small reductions in luminal diameter over the longer follow-up period. No perforations, aneurysms, pseudoaneurysms, or other safety issues were observed. At the 3-year time point, vascular radiotherapy continues to be a promising treatment for restenosis.

Reprint requests to Paul S. Teirstein, MD, Division of Cardiovascular Diseases, SW206, Scripps Clinic, 10666 N Torrey Pines Rd, La Jolla, CA 92037.

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint