Late Outcomes after Drug-Eluting Stent Implantation in “Real-World” Clinical Practice

ABSTRACT: Background. We report the late outcomes in 411 consecutive patients undergoing drug-eluting stent (DES) implantation by a single operator between 2003–2006. Methods. Prospective registry with continuous follow up. Patients with stable angina (SA) or acute coronary syndrome (ACS) received DES for long lesions, small vessels, chronic total occlusion, bifurcation, aorto-ostial, left main, post atherectomy or saphenous vein graft lesions, multivessel/multilesion single vessel (V) disease, in-stent restenosis (ISR) or diabetes. Results. Age range: 34–86 years. 166 (40.3%) had ACS, 98.3% hypercholesterolemia and 14.6% diabetes. 261 (63.5%) had percutaneous coronary intervention (PCI) to 1V and 150 (36.5%) to >1V. 607 V were treated. 259 patients (63%) had multilesion PCI, and 109 (26.5%) 1V multilesion PCI. 293 (71.3%) patients had long lesions and 224 had Vs < 2.75mm diameter. 75.5% of lesions were Type B2/C. 1–8 stents were implanted/patient. 822/883 stents were DES. 108 patients received ≥ 1stents ʺ 2.5 mm diameter and 246 patients received stents ≥ 20mm long. 25 patients developed late complications. ISR occurred in 23, 3.5–38 months after DES implantation. 3 had sudden late DES thrombosis (LST). 1/3 also had ISR. 20/25 required PCI, 1 CABG and 4 medical treatment. Subsequently, 3 of the 20 required further PCI for recurrent ISR and 1 CABG. 9 patients died during 1–5-year follow up. Conclusions. In “real-world” patients at increased risk of ISR after bare-metal stenting (BMS), “off-label” DES implantation has a low incidence of late complications. The most common is ISR which presents later than after BMS. Acute LST is serious but unusual and may be accompanied by ISR.

J INVASIVE CARDIOL 2008;20:493–500


Coronary artery stenting was shown to reduce the incidence of restenosis in two major landmark clinical trials.1,2 Although these trials had strict clinical and angiographic entry criteria, interventional cardiologists expanded the clinical use of intracoronary stenting to more complex patient subsets where the rates of restenosis after PTCA alone were amongst the highest. Intracoronary stenting reduced the incidence of restenosis but excess fibrointimal hyperplasia did occur inside some stents causing “in-stent” restenosis (ISR) which has proved extremely difficult to treat. The incorporation of a coating to the metal struts of stents and the application to the coating of a drug which could elute into the wall of the artery and limit the growth of smooth muscle cells would be the next step to improve outcomes further. Once again randomized clinical trials (RCTs) showed that “drug-eluting stents” (DES) reduced the incidence of ISR in specifically defined groups of patients compared to “bare-metal stents” (BMS).3,4 However, the evidence for benefit in more difficult subgroups is only just emerging from randomized clinical trials of specific subsets5–9 and multicenter registries.10–15 Simultaneously, however, the phenomenon of late (LST) and very late (VLST) stent thrombosis has temporarily slowed the enthusiasm for this technology, despite the fact that ISR within BMS has been shown not always to be a benign clinical entity.16,17

In an attempt to determine whether acceptable results are reproducible in “real-world” clinical practice by individual interventionalists, we report the prospective, long-term clinical follow-up of patients undergoing DES implantation by a single operator for a wide-spectrum of clinical and angiographic, “off-label” indications between 2003–2006 and compare the outcomes to those currently emerging from the RCTs and larger registries.

Patients and Methods

Between January 2003 and December 2006, DES were implanted by a single operator (DRR) in patients presenting with stable angina or an acute coronary syndrome if certain clinical or angiographic criteria existed. These mainly included lesions > 20mm long, vessel size < 3.0mm diameter, chronic total occlusion, bifurcation lesions, aorto-ostial lesions, left main coronary artery, post-rotational atherectomy, lesions in or beyond saphenous vein grafts, patients with multivessel disease, multilesions in a single vessel, patients with ISR and those with diabetes mellitus.

Patients were classified as having stable angina if they had previously recognized angina pectoris with predictable limitations without any recent change in frequency or severity. Patients with a recent increase in frequency of angina or with increase in severity of symptoms with or without ECG abnormalities but normal troponin levels were classified as unstable angina and those with ECG abnormalities (non-ST elevation) and a raised troponin level as “non-ST elevation MI.” Patients were classified as having MI if they presented with prolonged chest pain and ST-elevation, and were either treated by primary PCI or rescue PCI following failed thrombolytic therapy or for post-MI angina while still in-hospital — all with raised troponin and CKMB levels. These 3 groups comprised the acute coronary syndromes.

All patients were pretreated with 300 mg – 600 mg aspirin and 300 mg – 600 mg clopidogrel prior to PCI. The aspirin (75 mg daily) was continued indefinitely unless there were adverse effects or true allergy in which case clopidogrel was continued indefinitely. Otherwise, clopidogrel (75 mg daily) was continued for at least 12 months from the date of the PCI. The majority of procedures were performed via the femoral route. Unfractionated heparin (100 U/kg) was given immediately prior to PCI in all patients. Those with acute coronary syndromes treated with subcutaneous enoxaparin had no further enoxaparin after PCI.

Following DES implantation, clinical follow-up was carried out prospectively for all patients either by attendance at clinic or by letter from the referring physician. All patients were contacted after 1 year by telephone or postal questionnaire to confirm their symptom status, their need for hospital admission or need for further coronary angiography or intervention. The Academic Research Consortium definitions of early, late and very late thrombosis were used in this study.18

Complete follow-up was achieved for all patients at the end of December 2007 — one year after enrollment of the last patient in December 2006. Follow-up is planned for all patients up to 3 years post DES implantation.

Results

Between January 2003 and December 2006, 411 consecutive patients received DES during PCI. There were 291 males and 120 females. Ages ranged between 34-86 years, mean ± SD (standard deviation): 60.9±10.1 years. 59.7% of patients had stable angina and 40.3% had an acute coronary syndrome. The majority of patients had hypercholesterolemia and 14.6% had diabetes mellitus. The clinical details are shown in Table 1.

Table 2 shows that 261 patients (63.5%) had PCI to a single vessel and that 150 (36.5%) had PCI with DES to ≥ 2 vessels. One hundred and twelve patients had PCI with DES to 2 vessels, 31 patients to 3 vessels, 6 patients to 4 vessels and 1 patient to 5 vessels. Six hundred and seven vessels were treated in the 411 patients. These included 202 left anterior descending (33.3%), 168 right coronary (27.7%), 79 left circumflex (13%), 62 obtuse marginal (10.2%), 57 diagonal (9.4%), 15 intermediate (2.5%), 5 left main (0.8%), 3 right ventricular and 2 septal coronary arteries. Thirteen saphenous vein grafts and one radial artery graft were addressed in 13 patients.

Two hundred and fifty-nine patients (63.0%) had multiple lesions treated by PCI, and 109 (26.5%) single vessel multilesion PCI. Two hundred and ninety three patients (71.3%) had long lesions (> 20mm long) and 224 (54.5%) had vessels ʺ 2.75mm diameter. One hundred and sixty-four patients (39.9%) had combinations of long lesions and small diameter vessels. Of the lesions treated by DES, 75.5% were Type B2/C lesions (ACC/AHA Classification).

Table 3 shows that between 1–8 stents were implanted per patient. All patients received at least 1 DES. A total of 883 stents were implanted and 822 (93.1%) were DES. Forty-four patients (10.7%) had additional bare-metal stents (n = 61) implanted in a different vessel than the DES usually because of vessel size (≥ 3.0mm) and/or lesion length (≤ 20 mm). One hundred and eight patients (26.2%) received one or more stents ≤ 2.5 mm diameter and 246 patients (59.8%) one or more stents > 20 mm in length.

One hundred and seventy-seven patients (43.1%) had 1 DES and 234 (56.9%) >1 DES. One hundred and fourteen patients (27.7%) had 3 or more DES implanted. The majority of DES (97.4%) implanted were either Cypher® (Cordis, Miami, Florida) or Taxus® (Boston Scientific, Natick, Massachusetts) stents. One hundred and seventy-seven patients had Cypher stent(s) only, 148 Taxus stent(s) only, 8 Axxion® (Biosensors International, Singapore) and 2 CoStar® (Conor Medsystems) stents. Thirty-two patients (7.8%) had DES of different types implanted during the index procedure and 44 (10.7%) had combinations of DES and BMS implanted.

Follow up. All patients were followed for at least 12 months (range 12–59; mean 34.3 months). One patient being treated by rescue PCI died five hours after the procedure as a result of a ruptured hemorrhagic MI and eight other patients died during the 1–5 years of follow-up. Of these, there were 1 cardiac and 7 non-cardiac deaths. The details are shown in Table 4.

Of the 402 survivors, 326 (81.1%) patients were free of angina, 70 (17.4%) had very occasional or minimal angina and 6 (1.5%) frequent angina.

Table 5 shows the details of the 25 patients who developed late complications after DES implantation. ISR was the most common complication, occurring in 23 patients. Twenty-two patients presented with ISR as the initial complication and one patient who developed presumed acute LST at 3.5 months treated successfully with thrombolytic therapy, developed ISR 7.5 months later. One patient developed VLST in the LAD at 14 months and one silent occlusion of an SVG-LCX at 19 months. Patients presented between 3.5–38 months after DES implantation. Nineteen of the 25 patients (76.0%) had had multiple and overlapping stents inserted. Only 3 of the 25 patients (12.0%) developed ISR in DES < 20 mm in length, and one of these was at the origin of a diagonal within LAD/DG1 bifurcation stents. The mean ± SD length of DES in the 25 patients developing late complications was 55.1 ± 35.7 mm; range 12–143 mm and 52.4 ± 34.1 mm in the 22 patients developing ISR. Of the 198 patients receiving Cypher stents (including the 21 patients with additional BMS) 14 developed ISR (7.1%); of the 169 receiving Taxus stents (including 21 with additional BMS) 6 developed ISR (3.5%) and of the 32 patients receiving mixed DES, 2 (6.2%) developed ISR. Symptomatic ISR in BMS was not found in the 44 patients in this study.

One patient had early stent thrombosis at 12 hours after 3 Cypher stents had been implanted in series over a long segment of disease (74 mm long) in the RCA. The vessel was reopened promptly by PTCA and intracoronary abciximab but small Q-waves developed in leads III and AVF and the CKMB reached 43 IU/L.

One patient (patient 25) presented with presumed LST at 3.5 months with acute anteroseptal MI and was treated successfully with thrombolytic therapy. Angiography two days later showed a fully patent stent. He presented 7.5 months later with an acute coronary syndrome and was found to have ISR within the long Taxus stent. Two patients presented with very late, acute thrombosis of a DES-stented coronary artery. One patient (patient 17) presented with acute anterior myocardial infarction (MI) 10 days after stopping all antiplatelet therapy for a TURP procedure, 14 months after stent implantation. He was later shown to have LAD stent occlusion and was treated conservatively. The second patient (patient 12) who had developed ISR 7 months after DES stent implantation treated by 3 Cypher stents, subsequently developed acute stent thrombosis 17 months later. It was treated successfully by emergency thrombectomy.

A further patient (patient 24) was found to have an occluded SVG-LCX 19 months after being stented with multiple Cypher stents for ISR within 2 Ultra bare-metal stents. This “silent occlusion” was classified as a “possible thrombosis” although occlusion due to severe and extensive ISR is more likely.

Discussion

Several RCTs, some with follow-up out to 5 years have shown that DES implantation results in lower major adverse cardiac events (MACE) and the need for reintervention for ISR than BMS, in well-defined patient groups.19–23 A pooled-analysis of RAVEL, SIRIUS, E-SIRIUS and C-SIRIUS indicated that 91.7% of patients treated with a Cypher stent remained free from TLR compared to 75.4% of patients treated with a BMS equivalent.24 The NEW SIRIUS analysis (pooled E-SIRIUS and C-SIRIUS), showed that at 5 years 8% of the Cypher group required further PCI compared to 25.1% in the BMS group. Mortality was 7.6% vs 7.5% and Q-MI 3.1% and 1.8% respectively.25 A similar pooled-analysis of the results of TAXUS I, II, II-SR, IV and V indicated that after 4 years, 89.5% of DES patients remained free from ischemic TLR compared to 79.8% with BMS.26 Four year outcome data from TAXUS II confirmed a decrease in TLR rates (7.2% v 15.7%) over BMS, and cardiac death and MI rates of 1.6% v 1.5% and 4.7% v 6.7% respectively.27 At 5 years, TLR rates were 10.3% v 18.4% — a highly significant decrease of 44%, and MACE was 35% lower — 20.4% v 27.6%.22 Excellent clinical results can thus be expected for such “on-label” use of DES.

The studies that have randomized patients with certain specific complex lesion subsets have shown lower TLR rates in those treated with DES compared to BMS.5–9,28–31 However, as to what results can be expected in day-to-day clinical practice treating an even wider range of complex lesion subsets, we are dependent on registry studies and the analysis of “real-world” data following “off-label” use.12,13,32–40

In contemporary practice worldwide, “off-label” use of DES is common, and perhaps not surprisingly, workers have shown that compared with standard use, long-term effectiveness after DES is lower when they are used “off-label”. For example, Beohar et al reported that 1-year after DES implantation, the target vessel revascularization rate (TVR) was 7.6% for “off-label” use compared to 4.4% for standard use,10 and Win et al found that the composite clinical outcomes of death, MI or TVR/TLR were higher with “off-label” use compared to “on-label” use both during the index admission (10.9% vs 5.0%) and at 1-year (17.5% vs 8.9%).11 Our prospective study reports the results of DES in such a group of complex cases including those with single-vessel multilesion disease and/or multivessel disease, presenting with stable angina or an acute coronary syndrome. After a mean follow-up of 34.3 months, the majority of patients were free of angina or were symptomatically improved compared to pre-PCI. Only 3 patients had acute ST elevation/Q-wave MI after hospital-discharge due to LST or VLST (0.7%). They were treated successfully medically, with thrombolytic therapy or with PCI.

Twenty-two patients developed a recurrence of angina and underwent repeat angiography. Twenty one had angiographic evidence of significant ISR (5.1%) and one had an occluded SVGLCX which had been extensively re-stented with overlapping Cypher stents measuring 132mm in length. Eighty percent (20) of patients were treated successfully by further PCI — 17 with DES and 3 with cutting balloon angioplasty. Four patients required only medical treatment and one needed CABG surgery. Of the 4 treated medically, one subsequently developed ISR and was treated by PCI with DES. Three of the 20 patients (15%) treated by PCI for ISR by DES developed further ISR. Two were successfully treated by further PCI with DES and one required CABG surgery. One of the two treated by repeat DES (see above) presented with VLST 17 months later and was treated successfully with emergency thrombectomy. Since one patient died from a ruptured hemorrhagic MI within a few hours of a successful procedure (stents fully patent at autopsy), the target vessel failure rate following the first DES implantation procedure was calculated to be 6.1% (25 of 410). TVR was 5.1% (21 of 410). Q-wave MI post discharge after the DES implantation was 0.7% (3 of 410). Death occurred in 9 of 411 patients (2.2%), 8 of which were post-hospital discharge (1.9%). After 1 year of follow-up, total MACE was 3.9% (1 Q-wave MI (0.2%); 5 deaths (1.2%) — 1 cardiac, 4 non-cardiac; 13 TVR (3.1%); 1 LST (0.2%). Beyond 1 year after implantation, 2 patients had VLST — at 14 and 24 months respectively (0.5%). 

Our data are consistent with the recently completed RCTs of DES versus BMS in complex lesion subsets and with the findings of the other “off-label” studies noted above. They are also consistent with the results of published larger registries. The 2-year results from the 2,487 patients in the ARRIVE 1 registry showed that patients with complex lesions demonstrated a Taxus-related cardiac event rate of 9.9% (cardiac death: 1.3%; MI: 2.4%; TVR: 8.2%). Stent thrombosis rate was 2.9% and death from all causes (cardiac and non-cardiac) was 6%.12 The global OLYMPIA Registry includes a diverse and high-risk population of patients with complex lesions or clinical characteristics, multivessel disease, prior MI and diabetes who received Taxus Liberté stents. At 12 months, analysis of the first 7,124 patients showed a Taxus Liberté-related MACE of 3.9% including death (1.1%), MI (0.9%) and TVR (2.5%), along with a stent thrombosis rate of 0.8%.13 The MATRIX Registry reported a total and cardiac mortality of 3.3% and 1% respectively through 2 years of follow-up of patients receiving the Cypher stent; MI occurred in 4%, TVR in 10.7% and 0.7% definite and 1.1% probable stent thrombosis. The rates of MI and TVR were higher in “off-label” patients compared to “on-label” patients being 4.4% vs. 0.9% and 11.6% v 4.5% respectively. Stent thrombosis and death were not statistically different (1.1% vs. 0.5% and 3.4% vs. 2.7% respectively).37 The REAL Registry quoted a 2-year incidence for DES vs BMS for MI: 5.3% vs. 5.8%; TVR: 9.1% vs. 12.9%; MACE: 16.9% vs. 21.8%; death: 6.8% vs. 7.4%; stent thrombosis: 1% vs. 0.6%.32 Our results differ significantly from the recently published Spanish ESTROFA Registry which reported a cumulative stent thrombosis rate at 3 years of 2% and a one year mortality of 16%.41 These disturbing figures may be related to premature discontinuation of antiplatelet therapy and the use of DES in STEMI.

ISR remains a problem after DES although it is less frequent than what might be expected with BMS in such difficult cases. It most often occurs in areas involving long segments of stenting with multiple overlapping stents. The delayed presentation of ISR and VLST (78.3% occurred more than 6 months post implant; 43.5% more than 12 months and 34.8% more than 18 months in this study) has been partly explained by insight into the pathophysiology and may have implications for long-term antiplatelet therapy and follow-up in patients treated with these first-generation DES. However, significant ISR may be rendered less frequent or indeed may be abolished by future generations of DES which may limit late lumen loss and prevent LST/VLST by allowing fibrointimal and endothelial cell covering of stent struts within the first few months of implantation due to the non-inflammatory PC polymer carrying the sirolimus analogue. The results from the Endeavor series of trials (using the Zotarolimus-eluting stent) compared to the RAVEL/SIRIUS and TAXUS I, II, IV and V trials (LST at 3 years of 0.08% compared to 0.97% and 0.6% for Cypher and Taxus respectively) and of the SPIRIT I-III trials with the Xience V — everolimus-eluting stent are encouraging. However, although the Endeavor III trial has shown that MI (0.6% v 3.6%) and MI/death (2.2% v 7.1%) were less frequent with the Endeavor compared to the Cypher stent42 and SPIRIT III that MACE at 12 months including cardiac death, MI or TLR were 43% lower in Xience V (5.8%) than Taxus (9.9%) patients, interestingly stent thrombosis occurred in 5/647 (0.8%) vs 2/317 (0.6%) in the two groups respectively in this latter study.43 Moreover, the Endeavor V Registry of “real-world” patients reports a 2.0% cardiac death rate, Q-MI 0.4%, TLR 3.8%, MACE 7.0% and stent thrombosis 1.1% in 1,989 patients after 12 months of follow-up44 which is not too dissimilar to the results of our own study. The results of SPIRIT IV which will examine the Xience V stent’s effectiveness in patients with multilesion/multivessel disease and/or bifurcation lesions will be of further interest.

In conclusion, the current study shows that similar results to those obtained with DES in RCTs or in large registries are being achieved by individual interventionalists in a “real-world” group of patients presenting with stable angina or an acute coronary syndrome and a wide variety of complex coronary lesions. The results include a low incidence of MACE during long-term follow-up including a very low incidence of late/very late stent thrombosis. The results cannot be extrapolated to patients with acute ST elevation MI. Time will tell whether the next generation of DES has more to offer us.