Risk of perioperative coronary stent stenosis or occlusion in patients with previous percutaneous coronary intervention undergoing coronary artery bypass grafting surgery

Abstract

OBJECTIVES

There is an ongoing discussion about how to treat coronary stents during bypass surgery: Should patent stents be left alone and the stented vessels be ungrafted, or should every stented coronary artery receive a bypass graft? This study aims to determine the relevance of perioperative stent stenosis or occlusion on postoperative outcomes up to 3 years postoperatively.

METHODS

Patients undergoing coronary artery bypass grafting surgery (CABG) (±concomitant procedures) with previous percutaneous coronary intervention from 4 centres were prospectively included in this observatory study between April 2015 and June 2017. A coronary angiography was conducted between the fifth and seventh postoperative days. The preoperative and postoperative angiograms were assessed in a core laboratory, assessing the patencies of coronary stents and bypass grafts. The core lab investigators were blinded to the patients’ characteristics and perioperative course.

RESULTS

A total of 107 patients were included in the study. In the postoperative coronary angiography, 265 bypass grafts and 189 coronary stents were examined angiographically. Ninety-seven percent of preoperatively patent stents remained patent. New coronary stent stenoses were observed in 5 patients (4.7%). All 5 patients were asymptomatic and managed conservatively. Bypass stenoses were observed in 12 patients (11%), of whom were managed conservatively, 4 underwent percutaneous coronary intervention and 1 underwent redo-CABG. Two years postoperatively, 97% of patients were alive. Patients with new stent stenosis tended to have a better survival compared with patients with bypass stenosis (100% vs 73%; P = 0.09) up to 3 years postoperatively.

CONCLUSIONS

Perioperative coronary stent stenosis occurs rarely. It is safe to leave a patently stented coronary vessel without bypass grafting.

INTRODUCTION

Percutaneous coronary intervention (PCI) and coronary artery bypass grafting surgery (CABG) are well-established therapeutic options for revascularization of coronary artery disease (CAD) [1]. With increasing complexity of CAD, particularly in young diabetic patients, CABG is increasingly advantageous compared with PCI in terms of longevity and survival benefit. Contrarily, in less complex CAD, PCI and CABG show comparable results [2, 3]. Due to the progressive nature of CAD, patients who received interventional coronary revascularization through PCI are at risk of needing further coronary revascularization by means of CABG. In several analyses, these patients’ perioperative risk for mortality and morbidity was analysed and partially shown to be significantly elevated compared to patients undergoing CABG without previous PCI [4–6]. The underlying reasons are currently unclear. Also, patients with previous PCI pose a challenge from a surgical point of view: First, patients with recent, not-yet endothelialized stents require dual antiplatelet therapy to prevent acute stent thrombosis. This therapy increases the perioperative risk for bleeding complications [7, 8]. Second, it is unclear whether stented coronary arteries with patent stents should be bypassed or not: If not bypassed, an acute loss of coronary stent patency will cause acute myocardial infarction. This stenosis or occlusion could be caused by stent thrombosis because of dual antiplatelet therapy discontinuation and perioperative coagulation therapy or by mechanical manipulation of the heart during surgery. If a stented coronary vessel is bypassed and the stent remains patent, the bypass will not generate sufficient flow and can become occluded. Coronary stent thrombosis occurs rarely in the peri-interventional setting of PCI (1.3% in 9 months [9]), but is associated with a mortality of 45% [9]. So far, data on the rate and relevance of coronary stent thrombosis in the setting of CABG do not exist.

We hypothesize that perioperative acute coronary stent stenosis or occlusion leading to myocardial ischaemia and incomplete revascularization might be a relevant contributor to the elevated perioperative risk of patients with previous PCI undergoing CABG. This study aims to analyse the fate of coronary stents after CABG, also in order to give hints for the cardiac surgeon on how to proceed with stented vessels during CABG.

METHODS

Study design

The present study was a prospective, multicentre observational study.

Study population

Consecutive patients in 4 centres were prospectively included in the study in April 2015 and June 2017 according to the following inclusion and exclusion criteria: Inclusion criteria: age >18 years, previous PCI with stent implantation, first-time CABG or CABG plus concomitant procedure planned. Exclusion criteria: emergency procedures. Informed consent to study participation was obtained from all patients.

Perioperative antithrombotic therapy

All patients continued acetylsalicylic acid perioperatively. In patients with stents in the endothelialization phase (<12 months after implantation of a drug-eluting stent or a biodegradable scaffold, <4 weeks after implantation of a bare metal stent) on dual antiplatelet therapy, the adenosine diphosphate (ADP)-receptor antagonist (clopidogrel, prasugrel or ticagrelor) was discontinued 5 days before surgery. After 2 and 3 days of discontinuation, the platelet function was assessed using impedance aggregometry (Multiplate Analyzer, Roche, Mannheim, Germany). If the platelet function in the ADP-test was within the normal range, intravenous bridging with tirofiban was initiated and continued until 4 h before surgery. If the platelet function was still suppressed, no bridging was conducted.

Surgical procedure

All patients underwent CABG according to the institutional protocols. Thus, the operative strategy (off-pump or on-pump CABG), the choice of grafts and revascularization targets were at the discretion of the operating surgeon. Intraoperative coagulation management was also conducted according to local standards.

Postoperative coronary angiography

In order to obtain information about stent and bypass patencies, a coronary angiography was conducted between the fifth and seventh postoperative days. If during the postoperative coronary angiography incomplete revascularization (bypass- or stent-related) was detected, the further therapeutic concept was determined by an ad hoc heart team. For the scientific evaluation, the preoperative and postoperative coronary angiograms were again assessed by interventional cardiologists in a core laboratory. This assessment included angiographic evaluation of stenoses of coronary vessels, coronary stents, and bypass grafts, respectively (%). Additionally, preoperative and postoperative TIMI (thrombolysis in myocardial infarction) – flows for every coronary vessel >1 mm were evaluated (0 = no flow, 1 = flow without complete opacification of the entire coronary bed, 2 = flow with complete but delayed opacification of the coronary bed, 3 = rapid perfusion of the entire coronary bed) were evaluated [10]. The core laboratory investigators were blinded to the patients’ characteristics and perioperative course.

Outcomes

The primary end point was the prevalence of new coronary stent stenosis postoperatively. Other outcomes included survival at 30 days as well as perioperative myocardial injury, as reflected by the perioperative serum levels of myocardial necrosis parameters [creatine kinase (CK) and creatine kinase, isoform MB (CK-MB)].

Follow-up

The patients were contacted via telephone calls 1 and 2 years postoperatively. During these follow-up calls, the patients were asked to report upon postoperative myocardial infarctions, coronary catheterizations and strokes.

Ethics and regulatory authorities

The ethics committee of the Justus-Liebig-University approved the study as the leading ethics committee (approval number 03/13, 03 June 2014). For each study centre, the responsible ethics committees additionally approved the study. Each patient gave informed consent before enrolment. The study was conducted in accordance to the Declaration of Helsinki. The study was approved by the German Federal Office for Radiation Protection (Bundesamt für Strahlenschutz; approval number Z 5 - 22462/2 - 2013-012).

Statistical analysis

In this exploratory setting, no sample size calculation was conducted. Statistical analyses were conducted using SPSS Version 22 (IBM, Armonk, NY, USA) and GraphPad Prism software (GraphPad Software, Inc., La Jolla, CA, USA). Numeric parameters were analysed as mean ± standard deviation unless stated otherwise. Group comparisons were made using Fisher’s exact test or Student’s t-test, as appropriate. Results of the long-term follow-up were calculated as survival functions using Kaplan–Meier estimation and compared using the log-rank test. Statistical significance was assumed at the level of P-value <0.05.

RESULTS

A total of 107 consecutive patients were included in the study. Twenty-nine additional patients were initially included, but did not undergo postoperative coronary angiography. Twenty-five of the 29 patients without postoperative coronary angiography withdrew consent to study participation postoperatively while 4 patients had acute kidney injury prohibiting elective contrast agent application. These 29 patients were not included in the analysis, but their characteristics and outcomes are given in the Supplementary Material. The majority of patients (81%) had 3-vessel CAD. Thirty percent of patients were treated for diabetes mellitus, 41% of whom with insulin. Of note, 41% of patients had a recent myocardial infarction. The mean predicted operative risk, calculated using EuroSCORE II, was moderate (2.0%; Table 1).

The 107 patients had a median of 3 implanted coronary stents resulting in a total of 201 stents. The majority of stents (44%) were in the right coronary artery territory, followed by the circumflex (32%) and the left anterior descending (LAD) (24%) territories (Fig. 1A). Drug-eluting stents were most commonly used (Fig. 1B). Twenty-two percent of all stents were still in the endothelialization phase (Fig. 1C). The preoperative angiographic assessment showed that 89% of the stents had an insignificant lumen reduction of <50%, while 3.3% had a moderate stenosis (50–69%), 8.5% a severe stenosis (70–99%) and 3.4% were occluded (Fig. 1D).

All patients underwent CABG, 12% of whom had additional procedures (Table 2). Fifty percent of procedures were performed off-pump. Transfusions of erythrocytes were necessary in 33% of patients, and platelets were transfused in 14% of patients. One or both internal mammary arteries were used in 99% of patients. Additional vein grafts were used in 36% of patients, while radial artery grafts were rarely used (0.9%). The LAD was grafted in 98 patients (92%). The patients received a median of 3 bypass anastomoses, resulting in a total of 265 bypass anastomoses. These were predominantly distributed to the LAD (133 anastomoses; 47%) and the circumflex (100 anastomoses; 36%) territories with less bypass anastomoses in the right coronary artery territory (47 anastomoses; 17%). The grafting strategy for stented vessels during CABG showed that 95% (20/21) of vessels with preoperative stent stenoses were bypassed, while only 1.2% (2/168) of vessels with preoperatively patent stents were bypassed (Table 2).

Postoperatively, no patient required mechanical circulatory support. The patients were ventilated for a median of 7 h (interquartile range 4–12 h). They were discharged from the intensive care unit after 26 h (interquartile range 22–65 h). Two patients (1.9%) required surgical re-exploration, 1 of them because of bleeding and cardiac tamponade, 1 in the later postoperative course because of sternal wound-healing disorders.

Perioperative serum activities of CK and CK-MB showed a postoperative rise of these parameters until postoperative day 2 or 3, with a subsequent fall of enzyme activities over the following days. Compared with patients with regular postoperative angiographic result (mean postoperative CK activity 434 ± 969 U/l; mean postoperative CK-MB activity 24 ± 31 U/l), patients with postoperative new stent stenosis or occlusion showed no significantly elevated postoperative enzyme activity kinetics (mean postoperative CK activity 296 ± 163 U/l, P = 0.89; mean postoperative CK-MB activity 21 ± 28 U/l, P = 0.63). The enzyme activity kinetics of patients with bypass stenosis or occlusion were also not significantly higher compared with patients with regular postoperative angiographic result (mean postoperative CK activity 819 ± 739 U/l, P = 0.24; mean postoperative CK-MB activity 38 ± 41 U/l, P = 0.19; Fig. 2).

During postoperative coronary angiography, a total of 265 bypass grafts and 189 coronary stents were examined angiographically. Of the 168 preoperatively patent stents, 163 (97%) exhibited a postoperative stenosis grade of <50%. Five stents (3.0%) showed new stenoses of >50%, 2 (1.2%) with a stenosis grade between 50% and 90%, and 3 (1.8%) with a new stenosis grade of >90%. For 2 of these stents (1 in the LAD territory and 1 in the right coronary artery territory), the target vessel was bypassed. The remaining 3 stents were not deemed to be relevant intervention targets and the stent stenoses were managed conservatively. The 5 new stent stenoses affected 5 (4.7%; 95% confidence interval 2.0–10.5%) patients. None of the patients with stent stenoses showed signs of myocardial ischaemia. The stents (n = 21) that were already stenosed preoperatively showed postoperative stenoses similar to the respective preoperative angiographic findings (Table 3). The TIMI flow assessment in the stented vessels showed TIMI 3 flow in 172/189 (91%) vessels preoperatively compared with 182/189 (96%) vessels postoperatively. The mean ratio between postoperative and preoperative TIMI flow class was 1.01 ± 0.12. The angiographic examination of 265 bypass anastomoses showed stenosis or occlusion of 15 anastomoses in 12 patients. The reasons for stenoses or occlusions were narrowing of the anastomoses (n = 7), midway occlusion of the graft (n = 3), small target vessel without sufficient run-off (n = 3) and kinking of the bypass graft (n = 2). In 7 patients, these findings were managed conservatively, 4 patients underwent PCI of the native coronary arteries and 1 patient underwent redo-CABG.

The survival at 30 days postoperatively was 100%. Patients with stent or bypass stenoses showed 100% survival each (P = 0.88). The follow-ups at 1 and 2 years postoperatively were complete for 103 patients (96%) and 77 patients (72%), respectively. The median follow-up was 2.1 years postoperatively. The survival functions between patients with or without bypass or stent stenoses did not differ significantly until 3 years postoperatively [P (log rank) = 0.09]. However, the patients with bypass stenoses showed a trend towards reduced survival (1 year: 91%, 2 years: 91%, 3 years: 73%) compared with patients with stent stenosis (1–3 years: 100%) or patients without stent or bypass stenoses (1 year: 99%, 2 years: 98%, 3 years: 98%; Fig. 3A). After the ad hoc re-revascularization which occurred in 4 patients with bypass stenoses during the index hospital stay, 11 more patients underwent re-revascularization during the follow-up period, 3 of whom with postoperative bypass stenosis. No patient with postoperative stent stenosis reported re-revascularization during the follow-up period (Fig. 3B). Four patients reported recurrent myocardial infarction (Fig. 3C). Strokes were reported by 3 patients (Fig. 3D).

DISCUSSION

The main finding of this study is that the risk of new perioperative stent stenosis or occlusion in patients undergoing CABG is very low. Only 3.0% (5/168) of preoperatively patent stents in 4.7% (5/107) of patients with postoperative coronary angiography showed new stenoses or occlusion postoperatively. Additionally, 2 of these stents were in coronary vessels which were bypassed. None of these stent stenoses required interventional or surgical therapy. In total, the patients in the study exhibited excellent short- and mid-term outcomes without perioperative mortality and a survival of 97% after 2 years. These findings have several implications: First the clinical relevance of acute stent stenosis or stent occlusion in the context of CABG is low, although the rate of 4.7% found is somewhat uncertain due to the small number of patients studied, as shown by the wide 95% confidence interval for the rate of stent stenosis or occlusion (2.2–10.5%). On the one hand, the prevalence is lower than previously described for patients undergoing non-cardiac surgery, where acute stent thrombosis rates of 1–15% of surgical procedures have been described [11, 12]. Other than in the setting of non-cardiac surgery, where stent thrombosis usually has catastrophic consequences for the patient [13, 14], new stenosis or occlusion of coronary stents obviously induced no relevant myocardial ischaemia in the context of CABG, amongst others reflected by the non-pathological kinetics of postoperative cardiac enzymes in these patients. This differential effect of acutely reduced stent patency can probably be explained by the complete revascularization of the remaining coronary circulation achieved by CABG. Second, the rate of bleeding complications in our study was low with only one patient requiring re-exploration for bleeding. This suggests that the perioperative antiplatelet therapy management we used was appropriate to reduce bleeding complications and at the same time maintain stent patency. Discontinuation of ADP-receptor antagonists is recommended for patients undergoing elective or urgent CABG [15]. Although this guideline recommends bridging with glycoprotein IIa/IIIb inhibitors only for ‘extreme high-risk patients’ [15], bridging with glyocoprotein IIa/IIIb inhibitors seemed to be safe and effective in our patients. Previous reports described increased bleeding risk and lack of benefit for glycoprotein IIa/IIIb inhibitor bridging before cardiac surgery [16, 17]. The safety in our patients may result from the use of platelet function testing in order to optimize preoperative antiplatelet management [18]. Interestingly, dysfunctional bypass grafts were observed in 12 of 107 patients (11%) or 15 of 249 anastomoses (6.0%). Previous reports showed a mean failure rate of ∼12% of anastomoses [19]. Walker et al. [20] showed bypass dysfunction in 5% of anastomoses 12 days postoperatively, while Bigdeli et al. [21] showed 100% graft patency 2 weeks postoperatively. Jokinen et al. [22] showed graft stenosis or occlusion in 20% of grafts in asymptomatic patients 4 months after CABG. Interestingly, all patients with bypass dysfunction in our study were asymptomatic. Silent early graft failure has been described by Zientara et al. [23] in 3.6% of arterial anastomoses and 11.1% of venous anastomoses 4 days postoperatively. The extent of symptoms of bypass graft failure is probably determined by cardiac factors, e.g. collateral coronary flow, completeness of revascularization as well as patient-related factors (gender, presence of diabetes).

The low stenosis rate of coronary stents has a direct implication for the surgical planning of the procedure: The fear of not grafting a patently stented vessel is negligible. Furthermore, the low rate and low relevance of stent stenoses or occlusions make it safe to leave stented vessels with patent coronary stents ungrafted.

The findings of this study suggest that CABG in patients with previous PCI can be conducted with a lower risk than predicted by EuroSCORE II. Studies on the influence of previous PCI on perioperative risk after CABG have been inconclusive so far: while Massoudy et al. and Thielmann et al. reported an up to threefold increased risk in those patients compared with patients with native coronary vessels [5, 6], our group showed no influence of previous PCI on CABG risk in diabetic patients [4]. Besides the hypothesis of coronary stent failure which is refuted by the present study, incomplete revascularization due to the presence of coronary stents and perioperative bleeding complications has been thought to contribute to the elevated risk. Given the results of our study, it is probably safe to state that in the current era, CABG can be performed in patients with previous PCI without additional risk. However, a longer term follow-up will show if coronary stent patency might ultimately influence the outcome of these patients. Gaszewska-Zurek et al. [24] followed patients after CABG and previous PCI for 3 years and showed higher rates of angina and re-revascularization in these patients compared with patients after only CABG. However, our data show re-revascularization rates of 11% after 3 years which is comparable to contemporary trials with stent-naive CABG patients [2, 25]. The outcomes of the patients in this study suggest that the overlap between interventional and surgical revascularization strategies is smoother and safer than expected. This might be an additional argument in favour of coronary hybrid procedures which have been shown to be safe and effective in different previous studies [26–28].

Limitations

Some limitations of this study have to be mentioned. First, not all patients underwent postoperative coronary angiography. With increasing availability of high-resolution computed tomography, angiographic studies on coronary bypass grafts could be conducted more completely with less invasiveness. However, the radiation burden for the patients is also high using this technique. Second, the number of patients included is rather low and the results should be confirmed in larger scale studies. Nevertheless, this study shows for the first time the fate of coronary stents in the context of CABG.

CONCLUSION

In summary, these data suggest that CABG in patients with previous PCI can be performed with excellent outcomes non-inferior to CABG outcomes in patients with native vessels. The risk of perioperative coronary stent stenosis is low. Therefore, it is safe to leave patently stented coronary arteries ungrafted.

Presented at the 33rd Annual Meeting of the European Association for Cardio-Thoracic Surgery, Lisbon, Portugal, 3–5 October 2019.

ACKNOWLEDGEMENTS

The authors thank Monika Heinzel-Gutenbrunner for excellent assistance in statistical analysis.

Funding

This investigator-initiated study was conducted with support from the German Heart Foundation (Award number: F/22/12; 60.000€) and the Rhön Klinikum AG corporate funding (Award number 12/2014; 60.000€).

Conflict of interest: none declared.

Author contributions

Philippe Grieshaber: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Supervision; Validation; Writing—original draft; Writing—review & editing. Irina Oswald: Data curation; Formal analysis; Investigation; Project administration. Marc Albert: Investigation; Project administration; Writing—review & editing. Wilko Reents: Investigation; Project administration; Writing—review & editing. Michael Zacher: Investigation; Project administration. Peter Roth: Investigation; Writing—review & editing. Bernd Niemann: Investigation; Writing—review & editing. Oliver Dörr: Investigation; Writing—review & editing. Tobias Krüger: Investigation; Writing—review & editing. Holger Nef: Investigation; Writing—review & editing. Ayman Sodah: Investigation; Project administration; Writing—review & editing. Christian Hamm: Investigation; Writing—review & editing. Christian Schlensak: Investigation; Writing—review & editing. Anno Diegeler: Investigation; Writing—review & editing. Daniel Sedding: Investigation; Methodology; Validation; Visualization; Writing—review & editing. Ulrich Franke: Conceptualization; Investigation; Supervision; Writing—review & editing. Andreas Boening: Conceptualization; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Supervision; Writing—review & editing.

REFERENCES

ABBREVIATIONS

 
• CABG

  Coronary artery bypass grafting surgery

 
• CAD

  Coronary artery disease

 
• CK

  Creatine kinase

 
• CK-MB

  Creatine kinase, isoform MB

 
• LAD

  Left anterior descending

 
• PCI

  Percutaneous coronary intervention

 
• TIMI

  Thrombolysis in myocardial infarction