Comparison of outcomes of percutaneous coronary intervention for chronic total occlusion in patients with and without prior bypass grafting: A systematic review and meta-analysis

Objective: This review assessed evidence on the impact of prior coronary artery bypass grafting (CABG) on outcomes of percutaneous coronary intervention (PCI) for chronic total occlusions (CTO). Methods: PubMed, CENTRAL, Embase, ScienceDirect, and Google Scholar databases were searched from 1st January 1980 up to 10th January 2022 for studies assessing outcomes of CTO-PCI in patients with and without prior-CABG. Results: Eight studies were included. The meta-analysis demonstrated significantly reduced odds of procedural success in patients with prior history of CABG (OR: 0.51 95% CI: 0.41, 0.64 I2=84% p<0.00001). There was a tendency of increased in-hospital mortality (OR: 1.72 95% CI: 0.97, 3.04 I2=26% p=0.06) and major adverse cardiac events (MACE) (OR: 1.30 95% CI: 0.99, 1.69 I2=0% p=0.05), along with a significantly increased risk of myocardial infarction (MI) (OR: 2.56 95% CI: 1.65, 3.97 I2=0% p<0.0001) and coronary perforation (OR: 1.52 95% CI: 1.03, 2.24 I2=70% p=0.04) in patients with history of CABG. There was no difference in the risk of stroke, pericardial tamponade, major bleeding, vascular access complications, and renal failure. Conclusion: Our results suggest that patients with prior history of CABG undergoing PCI for CTO have a 49% reduced chance of procedural success. Such patients are at an increased risk of in-hospital mortality, MACE, MI, and coronary perforation.


INTRODUCTION
Chronic total occlusions (CTO) are amongst the most difficult coronary artery lesions to be treated with percutaneous coronary interventions (PCI). 1 Previously, PCI for CTO lesions was associated with the high complication and reduced success rates. 2 However, with continuous advances in CTO equipment and technology, clinical outcomes have greatly improved. 3 A recent trial has demonstrated that CTO-PCI is indeed feasible with good success rates and there is no difference in the risk of major adverse cardiovascular events (MACE) between CTO and non-CTO-PCI. 4 Research has also shown that successful PCI for CTO lesions leads to improved quality of life, better left ventricular function, improved survival, and a decreased risk of coronary artery bypass grafting (CABG). 5,6 Indeed, the use of PCI for CTO has increased significantly in the past decade with a corresponding increase in procedural success. 7 Evidence from randomized controlled trials suggests that in comparison to PCI, CABG results in significantly better longterm outcomes in complex coronary artery diseases. 8,9 However, it is also known that CABG itself precipitates atherosclerosis in the native coronary arteries. 10 CABG alters the blood flow and induces stasis, negative remodeling, and calcifications leading to accelerated development of atherosclerotic lesions in native coronary arteries proximal to the grafted site. 11 Indeed, prediction models for CTO-PCI have reported a prior history of CABG to be a risk factor for procedural failure. 12 Several studies have attempted to compare outcomes of CTO-PCI between patients with and without a history of prior CABG but with conflicting results. While some 13 report worse outcomes in patients with prior CABG others 14 suggest no such difference. Two meta-analysis studies have been conducted on this topic. 15,16 While one review 16 could only include studies published up to 2019, another recent review 15 included studies with overlapping data reporting outcomes from the same institute or PCI registries thereby resulting in biased reporting. Furthermore, more recently published studies 14,17 were not included in both these reviews. Therefore, to overcome the limitations of past reviews, we designed the current study to compare in-hospital and long-term outcomes of CTO-PCI in patients with and without prior CABG.

METHODS
This review was reported based on the PRISMA recommendations 18 and was pre-registered on the PROSPERO database (CRD42022299439). Literature search: PubMed, CENTRAL, Embase, ScienceDirect, Google Scholar were searched for English language studies from 1 st January 1980 up to 10 th January 2022. The search terms were: "chronic total occlusion", "CTO", "percutaneous coronary intervention", "PCI", "coronary artery bypass", and "CABG" (Supplementary Table-I). The search results were consolidated, deduplicated, and screened by title and abstracts by two reviewers separately. Articles of interest to the review were selected and downloaded for full-text analysis. They were cross-checked against the inclusion criteria for final selection. Disagreements between the two reviewers were cleared in consultation with another reviewer. Data extraction and quality assessment: The reviewers sourced author details, study year and type, study location and database, sample size, demographic details, smokers, comorbidities (Hypertension, diabetes mellitus, renal disease, dyslipidemia), previous MI or PCI, target vessel, the approach of PCI, prior failed attempts, Japanese CTO score (J-CTO), vessel calcification, blunt stump, procedural time, contrast volume, study outcomes, and follow-up. Procedural success as the primary outcome of our review: It was defined as residual stenosis <30% and TIMI flow grade ≥3 without any MACE in all studies, except for one wherein <50% residual stenosis was considered as a successful procedure. Other outcomes of interest were all-cause mortality, MACE, MI, stroke, coronary perforation, pericardial tamponade, major bleeding, vascular access complications, renal failure, and TVR. Based on the follow-up duration, outcomes were separated into in-hospital and late outcomes (>6 months of follow-up). We assessed the risk of bias using the Newcastle-Ottawa scale (NOS). 19 Statistical analysis: We used "Review Manager" (RevMan, version 5.3; Nordic Cochrane Centre [Cochrane Collaboration], Copenhagen, Denmark; 2014) for this study. Data was combined to compute odds ratios (OR) and 95% confidence intervals (CI). We also sourced adjusted hazard ratios (HR) of long-term outcomes and combined them to compute the total effect size. All meta-analyses were conducted using the random-effects model.
We assessed inter-study heterogeneity using the I 2 statistic. We assessed publication bias for the primary outcome by visual inspection of funnel plots. A sensitivity  analysis was conducted for procedural success and allcause mortality to assess if any study had an exaggerated effect on the pooled estimate.

RESULTS
Search: 2853 unique articles were found (Fig.1). Fulltexts of these 12 studies was reviewed. Three 20-22 were found to report overlapping data while one 23 was a review article. Finally, eight studies were included in our review. 13,14,17,[24][25][26][27][28] Baseline details are presented in Tables-I and II. All were retrospective observational studies. A total of 6182 PCI procedures for CTO were conducted after CABG in the included studies and these were compared with a control group of 25445 CTO-PCI procedures. Four studies reported only in-hospital data while the other four studies also reported long-term data. The NOS score of the studies ranged from six to eight. The majority of the studies lacked baseline matching of study cohorts. Procedural success: The meta-analysis showed significantly reduced odds of procedural success in patients with prior history of CABG (OR: 0.51 95% CI: 0.41, 0.64 I 2 =84% p<0.00001) (Fig.2). There was no publication bias ( Supplementary Fig.1). The results did not differ on sensitivity analysis. In-hospital mortality: Pooled analysis indicated a tendency of increased mortality in patients with prior history of CABG, but the results were not statistically significant (OR: 1.72 95% CI: 0.97, 3.04 I 2 =26% p=0.06) (Fig.3). On the exclusion of the study of Hernandez-Suarez et al, 17 Fig.3).

Long-term outcomes:
Limited long-term outcome data were available and a pooled analysis was possible only for mortality and TVR. Meta-analysis revealed an increased risk of all-cause mortality in patients with prior history of CABG on longterm follow-up (OR: 1.54 95% CI: 1.30, 1.84 I 2 =0% p<0.00001). Similarly, patients with prior CABG had a higher odds of TVR as compared to the CABGnaïve group (OR: 1.26 95% CI: 1.03, 1.54 I 2 =39% p=0.02) (Supplementary Fig.4).

DISCUSSION
Historically, limited number of CTO patients were treated by PCI. 29 However, technological improvements has changed the clinical scenario. Research shows that compared to medical therapy, PCI for CTO resulted in lower long-term mortality with no difference in the incidence of major complications. 30 Nevertheless, the success rates achieved by CTO-PCI have generally been approximately 15% lower as compared to PCI for non-CTO lesions. 30,31 Lower success rates have in turn been related with adverse long-term outcomes and reduced overall survival. 32 Several studies have shown that prior CABG could be a risk factor for lower success rates in CTO-PCI. 33,34 To better elucidate the influence of prior CABG on CTO-PCI outcomes, we conducted a meta-analysis. We noted that in patients with prior CABG the odds of procedural success are significantly reduced. Our results are similar to the prior meta-analysis of Liu et al. 16 However, the more recent review of Shi et al 15 did not analyze success rates. Lower procedural success with prior CABG could be attributable to several reasons. Firstly, the difference in the baseline demographic and patient characteristics are important contributors to PCI success.
It can be seen that in the majority of the included studies, patients with prior CABG were older, more male, had a higher incidence of comorbidities like diabetes, hypertension, dyslipidemia, renal disease along with the greater frequency of prior MI and PCI. Prior CABG patients also had complex lesions which makes the procedure challenging. Secondly, the association between prior CABG and aggressive development of atherosclerosis in native coronary arteries is quite well established. 10 Research suggests that CTO lesions in surgically revascularized patients have greater calcification, moderate negative remodeling, and excessive prevalence of blunt stumps compared to CABG naïve patients. 11 Supplementary Fig.4: Meta-analysis of long-term all-cause mortality and TVR between prior CABG (pCABG) and CABG-naïve patients (nCABG).
The escalated risk of atherosclerosis after CABG has been attributed to abnormal flow patterns resulting in altered shear stress. 35 The development of blood stasis with such atypical blood flow could contribute to the higher calcification rates found in native coronary arteries of post-CABG patients. 11 Such negative angiographic features could therefore impact the success rates of CTO-PCI. Thirdly, distortion and displacement of native vessel anatomy by prior CABG could also hinder CTO crossing attempts leading to higher failure rates. 16 Furthermore, a larger incidence of complications in prior CABG patients could reduce the procedural success.
Indeed, our meta-analysis revealed that there was an increased tendency of in-hospital mortality and MACE in patients with a history of CABG, but without statistical difference. This could be due to the low number of events in the included studies. Secondly, there was increased risk of MI and coronary perforation in patients with a history of CABG, but no difference in other complications. These results are similar to Liu et al. 16 However, Shi et al 15 in their meta-analysis have also noted the reduced incidence of cardiac tamponade and higher rates of contrast-induced nephropathy in patients with prior CABG.
It is pertinent to mention that the studies included in our review and that of Shi et al 15 13 and hence were excluded from our meta-analysis. Including the same patients twice in a pooled analysis can exaggerate the effect of the intervention thereby generating false results. We also added two recently published studies with large sample sizes thus providing updated evidence.
The higher incidence of coronary perforation in prior CABG patients could be attributable to the aggressive techniques used in such patients. The baseline data of the included studies show that prior CABG patients frequently underwent the retrograde and dissection reentry approach as compared to CABG-naive patients. The retrograde approach associated collateral channel damage and the requirement of aggressive balloon dilation in the severely atherosclerotic vessels of CABG patients could also increase the risk of coronary perforations. 27 Our review is the first to pool evidence on long-term outcomes of CTO-PCI between prior CABG and CABG naïve patients. Acknowledging the fact that long-term data is currently scarce, we noted higher mortality rates in prior CABG patients on the meta-analysis of crude data but the results were no longer significant on analysis of adjusted data. This indicates that the baseline differences in the prior CABG group are the primary drivers of poor survival. Crude data on TVR did indicate a higher need for revascularization in prior CABG patients, but we were unable to pool adjusted outcomes for TVR due to unavailability of data from the included studies.
Limitations: Firstly, the current evidence is derived from only retrospective observational studies which have an inherent risk of bias. Secondly, PCI in the included studies was performed over a large period ranging from 1999 to 2020. The current results do not take into account the technological developments and technique improvements for CTO-PCI occurring over such a long time duration. Also, the procedures were carried out at different centers worldwide by operators of varying experience.
PCI for CTO is a highly skilled procedure and success rates are directly proportional to operator and hospital experience. 36 Thirdly, our meta-analysis was a study level and not a patient-level meta-analysis. The latter would have provided better evidence. Lastly, maximum outcomes pooled in our study were from crude and not adjusted data. The majority of the included studies did not carry out baseline matching of the study groups and failed to report multivariable-adjusted data.

Future directions:
Further prospective studies with longterm follow-up are needed to generate better quality evidence on the effect of prior CABG on outcomes of CTO-PCI. Future studies should carry out baseline matching of patient characteristics and reported multivariable adjusted data for better interpretation of the available evidence.

CONCLUSION
Patients with prior history of CABG undergoing PCI for CTO have a 49% reduced chance of procedural success. Such patients are at an increased risk of in-hospital mortality, MACE, MI, and coronary perforation. The reduced success rates and higher complications in prior CABG patients are probably related to the unfavourable patient demographics, higher comorbidities, and increased complexity of CTO lesions.