Objectives Treatment options for coronary chronic total occlusions (CTO) are limited, with low historical success rates from percutaneous coronary intervention (PCI). We report procedural outcomes of CTO PCI from 7 centres with dedicated CTO operators trained in hybrid approaches comprising antegrade/retrograde wire escalation (AWE/RWE) and dissection re-entry (ADR/RDR) techniques.
Methods Clinical and procedural data were collected from consecutive unselected patients with CTO between 2012 and 2014. Lesion complexity was graded by the Multicentre CTO Registry of Japan (J-CTO) score, with ≥2 defined as complex. Success was defined as thrombolysis in myocardial infarction 3 flow with <30% residual stenosis, subclassified as at first attempt or overall. Inhospital complications and 30-day major adverse cardiovascular events (MACEs, death/myocardial infarction/unplanned target vessel revascularisation) were recorded.
Results 1156 patients were included. Despite high complexity (mean J-CTO score 2.5±1.3), success rates were 79% (first attempt) and 90% (overall) with 30-day MACE of 1.6%. AWE was highly effective in less complex lesions (J-CTO ≤1 94% success vs 79% in J-CTO score ≥2). ADR/RDR was used more commonly in complex lesions (J-CTO≤1 15% vs J-CTO ≥2 56%). Need for multiple approaches during each attempt increased with lesion complexity (17% J-CTO ≤1 vs 48% J-CTO ≥2). Lesion modification (‘investment procedures’) at the end of unsuccessful first attempts increased the chance of subsequent success (96% vs 71%).
Conclusions Hybrid-trained operators can achieve overall success rates of 90% in real world practice with acceptable MACE. Use of dissection re-entry and investment procedures maintains high success rates in complex lesions. The hybrid approach represents a significant advance in CTO treatment.
Statistics from Altmetric.com
Patients with coronary chronic total occlusions (CTO) are less commonly referred for percutaneous coronary intervention (PCI) than those with non-occlusive disease,1 potentially due to low expectations of procedural success and concerns regarding complications. Treatment options for patients with more complex CTO, particularly post coronary artery bypass grafting (CABG), are limited. Recent advances in wire technology, coupled with dissection re-entry and retrograde approaches for CTO, have markedly improved procedural success rates.2 Hybrid CTO PCI3 employs all of these recent advances. According to the hybrid algorithm (figure 1), an initial procedural approach is selected based on key angiographic variables, with early default to alternative approaches if there is failure of case progression.4 This approach has been widely disseminated throughout the UK via proctoring of dedicated CTO operators.5
CTO lesion complexity is commonly assessed using the five-point Multicentre CTO Registry of Japan (J-CTO) score (0 easy; 1 intermediate; 2 difficult; ≥3 very difficult). This was developed as a predictor of successful wire crossing within 30 min, predominantly using antegrade wire escalation (AWE) approaches.6 The score was derived from a cohort with lesions of low complexity (73% J-CTO ≤2) and low incidence of previous CABG (∼10%).7 Thus, J-CTO scoring may not predict wire crossing time or overall success in populations with more complex lesions or when hybrid approaches are employed.
We sought to define outcomes and predictors of CTO PCI success in a contemporary population of consecutive unselected patients treated by experienced CTO operators trained in hybrid approaches.5 In addition, we determined the efficacy of lesion modification (‘investment procedures’) at the end of unsuccessful CTO procedures for increasing success at subsequent attempts.
All data were derived from the UK Hybrid CTO registry, an anonymised audit of consecutive patients with CTO from seven UK centres between January 2012 and December 2014. All operators had received proctoring in AWE, antegrade dissection re-entry (ADR), retrograde wire escalation (RWE) and retrograde dissection re-entry (RDR) approaches used in the hybrid algorithm (figure 1) and had performed a career minimum of 300 CTO cases. Operators were trained in and used dedicated dissection re-entry devices (Crossboss and Stingray systems (Boston Scientific, Natick, Massachusetts, USA)) where required. Procedural and outcome data were collected and entered by operators without independent verification.
Data from 1211 patients were available. Only one CTO per patient was analysed. Patients who had two different CTO treated within one case (n=55) were excluded. Therefore, 1156 patients were included for outcome analysis (Belfast Health and Social Care Trust n=358, 31%; Edinburgh Heart Centre n=229, 19.8%; Freeman Hospital, Newcastle n=204, 17.6%; London Chest Hospital n=105, 9.1%; Golden Jubilee Hospital, Glasgow n=102, 8.8%; Bristol Heart Institute n=82, 7.1%; Ninewells Hospital, Dundee n=76, 6.6%). For patients with more than one CTO procedure (n=194), only the first procedure was included for the separate analysis of predictors of procedural success.
CTO were defined as angiographic evidence of a total occlusion with thrombolysis in myocardial infarction (TIMI) 0 grade flow and estimated occlusion duration of >3 months.8 Technical success was defined as restoration of TIMI 3 flow with a residual stenosis of <30% within the treated segment. This was subclassified as at first attempt or overall (per patient). Major adverse cardiovascular events (MACEs) during the index admission and at 30 days were calculated on a per procedure basis and defined as any of: death, myocardial infarction (MI, symptoms with ST elevation or troponin rise to >5× upper normal limit) and urgent target vessel revascularisation (TVR). Other complications recorded included: stroke, perforation requiring pericardiocentesis, acute kidney injury (AKI) requiring dialysis, or vascular injury requiring transfusion or surgery. The number and type of approaches (AWE/ADR/RDR/RWE) required to cross the CTO segment were recorded. Wire crossing time was from insertion of the guidewire into either the target vessel or retrograde donor vessel to when the wire successfully crossed the CTO segment and re-entered the true lumen.
CTO proximal cap location was defined according to American Heart Association (AHA) classification9 and additionally coded as either ostial or non-ostial and proximal (left main, proximal left anterior descending, proximal circumflex or proximal right coronary artery) or distal (all other sites). Cap morphology was coded as blunt or tapered, or as ambiguous when there was lack of clarity over the origin of the ongoing vessel. Calcification within the CTO was coded as none visible, mild (spots only), moderate (<50% of vessel circumference) or severe (>50% of vessel circumference). Tortuosity was assessed at baseline or after subsequent equipment passage and coded as straight (no bend or <45° single bend), slight (>45° single bend), moderate (2 bends >45° or 1 bend >90°) or severe (2 bends >90° or 1 bend >120°). Disease proximal and distal to the CTO was classified as absent, mild, moderate or severe. A proximal or distal cap side branch was considered present if occurring ≤3 mm from the respective CTO cap. Occlusion length was estimated by dual injections, or from apparent length after guidewire crossing. Collaterals were classified according to Cohen and Rentrop10 and deemed interventional if, on angiographic inspection, were thought to be crossable by the operator.11 Lesion complexity was judged using the J-CTO score and calculated by allocating 1 point each for: non-tapered proximal cap (ie, blunt or ambiguous), any calcification, any tortuosity, occlusion length >20 mm and any prior unsuccessful attempt.6 J-CTO score of ≥2 was defined as complex. An investment procedure at the termination of an unsuccessful attempt was defined as lesion modification of the proximal cap and/or CTO body by balloon angioplasty or passage of a microcatheter.
Overall patient success rates (irrespective of number of attempts) and per procedure success rates were calculated. Comparisons between first attempt success and first attempt failure groups were evaluated using Mann–Whitney U test for continuous variables and Pearson's χ2 test for categorical variables. In addition to lesion characteristics, candidate variables for predictors of first attempt success included patient demographic and clinical features. Variables significantly associated with success in univariate analysis or considered clinically important were assessed in stepwise logistic regression analysis within various novel models to assess the predictive capacity of different combinations of variables. Receiver operating characteristic curves were created and the C statistic defined to assess the performance of the original J-CTO score and the novel models as predictors of wire crossing time >30 min and of first attempt technical success for our cohort. Generalised estimating equations were used to account for the clustering of patients within hospitals. We used the Hosmer–Lemeshow goodness-of-fit test to assess model calibration. We compared the areas under the receiving operating characteristics curves using the non-parametrical method by DeLong et al.12 All statistical analyses were performed with Stata V.12 (Stata, College Station, Texas, USA).
Clinical and angiographic characteristics of the entire cohort and subdivided by first attempt success or failure are shown in table 1.
CTO PCI success
First attempt success rate by a hybrid-trained CTO operator was 79%. Forty-one per cent of first attempt failures (n=100) did not undergo repeat attempts due to subsequent referral for CABG or patient preference for medical therapy. For those undergoing repeat attempts (59%, n=143), 95% had a single repeat attempt. The success rate during the subsequent procedure was 87%. Thus, overall (per patient) technical success after all attempts was 90% (figure 2). First attempt technical success rates stratified by J-CTO score were 0=95%; 1=90%; 2=83%; 3=79%; 4=62%; 5=65% (figure 3).
Predictors of success
Cases with lower J-CTO scores were more likely to be successful at first attempt than higher complexity cases (J-CTO ≤1 92% vs J-CTO ≥2 74%; p<0.001). Independent predictors of first attempt success were: none/mild calcification, no proximal cap ambiguity, shorter lesion length, proximal lesion location, no prior CABG, none/mild proximal tortuosity, lower body mass index (BMI) and younger age (table 2).
The C statistic for the J-CTO score as a predictor of success was 0.68 (95% CI 0.64 to 0.71). The J-CTO score performed better in patients with a final AWE strategy (C statistic 0.73; 95% CI 0.67 to 0.79) than in those with a final strategy of ADR, RWE or RDR (ie, ‘non-AWE strategy’, C statistic 0.57; 95% CI 0.52 to 0.63). The J-CTO+ model (J-CTO variables plus prior CABG, non-proximal lesion position, proximal tortuosity (moderate/severe), distal cap ambiguity) and a novel model (age, BMI, lesion length categorised three ways (<15 mm, 16–29 mm, >30 mm), calcification categorised four ways (nil, mild, moderate, severe), proximal cap categorised three ways (tapered, blunt, ambiguous), prior CABG, non-proximal lesion position, proximal tortuosity (moderate/severe)) both produced small, but significant, improvements in the C statistic to 0.72 (p=0.0036 for comparison of J-CTO score to novel model) (table 2, figure 4). The Hosmer–Lemeshow goodness-of-fit test p value was 0.20, suggesting adequate model calibration.
Median wiring time in first attempt success cases was J-CTO 0=8 min; 1=13 min; 2=24 min; 3=40 min; 4=60 min; 5=76 min. The C statistic for the J-CTO score as a predictor of wiring time <30 min was statistically robust at 0.79 (95% CI 0.76 to 0.82).
Procedure time, radiation and contrast exposure were greater for patients with a failed first attempt compared with those with successful attempts (table 3). The median number of wires used was 4 (range 1–25). The mean number of strategies was 1.5 per procedure. ADR was the final approach in 21% of all cases. For dissection during ADR cases, CrossBoss was used in 73% while knuckle wires were used in 27%. Re-entry was achieved using Stingray balloon in 58%, directional wiring in 25%, true-to-true Crossboss passage in 15% and knuckle wire re-entry in 2%. Retrograde approaches were used in 30% of cases (54% as the primary strategy). Septal collaterals were used more commonly (65%) than epicardials (30%) or bypass grafts (5%).
The final approach used according to the J-CTO score is presented in figure 3. Complex lesions more commonly required multiple approaches (single approach 83% of J-CTO ≤1 vs 52% J-CTO ≥2) and dissection re-entry techniques (15% of J-CTO ≤1 vs 56% of J-CTO ≥2 and 67% of J-CTO >2). AWE as a primary and final strategy was highly successful in 94% (224/238) in J-CTO ≤1 cases, but less so in complex lesions (79% success, 247/313). AWE was used as the final strategy in 81% of J-CTO ≤1 lesions, but in only 37% of J-CTO ≥2 lesions. ADR as a final strategy was successful in 81% (26/32) J-CTO ≤1 lesions and 65% (140/214) of ≥2 lesions. Retrograde techniques were equally successful in J-CTO ≤1 and ≥2 lesions (77%, n=24/31 vs 77% n=251/328).
Procedural complication rates included: death 0%, MI 0.8%, acute vessel closure 0.4%, donor artery injury 0.3%, transient ischaemic attack (TIA)/stroke 0.4%, AKI requiring dialysis 0.3%, vascular access site injury 0.7% and coronary perforation (any) in 4.6%. Significant perforation (ie, Ellis grade 3 (1.0%) or tamponade requiring pericardiocentesis (0.7%)) occurred in 1.5% of the overall population and was more common in failed cases (2.8% vs 1.0%; p=0.02). No patient required urgent inhospital CABG or PCI.
At 30 days, death occurred in 0.3% (n=3) and MACE (death/MI/unplanned TVR) in 1.6% (n=18). MACE was more common in unsuccessful procedures (2.5% vs 0.4%; p<0.001). There were no reported cases of radiation dermatitis. The incidence of complications according to the CTO PCI approach is presented in figure 5.
Strategies for repeat procedures
Of patients with an unsuccessful first procedure, 59% underwent a repeat attempt at a median duration of 3 months (IQR 1.8–4.3). Re-attempt cases were more complex (original mean J-CTO score 3.1±1.2) yet the success rate was 87%.
When AWE was the final approach used during the first attempt, the final successful approach at re-attempt was AWE 42%, ADR 36% and RDR 22%. When ADR was the final approach used during the first attempt, the final successful approach at re-attempt was AWE 28%, ADR 34% and RDR 38%. If RWE was the initial failed approach, subsequent success usually required a dissection re-entry strategy (75%; half antegrade). When RDR was the final approach used during the first attempt, the majority of repeat attempts also used RDR.
Lesion modification (investment procedure) was performed in 62% of cases at the end of the unsuccessful first procedure. Although lesion complexity did not differ between groups (modification group J-CTO 3.0±1.2 vs no modification 3.3±1.2; p=0.24), prior investment was associated with very high success rates during subsequent attempts (96% vs 71%) and with favourable procedural characteristics, including shorter procedure time (105±42 min vs 154±48 min; p<0.001), lower contrast dose (291±116 mL vs 375±144 mL; p<0.001), shorter fluoroscopy time (41±19 min vs 64±24 min; p<0.001), lower radiation dose (dose area product (DAP) 13 323±7968 cGycm2 vs 20 187±10 377 cGycm2; p<0.001) and, if successful, shorter wire crossing time (35±29 min vs 97±43 min; p<0.001). For patients whose prior investment procedure involved balloon angioplasty of the proximal cap, success at repeat procedure was 95% with a median wire crossing time of 20 min.
Mechanisms of procedural failure
Failure modes were: subintimal passage and failure to re-enter (49%), inability to cross with a guidewire (29%), inability to deliver equipment (11%) and complication (11%). When ADR was the final strategy, subintimal haematoma was the predominant failure mode (47%). For retrograde failures (either RWE or RDR), an inability to cross collateral channels with a guidewire was the most common failure mode (67%), followed by the inability to cross the lesion with a guidewire (20%).
This cohort of patients with CTO with a preponderance of complex lesions represents one of the largest described experiences of contemporary CTO PCI. Despite high lesion complexity, we report a first attempt technical success rate of 79% and a final per patient success rate of 90%. We found a clear relationship between lesion complexity and increasing need for multiple approaches during CTO PCI to achieve success, which highlights the need for specialist CTO operators trained in retrograde and dissection/re-entry techniques. We also found that lesion modification during an initial unsuccessful attempt conveyed a higher likelihood of success at repeat attempt, and thus served as an ‘investment’.
Success rates and strategies used
Success rates for CTO PCI described in registries had remained relatively static at around 70–80% for many years;13–16 such registries typically included patients with lower complexity lesions (suggesting significant case selection) with a dominant use of AWE approaches. Recent technological advances in equipment as well as the evolution of the retrograde and dissection re-entry approaches have improved CTO PCI success rates.17 ,18 Our study included an all-comer population with patients treated on the basis of clinical need, rather than anatomical complexity. Three quarters of patients had J-CTO scores of ≥2 (difficult or very difficult), 22% had prior CABG, 26% had CTO segments >30 mm and 20% had previous failed attempts. Despite these unfavourable characteristics, we demonstrated that, in the hands of trained CTO subspecialists, patients and referring cardiologists can expect success rates close to those seen with conventional angioplasty, along with an acceptable MACE risk. We suggest that CTO PCI, performed by a dedicated CTO specialist, should be offered to the majority, rather than the current 10–15% of patients with CTO.18
For patients with non-complex lesions (J-CTO ≤1), AWE remains a highly successful strategy. However, its performance is limited in more complex anatomy and this undoubtedly contributed to the requirement for multiple approaches in our complex patient cohort. The hybrid approach encourages operators to avoid stalling in failure modes and to use all available strategies to improve the chances of success. This multimodality approach requires operators to be fully skilled in dissection re-entry and retrograde techniques. Such skill sets are not typically acquired during interventional training, but proctoring by a CTO specialist can rapidly lead to acquisition of necessary skills.5 Indeed, many of the operators contributing to this registry went through the UK proctoring process to become hybrid CTO specialists.
Use of the J-CTO score
Our data serve to validate the use of the J-CTO score for assessment of lesion complexity as judged by wire crossing time (area under the curve (AUC) 0.79), but suggest that when patients are treated by hybrid CTO specialists, the J-CTO score it is not as strong a predictor of first attempt success (AUC 0.68), particularly when the final strategy is not AWE. This likely reflects the fact that evolving techniques and technical approaches can overcome the challenges specific to patient anatomy on which the score is largely based. Therefore, CTO anatomy and higher J-CTO scores should not necessarily determine whether patients are offered a CTO PCI, but rather guide the likelihood of needing multiple strategies to achieve success.
Predictors of success
Calcification and lesion length were strong predictors of success as in the J-CTO model.6 However, proximal cap ambiguity was a better predictor than a tapered cap, and tortuosity proximal to the CTO was superior to insegment tortuosity. Success rates were significantly lower in our series in previously grafted vessels (an established contributor to case failure in other series7 ,19–22) and in lesions within the distal vessel.
High procedural success rates (87%) were achieved during subsequent procedures after prior failure by a hybrid-trained CTO specialist, despite higher lesion complexity. This was particularly evident in those who had undergone lesion modification during the index PCI, which can retrospectively be considered an ‘investment procedure’. Early recognition of impending procedural failure during the first attempt is important to avoid reaching radiation/time/contrast thresholds. Operators should identify modifiable factors contributing to failure and consolidate any progress that has been made. An example is balloon angioplasty at the proximal cap after failure to wire the distal vessel; negotiating the proximal cap commonly consumes procedural time, and ‘consolidation’ by angioplasty at this site improved the future success rate to 95% in our cohort, presumably by lesion modification or the creation of a track to the distal vessel architecture that could be exploited at a subsequent procedure.23 Achieving success across two procedures rather than one prolonged procedure may also reduce overall radiation and contrast exposure, thereby improving patient safety.
Demographic, procedural and outcome data were entered by the respective CTO operators and angiograms were not reviewed by a core laboratory. The data thus lack external validation. It is unlikely that the results of this study will be generalisable to other, less experienced CTO PCI operators, or to those who use only AWE approaches. However, the majority of CTO operators were proctored in retrograde and dissection/re-entry CTO techniques within the last 5 years, demonstrating that the necessary skills may be rapidly acquired and translated into direct patient benefits. Although we aimed to take an all-comer population, the majority of operators were reliant upon referrals from other cardiologists, therefore, we cannot determine the degree of case selection that may have occurred upstream of referral.
In real world contemporary practice, the first attempt success rate of dedicated CTO operators using hybrid approaches is 79%. Twelve per cent of the original cohort (59% of first attempt failures) underwent subsequent CTO PCI procedures to achieve an overall per patient success rate of 90% with a 30-day MACE rate of 1.6%. The requirement for multiple approaches including retrograde and dissection/re-entry techniques increased with lesion complexity. Lesion modification at the termination of a failed attempt facilitated future success. The hybrid approach represents a significant advance in the treatment of coronary CTO.
What is already known on this subject?
Chronic total occlusions (CTO) are seen in 15% of patients undergoing coronary angiography, but revascularisation is attempted in the minority. There is good evidence for an improvement in quality of life and a reduction in symptoms of angina and dyspnoea after successful CTO percutaneous coronary intervention (PCI). Increased anatomical complexity including longer lesions, blunt proximal cap and diffuse disease typically negatively influence the decision to attempt revascularisation, owing to a perception of futility and/or higher procedural risk.
What might this study add?
Overall patient success rates of 90% with an acceptable major adverse cardiovascular event rate (1.6%) are achievable in an all-comer population when PCI is performed by hybrid-trained CTO operators. Although anatomical complexity increased the likelihood of requiring multiple PCI techniques, it was neither a strong predictor of overall case failure nor a risk of complications. In cases of failure at first attempt, investment procedures increase the chances of subsequent success.
How might this impact on clinical practice?
These results should encourage increased adoption of PCI for symptomatic patients with CTO lesions. Hybrid-trained CTO operators may be best placed to assess the likelihood of successful revascularisation, particularly if anatomical complexity is high.
The authors thank the following for assistance with data entry: Jane Blair, Jane Geddes, Roshan Paranamana, Aadil Shaukat, Murugapathy Veerasamy. The authors also thank Optima education and Adrian Brown for illustrative inputs.
Contributors WMW was involved in study design, manuscript preparation. SJW was involved in study design, data contribution and manuscript preparation. CGH was involved in study design, data contribution and manuscript preparation. AJB was involved in study design, data contribution and manuscript preparation. ES was involved in data contribution and manuscript preparation. JI was involved in data contribution and manuscript preparation. JS was involved in data contribution and manuscript preparation. HD was involved in data contribution. KGO was involved in data contribution. MME was involved in data contribution. JCS was involved in study design, data contribution and manuscript preparation.
Competing interests AJB and JS have received income for proctoring from Abbott Vascular. SJW is a consultant to Abbott Vascular, Boston Scientific, Medtronic and Vascular Solutions. He has also received research funding from Abbott Vascular, Boston Scientific and Nitiloop. CGH is a consultant to Abbott Vascular, Boston Scientific, Medtronic and Vascular Solutions. ME has received consultancy and honorarium from Abbott Vascular, Boston Scientific, Volcano, Vascular Perspectives and Spectranetics. ES has received income for proctoring from Boston Scientific and Vascular Perspectives. JI has received income for proctoring from Boston Scientific and Vascular Perspectives. JCS is a consultant to Abbott Vascular, Boston Scientific, Asahi Intecc, Vascular Perspectives and Vascular Solutions. He has received research funding from Abbott Vascular, Boston Scientific and Nitiloop.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.