Author + information
- Received September 24, 2014
- Revision received April 14, 2015
- Accepted July 2, 2015
- Published online October 1, 2015.
- Giuseppe Alessandrino, MD,
- Bernard Chevalier, MD,
- Thierry Lefèvre, MD,
- Francesca Sanguineti, MD,
- Philippe Garot, MD,
- Thierry Unterseeh, MD,
- Thomas Hovasse, MD,
- Marie-Claude Morice, MD and
- Yves Louvard, MD∗ ()
- Institut Cardiovasculaire Paris Sud, Hôpital Privé Jacques Cartier, Générale de Santé, Massy, France
- ↵∗Reprint requests and correspondence:
Dr. Yves Louvard, Hopital Jacques Cartier, L’angio, 6 avenue du Noyer Lambert, 91300 Massy, France.
Objectives This study sought to develop a scoring model predicting percutaneous coronary intervention (PCI) success in chronic total occlusions.
Background Coronary chronic total occlusion is the lesion subtype in which angioplasty is most likely to fail. Chronic total occlusion for PCI (CTO-PCI) failure is associated with higher 1-year mortality and major adverse cardiac events compared with successful CTO-PCI. Although several independent predictors of final procedural success have been identified, no study has yet produced a model predicting final procedural outcome.
Methods Data from 1,657 consecutive patients who underwent a first-attempt CTO-PCI were prospectively collected. The scoring model was developed in a derivation cohort of 1,143 patients (70%) using a multivariable stepwise analysis to identify independent predictors of CTO-PCI failure. The model was then validated in the remaining 514 (30%).
Results The overall procedural success rate was 72.5%. Independent predictors of CTO-PCI failure were identified and included in the clinical and lesion-related score (CL-score) as follows: previous coronary artery bypass graft surgery +1.5 (odds ratio [OR]: 2.49, 95% confidence interval [CI]: 1.56 to 3.96), previous myocardial infarction +1 (OR: 1.6, 95% CI: 1.17 to 2.2), severe lesion calcification +2 (OR: 2.72, 95% CI :1.78 to 4.16), longer CTOs +1.5 (≥20 mm OR: 2.04, 95% CI: 1.54 to 2.7), non–left anterior descending coronary artery location +1 (OR: 1.56, 95% CI: 1.14 to 2.15), and blunt stump morphology +1 (OR: 1.39, 95% CI: 1.05 to 1.81). Score values of 0 to 1, >1 and <3, ≥3 and <5, and ≥5 identified subgroups at high, intermediate, low, and very low probability, respectively, of CTO-PCI success (derivation cohort: 84.9%, 74.9%, 58%, and 31.9%; p < 0,0001; validation cohort: 88.3%, 73.1%, 59.4%, and 46.2%; p < 0.0001).
Conclusions This clinical and angiographic score predicted the final CTO-PCI procedural outcome of our study population.
Several studies have shown that successful percutaneous coronary intervention (PCI) for coronary chronic total occlusion (CTO) is associated with a better outcome in terms of reduced mortality and 1-year major adverse cardiac events compared with CTO-PCI failure (1–5).
Although new dedicated devices and guidewires have had a favorable impact on procedural success (6,7), CTO remains the type of lesion in which angioplasty is most likely to fail. In a multicenter study, Morino et al. (8), developed an angiographic scoring system to stratify CTO-PCI complexity. Recently, Nombela-Franco et al. (9) confirmed that the Japanese chronic total occlusion (J-CTO) score is a useful tool for predicting successful guidewire crossing of the CTO within 30 min, but they failed to demonstrate the value of such a model in predicting final procedural success. Although the predictive value of angiographic parameters has been thoroughly investigated, it has been recently reported that even clinical characteristics may have a predictive role (10,11).
To obtain a scoring model able to predict final CTO-PCI success, we prospectively analyzed a cohort of patients who underwent CTO-PCI, taking into account both clinical and angiographic parameters.
We analyzed 1,671 consecutive patients with CTO treated by PCI from January 2004 to December 2013. All patients provided written informed consent. The study protocol was approved by the local ethics committee, and all procedures were performed according to current international guidelines (12,13).
We reviewed the database of the 1,671 patients who underwent CTO-PCI procedures at ICPS (Massy and Quincy hospitals, France).
Definitions and patient selection
A CTO was defined by angiography as a coronary occlusion without antegrade filling of the distal vessel other than via collaterals. The duration of the occlusion had to be more than 3 months, as estimated from the onset of clinical events including myocardial infarction (MI), sudden onset or worsening of chest symptoms, or evidenced by angiography. When the duration of the occlusion was uncertain, and the investigators had no clear evidence that it was <3 months, the patient was included in the analysis.
Quantitative assessment was performed using the Quant-Cor QCA software package (CAAS II, V.5.0, Pie Medical Imaging, Maastricht, the Netherlands).
Previous MI was defined as an MI (non–ST-segment elevation MI or ST-segment elevation myocardial infarction) with a culprit lesion other than the CTO or an MI that had occurred at least 3 months before the CTO-PCI.
Previous coronary artery bypass graft surgery (CABG) was defined as a previous CABG of the CTO vessel carried out at least 3 months before the CTO-PCI procedure.
Blunt morphology of the lesion: the lesion was classified as a blunt morphology type if the occluded segment did not end in the shape of a funnel.
Calcified lesions: presence of calcification was classified according to 3 categories—mild, moderate, and severe.
The measurement of the degree of retrograde collateral supply was performed using a collateral grading system, according to the Rentrop classification (14).
Bending lesion: this was defined as any lesion with at least 1 bend of >45° assessed by angiography along the whole length of the occluded segment.
Length of occlusion: occlusion length was categorized as either <20 or ≥20 mm according to the EuroCTO Club consensus (15).
Successful procedure was defined as the achievement of <30% residual diameter stenosis as assessed by quantitative coronary angiography and associated with Thrombolysis In Myocardial Infarction flow grade 3.
To improve the accuracy of lesion analysis, all cases were reviewed by our core laboratory before and after the procedure.
Patients in whom CTO-PCI was attempted in the setting of acute cardiogenic shock were excluded. In addition, to minimize confounders, only the first CTO-PCI attempts performed during the enrolling period were included in the analysis. If 2 or more CTO PCIs were attempted during the same procedure, the patient was excluded. If a single CTO lesion or 2 different CTO lesions were attempted in 2 or more separate procedures during the enrolling period, only the first procedure was considered as a first attempt and subsequently included in the analysis.
Definition of complications
MI was defined as >3 times the upper limit of normal creatine phosphokinase release, in the absence of escalating creatine phosphokinase before PCI.
Ostial dissection and major pericardial effusion were defined, respectively, as angiographic evidence of significant staining of medium contrast in subintimal strata of the coronary wall involving the coronary ostium and pericardial effusion requiring pericardiocentesis.
Continuous variables were presented as mean ± SD or median (25th to 75th percentiles) according to distribution. Categorical variables were expressed as absolute numbers and percentages. Chi-square test was applied to evaluate the association between qualitative variables. The Student t test was used to compare quantitative variables between the 2 groups. A predictive score was developed on the basis of a derivation set of 1,143 randomly selected patients (70%). The score was then validated in the remaining 514 patients (30%; validation cohort). In the derivation cohort, a backward stepwise multivariable logistic regression was used to select a set of independent predictive variables. Independent variables were then scored according to the corresponding odds ratio (OR). The score validation was performed by logistic regression analysis of the validation cohort. Comparison of the success rates among score classes was also performed with a chi-square test for trend. A receiver-operating characteristic curve analysis was performed to compare the predictive impact of the new score with that of the J-CTO score.
To assess whether technical changes and improvement during the study period had an impact on our predictive variables, procedures were classified in chronological order and divided into 2 subgroups according to the procedural period: group A (January 2004 to December 2010) and group B (January 2011 to December 2013).
Between January 2004 and December 2013, 14 operators performed 1,833 CTO-PCI procedures in 1,671 patients of whom 14 were excluded because CTO-PCI was attempted in the setting of cardiogenic shock. A total of 1,657 patients with 1,657 first CTO-PCI attempts were enrolled. Procedural success was achieved in 1,202 of 1,657 procedures, accounting for an overall success rate of 72.5%.
The baseline characteristics of the 1,657 study patients are reported in Table 1.
The median age of the population was 64 years; 27% of patients had diabetes, and 21.2%, 37.7%, and 7.5% had prior MI, prior PCI, and prior CABG, respectively. The most common indication for PCI was stable angina (56.3%), and the majority of patients had multivessel disease (55.9%).
The right coronary artery was the most common site of CTO (46.1%), followed by the left anterior descending coronary artery (LAD) (30.5%), circumflex (23.0%), and left main (0.4%); 56.8% of lesions were calcified, 15.1% were tortuous, and 35.0% were long (≥20 mm). Class 3 Rentrop collateralization was reported in 44.1% of cases.
The majority of procedures were transradial (65.9%) and were carried out via an anterograde approach (90.6%). Contralateral injection was performed in 36.2% of patients. Attempts at achieving guidewire support were made using a microcatheter in the majority of cases (51.2%), followed by a coaxial balloon (22.5%) and by a monorail balloon (5.7%). The mean J-CTO score was 1.62 ± 0.89.
Each patient was randomly assigned to the derivation cohort (70%, 1,143 patients), or to the validation cohort (30%, 514 patients). Except for smoking, no significant differences in terms of clinical and lesion-related characteristics were reported between the 2 groups (Table 2).
Derivation model univariable analysis
The derivation cohort was divided according to procedural outcome, and clinical and lesion-related characteristics were analyzed (Table 3). Patients in whom CTO-PCI failed had a significantly higher body mass index (28.0 ± 4.4 vs. 27.3 ± 3.9; p = 0.024), higher incidence of dyslipidemia (69.3% vs. 61.8%; p = 0.018), more frequent history of MI (27.8% vs. 18.0%; p < 0.0001), PCI (44.3% vs. 33.4%; p = 0.001), and coronary artery bypass surgery (12.3% vs. 5.6%; p < 0.0001) Stable angina was a less frequent indication of CTO-PCI in patients in whom PCI was unsuccessful (52.2% vs. 58.8%; p = 0.045).
Concerning lesion-related variables, nonrevascularized lesions were less frequently located in the LAD (20.9% vs. 33.6%; p < 0.0001), more often had a blunt stump (62.1% vs. 51.5%; p = 0.001), and were tortuous (19.4% vs. 13.3%; p = 0.01), severely calcified (16% vs. 6.7%; p < 0.0001), and longer (≥20 mm: 47.2% vs. 29.9%; p < 0.0001).
Using a backward stepwise approach, severe lesion calcification (OR: 2.72; 95% confidence interval [CI]: 1.78 to 4.16; p < 0.0001), previous CABG (OR: 2.49; 95% CI: 1.56 to 3.96; p < 0.0001), lesion length ≥20 mm (OR: 2.04; 95% CI: 1.54 to 2.70; p < 0.0001), previous MI (OR: 1.60, 95% CI: 1.17 to 2.19; p = 0.003) non-LAD location of the lesion (OR: 1.56, 95% CI: 1.14 to 2.15; p = 0.006), and blunt morphology of the lesion (OR: 1.39; 95% CI: 1.05 to 1.81; p = 0.028) were all independent predictors of unsuccessful CTO-PCI (Table 4).
Independent predictors of unsuccessful CTO-PCI were scored according to the corresponding OR size (Table 5). Four subgroups were identified on the basis of the success rate: score values of 0 to 1 (class 0), >1 and <3 (class 1), ≥3 and <5 (class 2), and ≥5 (class 3) corresponding to high, intermediate, low, and very low probability, respectively, of CTO-PCI success in the derivation cohort (p < 0.0001) (Figure 1). According to our score ranking, the probability of CTO-PCI failure increased significantly from class 0 to class 1 (15.1% to 25.1%; p = 0.0004), from class 1 to class 2 (25.1% to 41.1%; p < 0.0001), and from class 2 to class 3 (41.1% to 68.1%; p = 0.0005), respectively.
Validation model analysis
The predictive score was then applied to the validation cohort. The CL-score was also shown to be a strong predictor of final procedural success in the validation cohort (88.3%, 73.1%, 59.4%, and 46.2% in classes 0 to 3, respectively; chi-square for trend: p < 0.0001). In a logistic regression, the ORs of class 1 to 3 with respect to class 0 were: 2.78 (95% CI: 1.53 to 5.0), 5.15 (95% CI: 2.8 to 9.3), and 8.78 (95% CI: 3.5 to 21.7), respectively (all p values were <0.0001).
A receiver-operating characteristic curve analysis of the clinical and lesion-related score (CL-score) versus J-CTO score carried out in the validation cohort demonstrated the superior performance of the CL-score: area under the curve 0.68 (95% CI: 0.63 to 0.73) versus area under the curve 0.60 (95% CI: 0.54 to 0.65), respectively (Figure 2).
The complication rate was low. No periprocedural deaths were recorded (<24 h). Post-procedural MI was reported in 34 patients (2%). Ostial dissection and pericardiocentesis were reported in 17 (1.0%) and 20 (1.3%) patients, respectively (Table 6). Patients who experienced coronary ostium dissection were all asymptomatic and did not require any surgical treatment. Two of 20 patients with pericardiocentesis required surgical pericardial drainage.
Our population was divided into 2 subgroups according to the date of the procedure. Group A (n = 1,066) and group B (n = 591) included patients who underwent CTO-PCI between January 2004 and December 2010, and between January 2011 and December 2013, respectively (Figure 3).
The impact of the subgroup variable on final procedural success was found to be statistically significant (group A, 70.8%; group B, 75.6%; p = 0.036). Mean J-CTO score of group A was not statistically different compared with group B (1.59 ± 0.88 vs. 1.62 ± 0.89; p = 0.27).
As expected, the availability of new dedicated devices changed the operators’ approach to CTO-PCI during the enrolling period (Table 7). In group B, a larger proportion of transradial procedures (73% vs. 62%; p < 0.001), retrograde approach (15.4% vs. 6.0%; p < 0.0001), and concomitant collateral injection (50.2% vs. 28.4%; p < 0.0001) was performed. Coaxial (32.4% vs. 3.9%; p < 0.0001) and monorail balloon (7.9% vs. 1.8%; p < 0.0001) support and parallel wire techniques (19.6% vs. 14.4%; p = 0.001), frequently implemented in group A, were substituted by use of the microcatheter in the second part of the study (32.1% vs. 85.6%; p < 0.0001). The Tornus microcatheter (Asahi Intecc, Nagoya, Japan) was more frequently used in the first part of the study (2.8% vs. 0.8%; p = 0.008). No differences in the use of rotational atherectomy, anchor wire, or anchor balloon technique were reported. In group B, a significant reduction in the myocardial infarction rate was observed (0.8% vs. 2.7%, p = 0.01). On the other hand, a significant increase in the rate of coronary ostium dissection was reported in group B compared with group A (2.5% vs. 0.2%; p = 0.0001). No difference in pericardiocentesis was observed between the groups.
The score system presented here (CL-score) was developed on the basis of a prospective single-center study of 1,657 patients who underwent a first CTO-PCI attempt. To the best of our knowledge, this is the first predictive model of final procedural CTO-PCI success including both clinical and angiographic variables.
Notwithstanding the significant increase in success rates up to more than 85% to 90% reported by the majority of recent studies, especially in patients treated using the retrograde or hybrid approach (16,17), CTO is still the setting in which PCI is associated with the highest probability of failure. Previously reported studies have already underlined the impact of several lesion-related characteristics on the procedural outcome (18–23). More recently, calcification, bending, blunt stump, occlusion length ≥20 mm, and prior failed attempts were included in the J-CTO validated score as predictors of probability of guidewire crossing and predictors of longer procedural time (8,9). In addition, clinical variables have been shown to have a negative impact on final procedural success as reported in the most recent CTO-PCI series even when the hybrid approach was implemented (11,24,25).
In the study reported here, 2 clinical variables (history of CABG and history of MI), and 4 lesion-related variables (blunt stump, lesion calcification, non-LAD CTO, and lesion length ≥20 mm) were identified as independent predictors of unsuccessful CTO-PCI. Given that only first-attempt CTO-PCIs were included in the study, the majority of these procedures were performed via the antegrade approach. In instances where the antegrade approach failed, a new procedure was generally attempted via the retrograde approach. The fact that second attempts did not meet the inclusion criteria may account for the low rate of retrograde approach in our study. This, along with the low rate of contralateral injection and the relatively long enrollment period, may explain the relatively low success rate observed in our population.
A subgroup analysis was carried out to compare procedures performed in the first period (group A) and in the second period (group B) to minimize the impact of newly available devices on final procedural success during the enrolling period. The higher success rate observed in the second part of the study suggested a favorable impact of new dedicated devices and increasing experience on the final procedural success rate. In addition, the increased incidence of coronary ostium dissection in group B was not clinically relevant (no need for surgical treatment and spontaneous resolution), whereas a significant reduction in post-procedural MI was observed in the second part of the enrolling period.
The effect of operator expertise on procedural success was also taken into account. Consensus reports on CTO management specify that patients scheduled for CTO-PCI should be referred to a skilled operator. Although a minimum of 75 CTO-PCIs per operator has been considered a reasonable threshold to identify a skilled operator (26), no definite cutoff value has yet been validated. Although the operators involved in the study had varying degrees of expertise, all of them were, nevertheless, skilled operators. Extrapolation analysis of data pertaining to 2 highly experienced CTO-PCI operators (>300 CTOs each in this study) who performed CTO-PCIs from 2010 to 2013 (213 procedures, contralateral injection 60%, retrograde approach 21%) showed an 87.8% success rate.
An increasing number of operators and centers are currently building up CTO PCI expertise owing to the availability of new dedicated devices, techniques, and training courses organized by expert operators and to mounting evidence of the benefits of CTO revascularization compared with medical therapy (25,27).
Given that the present predictive CL-score was mainly established on the basis of CTO-PCIs attempted via the antegrade approach, it seems to be applicable in contemporary settings, especially in centers where the retrograde/hybrid approach has not yet been adopted. Pre-procedural assessment of the probability of success may support the decision to attempt PCI in a CTO lesion or to refer the patient to a high-volume CTO-PCI center where the hybrid/retrograde approach is performed. This particularly applies to patients with a high CL-score, especially those with chronic renal insufficiency, mild angina (low Canadian Cardiovascular Society classes), or a modest amount of ischemic burden.
A few limitations of this analysis should be taken into account. 1) Given that this score system was developed from data collected in a study carried out in a single center on first-attempt CTO-PCI procedures mainly performed via the antegrade approach with a high proportion of experienced operators, our predictive model cannot be automatically extrapolated to settings with a different level of operator experience, previously attempted procedures, or a different procedural algorithm. 2) Another limitation of our study is the absence of anatomic evaluation by an external core laboratory. 3) Other clinical variables, such as the degree of renal impairment, could have a predictive impact. 4) As the techniques and CTO-speciﬁc medical technologies are evolving, the score will require updating in the future.
In this single-center study, clinical and angiographic parameters predicting CTO-PCI procedural success were identified and included in a scoring model system using a backward stepwise approach. This model allows the identification of 4 subgroup score values corresponding to high, intermediate, low, and very low probability of CTO-PCI success. The increasing score values correlate with low probability of CTO-PCI success ranging from <50% to more than 80%. The CL-score outperformed the J-CTO score in predicting procedural success. The applicability of the score needs to be validated in other centers.
WHAT IS KNOWN? Multiple nonrandomized studies reported the long-term clinical benefits of successful CTO-PCI compared with CTO-PCI failure. However, these results are limited in their application because it is difficult to predict the success of a CTO-PCI.
WHAT IS NEW? We report here the first predictive score of final procedural CTO-PCI success. This model seems to be applicable in centers where the retrograde/hybrid approach has not yet been implemented.
WHAT IS NEXT? Validation of this predictive model of CTO-PCI success in a population of patients treated using a hybrid approach is needed.
Dr. Chevalier is a consultant for Abbott Vascular, Medtronic, and Terumo. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- coronary artery bypass graft surgery
- confidence interval
- clinical and lesion-related score
- chronic total occlusion
- chronic total occlusion for percutaneous coronary intervention
- Japanese chronic total occlusion score
- left anterior descending coronary artery
- myocardial infarction
- odds ratio
- percutaneous coronary intervention
- Received September 24, 2014.
- Revision received April 14, 2015.
- Accepted July 2, 2015.
- American College of Cardiology Foundation
- Rathore S.,
- Matsuo H.,
- Terashima M.,
- et al.
- Prasad A.,
- Rihal C.S.,
- Lennon R.J.,
- Wiste H.J.,
- Singh M.,
- Holmes D.R. Jr..
- Olivari Z.,
- Rubartelli P.,
- Piscione F.,
- et al.,
- TOAST-GISE Investigators
- Suero J.A.,
- Marso S.P.,
- Jones P.G.,
- et al.
- Ramanathan K.,
- Gao M.,
- Nogareda G.J.,
- et al.
- Morino Y.,
- Kimura T.,
- Hayashi Y.,
- et al.
- Nombela-Franco L.,
- Urena M.,
- Jerez-Valero M.,
- et al.
- Teramoto T.,
- Tsuchikane E.,
- Matsuo H.,
- et al.
- Michael T.T.,
- Karmpaliotis D.,
- Brilakis E.S.,
- et al.
- Levine G.N.,
- Bates E.R.,
- Blankenship J.C.,
- et al.
- Wijns W.,
- Kolh P.,
- Danchin N.,
- et al.
- Cohen M.,
- Rentrop K.P.
- Lamas G.A.,
- Flaker G.C.,
- Mitchell G.,
- et al.,
- for the Survival Ventricular Enlargement Investigators
- Tan K.H.,
- Sulke N.,
- Taub N.A.,
- Watts E.,
- Karani S.,
- Sowton E.
- Brilakis E.S.,
- Banerjee S.,
- Karmpaliotis D.,
- et al.
- Thompson C.A.,
- Jayne J.E.,
- Robb J.F.
- Jang W.J.,
- Yang J.H.,
- Choi S.H.,
- et al.