Article Text
PDF
Abstract
Objectives To determine predictors of failure of transradial approach (TRA) in patients with ST elevation myocardial infarction (STEMI) undergoing primary percutaneous coronary intervention (PCI), and develop a novel score specific for this population.
Methods Consecutive patients with STEMI undergoing primary PCI in a tertiary care high-volume radial centre were included. TRA-PCI failure was categorised as primary (primary transfemoral approach (TFA)) or crossover (from TRA to TFA). Multivariate analysis was performed to determine independent predictors of TRA-PCI failure, and an integer risk score was developed. Clinical outcomes up to 1 year were assessed.
Results From January 2006 to January 2011, 2020 patients were studied. Primary TRA-PCI failure occurred in 111 (5%) patients and crossover to TFA in 44 (2.2%) patients. Independent predictors of TRA-PCI failure were: weight ≤65 kg (OR: 3.0; 95% CI 1.9 to 4.8, p<0.0001), physician with ≤5% TFA conversion (OR: 0.45; 95% CI 0.2 to 0.9, p=0.033), and physician with ≥10% conversion to TFA (OR: 2.2; 95% CI 1.2 to 3.7, p=0.005), intra-aortic balloon pump (OR: 2.0; 95% CI 0.9 to 4.3, p=0.066), cardiogenic shock (OR: 2.8; 95% CI 1.4 to 5.6, p=0.0035), endotracheal intubation (OR: 107; 95% CI 42 to 339, p<0.0001), creatinine >133 μmol/L (OR: 3.6; 95% CI 1.9 to 6.8, p<0.0001), age ≥75 (OR: 1.7; 95% CI 1.0 to 2.9, p=0.031), prior PCI (OR: 2.6; 95% CI 1.5 to 4.5, p=0.0009), hypertension (OR: 1.8; 95% CI 1.2 to 2.9, p=0.009). An integer risk score ranging from −1 to 12 was developed, and predicted TRA-PCI failure from 0% to 100% (c-statistic of 0.868; 95% CI 0.866 to 0.869). Mortality at 1 year remained significantly higher after TRA-PCI failure (adjusted OR 2.2; 95% CI 1.2 to 3.9, p=0.011).
Conclusions In a high-volume radial centre, the incidence of TRA-PCI failure is low and can be accurately predicted using a 9-variables risk score. Since outcomes after TRA-PCI failure remained inferior, further effort to maximise the use of radial approach for primary PCI should be investigated.
Statistics from Altmetric.com
Introduction
Primary percutaneous coronary intervention (PCI) in patients with ST elevation myocardial infarction (STEMI), reduces mortality compared with thrombolysis alone,1 and has become the treatment of choice. The use of multiple antithrombotic drugs in patients with acute coronary syndrome (ACS) increases their bleeding risk.2 ,3 Periprocedural bleeding has emerged as an independent predictor of adverse clinical outcomes.4–7 It is thus prudent to pursue all potential strategies to minimise bleeding risk, including choice of access site for PCI. Major bleeding occurs in approximately 5% of patients with ACS and STEMI,8 and a substantial proportion of bleeding occurs at the vascular access site.9 ,10 Several meta-analyses have shown a significant reduction in bleeding and access site complications with the transradial approach (TRA) compared with the transfemoral approach (TFA).11 ,12 Concordant with such data, the RIVAL trial demonstrated a 60% reduction in access site complications with TRA compared with TFA, and, in the STEMI subset of RIVAL, a reduction in 30-day mortality.13 The RIFLE-STEACS trial showed that, in patients undergoing primary/rescue PCI for STEMI, TRA was associated with a significantly lower rate of bleeding, cardiac mortality and shorter hospital stay, compared with TFA.14 Recent observational data from different countries have also suggested that TRA in patients with STEMI is associated with better clinical outcomes.15–17
Nonetheless, there are currently limited data on incidence and predictors of TRA-PCI failure in patients presenting with STEMI.18–20 Accordingly, we aimed to examine the incidence and predictors of TRA-PCI failure in the setting of all-comers with STEMI in a high-volume tertiary TRA centre, and second, to derive and validate a simple clinical scoring system predictive of TRA-PCI failure in patients with STEMI. We also compared early and late clinical outcomes in patients with success or failure of TRA-PCI.
Methods
Study population
During the 5-year study period, 2144 consecutive patients with STEMI were referred to the Quebec Heart-Lung Institute (QHLI) for primary PCI within 12 h of symptoms onset. QHLI uses a radial approach since 1994 and is a tertiary care university hospital which performs >3200 PCIs/year (including >600 primary PCIs). Study inclusion criteria were: chest pain lasting >30 min, ST segment elevation ≥1 mm in ≥2 adjacent ECG leads, new left bundle branch block or true posterior myocardial infarction. Patients with prior coronary artery bypass grafting (CABG) (n=73) and those with normal coronaries or no significant coronary disease (n=51) were excluded from the study. All the remaining 2020 patients comprised the study population.
TRA-PCI technique
During the study period, operators included 6 interventional fellowship trainees and 12 full-time interventional cardiologists (with ≥10 years experience in TRA). In accordance with our institutional protocol, patients underwent assessment of radial artery patency and adequacy of dual hand blood supply using oximetry test prior to procedure. The radial artery catheterisation technique was performed as previously described.21–23
All the patients were pretreated with a loading dose of aspirin and thienopyridine. After TRA sheath insertion, an initial bolus of 70 IU/kg of unfractionated heparin was administered intravenously when used with a glycoprotein IIb/IIIa inhibitor (GPI), or completed to 100 IU/kg when used alone. Use of bivalirudin, or administration of GPI was at the operator's discretion. Selective angiography and TRA-PCI were performed with standard techniques. The radial sheath was removed in the cath lab immediately after completion of the procedure, and haemostasis achieved by application of an adjustable plastic bracelet.21 Study participation required written informed patient's consent before enrolment, and was approved by our institutional ethics committee.
Definitions and outcomes
TRA-PCI failure was divided into two groups: primary TRA-PCI failure when TFA access was chosen as initial access for any clinical reason (no radial puncture attempted) or crossover TRA-PCI failure, due to inability to complete PCI procedure via TRA, requiring access site crossover to TFA.
Standard definition was used for cardiogenic shock.24 Angiographic success was defined as successful completion of PCI with <20% residual stenosis, with Thrombolysis In Myocardial Infarction (TIMI) grade 3 flow in the treated vessel.
Non-CABG related bleeding was defined according to major and minor TIMI classifications.25 ,26 Major vascular complications were defined as large (>10 cm) haematoma (≥grade 2 EASY haematoma scale), retroperitoneal bleeding or access site complication requiring surgical intervention.
Clinical follow-up information was obtained from the referring physician or by telephone contact with the patient. Information on 1-year vital status of patients lost to follow-up was collected from the Quebec death registry ‘Directeur de l’Etat Civil’ service. Vital status was completed in 100% of patients.
The primary end point of our study was incidence and predictors of TRA-PCI failure, and secondary outcomes were incidence of TIMI major and minor bleeding, and 30-day and 1-year outcomes.
Statistical analysis
Continuous variables are summarised as mean±SD or median (IQR), and categorical variables as absolute numbers and percentages. Statistical comparisons were performed using Fisher's exact test or Pearson test for categorical variables, while Student's t test, Wilcoxon rank-sum test, Kruskal–Wallis test and analysis of variance were used for continuous variables.
To determine independent predictors of TRA-PCI failure, all baseline demographic and clinical parameters in table 1, in addition to the operator's factor, were prescreened by univariate logistic regression analysis. Subsequently, significant univariate predictors of TRA-PCI failure were subjected to multivariable regression analysis with stepwise forward and backward selections, where candidate predictors were entered into the model at p<0.10 and retained at p<0.10. A probability value of <0.05 was considered statistically significant.
Baseline characteristics
Development and validation of risk score
Independent predictors of TRA-PCI failure formed the basis of the risk score, and we attributed a weight to each variable according to the regression coefficient. Each integer amount is a rounding of the figure obtained from the logistic model. The area under the receiver operating characteristics (ROC) curve for the integer score was determined by calculating the c-statistic in logistic regression analyses with TRA-PCI as the dependent variable and integer score as the independent variable. Model discrimination was assessed using the area under the ROC curve or the c-statistic. The internal validity of the risk score model was examined using the bootstrap resampling method. For this purpose, 1000 subgroups of patients from a derivation data set were sampled, and the model was refitted within each sample. The model's goodness-of-fit was evaluated using the Hosmer–Lemeshow method. All calculations and statistical analyses were performed using JMP V.9.0 and SAS V.9.3 (SAS Institute, Cary, North Carolina, USA).
Results
From January 2006 to January 2011, a total of 2020 patients underwent primary PCI for STEMI. These were divided into three groups based on the success or failure of TRA for PCI. The procedure was started via TRA in 1909 (95%) patients, and successfully completed via initial attempted TRA in 1865 (92%) patients (TRA-PCI success). In 111 patients (5%), TFA was chosen as primary access by the operator for clinical reasons (primary TRA-PCI failure). Crossover to TFA was required in 44 (2.2%) patients to complete the procedure (crossover TRA-PCI failure). Crossover from radial to femoral access varied little over time: 1.04% in 2006, 1.62% in 2007, 3.36% in 2008, 2.58% in 2009 and 1.98% in 2010. Baseline characteristics are shown in table 1.
Patients in whom primary access was TFA (primary TRA-PCI failure) or in whom crossover to TFA was required (crossover TRA-PCI failure) were more likely to be older (p<0.0001) and be female (p=0.0002), and had more frequently a body weight ≤65 kg (p<0.0001), compared with the TRA-PCI success group. In addition, both failure groups had a higher incidence of hypertension (p<0.0001) and higher baseline creatinine (p<0.001), compared with the TRA-PCI success group.
Furthermore, there was a higher incidence of cardiogenic shock at presentation in both primary and crossover TRA-PCI failure groups compared with the TRA-PCI success group (p<0.0001); they were more likely to receive an intra-aortic balloon pump (IABP) (p<0.0001) and more likely to be intubated (p<0.0001).
At a procedural level, the infarct-related artery was more likely to be the left anterior descending artery in the primary and crossover TRA-PCI failure groups (p=0.004) compared with the TRA-PCI success group (table 2). Patients in both the TRA-PCI failure groups were less likely to receive a GPI (p<0.0001) and more likely to receive bivalirudin (p=0.002). There was no significant difference in the initial TIMI 0 flow, thrombus burden or lesion characteristics between the three groups.
Procedural characteristics
Both median door-to-balloon (p=0.031) and procedural times (p<0.0001) were longer in primary and crossover TRA-PCI failure groups compared with TRA-PCI success. Door-to-balloon time remained stable over the 5-year period (33 (25–56) min in 2006 and 36 (26–63) min in 2010). Final TIMI 3 flow was achieved less often in the primary TRA-PCI failure group compared with the TRA-PCI success and crossover failure groups (p=0.0047). Angiographic success was achieved in 97% of patients, but only in 91% in the primary TRA-PCI failure group compared with 98% in the TRA-PCI success and crossover failure groups (p<0.0001). From 1865 cases performed by TRA, 1858 were accessed through the right radial (n=1813) artery or the left radial (n=45) artery. In seven cases, the contralateral radial artery was used after initial puncture failure.
Post-PCI outcomes
Inhospital, 30-day and 1-year outcomes are outlined in table 3. Infarct size was significantly higher in both primary and crossover TRA-PCI failure groups compared with the TRA-PCI success group, and significantly more patients after TRA-PCI failure had a reduced left ventricular ejection fraction. Both primary and crossover TRA-PCI failure groups had a higher 30-day mortality and 1 year (p<0.0001) mortality, compared with the TRA-PCI success group (table 3 and figure 1). There was a significantly higher incidence of non-CABG related TIMI major bleeding (p<0.0001) in the primary and crossover TRA-PCI failure groups (table 3). Furthermore, patients in primary and crossover TRA-PCI failure groups experienced a significantly higher incidence of major vascular complications (6.5% and 2.3% vs 0.1%, respectively, p<0.0001). In the primary TRA-PCI failure group, there were four large femoral haematomas, a retroperitoneal bleed and two ischaemic lower limb complications, both requiring surgical intervention.
Clinical outcomes
Kaplan–Meier curve for survival according to access site. TFA-PCI, transfemoral approach- percutaneous coronary intervention; TRA-PCI, transradial approach-percutaneous coronary intervention.
The multivariate analysis identified the following independent predictors of 1-year mortality: age ≥75 years (OR: 2.8; 95% CI 1.5 to 5.3, p=0.0009), shock (OR: 6.7; 95% CI 3.7 to 12.4, p<0.0001), IABP (OR: 2.6; 95% CI 1.3 to 5.0, p=0.0059), renal insufficiency (glomerular filtration rate <60) (OR: 3.8; 95% CI 2.1 to 7.0, p <0.0001), GPI (OR: 0.5; 95% CI 0.3 to 0.8, p=0.0018), TIMI 0/1 final (OR: 4.0; 95% CI 1.3 to 11.8, p=0.014), left ventricular ejection fraction <40% (OR: 3.6; 95% CI 2.2 to 5.9, p<0.0001) and femoral access (after TRA-PCI failure) (OR: 2.2; 95% CI 1.2 to 3.9, p=0.011).
Operator factor
Despite the fact that primary PCIs were performed by experienced radial operators±interventional fellows, there was large variability between operators. The use of TFA after TRA-PCI failure varied from 2.6% to 15.2% among the 12 operators.
Mechanisms and predictors of TRA-PCI failure
The two main reasons for primary TRA-PCI failure (n=111) were presentation with cardiogenic shock with or without intubation in 40 patients (36%) and resuscitated sudden cardiac death with intubation in 33 (30%) (table 4). The most common reason for crossover TRA-PCI (n=44) was inadequate radial arterial puncture in 28 patients (64%).
Causes of primary and crossover TRA-PCI failure
The multivariable model selected six baseline, one laboratory and two treatment-related variables as independent predictors of TRA-PCI failure: age ≥75 years, weight ≤65 kg, creatinine >133 μmol/L, hypertension, prior PCI, cardiogenic shock at presentation, endotracheal intubation at or prior to presentation, need for IABP, physician with ≤5% conversion to TFA and physician with ≥10% conversion to TFA (table 5). The model c-statistic value was 0.88.
Multivariable predictors of TRA-PCI failure in patients with STEMI
Risk score of TRA-PCI failure
The integer risk score derived from this model appears in table 5. Except for operator and intubation, all other variables were assigned a score of 1. The ‘operator’ factor was divided into three groups based on the total percentage of femoral cases (primary and crossover) he performed in relation to his total number of cases of primary PCI performed during the study period: ≤5% TFA cases (score −1), 5.1–9.9% (score 0) and ≥10% TFA cases (score +1). Intubation was assigned a score of 5. The total score consists of the summation of the nine variables representing the patient's individual risk of TRA-PCI failure.
Figure 2 shows the risk score with the observed versus predicted probabilities of TRA-PCI failure in primary PCI. The model provided good calibration as indicated by the non-significant Hosmer–Lemeshow goodness-of-fit test (p=0.81). Internal validation with the bootstrap method provided a c-statistic of 0.868 (95% CI 0.866 to 0.869).
Observed versus predicted probability of TRA-PCI failure in STEMI according to risk score. WRIST-CASE risk score consisting of nine independent predictors of TRA-PCI failure in STEMI. The vertical blue bars represent the observed per cent of patients who failed TRA-PCI and the red line represents the predicted probability value of failure for TRA-PCI. TRA-PCI, transradial approach-percutaneous coronary intervention; STEMI, ST elevation myocardial infarction; WRIST-CASE, Weight, Radial proficiency, IABP (intra-aortic balloon pump), Shock, Tube, Creatinine, Age, prior Stent, and Elevated blood pressure.
Based on the nine independent predictors, our novel score was assigned the acronym WRIST-CASE to aid easy memorisation in clinical use: Weight, Radial proficiency, IABP, Shock, Tube, Creatinine, Age, prior Stent, and Elevated blood pressure.
Discussion
The main findings from the present study are: (1) The incidence of overall TRA-PCI failure during primary PCI in a default radial centre setting is very low; (2) The need for access site crossover from TRA to TFA was only 2%; (3) A novel simplified risk score for TRA-PCI failure has been developed, and accurately predicts TRA-PCI failure in all patients with STEMI; (4) Patients undergoing primary PCI for STEMI via TFA after TRA-PCI failure had a significantly higher incidence of bleeding, major access site vascular complications, 30-day mortality and 1-year mortality, compared with TRA-PCI.
Periprocedural major bleeding has emerged as a powerful and independent predictor of mortality and ischaemic outcomes after PCI.4–7 Bleeding risk is contributed to by patient-related factors,8 as well as procedure-related factors. The latter includes indication for intervention, complexity of procedure, choice of access site and, procedural and periprocedural pharmacotherapy.10 ,13 ,14
Due to the need for potent antithrombotic therapy in the setting of ACS and STEMI, patients undergoing PCI in these subsets would arguably benefit more from access site optimisation to reduce their risk of bleeding. Such a benefit was noted in the seminal, multicentre the RadIal Vs femorAL access for coronary intervention (RIVAL) trial, which demonstrated that TRA significantly reduced vascular access complications compared with TFA, with similar PCI success rates.13 In the Radial Versus Femoral Randomized Investigation in ST-Elevation Acute Coronary Syndrome (RIFLE-STEACS) trial which included patients undergoing primary/rescue PCI, TRA was associated with significantly lower rates of cardiac mortality, bleeding and shorter hospital stay compared with TFA.14
Although early studies demonstrated that the risk of access failure and need for crossover was higher with TRA versus TFA (7.3% vs 2%, p<0.01),11 recent registries have shown relatively lower need for crossover with improved materials and technical expertise.18 ,19 ,27 Crossover from TRA to TFA in primary PCI for STEMI in recent studies has been reported to range from 3.7% to 9.6%.13 ,14 ,28 ,29 However, current real world data on predictors of TRA-PCI failure in the STEMI population are very scarce, and limited to selected cohorts.19 ,20
Some retrospective studies reporting on TRA failure rate and predictors were limited by heterogeneity of patients, in addition to overall TRA volume,18 while others were selective, and excluded patients presenting with cardiogenic shock complicating STEMI.19 ,20 ,29 Vink et al19 recently examined the feasibility of routine TRA in primary PCI (excluding cardiogenic shock) in a single centre where TRA was the default for >10 years. TRA was the primary access in 96.1% of procedures, and site crossover was required in 3.8%, with improving trend over time. In the RIVAL trial, crossover from TRA to TFA was required in 5.3% of patients with STEMI.13 When radial PCI centres were categorised into tertiles according to TRA-PCI volume, crossover rates were 8.0% for the lowest, 9.7% in the intermediate and 4.4% in the highest tertile.13
The present study reports on the incidence and predictors of TRA-PCI failure among all patients with STEMI (including cardiogenic shock), and introduces a novel risk score predictive of such failure in this high-risk population, using readily available criteria. In accordance with previous findings,18–20 we report that reasons for access site crossover fell into three broad categories: inadequate or failed radial artery puncture (which remains the dominant factor), failure to advance or manipulate guidewires and catheters due to arterial spasm or tortuosity, and inadequate guide catheter support.
Our results may help operators to risk stratify patients with STEMI with regard to access site selection and procedure pharmacotherapy, recognising the higher incidence of bleeding and vascular complications associated with TFA, as we have shown in this study. Although we identified age, creatinine and cardiogenic shock as independent predictors of TRA-PCI failure, these criteria, particularly when combined, represent a high risk of periprocedural bleeding.4 ,8 It is therefore, in our opinion, prudent to consider TRA as first approach, particularly in these settings. However, choice of TRA must also be balanced against the individual's risk of failure, as well as the imperative of timely reperfusion in STEMI.
Our results strongly show that outcomes of patients undergoing primary PCI after failed TRA-PCI are worse. Hence, this suggests that team approach and all efforts should be put in place to maximise the use of TRA in such patients. The use of ultrasound-guided puncture may potentially help operators gain radial access in case of haemodynamic compromise. Conversely, if TRA remains impossible in such circumstances, particular attention should be paid to technical aspects in obtaining the safest possible femoral puncture (sheath sizes and femoral puncture technique), as well as pharmacological therapies and haemostasis management.
Limitations
Due to the observational nature of our study, residual confounding factors may still influence results. Nonetheless, a major strength of the current risk score is that it was derived from a ‘real world’ unselected population, and that our risk model performed well in derivation set, calibration and internal validation, suggesting it is robust and generalisable. However, further external validation of our score is warranted in other STEMI registries with unselected cohorts in high-volume radial centres.
Key messages
What is already known on this subject?
Transradial approach offers significant advantages and benefits over standard femoral approach. However, there are more failures and crossovers with radial approach.
What might this study add?
Our study identified nine independent predictors of radial access failure in patients referred for primary percutaneous coronary intervention (PCI). It also describes a simple scoring system that is a practical tool for clinicians to enable immediate evaluation of patient suitability for radial access. This integer risk score ranges from −1 to 12, predicts radial access failure from 0% to 100%, and was assigned the acronym WRIST-CASE: Weight, Radial proficiency, IABP (intra-aortic balloon pump), Shock, Tube, Creatinine, Age, prior Stent, and Elevated blood pressure.
How might this impact on clinical practice?
Our novel risk score might help practising interventional cardiologists to adjust their judgement in selecting the initial access site for primary PCIs. Yet, our results clearly show that radial access should be privileged in most cases as patients with radial access failure bear worse clinical prognosis.
Acknowledgments
The authors thank Mr Serge Simard for his statistical advice and Dr Sunil Rao who helped them with the acronym for the risk score.
References
Footnotes
Collaborators Interventional cardiologists at the Quebec Heart-Lung Institute.
Contributors EA, JM, OC, OFB designed the study. EA, GP, ABP, M-PB, MR, J-PD, UD maintained the database, performed data extraction and did some analyses. EA, JM, OC and OFB completed the data analyses and developed the risk score. All authors have reviewed the data, suggested modifications to text and/or analyses. All authors reviewed and approved the final version of the manuscript as submitted in its present form.
Funding EA is supported by the Laval University International Chair on Interventional Cardiology and Transradial Approach. This study was conducted with internal funds and did not receive external sources of funding.
Competing interests None declared.
Patient consent Obtained.
Ethics approval Quebec Heart-Lung Institute IRB.
Provenance and peer review Not commissioned; externally peer reviewed.