Predictive factors of difficult implantation procedure in cardiac resynchronization therapy

Abstract

Aims

The usefulness of cardiac resynchronization therapy (CRT) in patients with congestive heart failure is offset by its long, user-dependent, and technical procedure. No studies have been published regarding factors related to CRT implantation procedure duration and X-ray exposure. Additionally, only a few studies have investigated the predictive factors of primary left ventricular (LV) lead implant failure. The aim of this prospective study was two-fold: (i) to evaluate the prevalence and predictive factors of prolonged CRT implantation procedure and (ii) to identify the predictive factors of primary LV lead implantation failure.

Methods and results

Between November 2008 and September 2009, 128 consecutive patients underwent CRT implantation; of these, 22 patients (17.2%) were excluded because of CRT generator replacement. Population characteristics were a mean age of 69 ± 10 years, 28.3% female, New York Heart Association class 3.2 ± 0.3, LV ejection fraction (LVEF; 29 ± 6%), and QRS width 146 ± 23 ms. Cardiac resynchronization therapy implantation was attempted in 106 patients, and first LV lead implantation was obtained in 96 of 106 patients (90.5% primary success). Ten primary implantations failed (9.5%), due to unsuccessful LV lead implants. A second procedure was successfully attempted in six patients with a second more experienced operator (5.7%). Among the remaining four patients, one patient required a surgical epicardial LV lead implantation, and the implantation was not reattempted in the other three patients. The overall success rate of CRT system implantation was 96.2% (102 of 106 patients). Procedure parameters were as follows: LV threshold (1.4 ± 0.9 V); LV wave amplitude (15 ± 8 mV); LV impedance (874 ± 215 ohm); median procedure time (skin to skin), 55 min (45–80); and median of procedure fluoroscopy time, 11 min (6.2–29). In 24 patients (22.6%), difficult procedures requiring ≥85 min of implantation duration occurred. By univariate analysis, predictive factors of difficult implantation were LV ejection fraction (25.6 ± 6 vs. 30.2 ± 8%; P = 0.02), LV end-diastolic diameter (72.4 ± 11 vs. 66 ± 11 mm; P = 0.01), LV end-systolic diameter (LVESD; 62 ± 12 vs. 56 ± 12 mm, P = 0.04), and the operator's experience (very experienced operator vs. less experienced operator, P = 0.006). By multivariate analysis, only primary LV lead implantation failure, LVESD, and operator's experience were independently associated with difficult procedures. In this patient subset with primary LV lead implant failure ( n = 10), the only independent predictive factor was the LV end-systolic volume ( P = 0.03).

Conclusion

In this study, the rate of difficult CRT device implantation procedures approached 25%. Both the degree of LV dysfunction and the operator's experience were independent predictors of surgical difficulties. Left ventricular end-systolic volume was the only independent predictor of primary LV lead implant failure.

Introduction

The purpose of cardiac resynchronization therapy (CRT) is to restore ventricular relaxation and contraction sequences by simultaneously pacing both ventricles. ^1–6 Several studies reported positive long-term effects in terms of symptoms, exercise tolerance, well-being, and CRT prognosis. ^3–10 Therefore, the increase in indications for CRT has logically resulted in an increased number of device implantations. However, potential clinical benefits of CRT depend on the proper implantation of the left ventricular (LV) lead, which remains technically challenging. ^{11 , 12} Cardiac resynchronization therapy needs transvenous pacing of the LV lead via the coronary sinus (CS) branches. Procedures may be long and difficult, due to CS access, CS vein anatomy, need for CS angiography, and lack of operator's experience. ^{6 , 10–12} Although the availability of multiple lead designs and LHDS options may improve the procedure's success, problems still remain, with failure occurring in 5–15% of cases, particularly for a primary LV lead implant. ^{6 , 12} Moreover, despite the use of systematic antibioprophylaxis, ¹³ a 124% increase in the rate of device-related infections (DRI) was reported in a published series, ¹⁴ and most investigators agree that infections occurring within 1 year of implantation are usually due to contamination during prolonged surgery. ¹⁵ More recently, our group has found that the prevalence of CRT DRI was 4.3% at 2.6 years and that both procedure duration and surgical re-intervention were the independent predictors of DRI. ¹⁶ Owing to the lack of published studies, more data are required regarding the effects of prolonged CRT implantation procedure duration and X-ray exposure. Additionally, only few published studies have investigated the predictive factors of primary LV lead implant failure. ^17–19 These studies, however, provided debatable results, due to several methodological issues: the use of retrospective study designs, number of variables tested, type of parameters studied, and absence of right anatomical evaluation. ^17–19

Consequently, the aim of this prospective study was two-fold: (i) to evaluate both the prevalence and the predictive factors of difficult CRT implantation procedures and (ii) to identify the predictive factors of primary LV lead implant failure.

Methods

Patient selection

Between November 2008 and September 2009, 128 consecutive patients underwent CRT implantation in our centre. Among these, 22 patients (16.4%) were referred for CRT generator replacement and excluded from the study. All patients included in the trial provided written informed consent. Inclusion criteria required that all of the following conditions be met: (i) New York Heart Association (NYHA) functional class III or IV heart failure (HF); (ii) QRS width ≥120 ms with a left bundle branch block pattern or QRS ≥ 180 ms for paced patients [measured on three or more surface electrocardiographic (ECG) leads]; (iii) chronic LV systolic dysfunction defined by an LV ejection fraction (LVEF) ≤35% and LV end-diastolic diameter (LVEDD ≥60 mm) on echocardiography; (iv) optimal medical treatment for HF including angiotensin-converting enzyme inhibitors or AT1 receptor antagonists, diuretics, β-receptor blockers, and spironolactone; and (v) patients requiring atrioventricular (AV) node ablation for atrial fibrillation (AF) with moderate LV dysfunction (LVEF ≤ 40%) with or without QRS ≥ 120 ms in accordance with the PAVE study results. ²⁰ The exclusion criteria were: (i) hypertrophic or restrictive cardiomyopathy; (ii) suspected acute myocarditis; (iii) correctable valvulopathy; (iv) acute coronary syndrome; (v) recent coronary revascularization (previous 3 months) or planned revascularization; (vi) treatment-resistant hypertension; (vii) severe obstructive lung disease; and (viii) reduced life expectancy not associated with cardiovascular disease.

Study design

Between 5 and 30 days prior to implantation, the patients underwent a clinical examination, 12-lead ECG, transthoracic echocardiography, and Doppler evaluation.

Echocardiographic measurements

Transthoracic Doppler echocardiography was performed at least 24 h before the procedure by an observer (L.B. or L.K.) who was blinded to the patient's status. Ultrasound studies were performed with a Vivid 7 imaging system using a 2.5 MHz transducer (second harmonic activation). M-mode and bi-dimensional measurements were made according to the recommendations of the American Society of Echocardiography. Right atrial anatomical two-dimensional echocardiographic measurements were performed as well as the septal isthmus length measurement. Right and left atrial surface were measured by planimetry (in cm ² ). All two-dimensional echocardiographic and Doppler-derived pulmonary pressure measurements were averaged over five cardiac cycles. Tricuspid regurgitation was evaluated and considered significant when the grade was 3 or more.

Implant procedure

The method of implantation was published elsewhere. ²¹ Patients were systematically shaved and had an antiseptic shower with a povidone iodine 10% aqueous solution on the night before the operation. Antisepsis was performed immediately before surgery, and the skin was painted with two solutions: an aqueous povidone iodine 10% solution followed by a povidone iodine 7.5% solution. Implantations were performed by the same two operators (A.C. and C.R.B.) under local anaesthesia and conscious sedation. All patients received the same antibiotic prophylaxis. ¹³ Venous administration was performed 30 min prior to the incision using a single dose of cefazolin (1.5 g). No local antibiotic pocket wash was used. Patients were examined daily until hospital discharge. All devices were implanted subcutaneously. Total implant procedure duration (skin to skin), total X-ray time, X-ray time for LV lead implantation, and pacemaker lead-related data (threshold, impedance, and amplitude signal) were recorded, as were complications during the procedure and postoperative period. In all patients receiving an implantable cardioverter-defibrillator (ICD) CRT, a defibrillation test was performed. All patients received a biventricular DDD-R device in the presence of sinus rhythm (SR), and a biventricular VVI-R device in the presence of permanent AF.

Endpoints

The primary aim was to evaluate both the prevalence and the predictive factors of difficult CRT implantation associated with long procedure duration. The definition of long procedure duration was based on the cut-off value associated with a higher risk of CRT DRI in our previous study. ¹⁶ A median of procedure duration ≥85 min was found to be independently correlated with a higher risk of CRT DRI. ¹⁶ The secondary aim was the identification of predictive factors for primary LV lead implant failure.

The following factors were tested: age, gender, body mass index (BMI), cardiovascular risk factors, chronic kidney failure, device procedure category (CRT-P, CRT-D, or upgrading), aetiology of cardiomyopathy, NHYA class, LVEF, LV volumes and dimensions, the presence of AF, systolic arterial pulmonary pressure level, right atrial dimensions, tricuspid regurgitation graduation, right ventricular dimensions, the septal isthmus (length between the CS ostium and the septal tricuspid leaflet, reflecting the CS insertion level), lead manufacturer, and lead position (posterior, posterolateral, lateral, and anterolateral). The operator's experience in a high-volume centre was also examined, with CRT implantation experience of at least 10 years (ADC) vs. experience <10 years and more than 5 years (CRB).

Re-intervention

Surgical re-intervention was defined as a surgical procedure required for a non-infectious or infectious complication of the implant, and was further stratified as occurring early (≤5 days) or late (>5 days). The overall success of implantation was determined after the second LV lead implantation attempt.

Follow-up

Prospective data includes: (i) patients' demographic and clinical characteristics; (ii) echocardiographic measurements; (iii) CRT characteristics (pacemaker or defibrillator implantation, and primo-implants or system upgrade of pulse generators or leads); (iv) type of implanted device and number of leads; (v) procedure duration (skin to skin) and total X-ray exposure as wells as LV lead procedure duration estimated by X-ray exposure; and (vi) occurrence of complications requiring re-intervention.

Patients' follow-ups were conducted by three cardiologists (A.C., L.B., and C.R.-B.). Electrocardiograms, device controls, and chest X-rays were performed the day after the procedure (Day 1) and on Day 5, just prior to discharge. The patients' scars and sutures, and chest X-rays, were examined in the outpatient clinic on Day 10 and at 3 months; a physical examination and device interrogation were also carried out on these occasions.

Statistical analysis

All clinical variables were assessed at the time of the device implantation. The echo parameters were evaluated during the hospitalization by two operators who were blinded to the patients' status (L.B. and L.K.).

Continuous variables were presented as mean ± SD or median with inter-quartiles as appropriate. Comparisons between the patients' groups were performed on continuous variables using, respectively, unpaired Student's t -test or the Mann–Whitney test. Categorical variables were compared using χ ² tests or Fisher's exact test as appropriate.

Univariate logistic regression models were fitted to study the relationship between each covariate and the risk of a prolonged procedure time on the one side, and the risk of primary LV lead implantation failure on the other side. Backward elimination was used, removing the least significant variables at each step to elaborate multivariate models. Risks were presented as odds ratio and their 95% confident interval (CI).

A value of P < 0.05 was considered statistically significant. All analyses were performed using StatView ^® 5.0 (StatView IV, Abacus Concept, Berkeley, CA, USA).

Results

Baseline population characteristics

Baseline clinical data is summarized in Table  1 . The CRT implantation was attempted in 106 patients, and first LV lead implantation was achieved in 96 of 106 patients (90.5% primary success). Ten primary implantations failed (9.5%) due to unsuccessful LV lead implants: four patients failed due to the inability to catheterize the CS and six patients due to the inability to find tributary vein. A second procedure was attempted with a more experienced operator (A.C.) in six patients (5.7%), and the success was achieved in all six patients. Among the remaining four patients, one required a surgical epicardial LV lead implantation, and implantation was not reattempted in the remaining three patients, due to the investigators' decision not to proceed ( n = 2) or patient decision ( n = 1). The overall success rate of CRT system implantation was 102 of 106 (96.2%) patients. Of the 106 patients, 77 were in SR, 29 were in AF, and 12 had a previous device [pacemaker (PM) in 2 patients and ICD in 10 patients]. The medical treatment was as follows: 97 patients were treated with angiotensin-converting enzyme inhibitors or AT2 receptor antagonists (91.5%), 106 patients with diuretics (100%), 80 patients with β-receptor blocker (75.5%), 52 patients with aldosterone antagonists (49.1%), 48 patients with vitamin K antagonists (45.3%), and 58 patients with aspirin (54.7%).

Implantation and technical aspects

Procedure parameters were as follows: LV threshold (1.4 ± 0.9 V); LV wave amplitude (15 ± 8 mV), LV impedance (874 ± 215 ohm), procedure time (skin to skin) (median of 55 min), and overall procedure fluoroscopy time (median of 11 min). The direct LV lead was successfully positioned as follows: lateral vein in 65 patients (63.7%), posterolateral vein in 25 patients (24.6%), posterior in 7 patients (6.8%), and anterolateral in 5 patients (4.9%). The anterolateral vein was used in the primary implantation in four patients and only one patient required this target vein during the second procedure. The mean fluoroscopy time for LV implantation was 17 ± 16 min. The 102 devices implanted were CRT PMs in 16 patients (15.7%), CRT ICDs in 75 patients (73.5%), and upgrading to CRT in 11 patients (10.8%) (PM, n = 2; ICD, n = 9). As mentioned, one patient required surgical LV lead implantation with a CRT-D device.

Predictive factors of difficult cardiac resynchronization therapy implantation

In 24 patients (22.6%), difficult procedures requiring ≥88 min of implantation duration occurred. By univariate analysis, predictive factors of difficult implantation were LV ejection fraction (25.6 ± 6 vs. 30.2 ± 8%; P = 0.02), LVEDD (72.4 ± 11 vs. 66 ± 11 mm, P = 0.01), LV end-systolic diameter (LVESD; 62 ± 12 vs. 56 ± 12 mm, P = 0.04), and the operator's experience (very experienced operator vs. less experienced operator, P = 0.006) ( Table  2 ). By multivariate analysis, only primary LV lead implantation failure, LVESD, and operator's experience were independently associated with difficult procedures ( Table  3 ). Right anatomic ventricular parameters were not associated with difficult CRT procedures, neither was left atrial or right atrial dimension (n.s.). Several other variables were tested but did not correlate with CRT implantation difficulties (n.s.): age, presence of AF, NYHA class, cardiovascular risk factors, aetiology of myocardiopathy, QRS width, device type, LV lead brand, inferior vena cava diameter, tricuspid, LV lead definitive position, and BMI.

Predictive factors of successful vs. unsuccessful primary left ventricular lead implantation

A trend towards a correlation was found for three variables: LVEDD ( P = 0.053), LV end-diastolic volume ( P = 0.084), and the operator's experience ( P = 0.09). The operator's experience affected the procedure duration (57.5 ± 20 vs. 68 ± 32 min, P = 0.05) and X-ray exposure (14 ± 13 vs. 22 ± 19 min, P = 0.02), mainly due to the LV placement (12 ± 12 vs. 18 ± 18 min, P = 0.07). By univariate and multivariate analyses, LV end-systolic volume (LVESV) ( P = 0.03) was the only predictive factor of primary LV lead failure ( Tables  4 and 5 ).

Follow-up

During the follow-up period lasting at least 3 months, re-intervention was necessary in 11.3% patients (12 of 106): a secondary LV lead placement was attempted in six patients with a more experienced operator (A.C.) (5.6%), a surgical LV lead was implanted in one patient (0.9%), a local infection was observed in two patients requiring an extraction without any specific tool (1.9%), and three lead dislodgements were observed (2.8%). Phrenic nerve stimulation was resolved with a lower output LV pacing.

Discussion

Major findings

This study found that the rate of difficult procedures in CRT device implantation was close to 25%, mostly due to LV placement. Left ventricular dysfunction severity and operator's experience were independently associated with both prolonged procedure time and X-ray exposure duration. Moreover, LVESV was the only predictor of primary LV lead implant failure. Lastly, despite technical difficulties, CRT devices were successfully implanted via the CS in 96.2% of the entire study population.

Technical aspects and complication related to cardiac resynchronization therapy implantation

The increasing number of indications for CRT has logically led to an increased number of device implantations. However, the potential clinical benefits of CRT depend on the proper implantation of the LV lead, which remains technically challenging. ^{10–12 , 18 , 19 , 22 , 23} It has been suggested that the successful outcome of LV lead placement depends on several factors, such as CS anatomy, presence of the thebesian valve, presence of tributary CS veins, and operator's experience. ^{12 , 18} Despite several hypotheses, few studies have been published in this field to date. ^17–19 Cardiac resynchronization therapy implant procedure duration and X-ray exposure are still too long, and prolonged surgery times are correlated with a high rate of complications, particularly the risk of DRI. ^{9–12 , 18 , 19 , 22 , 23} Mean durations of both CRT procedure and X-ray exposure vary widely in the literature; they were reported to be 143 ± 50 min for CRT procedure and 34 ± 24 min for X-ray exposure. ^{10–12 , 23 , 24} High complication rates ranging from 5 to 18% were reported to be associated with CRT implantations. ^{10–12 , 18 , 19 , 22 , 23} Reported complications included death, CS perforation or dissection, tamponade, inadvertent occurrence of third-degree AV block, ventricular arrhythmias, LV lead dislodgement, high LV pacing threshold, phrenic nerve stimulation, ventricular oversensing, and infections requiring removal of the entire system. Kautzner et al ., ¹⁰ for example, reported 4.3% subintimal CS dissection, 1.1% leak of contrast liquid into the pericardial space (n.s.), 5.4% lead dislodgement or phrenic nerve stimulation, 4.3% third-degree AV block during introducer handling, and 2.2% rate of infection. In the Multicenter Insync Randomized Clinical Evaluation (MIRACLE) study, the authors reported two deaths, 4% coronary dissection, 2% CS perforation requiring pericardiocentesis, 5.7% lead dislodgement, and 1.3% rate of infection after a 6-month follow-up. ⁹ More recently, our group demonstrated that the risk of DRI was 4.3% at 2.6 years, and we found that the risk of CRT DRI was correlated with CRT implant procedure duration. ¹⁶ Therefore, the knowledge of predictors for both procedure duration and X-ray exposure may allow for a higher rate of CRT implant success and lower rate of complications.

Predictive factors of difficult cardiac resynchronization therapy implantation procedure and primary left ventricular lead failure

Considering the technical complexity of CRT system implantation, and the relative fragility of patients suffering from advanced HF, as well as the risk of procedure complications, studies investigating predictive factors of prolonged CRT implant procedure were required. Although three studies have been published in this clinical setting, their conclusions remain debatable or their results incomplete, owing to several technical shortcomings. ^17–19 In a retrospective study involving 212 patients, a mean procedure time of 120 min and a mean X-ray exposure of 35.6 min were reported by Macías et al . ¹⁷ Cardiac resynchronization therapy devices were successfully implanted in 87.7% of patients, and two independent predictors of unsuccessful implantation were found: presence of chronic AF and left atrial dimensions. ¹⁷ The major limitations of these studies were their retrospective design, the operators' learning curve inclusion, and the use of old design leads for guiding catheters. ^17–19 Major results in this field of research were reported by Gras et al . ¹⁸ regarding patients enrolled in the CARE-HF trial. Successful LV lead implant was achieved in 95.6% of cases but the success rate for primary implantation was only 86.7%. This study allowed for a maximum of three attempts, and in order to increase the likelihood of success, referral to more experienced centres was possible, and even encouraged. ¹⁸ The only variable predictive of implant success was related to the experience of the participating medical centres, which is in line with previous observations indicating that success rates depend on operators' experience. This suggests that the likelihood of successful implantation after the first attempt may be significantly increased if the procedure is repeated by a more experienced operator. This should encourage the development of centres of excellence. Lastly, Fatemi et al . ¹⁹ reported a primary LV implant success rate in 90% of cases, but a second procedure was successful in four patients, resulting in an overall success rate of 94%. No predictors of primary LV lead placement were examined. Our study is the first prospective report to look for predictive factors of prolonged CRT procedure. In a previous work, our group demonstrated that prolonged implantation duration increases infection risk, ¹⁶ emphasizing that CRT implantation is not a benign intervention and that experienced physicians as well as material improvements are needed. Our study confirmed the predictive role of the operator's experience, even in high-volume centres, but was also the first to emphasize the role of LV dysfunction severity, which is usually accompanied by LV dilatation during CRT implant procedures. Indeed, advanced cardiomyopathy is usually associated with cavity enlargements, which may severely hamper access to the CS ostium or LV lead stability, due to changes in anatomical landmarks. The same observation was made when analysing primary LV lead implant success rates, with the predictive role of LV dilatation based on LVESV and LVESD. In conclusion, the rate of CRT implant success, as well as the need for re-intervention, procedure duration, and X-ray exposure, may be improved by early selection of patients, prior to severe and definitive LV deterioration.

Clinical implications

Implantation of special pacing leads for LV-based resynchronization therapy for congestive HF is a rather technical procedure, usually associated with a high rate of complications, including primary LV lead failure, prolonged CRT procedure time, prolonged X-ray exposure, LV lead dislodgement, and device-related infection. Prevention of these complications depends on multiple factors, notably the identification of predictors of primary LV lead implant failure and prolonged CRT implant procedure time. In this clinical setting, our study confirms the need for experienced centres, as well as the independent predictive role of the operator's experience. Moreover, our study also demonstrates that the heart dimensions and the severity of the cardiomyopathy are the main factors leading to difficult procedures as well as primary LV lead implant failures. Knight et al . ¹¹ reached similar conclusions in 443 patients requiring CRT device, with long-term unsuccessful results observed in 20 patients (5%). The only predictive factor associated with CRT LV lead implant failure was LVEF. ¹¹ Therefore, in order to avoid unsuccessful LV lead implantation or prolonged procedure duration, patients requiring CRT device implantation may be selected at an earlier stage prior to severe LVEF degradation or LV dilatation. ¹¹ It can be assumed that implantation in less severe patients would be associated with higher rates of success, as was shown in the REVERSE study, publishing the highest implant success rate ever reported (97%), ²⁴ as well as in the MADIT-CRT study (98.4%). ²⁵ On the other side, our patients were already with a relatively high LVEF compared with CRT compatible patients and the LVEF is probably in part a surrogate marker for larger and pathological dimensions found in more severely compromised cardiomyopathic hearts. Moreover, there is still debate of the benefit and specifically cost regarding implant in mildly symptomatic HF patients.

Study limitations

In our study, right atrial dimensions were not associated with prolonged procedure duration or primary LV lead implant failure. On the other hand, right cavity dimensions are known to be difficult to measure, and reproducibility is usually low and might explain these results. ^{26 , 27} Left ventricular ejection fraction is the main variable which is significant as expected but the fact that all the other factors regarding anatomic dimensions are not significant is probably the result of sample size. There is no mention of CS anatomy which may play a part in success, but again with good experience and small sample size, it would not be expected to be prominent. Operators report CS implantation level or shape as a main factor of LV lead instability; consequently, this variable may also affect procedure duration or LV lead implant failure. A study using intracardiac echocardiography or electron beam computed tomography could be complementary to standard fluoroscopic imaging and could be an optimal modality for imaging structures that are highly relevant for the assessment of CS anatomy. ^{27 , 28}

Conclusions

This study found that the rate of difficult procedures in CRT device implantation approached 25%. Both the degree of LV dysfunction and the operator's experience were independent predictors of surgical difficulties. Left ventricular dimensions were the only predictors of primary LV lead implant failure.

Conflict of interest: none declare.

References