Left Ventricular Lead Electrical Delay Is a Predictor of Mortality in Patients With Cardiac Resynchronization Therapy

We retrospectively analyzed data from a prospective database of patients with an implanted biventricular pacemaker or defibrillator at the Regional Hospital Liberec, Czech Republic, between June 2005 and June 2013. CRT was indicated according to current guidelines of the European Society of Cardiology: symptomatic chronic HF despite optimal medical therapy, with LV ejection fraction ≤35% and QRSd≥120 ms.¹³ Patients were included when they had spontaneous atrioventricular conduction, left bundle branch block (LBBB) or intraventricular conduction delay (IVCD) according to the Strauss criteria,¹⁴ and when QLV measurement was available at the index CRT procedure. Several patients (n=14) with a surgically implanted LV lead were also included; these had either an LV lead implanted during previous open heart surgery or a video-thoracoscopically implanted LV lead because of technical failure of the transvenous approach. We excluded patients who underwent early (<1 year after the enrolment) reimplantation of the LV lead or electric reposition of the multipolar lead. Patients with delayed reimplantation of the LV lead (n=10) were included, and original QLV at the index implantation was considered relevant. All patients signed an informed consent with the procedure. The study was performed in accordance with the Declaration of Helsinki guidelines, and the analysis was approved by the local Ethics Committee.

Commercially available CRT systems were implanted at our center by 3 operators. The right ventricular (RV) lead was commonly placed in the midseptum region. The LV lead was inserted transvenously with a preference for lateral cardiac veins, followed by a posterolateral position. Whenever possible, attempts were made to maximize the QLV interval at implant.

Empirical atrioventricular delay of 120 ms and zero V–V delay were programmed at implant and were not routinely optimized with echocardiography. When no clinical improvement was observed during follow-up visits, at least 1 echocardiographically guided optimization was performed.

The QLV interval was measured from the beginning of the native QRS complex to the local LV electrogram (EGM) from the implanted lead. A unipolar EGM from the electrode that was subsequently used for LV stimulation was preferentially used for this measurement. Alternatively, a bipolar EGM was used when a high quality unipolar EGM was not available. When a quadripolar lead was used, only unipolar EGMs were accepted for the measurement. The timing of local activation was considered to be at the fastest deflection (dV/dt minimum) of the unipolar EGM. For bipolar EGMs, local activation was annotated at the largest bipolar deflection crossing the isoelectric line or the first sharp spike of a signal if the crossing was not noticeable in case of more fragmented EGMs.¹⁵ The local EGM was displayed simultaneously with a surface 12-lead ECG at a sweep speed of 200 mm/s on the electrophysiological recording system (Biotronik EP Control, Germany or Siemens Axiom Sensis XP, Germany). The QRSd and QLV interval were measured using electronic calipers, and the QLV ratio was assessed (Figure 1). All these measurements were performed prospectively at the time of implant.

All patients were followed in the local outpatient department every 6 months. During office visits, the results of clinical examination, standard ECG, CRT device settings, medical treatment, and echocardiographic findings were recorded prospectively into a dedicated database. Clinical outcome data were collected from the medical records and by contacting primary care physicians. Patients were classified as New York Heart Association (NYHA) responders when NYHA class decreased by ≥1 grade. Echocardiographic response was defined as a reduction of the LV end-systolic diameter ≥10%. When the proportion of ventricular pacing in patients with atrial fibrillation was <90% according to the device memory despite the rate control medical therapy, radiofrequency atrioventricular junction ablation was performed.

We defined 4 study end points for the follow-up: first HF hospitalization, HF mortality, cardiac mortality, and all-cause mortality. HF hospitalization was defined as an inpatient admission with overnight stay because of signs or symptoms of HF, including shortness of breath, peripheral edema, or congestion on chest radiograph, with subsequent improvement of these signs and symptoms with medical therapy. The cause of death was assessed by the consensus of 2 physicians, who were blinded to QLV interval. This was done by careful review of clinical, death, and necropsy reports and CRT device memory when available. Any sudden death of uncertain cause was considered sudden cardiac death. All the data were collected throughout the follow-up period in a dedicated database until the final lock. Patients with insufficient clinical data were considered as lost to follow-up.

The follow-up database was locked in May 2014. Continuous variables were expressed as a mean±SD or median (25th–75th percentiles, ie, interquartile range [IQR]) and compared by 2-tailed t test for independent samples or Mann–Whitney U test for clearly non-normally distributed data. Categorical variables were expressed as percentages and compared by χ² test. Based on our previous experience from the short-term CRT response study,⁹ the QLV ratio cutoff value of 0.70 (approximately lower tertile boundary) was considered a reasonable risk stratifier. Occurrence of study end points in empirical categories by QLV ratio were expressed as percentages with 95% confidence intervals (modified Wald method). The effect of selected baseline factors on QLV ratio was analyzed by main effects ANOVA with full-factorial scheme of interactions. Associations of baseline variables with all study end points were investigated by Cox proportional hazards regression analysis. Likewise QLV ratio, other continuous and nonbinary categorical variables were dichotomized by their high-risk tertile boundaries or other clinically meaningful/suitable values. All factors that were univariably associated (P<0.10) with at least 1 study end point were entered into the multivariable Cox regression models and investigated by stepwise-forward method. Schoenfeld residuals were used to test for proportionality assumption. Kaplan–Meier curves were used to display cumulative event-free survival. Multivariable models were used to demonstrate the hypothetical benefit from a better LV lead position. Life months gained were computed by comparing the survival curves in high-risk subgroup (QLV ratio ≤0.7) with those when QLV ratio was fixed to low-risk value (>0.7). A P value ≤0.05 was considered significant. All analyses were performed using the STATISTICA Version 12 software (Statsoft Inc.).

During the mean follow-up period of 3.3±1.9 years (range, 0.5–8.8 years; median, 3.1 years; IQR, 1.9–4.7 years), a total of 83 (25.2%) patients died from cardiac (n=49, 14.9%) or noncardiac (n=34, 10.3%) causes. The majority (82%) of cardiac deaths occurred because of HF (n=40, 12.2%). Other cardiac deaths included sudden death (n=7, 2.1%), myocardial infarction (n=1), and aortic stenosis (n=1). Noncardiac deaths occurred because of malignancy (n=16), infection (n=12), aortic aneurysm rupture (n=2), aortic bifurcation thrombosis (n=1), stroke (n=1), myasthenia gravis (n=1), and suicide (n=1). Event rates for all-cause mortality, cardiac mortality, noncardiac mortality, HF mortality, and sudden death were 25.2%, 14.9%, 10.3%, 12.2%, and 2.1%, respectively. Cumulative survival by the mode of death during 7-year follow-up is shown in Figure 2. A total of 83 (25.2%) patients were hospitalized for HF. All but one patient with HF death had prior hospitalization for HF.

Differences between subgroups of patients defined by clinical outcome are shown in Table 2. For all study end points, higher NYHA class, lower LVEF, and lower QLV ratio at implant were consistently associated with adverse outcomes. Survivors were more likely to have received an implantable cardioverter defibrillator (ICD); however, cardiac mortality was comparable between ICD and non-ICD groups.

In order not to miss clinically relevant QLV ratio dichotomies, we investigated the occurrence of study end points in 4 subgroups defined by QLV ratio boundaries of 0.60 (≈15% percentile), 0.70 (approximately lower tertile), and 0.80 (approximately median). Visual analysis of the corresponding data (Figure 3) suggested again an optimum QLV ratio cut-off of 0.70 for predicting study end points, and thus this value was used for further analysis.

There were 97 patients (30%) with a QLV ratio ≤0.70 (approximately lower tertile boundary). The differences between patients with a QLV ratio ≤0.70 and a QLV ratio >0.70 are shown in Table 1. Compared with the rest of the population, patients with a QLV ratio ≤0.70 had significantly worse event-free survival for all study end points (Table 3; Figure 4). With the QLV ratio dichotomy of ≤0.70, survival curves separated early after CRT implantation for HF hospitalization. This occurred no earlier than after 2 years for mortality end points.

All continuous or ordinal baseline factors were dichotomized as follows: age>72 years (n=113; 34%), LVEF<25% (n=111; 34%), QRSd ≥170ms (n=107; 33%) or QRSd ≤150ms (n=99; 30%), biventricular pacing <98% (n=95; 29%), mitral regurgitation grade ≥2 (n=141; 45%), NYHA class=4 (n=59; 18%), serum creatinine >107 µmol/L (n=109; 33%), LV end-diastolic diameter >68 mm (n=108; 33%), and LV end-systolic diameter >60 mm (n=82; 25%). The results of both univariable and multivariable Cox regression analysis are shown in Tables 4 and 5 only for those factors that were univariably associated (P<0.10) with at least 1 study end point. All factors conformed to an assumption of proportionality. A QLV ratio ≤0.70 was the only factor independently associated with all study end points. Other factors (age, coronary artery disease, NYHA class, QRSd, percentage of biventricular pacing, serum creatinine, and LVEF) were not (or not consistent) predictors of study end points either in univariable or multivariable analysis. Sudden death occurred in 7 patients (including 5 patients with CRT-P). The only predictor of sudden death was the absence of ICD (relative risk, 5.5; 95% confidence interval, 1.0–28.9; P=0.046).

After adjustment for other clinical confounders, it was demonstrated that better LV lead position alone (>0.70 versus ≤0.70), the only modifiable baseline factor, is associated with the gain of 3.0, 2.8, 3.2, and 3.3 event-free months of life for all-cause mortality, cardiac mortality, HF mortality, and HF hospitalization, respectively.

This study expands the evidence that suboptimum LV lead position in CRT patients is associated with adverse outcome. Importantly, the QLV ratio cutoff value of 0.70 used to define high-risk patients is considerably higher than cutoff values used in previously published studies. Our data suggest that patients with a QLV ratio ≤0.70 may be considered candidates for reintervention. Since transvenous LV lead reimplantation is frequently not feasible because of the anatomy of the coronary sinus and its tributaries, the benefits and risks of a surgical procedure should be considered.¹⁸ We have recently developed a simple technique of thoracoscopic epicardial mapping during LV lead implantation,¹⁹ which seems safe and efficacious for improvement of LV lead electric position. Whether such an intervention should be performed and when—that is, either early after suboptimum transvenous LV lead implant in all patients or reserved for those with subsequent absence of clinical and echocardiographic response to CRT—remains to be established in future prospective studies.

Although the data in our CRT database were collected prospectively, the hypotheses were defined post hoc and data analyzed retrospectively. Therefore, the results should be interpreted with caution. A cutoff value of 0.70 for the QLV ratio was chosen more or less empirically and may not be the optimal choice. However, an appropriate search for the most applicable dichotomy would require a much larger patient population. In addition, the LV lead positions were not recorded on cine-loops systematically. Therefore, we could not evaluate the potential impact of a given LV lead anatomic position on outcome measures. The amount and the distribution of LV scar tissue, which may interfere with the CRT response, were also not assessed. We did not collect data on Echo-guided CRT optimization which was performed only in early nonresponders to CRT, more likely in those with QLV ratio <0.70. This consequently might influence the results of our study, but with the effect of diminishing rather than strengthening the association of LV lead electric position with clinical end points.

Electric LV lead position assessed by QLV ratio was found to be a significant predictor of mortality in CRT patients with LBBB/IVCD. Therefore, maximum effort should be made to optimize the LV lead electric position during the implant procedure. Patients with suboptimal transvenous LV lead position may be considered candidates for surgical intervention.