Electrocardiographic Patterns during Left Ventricular Epicardial Pacing MAREK JASTRZEBSKI, M.D., PH.D.,* KAMIL FIJOREK, M.SC.,† and DANUTA CZARNECKA, M.D., PH.D.‡ From the *First Department of Cardiology and Hypertension, University Hospital, Cracow, Poland; †Department of Statistics, Cracow University of Economics, Cracow, Poland; and ‡First Department of Cardiology and Hypertension, Jagiellonian University, College of Medicine, Cracow, Poland
Background: There is a paucity of data concerning the use of QRS morphology patterns for identifying pacing sites during left ventricle (LV)-only epicardial pacing in patients with a biventricular device. The objective of this study was to identify QRS patterns during LV-only pacing, and to establish their relationship with LV lead position. In addition, to validate the diagnostic performance of such electrocardiogram (ECG) patterns for predicting posterolateral versus anterior and apical versus nonapical LV pacing site. Methods: The study retrospectively analyzed data from 376 cardiac resynchronization therapy device patients. Data analyzed included ECGs registered during LV-only VVI pacing, fluoroscopic projections, and lateral chest roentgenograms that documented postimplantation LV lead position. Phase one of the study involved categorization of the ECG patterns of the first 66 study cases. Phase two of the study examined the association between ECG pattern and different LV lead positions. Results: As the LV epicardial pacing site became more anteroapical, the LV-only paced QRS complexes in the precordial leads became more negative. Three ECG patterns were identified (posterolateral, intermediate, and anteroapical), and their distribution was found to be associated with LV lead position (P < 0.001). The posterolateral ECG pattern was mostly observed in cases where the LV lead was in the posterolateral area (diagnostic accuracy of 89.1% for predicting a nonapical LV lead position). The anteroapical ECG pattern was associated with LV leads in anteroapical segments (specificity of 98.5%, accuracy of 89.1% for predicting an anteroapical pacing site). Conclusions: Posterolateral and anteroapical ECG patterns are highly predictive of LV lead position. (PACE 2012;XX:XXX–XXX) CRT, Electrocardiogram, left ventricular lead position, QRS morphology, left ventricular pacing Introduction Pacing of the delayed posterolateral segments of the left ventricle (LV) in left bundle branch block (LBBB) patients with heart failure forms the theoretical and practical cornerstone of cardiac resynchronization therapy (CRT). Electrocardiogram (ECG) could be helpful in identifying the LV pacing site both during implantation and follow-up.1,2 However, there is a paucity of data concerning the use of QRS morphology patterns for identifying pacing sites during LV-only pacing. Specifically, the sensitivities, specificities, and predictive values of various LV-paced QRS patterns for determining various LV lead positions remain unknown. A number of studies have reported on the impact of LV lead position on clinical outcomes,
and many, although not all, found that a nonposterolateral LV lead position, especially an anterior and/or apical position, was associated with worse results.3–12 Therefore, it seems important to be able to differentiate between a posterolateral nonapical position and other LV lead positions. We hypothesized that apical and/or anterior LV epicardial pacing sites should result in QRS patterns distinct from those related to pacing posterolateral nonapical sites. Our goal was to categorize ECG patterns during LV-only pacing, and to establish the relationship between various patterns and the LV lead position. In addition, to validate the diagnostic performance of such ECG patterns for predicting posterolateral versus anterior and apical versus nonapical LV pacing site.
Conflicts of Interests: None.
Methods We retrospectively examined data from 376 consecutive patients implanted with CRT devices in our institution. Data examined included ECGs registered during LV-only VVI pacing, fluoroscopic projections demonstrating postimplantation LV lead position, lateral chest roentgenograms documenting postimplantation LV lead position,
Address for reprints: Marek Jastrzebski, M.D., Ph.D., I Department of Cardiology and Hypertension, ul. Kopernika 17, 31-501 Cracow, Poland. Fax: 048-124247320; e-mail:
[email protected] Received March 21, 2012; revised June 14, 2012; accepted July 02, 2012. doi: 10.1111/j.1540-8159.2012.03504.x
C 2012 Wiley Periodicals, Inc. C 2012, The Authors. Journal compilation
PACE, Vol. 00
2012
1
JASTRZEBSKI, ET AL.
of the heart within the chest cavity. Therefore, as shown by Rickard et al., an apparently apical position according to a PA roentgenogram is indeed more often nonapical.13
and basic clinical and demographic data. Cases in which a patient had multiple LV lead positions because of either LV lead reposition/replacement or an upgrade to a dual-LV-site pacing system was analyzed as separate entries. Phase one of the study involved the initial 66 cases. Data from those 66 cases were used to identify ECG pattern categories based on ECG and roentgenogram analysis. Phase two of the study involved the remaining subsequent patients. During this phase, a “blind” analysis was performed to determine the diagnostic value of various ECG patterns.
Electrocardiographic Assessment All electrocardiograms for analysis were registered as standard 12-lead ECGs with a paper speed of 25 mm/s during VVI 100 beats per minute, LV-only pacing. All ECGs in phase two of the study were analyzed by a cardiologist blinded to the results of fluoroscopy and chest roentgenograms. In all cases, the precordial lead at which there was either transition from a positive to a negative QRS complex or a concordant pattern (positive or negative) was determined.
LV Lead Position Assessment Using Fluoroscopy Fluoroscopic LV lead position classification was based on analysis of two postimplantation images (i.e., the left anterior oblique [LAO] and right anterior oblique [RAO] at 30◦ –40◦ ), as per previous studies.7 RAO view was used to classify the lead position as apical or nonapical (basal, midventricular-1, midventricular-2). Using that view, the heart silhouette was divided into four equal parts from the apex to the base. The base of the heart was approximated by the course of the lead in the coronary sinus. Using the LAO view, the LV free wall was divided into four equal parts, namely anterior, anterolateral, lateral, posterolateral, and posterior. The RAO and LAO data were combined to categorize the LV lead position as anteroapical (apical on RAO plus anterior on LAO) or nonanteroapical (all other positions).
Statistical Methods Continuous variables are presented as mean and standard deviation. Categorical variables are presented as numbers and percentages. The equality of multinomial proportions was tested using exact goodness-of-fit chi-square tests. The performance of binary (or dichotomized) decision rules was described by the following measures (with a 95% confidence interval): diagnostic accuracy, sensitivity (SN), specificity (SP), positive predictive value (PPV), and negative predictive value (NPV), which were calculated using a selfwritten code and the BDT comparator program.14 Statistical analysis was performed in STATISTICA 10.0 and R (StatSoft Inc., Tulsa, OK, USA), which is a language and environment for statistical computing.15 P values < 0.05 were considered to indicate statistical significance.
LV Lead Position Assessment Using Chest X-Rays Chest roentgenograms are more applicable in follow-up than fluoroscopy, and provide more information with regard to the LV lead position in the sagittal plane within the chest cavity. A lateral chest roentgenogram was used to categorize the LV lead position as: (1) posterolateral, (2) intermediate, or (3) anterior/apical. For that categorization, the posteroanterior (PA) ratio on lateral film was used. The PA ratio was calculated by dividing the distance from the posterior heart border to the LV lead tip by the distance between the posterior and anterior heart border. A posterolateral position was defined as a PA ratio < 0.3, an intermediate position was defined as a PA ratio of 0.3–0.4, and an anteroapical position ratio was defined as a ratio > 0.4. PA roentgenograms were not used for LV lead position assessment because they neither differentiate posterior from anterior nor midventricular from apical positions. Although that second point might seem less obvious, a posterior midventricular LV lead position often appears as apical because of the oblique position
2
Results Of the 376 screened CRT patients, ECG/LV lead position data were incomplete for 20 patients. More than one set of LV lead position data was available for 19 patients. Therefore, the study ultimately analyzed 378 ECG/LV lead position pairs derived from 356 patients. Basic patient clinical and demographic data are summarized in Table I. Identification and Definition of ECG Patterns Phase one of the study involved defining ECG patterns based on LV-paced QRS morphologies and lateral roentgenogram data. Three ECG pattern categories were identified: 1. A posterolateral pattern was defined as QRS complexes predominantly positive in at least leads V1–V3, and without an S wave larger than 50% of the R wave in V3. This was by far the predominant pattern, and was observed when the lead tip was close to the posterior heart border
2012
PACE, Vol. 00
ECGS DURING LEFT VENTRICULAR PACING
usually with a qR complex in V6 (Fig. 1, panel B) and rarely with an R complex in V6 (Fig. 1, panel A). 2. An anteroapical pattern was defined as QRS complexes predominantly negative in leads V2–V4. Typically, that pattern presented with a positive V1 and a negative V2–V6 (Fig. 1, panel H). However, when the LV lead was in the anterior interventricular vein, there was a negative concordancy (Fig. 1, panel I) or even a LBBB pattern (panel H). 3. An intermediate pattern was defined as QRS complexes predominantly positive in leads V1–V2, and either predominantly negative in leads V3–V6 or with an S wave >50% of the R wave in V3 (Fig. 1, panel F). This pattern was seen when the LV lead tip was in the area intermediate between the posterior and anterior positions, and usually slightly posterior to a vertical line formed by the course of the leads coming from the superior vena cava according to the lateral chest roentgenogram.
Table I. Basic Clinical Characteristics of the 356 Study Patients Age [Years] ± SD Women/Men [%] Etiology: ischemic heart disease/DCM [%] LV ejection fraction [%] ± SD LVEDD [mm] ± SD QRS width [ms] ± SD
67.1 ± 9.9 14.3/85.7 (n = 51/305) 61.5/38.5 (n = 219/137) 24.8 ± 8.0 70.2 ± 9.9 169.9 ± 33.6
DCM = diastolic cardiomyopathy; LVEDD = Left ventricle end-diastolic dimension; SD = standard deviation.
(Fig. 1, panels A–E). In the majority of cases, the QRS complexes of this pattern were positive up to leads V4–V5, and were then abruptly negative in leads V5 and/or V6 (Fig. 1, panels C–D). In some cases, usually with a very basal LV lead position, the positive QRS complexes extended to V6, giving rise to a positive concordancy,
Figure 1. LV-only paced QRS patterns in precordial leads. Note that as the LV lead was positioned progressively more anteriorly (photos), the positive QRS complexes were replaced by negative QRS complexes commencing from lead V6 and moving toward lead V1 (tracings). Finally, there was an entire reversal of polarity from a right to a left bundle branch block (LBBB) pattern. Panels A and B: Posterolateral basal patterns characterized by positive concordancy. Panels C–E: The most commonly observed posterolateral patterns characterized by negative QRS complexes limited to leads V4–V6. Panel F: Intermediate pattern characterized by a QRS transition at lead V3. Panel G: Apical pattern characterized by a positive QRS only in lead V1. Panels H–I: Anterior interventricular vein patterns: negative concordancy (Panel H) and LBBB (Panel I) patterns.
PACE, Vol. 00
2012
3
JASTRZEBSKI, ET AL.
Table II. Distribution of ECG Patterns with Regard to LV Lead Position and Heart Failure Etiology LV Lead Position RTG: posterolateral RTG: intermediate
RTG: anteroapical
RAO: nonapical RAO: apical LAO: anterior LAO: anterolateral LAO: lateral LAO: posterolateral LAO: posterior RAO + LAO: anteroapical RAO + LAO: nonanteroapical
Posterolateral ECG
Intermediate ECG
Anteroapical ECG
P
DCM (n = 100) CHD (n = 140) DCM (n = 11) CHD (n = 14) All (n = 25) DCM (n = 17) CHD (n = 30) All (n = 47)
99.0% 90.7% 81.8% 35.7% 56.0% 23.5% 23.3% 23.4%
1.0% 8.6% 9.1% 50.0% 32.0% 29.4% 20.0% 23.4%
0.0% 0.7% 9.1.0% 14.3% 12.0% 47.1% 56.7% 53.2%
