Author + information
- Received December 31, 2009
- Revision received March 10, 2010
- Accepted April 13, 2010
- Published online October 26, 2010.
- Roderick W.C. Scherptong, MD⁎,
- Mark G. Hazekamp, MD, PhD†,‡,
- Barbara J.M. Mulder, MD, PhD§∥,
- Olivier Wijers, MSc⁎,
- Cees A. Swenne, PhD⁎,
- Ernst E. van der Wall, MD, PhD⁎,
- Martin J. Schalij, MD, PhD⁎ and
- Hubert W. Vliegen, MD, PhD⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Hubert W. Vliegen, Department of Cardiology, Leiden University Medical Centre, C5-P, P.O. Box 9600, Leiden 2300 RC, the Netherlands
Objectives The aim of this study was to analyze whether QRS duration, before and after pulmonary valve replacement (PVR), is related to long-term outcome in patients with tetralogy of Fallot (TOF).
Background Key factors that determine outcome after PVR in adult TOF patients are largely unknown. Recognition of such factors assists the identification of patients at increased risk of adverse events.
Methods Adults who previously underwent total correction for TOF (n = 90; age 31.4 ± 10.3 years) and required PVR for pulmonary regurgitation were included. The QRS duration was measured pre-operatively and 6 months after PVR. The post-operative changes in QRS duration were calculated. Adverse events (death, re-PVR, ventricular tachycardia, and symptomatic heart failure) were noted during follow-up.
Results During 5.5 ± 3.5 years of follow-up, 13 adverse events occurred. The 5-year event-free survival rate was 76% for patients with a pre-operative QRS duration >180 ms and 90% in patients with a QRS duration ≤180 ms (p = 0.037). For patients with a post-operative QRS duration >180 ms, 5-year event-free survival was 71%, whereas it was 91% for patients with a post-operative QRS duration ≤180 ms (p = 0.004). After multivariate correction, a post-operative QRS duration >180 ms (hazard ratio: 3.685, 95% confidence interval: 1.104 to 12.304, p < 0.05) and the absence of a reduction in QRS duration post-PVR (hazard ratio: 6.767, 95% confidence interval: 1.704 to 26.878, p < 0.01), was significantly associated with adverse outcome.
Conclusions Severe QRS prolongation, before or after PVR, and the absence of a reduction in QRS duration after PVR, are major determinants of adverse outcome during long-term follow-up of patients with TOF.
Progressive pulmonary regurgitation (PR) is a common complication after total surgical correction of tetralogy of Fallot (TOF) (1). Long-standing PR leads to right ventricular (RV) dilation, which in turn causes RV dysfunction and reduced exercise tolerance (2,3). Pulmonary valve replacement (PVR) provides an adequate surgical therapy for PR as it leads to improvement of RV function and the patient's functional class (4). Numerous studies have indicated the beneficial effects of PVR in terms of improvement of RV volume and ejection fraction shortly after PVR (5–7). In addition, it was demonstrated that RV depolarization and repolarization characteristics improve after PVR, partly in response to the reduction in RV volumes (8,9). Specifically, QRS duration, which is strongly associated with RV function and prognosis in TOF (10), tends to reduce after PVR (8). In addition, it is known that TOF patients with severely prolonged QRS duration (>180 ms) are at risk for adverse events (11). Nevertheless, it is unclear whether an association exists between, on the one hand, pre-PVR QRS duration and post-PVR changes in QRS duration and, on the other hand, long-term outcome after PVR. Consequently, the clinical relevance of pre- and post-PVR QRS duration, and of changes in QRS duration after PVR, is presently unknown. The current study aimed to assess how pre-operative QRS duration and post-operative changes in QRS duration are related to outcome during long-term follow-up after PVR in TOF patients.
In a prospective, multicenter cohort study, adult patients after prior total repair for TOF, who had undergone a PVR for PR, were followed up (12,13). In all patients, primary surgical repair was performed during childhood. The indications for PVR were moderate to severe PR in combination with RV dilation and impaired New York Heart Association (NYHA) functional class (4). A standard 12-lead electrocardiogram (ECG) was obtained immediately before and 6 months after PVR. If an ECG was not available at 1 of the time points, patients were excluded from further analysis. All patients were followed up annually at the out-patient clinic, and the occurrence of adverse events was noted during follow-up after PVR. Adverse events were defined as death due to any cause, reoperation for recurrent pulmonary regurgitation, symptomatic heart failure, and ventricular arrhythmias.
The PVR was performed through median sternotomy. After release of adhesions, total cardiopulmonary bypass was started. A cryopreserved pulmonary homograft was implanted, usually on beating heart. If necessary, small residual ventricular septal defects were suture closed, and right ventricular outflow tract (RVOT) reconstruction was performed as described previously (12). In addition, when pre-operative assessment revealed significant tricuspid regurgitation, tricuspid annuloplasty was performed (14).
Electrocardiograms were stored digitally and exported from the ECG database management system for off-line analysis with the MATLAB-based (The MathWorks, Natick, Massachusetts) computer program LEADS (Leiden, the Netherlands). Details about the calculation methods used in LEADS have been reported previously (15). The software averages a digitized version of the standard 12-lead surface ECG into a representative single-beat ECG in which all 12 leads are superimposed. The software is equipped with a cross-hair editor interface that allows magnification of the ECG, and thereby facilitates accurate measurement of QRS duration according to the Minnesota criteria for population-based ECG studies (8). All ECGs were analyzed by a single observer blinded to the patient's clinical status and operation results. The cut-off for severely prolonged QRS duration was set at 180 ms (10,11).
The SPSS version 16.0.2 software (SPSS, Inc., Chicago, Illinois) was used for statistical analysis. GraphPad Prism 4.00 for Windows (GraphPad Software, La Jolla, California) was used to obtain life tables and corresponding Kaplan-Meier survival curves.
First, pre- and post-operative QRS duration and NYHA functional class were compared using a paired ttest. Second, the association between QRS duration and the occurrence of adverse events during post-operative follow-up was assessed using Cox's proportional hazards regression analysis. Univariate hazard ratios (HR) were calculated with corresponding 95% confidence interval (CI) for the following variables: 1) pre- and post-operative QRS duration as a continuous variable, in ms; 2) pre- and post-operative QRS duration as a categorical variable, ≤180 or >180 ms; 3) changes in QRS duration, from pre- to post-operative, as a continuous variable, in ms; and 4) changes in QRS duration, from pre- to post-operative, as a categorical variable, expressed as the presence or the absence of reduction in QRS duration. Afterward, the HRs were adjusted for age at PVR (below or above the median age of 30 years) and the presence of concomitant procedures during PVR, to obtain the multivariate HRs.
Third, the annualized event rates were assessed in patient subgroups on the basis of the pre- and post-operative QRS durations (≤180 or >180 ms) in addition to the post-operative changes in QRS duration (reduction or no reduction). The number of observed events was normalized to 100 patient-years in each subgroup to facilitate comparison between groups. Afterward, the 95% CI of the observed event rate was calculated, and the difference in event rate was compared to the overall event rate by assuming that the number of observed events has a Poisson distribution.
Finally, to visualize the difference in event-free survival between different QRS duration patient categories, Kaplan-Meier curves were drawn and log-rank statistics were calculated for the following 3 patient subgroups: 1) pre-operative QRS duration ≤180 or >180 ms; 2) post-operative QRS duration ≤180 or >180 ms; and 3) changes in QRS duration, expressed as the presence or absence of a post-operative reduction in QRS duration. Separate Kaplan-Meier curves were obtained in which death, symptomatic heart failure, and ventricular arrhythmias were included as an end point and re-PVR was excluded.
Where appropriate, data are reported as mean ± SD, n (%), or as HR (95% CI). All p values <0.05 were considered statistically significant.
Ninety-nine consecutive corrected TOF patients who underwent PVR at adulthood were considered for the study. Nine patients were excluded because of missing ECGs, and the remaining 90 patients were all included in the study (Table 1).The PVR was performed at an age of 31.4 ± 10.3 years, and pre-operative NYHA functional class was 2.4 ± 0.7. The size of the implanted homografts was 25.6 ± 1.6 mm, and concomitant surgical procedures, mostly RVOT reconstructions, were required in 42 patients (47%) (Table 1).
No patients were lost to follow-up. After PVR, NYHA functional class improved from 2.4 ± 0.7 to 1.3 ± 0.6 (p < 0.001). During a mean follow-up duration of 5.5 ± 3.5 years (range 1.1 to 17.9 years), 13 adverse events occurred (Fig. 1A).The observed event-free follow-up was 100% at 1 year, 96% at 2 years, 89% at 5 years, and 78% at 10 years (Fig. 1B). During follow-up, 2 patients died after 1.4 and 8.5 years, respectively, of sudden cardiac death (first patient) and fast-progressive, therapy refractory heart failure (second patient). Re-PVR was performed in 5 patients, and ventricular tachycardia requiring implantation of an implantable cardioverter defibrillator occurred in 4 patients.
Prediction of patient outcome using QRS duration
Before PVR, QRS duration was 158 ± 29 ms, which reduced 4 ± 17 ms to 154 ± 32 ms 6 months after surgery (p = 0.027) (Table 1).
Pre-operative NYHA functional class was not associated with pre- or post-operative QRS duration or the post-operative changes in QRS duration (black bars in Fig. 2)and demonstrated to improve in all patient groups based on QRS duration (Fig. 2). However, in patients with a QRS duration >180 ms, either pre- or post-operative, a less prominent improvement in NYHA functional class was observed as compared to patients with a QRS duration ≤180 ms pre-operatively (p = 0.002) or post-operatively (p = 0.045).
Both pre- and post-operative QRS durations, as well as post-operative change in QRS duration, were significantly associated with adverse outcome after PVR (Table 2).The crude HRs of pre- and post-operative QRS duration were 1.025 and 1.026, respectively, per ms increase in QRS duration. This indicated that a longer QRS duration either pre- or post-operatively was associated with a higher incidence of adverse events. Furthermore, the change in QRS duration from pre-operatively to 6 months after PVR was also related to post-operative outcome. The crude HR for the occurrence of adverse events was 0.978 per ms change in QRS duration, demonstrating that a post-operative reduction in QRS duration was associated with a lower incidence of adverse events during follow-up. After multivariate correction, pre- and post-operative QRS duration and post-operative changes in QRS duration remained significantly related to the incidence of adverse events. Specifically, a post-operative QRS duration >180 ms or the lack of a reduction in QRS duration post-operatively were strong predictors of adverse events, with multivariate HRs of 3.685 and 6.767, respectively.
To further explore the association between QRS duration and event-free survival after PVR, the study population was subdivided based on QRS duration (pre- and post-operative; ≤180 or >180 ms), and post-operative change in QRS duration (reduction or no reduction), which yielded 8 subgroups (Table 3).Subsequently, the annualized event rate, normalized to 100 patient-years, was calculated for the overall study population and for the 8 subgroups (Table 3). The overall event rate in the study population was 2.7 (95% CI: 1.4 to 4.6) per 100 patient-years. No events were observed among patients with a pre-operative QRS duration ≤180 ms who demonstrated a post-operative reduction in QRS duration (n = 44). On the contrary, a significantly higher event rate was observed among patients who did not demonstrate a reduction in QRS duration and had a QRS duration >180 ms, either pre-operatively or post-operatively. In those groups, the event rates were 20.6 events (95% CI: 2.5 to 74.5; p = 0.007 vs. the overall population) and 9.3 events (95% CI: 2.5 to 23.8; p = 0.030 vs. the overall population) per 100 patient-years (Table 3).
The Kaplan-Meier survival curves (Fig. 3)confirmed that patients with significantly prolonged QRS duration (>180 ms) demonstrated worse event-free survival as compared with that of patients with a QRS duration ≤180 ms. The observed 5-year event-free survival was 76% for patients with a pre-operative QRS duration >180 ms as compared with 90% for patients with a QRS duration ≤180 ms (p = 0.037) (Fig. 3A). The observed 5-year event-free survival was 71% for patients with a post-operative QRS duration >180 ms as compared with 91% for patients with a QRS duration ≤180 ms (p = 0.003) (Fig. 3E). Furthermore, 5-year event-free survival of patients in whom no reduction in QRS duration was observed 6 months after PVR was worse as compared with that of patients who had a reduction in QRS duration after PVR. The observed difference in event-free survival between those 2 groups was 24.5% at 5 years (p = 0.004) (Fig. 3C). Exclusion of re-PVR as an end point in the Kaplan-Meier survival curve demonstrated that event-free survival was 4.1% lower in patients with a pre-operative QRS duration >180 ms as compared to ≤180 ms, which was not statistically significant (p = 0.61) (Fig. 3B). The Kaplan-Meier curve, in which the patient population was stratified according to post-operative QRS duration (≤180 or >180 ms) or post-operative changes in QRS duration (reduction or no reduction), was similar whether or not re-PVR was excluded as an end point (Figs. 3D and 3F).
The main findings of this study were: 1) a QRS duration >180 ms, either pre- or post-operatively, or the lack of a reduction in QRS duration after PVR, was significantly associated with adverse outcome after PVR; and 2) the highest incidence of adverse events was observed among patients with a severely prolonged QRS duration either pre- or post-operatively, without a reduction in QRS duration after PVR.
Progressive PR is one of the most common complications after total surgical repair of TOF. In response to the regurgitated volume, the RV dilates to overcome the increased loading conditions (1,16). Long-standing PR will eventually lead to dysfunction of the RV, and patients may experience reduced exercise tolerance (17). In such cases, PVR with a competent pulmonary valve provides a satisfactory solution, with excellent short-term results in terms of ventricular function and NYHA functional class (4,18). Although long-term results are also good for PVR, a substantial number of patients still have adverse events during follow-up (7). In the current study, event-free survival was 76% after 10 years, which is similar to that of other studies in which long-term survival after PVR was reported (12,19). A number of factors that are important for post-PVR outcome have been identified. These mainly include post-operative homograft-related factors such as early recurrence of PR or pulmonary stenosis (12,20). Interestingly, in the study by Oosterhof et al. (12), which was a large follow-up study that consisted of 158 adult TOF patients who underwent PVR, neither the type of initial TOF correction nor the PVR surgical characteristics were predictive of outcome after PVR. In that study, RV volumes and ejection fraction, assessed pre-operatively with cardiovascular magnetic resonance imaging, were not predictive of post-PVR outcome. Although it is known that QRS duration is strongly associated with outcome in TOF patients (10), no studies on the clinical application of QRS duration for the assessment of post-PVR outcome are available. The current study demonstrates that pre-operative QRS duration is associated with post-operative outcome. Longer QRS duration was significantly associated with increased incidence of adverse outcome after PVR. Furthermore, severe QRS prolongation (>180 ms) after PVR and the absence of a reduction in QRS duration demonstrated an even stronger association with post-operative outcome. As re-PVR has an indirect association with ventricular function (12), separate Kaplan-Meier survival curves were obtained in which re-PVR was excluded as an end point. That confirmed that QRS duration has a strong association with post-PVR outcome in patients with TOF.
The association between prolonged QRS duration and patient outcome in TOF was first recognized by Gatzoulis et al. (10). In a study on 178 adults with a previous total repair for TOF, it was demonstrated that ventricular tachycardia was associated with QRS prolongation beyond 180 ms. In addition, it was shown that patients who had an enlarged RV had a longer QRS duration. The relation between RV enlargement and QRS duration was confirmed thereafter in study by Book et al. (21). They demonstrated that prolonged QRS duration (>150 ms), had a positive predictive value of 92% for the detection of RV dilation on cardiovascular magnetic resonance imaging. Moreover, Neffke et al. (22) demonstrated that QRS duration was also correlated to RV mass. In addition to the relation between QRS duration and ventricular function, it was shown in a study Budts et al. (23) that lack of reduction in QRS duration during exercise is significantly related to reduced exercise capacity in adult TOF patients. Furthermore, previous studies have indicated reduction of QRS duration in response to PVR (8). This reduction in QRS duration was related to the reduction in RV end-diastolic volume, which demonstrated that PVR improved the electrophysiological characteristics of the RV. However, to date, it remained to be investigated whether these changes in electrophysiological characteristics were of relevance in terms of clinical outcome.
The current study demonstrates for the first time that severe QRS prolongation (>180 ms) either before or after PVR, or the absence of a reduction in QRS duration after PVR, is associated with the increased incidence of adverse outcome during long-term follow-up. It could be postulated that the patients who lack improvement in QRS duration have a poorer RV condition and thus exhibit a higher risk for adverse events as compared with patients who improve in QRS duration after PVR. Analogous with this, the observation that longer QRS duration before PVR was associated with increased incidence of adverse events after PVR may also be related to a more deteriorated RV condition already before PVR. That no events were observed in patients with a QRS duration <180 ms who demonstrated improvement of QRS duration after PVR may reflect the other end of the same pathophysiological spectrum.
As with most studies on patients with congenital heart disease, a limitation of the current study is the relatively small number of patients. Therefore, a comprehensive multivariate Cox's regression analysis, in which the HRs are adjusted for all possible confounders, could not be obtained. Furthermore, the subgroup analysis based on before and after PVR QRS duration and changes after PVR in QRS duration would benefit from a larger sample size, as that increases the precision of the event rate estimations. A large multicenter trial with sufficient power for a complete analysis of possible confounders is warranted to improve the quantification of the event rates and to confirm the differences between the subgroups. In addition, that QRS duration was tested continuously as well as categorically using a predefined cut off may introduce type I statistical error in the current evaluation. However, a consistent association between QRS duration and patient outcome was observed whether QRS duration was entered continuously or categorically in the regression equation.
The current study demonstrates that both QRS duration itself and the changes in QRS duration after PVR can be used to identify patients who are at increased risk for adverse events during follow-up after PVR. Therefore, the patients who have a QRS duration >180 ms either before or after PVR, and the patients who do not demonstrate improvement of QRS duration after PVR, may require a vigilant follow-up regimen, as they exhibit a high risk for adverse outcome. Currently, PVR is recommended when important PR and severe RV dilation occur, specifically when this coincides with reduced NYHA functional class (18,24). As the current study demonstrates that severely prolonged QRS duration (>180 ms) is associated with increased incidence of adverse outcome, it could be postulated that PVR should be performed before severe QRS prolongation occurs. The latter point, however, needs to be confirmed in a sufficiently powered clinical trial.
The QRS duration either before or after PVR and the changes in QRS duration after PVR are important determinants of long-term outcome in adults with TOF and may serve as an easy screening tool to identify patients at high risk for adverse events after PVR.
Dr. Scherptong is supported by an unrestricted educational grant from Actelion Pharmaceuticals Nederland BV. Dr. Schalij receives research grants from Biotronik, Medtronic, and Boston Scientific. All other authors have reported that they have no relationships to disclose.
- Abbreviations and Acronyms
- confidence interval
- hazard ratio
- New York Heart Association
- pulmonary artery
- pulmonary regurgitation
- pulmonary valve replacement
- right ventricular
- right ventricular outflow tract
- tetralogy of Fallot
- Received December 31, 2009.
- Revision received March 10, 2010.
- Accepted April 13, 2010.
- American College of Cardiology Foundation
- Norozi K.,
- Buchhorn R.,
- Bartmus D.,
- et al.
- Vliegen H.W.,
- van Straten A.,
- de Roos A.,
- et al.
- Adamson L.,
- Vohra H.A.,
- Haw M.P.
- Gengsakul A.,
- Harris L.,
- Bradley T.J.,
- et al.
- van Huysduynen B.H.,
- van Straten A.,
- Swenne C.A.,
- et al.
- Gatzoulis M.A.,
- Till J.A.,
- Somerville J.,
- Redington A.N.
- Oosterhof T.,
- Meijboom F.J.,
- Vliegen H.W.,
- et al.
- Scherptong R.W.,
- Vliegen H.W.,
- Winter M.M.,
- et al.
- Therrien J.,
- Siu S.C.,
- McLaughlin P.R.,
- Liu P.P.,
- Williams W.G.,
- Webb G.D.
- Troost E.,
- Meyns B.,
- Daenen W.,
- et al.
- Neffke J.G.,
- Tulevski I.I.,
- van der Wall E.E.,
- et al.
- Oosterhof T.,
- van Straten A.,
- Vliegen H.W.,
- et al.