Aims Heart failure is common late after Senning or Mustard palliation of transposition of the great arteries (TGA). Although cardiac magnetic resonance (CMR) is the gold standard for evaluating systemic right ventricular performance, additional information regarding heart failure status might be gleaned from the surface ECG and circulating N-terminal pro-brain natriuretic peptide (NT-proBNP) levels. The interrelationships between these heart failure markers were examined in adults late after Mustard and Senning surgery.
Methods Thirty-five consecutive adults with Senning or Mustard repair of TGA attending a dedicated congenital heart failure clinic were studied. Assessment included symptom assessment, venous blood sampling for measurement of circulating NT-proBNP levels, surface 12-lead ECG and CMR for the assessment of right ventricular systolic function and determination of indexed right ventricular volumes.
Results Mean age was 29±6.5 years, 54% had undergone Mustard surgery. Compared with those with uncomplicated surgery, patients with complex surgical history had higher NT-proBNP levels (55±26 vs 20±35 pmol/l; p=0.002) and longer QRS duration (116±28 ms vs 89±11 ms; p=0.0004) while showing no difference in New York Heart Association class and right ventricular function. There was a significant relationship between diastolic and systolic right ventricular volumes and both NT-proBNP levels (r=0.43, p=0.01; r=0.53, p=0.001, respectively) and QRS duration (r=0.47, p=0.004; r=0.53, p=0.001, respectively).
Conclusions Circulating NT-proBNP levels and several surface ECG parameters constitute safe, cost-effective and widely available surrogate markers of systemic right ventricular function and provide additional information on heart failure status. Both measures hold promise as prognostic markers and their association with long-term outcome should be determined.
- transposition of the great arteries
Statistics from Altmetric.com
Until the widespread uptake of arterial switch surgery for classic transposition of the great arteries (TGA), palliation with atrial redirection surgery (Mustard or Senning procedure) was the norm. Many patients managed in this way survive today but with the right ventricle in the subaortic position there is a high burden of morbidity and premature mortality from attritional heart failure.1 Survival has been calculated at 76% by 20 years of age2 with a mean age at death of 27 years.3
Outcome in this population is related to the function of the systemic right ventricle.4 Cardiac magnetic resonance (CMR) imaging is the most accurate and reproducible tool for assessing this;5 6 however, this method is time consuming and availability is often restricted to specialist centres. The high prevalence of implanted pacing/defibrillator devices in this population and the small but significant proportion of patients with claustrophobia also limit its use. CMR imaging may be prompted following clinical assessment, yet patients notoriously underreport functional ability and symptoms despite marked impairment on objective testing.7 Cost-effective and more readily available surrogate clinical markers that closely relate to ventricular function would therefore be highly desirable.
Elevated circulating N-terminal pro-brain natriuretic peptide (NT-proBNP) levels help identify patients with impaired systemic ventricular function8 and relate to outcome in those with acquired heart disease.9 In mixed cohorts of patients with congenital cardiac disease, brain natriuretic peptide (BNP) has been shown to relate to systemic ventricular function as determined by echocardiography10 and to function of the pressure overloaded systemic and subpulmonary right ventricle as determined by CMR.11 Surface ECG measures, including QRS width, have also been found to relate to right ventricular size and function in the setting of pressure overload.12 The interrelationship of systemic right ventricular size and function with circulating NT-proBNP levels and surface ECG measures in adults with TGA late after atrial switch repair is unknown.
Data routinely collected from consecutive adults with previous Senning or Mustard surgery for TGA attending a dedicated congenital heart disease heart failure clinic at our institution between January 2008 and March 2009 were included in this study. All patients had undergone assessment of symptoms by New York Heart Associaion (NYHA) functional class, measurement of circulating NT-proBNP levels and surface ECG as part of their routine clinic work-up. CMR imaging was performed within 3 months of clinic attendance in accordance with clinic protocols.
Standard original hard copies were scanned and analysed using the CardioCalipers program (version 3.3 for Windows, http://www.iconico.com/), which enabled a zooming function for greater measurement accuracy. ECG were analysed by one observer (CP) who was blinded to NT-proBNP and CMR findings. QRS duration was averaged after analysis of all 12 leads and defined as the first positive/negative deflection to the last positive/negative deflection across the isoelectric line. QRS, QT and JT dispersions (d) were determined by subtracting the narrowest interval from the longest across any of the 12 leads. The end of the T wave was defined as the point of return to the isoelectric line. QT intervals were averaged across the 12 leads and corrected for heart rate to obtain QTc.
N-terminal pro-brain natriuretic peptide
Peripheral venous blood was collected from each patient after they had rested for a minimum of 20 min. Samples were centrifuged and immediately analysed by sandwich immunoassay using electrochemiluminescence (E 170 Module; Roche Diagnostics, Basel, Switzerland).
CMR imaging was performed using a 1.5T scanner (Avanto, Siemens Medical Systems, Erlangen, Germany) with assessment of ventricular volumes as previously described.13 14 Manual segmentation of the ventricles with exclusion of the major right ventricular trabeculae from the blood pool volume was undertaken. Quantification of aortic forward flow volume using phase contrast volumetry provided an internal quantitative guide to the right ventricular stroke volume and improved the accuracy of volumetric quantification in the presence of tricuspid valve regurgitation.15 Volumes were indexed to body surface area.
The Kolmogorov–Smirnov test was used to confirm or exclude normal distribution for each variable. Values are expressed as mean±SD, median (25–75th percentile values) or percentages, as appropriate. Univariate comparisons were performed by Student's t test, Mann–Whitney Utest, and χ2 test for normally distributed, non-normally distributed and categorical variables, respectively. Pearson's or Spearman's correlation coefficients were used as appropriate to look for univariate associations between continuous variables. Differences were considered to be significant at a p value of less than 0.05.
Thirty-five adults were included in the study. None had decompensated heart failure nor compromising arrhythmia between the time of clinic attendance and CMR imaging. Mean age was 29±6.5 years at the time of clinic attendance and 14±13.6 months at the time of surgery (54% Mustard operation). Five (14%) patients had associated ventricular septal defects repaired at original palliation—‘complex’ anatomy. Seven further patients (20%) underwent later surgical revision of atrial pathways. For the purpose of statistical analysis these 12 patients comprised the ‘complex surgical’ group. The remaining 23 patients had simple TGA and single-stage surgery. Ten patients were taking β-blockers (29%), 18 ACE inhibitors/angiotensin receptor blockers (51%) and two were on diuretic therapy (6%). Table 1 summarises baseline measurements of all investigations.
Twelve patients were in NYHA class II (34%) with the remainder in functional class I. Patients in NYHA class II had significantly longer QRS duration (90±3 vs 113±8 ms; p=0.003), QRSd (30±2 vs 39±4 ms; p=0.03) and QTc interval (390±8 vs 426±17 ms; p=0.03) compared with those in functional class I. Right ventricular end-systolic indexed volume (RVESVI) (48±3 vs 64±8 ml/m2; p=0.04) and NT-proBNP (29±6 vs 56±10 pmol/l; p=0.02) were also higher in patients in NYHA class II. There was no statistical difference between functional classes with respect to palliation method, age at follow-up, age at repair or right ventricular ejection fraction (RVEF).
Complex versus simple surgical course
Compared with those with uncomplicated surgery, patients with complex anatomical and surgical history had higher NT-proBNP levels (55±26 vs 20±35 pmol/l; p=0.002) and longer QRS duration (116±28 ms vs 89±11 ms; p=0.0004; table 1) while showing no difference in NYHA class, right ventricular volumes or right ventricular function. Four of those with a ‘complex’ history had undergone Senning palliation compared with eight who had undergone Mustard palliation.
For the cohort as a whole, mean right ventricular end-diastolic indexed volume (RVEDVI) was 107±27 ml/m2, mean RVESVI 53±21 ml/m2 and mean RVEF 51±8%. RVEDVI did not relate to age at follow-up (p=0.8) or age at palliative surgery (p=0.5). There was also no statistical difference when comparing CMR measures between those with Mustard and Senning palliation.
The mean plasma NT-proBNP concentration was 38±34 pmol/l. There was no relationship between NT-proBNP and age at clinic attendance or age at palliation. There was a trend towards higher NT-proBNP levels in women (p=0.054) but values were unrelated to serum creatinine or body weight. NT-proBNP levels were significantly higher in those with Mustard palliation compared with those who underwent Senning surgery (50±38 vs 23±22; p=0.01), but this significance was lost when adjusting for age (p=0.09).
NT-proBNP and CMR
There was a significant negative correlation between NT-proBNP levels and RVEF (r=−0.54, p=0.0007; figure 1) corresponding with significant positive correlations between NT-proBNP levels and both indexed end-diastolic and systolic volumes (r=0.43, p=0.01 and r=0.53, p=0.001, respectively; table 2). No association was found between NT-proBNP levels and left ventricular ejection fraction (LVEF) or left ventricular volume measurements.
The mean QRS duration was 98±23 ms, QTc interval 402±47 ms, QRSd 33±12 ms, JTd 53±18 ms and QTd 57±22 ms. No ECG parameter related to age at follow-up, age at palliation or the palliative subtypes.
ECG and CMR
QRS duration, QRSd, QTc and QTd all correlated significantly with both RVEDVI and RVESVI (table 2), the most significant correlation being between QRSd and RVEDVI (r=0.60, p≤0.0001; figure 2). QRS duration, QRSd and QTc all had inverse relationships to RVEF. Only QTd showed no relation with RVEF and it was the only ECG parameter to show a correlation with the LVEF (r=−0.39, p=0.02).
This is the largest prospective study examining the interrelationship between NT-proBNP, ECG parameters and systemic right ventricular function in a homogenous cohort of adults late after atrial switch redirection surgery for d-TGA. It demonstrates that larger systemic right ventricular volumes and lower ejection fractions are associated with higher circulating levels of NT-proBNP and longer ECG measures of ventricular depolarisation and repolarisation. This pattern of falling systemic ventricular function, increasing circulating natriuretic peptide levels and prolongation of ECG measures characterises the heart failure syndrome in acquired heart disease and adds further support to the view that many congenital heart diseases are, in essence, heart failure syndromes.16
It is interesting to note that NT-proBNP levels were higher and QRS width longer in the group with complex anatomical and surgical history compared with those without, while at the same time NYHA class and CMR measures of right ventricular volumes and function showed no difference between these groups. These findings suggest that NT-proBNP levels and ECG parameters provide information about heart failure status over and above that determined by symptom assessment and CMR.
The subjective nature of symptom scores and the poor correlation between symptoms and cardiopulmonary exercise capacity, a powerful predictor of prognosis,7 is well known. The relatively poor prognostic power of LVEF in patients with acquired heart failure compared with ventilatory efficiency and natriuretic peptide levels is also well documented.17 Taken together with these findings, this suggests that symptom assessment and measurement of right ventricular volumes and function alone may not provide the optimum description of pathological processes in a particular patient and may not be the best measures to use when determining future risk. It would be helpful to clarify the relative prognostic power of these various measures in futures studies.
Natriuretic peptides and systemic right ventricular function
NT-proBNP levels were higher in the Mustard cohort compared with the Senning cohort. This perhaps reflects developing surgical practice, with earlier age at palliation using the Senning technique superseding the relatively later palliation age with the former technique. The trend towards significantly higher levels after adjusting for these variables, however, implies that an intrinsic disadvantage of Mustard surgery over the Senning operation cannot be excluded with confidence on these data.
NT-proBNP levels correlated significantly with RVEDVI and RVESVI. BNP levels have also been shown by others to relate to the severity of tricuspid regurgitation and right ventricular volume overload following Mustard and Senning surgery.18 19 Tulevski and colleagues11 included nine patients with TGA palliated by the Mustard or Senning procedure in a heterogenous group and found that RVEF, measured by CMR, correlated negatively to BNP levels (r=−0.65). Garg and coworkers20 found no correlation between RVEF and BNP in patients with systemic right ventricular physiology, although there was a strong correlation with atrial natriuretic peptide; a somewhat surprising result as both atrial natriuretic peptide and BNP correlate exceptionally closely (r=0.91, p<0.0001) in adults with congenital heart disease.10
Although similar relationships between BNP levels and right ventricular volumes have also been described in the context of subpulmonary right ventricular anatomy,21 Neffke et al12 found only a weak correlation between BNP and RVEDVI in a heterogeneous cohort with systemic or subpulmonary right ventricular physiology. This suggests that pathophysiological responses differ between various anatomical substrates, reflected in the mechanism of neurohormone production, and that these substrates might be better investigated in isolation.
Electrocardiographic markers and systemic right ventricular function
In patients with acquired heart disease, QRS duration is a measure of myocardial electromechanical dissociation, helps guide the prescription of resynchronisation therapy and is useful in assessing the risk of malignant arrhythmias.22 Prolongation of QRS duration and dispersion are likely to reflect the effects of chronic pressure overload and myocardial stretch on the systemic right ventricle and indicate electromechanical uncoupling here also. This study demonstrates the close relationship between measures of ventricular depolarisation and repolarisation with the volume and function of the systemic right ventricle, and indeed Gatzoulis and colleagues23 reported similar associations with respect to QRS duration in a comparable cohort. Neffke et al12 found that QRS dispersion manifested a greater rate of change over time than QRS duration in this context, describing that the latter measure did not significantly increase over 5 years follow-up. This is in contrast to Roos-Hesselink and colleagues4 who confirmed significant increases in QRS duration over long-term follow-up in a large cohort accompanying deterioration in systemic right ventricular function as measured by echocardiography.
Prolonged QTd has been shown to be an independent predictor of late arrhythmia in patients with Mustard palliation and, in combination with loss of sinus rhythm on the surface ECG, it has been associated with sudden cardiac death in adults with either Mustard or Senning palliation.24 Recently, Schwerzmann and colleagues25 suggested that a cut-off QRS duration of over 140 ms defines a group of adults after Mustard palliation at increased risk of malignant arrhythmia or sudden death. By contrast, Kammeraad and coworkers26 were unable to find differences in QRS duration, QTc or QTd between patients with TGA who had died suddenly and those who had not.
The relationships of these various ECG parameters to electromechanical dyssynchrony in the systemic right ventricle and to prognosis in palliated TGA therefore need further clarification. They may yet provide useful insights into the mechanisms that drive outcome and could help guide resynchronisation therapy in a group in which the evidence base for this therapy remains narrow.
Patients were only investigated at a single time point, and serial investigation may provide deeper insights into the dynamic interaction of the variables over time. Longitudinal follow-up would also provide information on the prognostic implications of these findings. Greater numbers of patients may also help bring further clarity to the effects of more complex anatomy and re-do surgery, factors that appear to presage worse outcome. Finally, CMR imaging was undertaken within 3 months of clinical assessment rather than contemporaneously, and it is possible that the relationships between CMR findings and ECG parameters and BNP levels were distorted as a result. This is a relatively brief period in the natural history of the condition, however, and no patients had a heart failure or arrhythmic event between these two time points, thus minimising this possible confounder.
This study of patients late after surgical palliation of TGA describes the close relationships between NT-proBNP, ECG parameters and measures of systemic right ventricular function. Circulating peptide levels and surface ECG parameters constitute safe, cost-effective and widely available surrogate markers of systemic right ventricular function and provide useful additional information on heart failure status. Both measures hold promise as prognostic markers and their association with long-term outcome should be determined.
Funding CMP is funded by a fellowship grant from the British Heart Foundation. AMT and PDL receive funding from the Higher Education Funding Council for England. This research was supported by the National Institute for Health Research.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.