Author + information
- Received April 4, 2002
- Revision received July 17, 2002
- Accepted August 19, 2002
- Published online December 4, 2002.
- Periklis A. Davlouros, MD*,
- Philip J. Kilner, MD, PhD†,
- Tim S. Hornung, MD*,
- Wei Li, MD, PhD*,
- Jane M. Francis, DCR(R)†,
- James C.C. Moon, MD†,
- Gillian C. Smith, BSe†,
- Tri Tat, PhD‡,
- Dudley J. Pennell, MD, FACC† and
- Michael A. Gatzoulis, MD, PhD, FACC*,* ()
- ↵*Reprint requests and correspondence:
Dr. Michael A. Gatzoulis, Royal Bromptom Hospital, Sydney Street, London, SW3 6NP, United Kingdom.
Objectives We examined the relationship among biventricular hemodynamics, pulmonary regurgitant fraction (PRF), right ventricular outflow tract (RVOT) aneurysm or akinesia, and baseline and surgical characteristics in adults with repaired tetralogy of Fallot (rTOF).
Background The precise relationship of pulmonary regurgitation with biventricular hemodynamics has been hampered by limitations of right ventricular (RV) imaging.
Methods We assessed 85 consecutive adults with rTOF and 26 matched healthy controls using cardiovascular magnetic resonance imaging.
Results Patients had higher right ventricular end-diastolic volume index (RVEDVi) (p < 0.001), right ventricular end-systolic volume index (RVESVi) (p < 0.001), right ventricular mass index (RVMi) (p < 0.001), and lower right ventricular ejection fraction (RVEF) (p < 0.001) and left ventricular ejection fraction (LVEF) (p = 0.002) compared to controls. The PRF (range 0% to 55%) independently predicted RVEDVi (p < 0.01) and the latter predicted RVESVi (p < 0.01) and RVMi (p < 0.01). The RVOT aneurysm/akinesia was present in 48/85 (56.9%) of patients and predicted RV volumes (RVEDVi, p = 0.01, and RVESVi, p = 0.03). There was a negative effect of RVOT aneurysm/akinesia and RVMi on RVEF (p < 0.01 and p = 0.02, respectively). There was only a tendency among patients with transannular or RVOT patching toward RVOT aneurysm/akinesia (p = 0.09). The LVEF correlated with RVEF (r = 0.67, p < 0.001).
Conclusions Pulmonary regurgitation and RVOT aneurysm/akinesia were independently associated with RV dilation and the latter with RV hypertrophy late after rTOF. The RVOT aneurysm/akinesia was common but related only in part to RVOT or transannular patching. Both RV hypertrophy and RVOT aneurysm/akinesia were associated with lower RVEF. Left ventricular systolic dysfunction correlated with RV dysfunction, suggesting an unfavorable ventricular-ventricular interaction. Measures to maintain or restore pulmonary valve function and avoid RVOT aneurysm/akinesia are mandatory for preserving biventricular function late after rTOF.
Repair of tetralogy of Fallot (rTOF) is associated with excellent prognosis (1–3). However, morbidity and mortality rise over the long term (4). Pulmonary regurgitation (PR) is the most common lesion postrepair and has been associated with exercise intolerance, atrial and ventricular arrhythmia, and sudden cardiac death (2,5–9). Pulmonary regurgitation relates to right ventricular outflow tract (RVOT) reconstruction and, in particular, the usage of a transannular patch during repair (7). As a result, there has been a modification of early surgical management toward: 1) preservation of pulmonary valve function, whenever possible; and 2) limiting the extent of patching when a transannular type of repair is necessary. It remains a surgical challenge, however, to adequately relieve RVOT obstruction without inducing significant PR, aiming for the optimal balance between the two. Furthermore, the wide spectrum of RVOT and pulmonary artery anatomy encountered in tetralogy clearly necessitates an individualized approach for each patient at repair. For the older patient with previous rTOF, pulmonary valve replacement (PVR) may lead to improved right ventricular (RV) volumes and function, improved functional class, stabilization of QRS duration, and a reduction in atrial and ventricular arrhythmia (8,10–13). However, the optimal time for late PVR remains unclear, and this has been hampered by limitations in serial quantification of PR and RV function (14). We employed cardiovascular magnetic resonance (CMR), the gold standard noninvasive imaging technique (15), to examine the relationship among biventricular hemodynamic indices, pulmonary regurgitation fraction (PRF), RVOT aneurysm or akinesia, and baseline and surgical characteristics in adults with rTOF.
We studied prospectively 101 consecutive patients attending the Royal Brompton Adult Congenital Heart Programme age ≥15 years, following ethics approval and informed consent. Surgical details were obtained from operative notes. Patients who underwent pulmonary valve implantation for severe PR and/or RVOT obstruction after rTOF (n = 16) were excluded. All patients underwent clinical examination, transthoracic echocardiogram, and CMR imaging on the same day. Twenty-six age- and gender-matched healthy volunteers were similarly studied with CMR.
Scans were performed using a 1.5-tesla Siemens Sonata system (Siemens Medical Solutions, Erlangen, Germany). A TrueFISP cine was acquired in an oblique sagittal plane aligned with the RVOT before velocity mapping and short-axis cine acquisitions. Volume and mass measurements were made on TrueFISP breath-hold cine acquisitions in multiple short-axis slices covering both ventricles from base to apex (in-plane resolution 1.4 × 2.2 mm, temporal resolution 35 ms). We used 7-mm-slice thickness, starting with a slice aligned with the most basal myocardium of left ventricle (LV) and RV at end-diastole as visualized on four-chamber and vertical long-axis cines, then working down to the apex. This resulted in 12 to 16 relevant slices at end-diastolic, and usually one less at end-systole. This approach is usual for ventricular volume analysis by magnetic resonance imaging in adults (16).
Regurgitant fraction (diastolic reversed flow expressed as a percentage of forward flow) was measured from phase-velocity maps in the main pulmonary artery and aortic root. Two experienced cardiologists reviewed the RVOT cines in the sagittal and short-axis cine images of the RV for assessment of akinesia or dyskinesia. Akinesia was defined as lack of thickening during systole in >10% of the RV muscle perimeter. Dyskinesia (“aneurysm”) was defined as outward movement during systole of part of the ventricular wall or its reconstructed outflow tract (Fig. 1). Image analysis was performed by manual segmentation using CMR Tools (Imperial College, London, UK) and areas in adjacent slices summated to give volume measurements according to Simpson’s rule. Ventricular mass was calculated by outlining the myocardium in multiple slices and multiplying the summed volume in cc by 1.05 (myocardial specific gravity) to give the mass in grams. The RV was defined as the mass of the RV myocardium, measured from the junction between the RV free wall and the interventricular septum on each slice from the base to the apex. The RV muscular trabeculations were traced separately and included in RV mass calculations; LV mass was defined as the sum of LV free wall and interventricular septum mass. All values for the CMR-derived volume and mass indices were indexed to body surface area (m2).
Values are expressed as mean ± SD. Independent two-tailed Student ttest was used for comparison between patients and controls and between patient subgroups. For comparison of more than three subgroups, one-way analysis of variance with Bonferroni correction (or Dunnett’s T3for variables with unequal variances) was used. Chi-square (χ2) test was used for comparison of categorical variables. Pearson’s correlation was used for variables with normal distribution. Variables not normally distributed were logarithmically transformed. Multivariate linear regression analysis was employed for definition of possible independent predictors for outcome variables. Only variables significant in univariate analysis were included in the multivariate models. Parameter estimates (partial regression coefficients-b) with 95% confidence intervals and level of statistical significance (p) are presented. SPSS for windows (version 10.0.1, SPSS Inc., Chicago, Illinois) was used for data analysis.
Demographics, surgical data, and current hemodynamics for patients and controls are presented in Table 1. In 15 patients details on the surgical technique used for RVOT reconstruction were not available. Biventricular indices of function are presented in Table 2, whereas independent predictors of CMR-derived indices are presented in Table 3.
PR and peripheral pulmonary stenosis
The PRF ranged from 0% to 55% (24.4 ± 16.4%). Nine patients (10%) had no detectable PR by CMR. In patients with transannular patch, PRF was significantly higher compared to those without any patch (33 ± 9.5% vs. 18.6 ± 18.2%, Dunnett’s T3, p = 0.003), whereas no significant difference existed in PRF between the two patch subgroups (transannular vs. RVOT; Fig. 2A). Transannular patching was the only significant independent predictor of PR severity (Table 3). Twelve patients (14.1%) had moderate to severe peripheral pulmonary artery stenosis, six at the site of a previous Waterston shunt. The PRF tended to be higher in these patients, but the difference was not significant (27 ± 18% vs. 23 ± 16%, p = NS), whereas right ventricular mass index (RVMi) was higher compared to the remainder (60.7 ± 16.7 g/m2vs. 48.9 ± 12.8 g/m2, p < 0.01).
RVOT aneurysm and akinesia versus transannular and RVOT patching
A total of 48/85 (56.4%) patients had either RVOT aneurysm (n = 16, 18.8%) or akinesia (n = 32, 37.6%). No statistically significant difference existed in the incidence of either RVOT aneurysm or akinesia among patients who underwent RVOT reconstruction with transannular patch, RVOT patch, or without usage of a patch (Fig. 3A). The cumulative incidence of RVOT aneurysm and akinesia in patients repaired without patch was marginally lower compared to patients repaired with transannular patch (17/35 vs. 14/19, χ2p = 0.07, Fig. 3A), and lower but not significantly different compared to patients with any patch repair (17/35 vs. 24/35, χ2p = 0.09, Fig. 3B). Significant differences existed in the mean right ventricular end-diastolic volume index (RVEDVi) and right ventricular end-systolic volume index (RVESVi) in the three RVOT reconstruction subgroups (RVOT, transannular, and no patch), but no difference was seen between the two patch subgroups (Fig. 2A). The combined subgroup of RVOT and transannular patch had significantly higher PRF, RVEDVi, and RVESVi from the nonpatched group (Fig. 2B).
All RV indices were deranged in patients compared to controls. Both PRF and RVOT aneurysm/akinesia predicted RVEDVi. The RVOT aneurysm/akinesia and RVEDVi predicted RVESVi (Table 3). Patients with RVOT aneurysm/akinesia had higher RVEDVi and RVESVi and lower right ventricular ejection fraction (RVEF) from the remainder (Fig. 2C).
No significant difference was seen in LV volumes between patients and controls. However, the left ventricular end-systolic volume index (LVESVi) to end-diastolic volume index (LVEDVi) ratio was significantly higher in patients. Current age, aortic regurgitant fraction, and RVEDVi were predictive of LVEDVi. The latter, along with RVEF and period that patients remained palliated with an arterial shunt, predicted LVESVi. Fifteen patients had a residual, small, and restrictive ventricular septal defect. All 15 had a left-to-right shunt with a Doppler-derived peak velocity ≥4 m/s. Furthermore, there were no differences in means of LV-CMR indices between patients with a ventricular septal defect and the remainder.
The RVMi correlated weakly with PRF (r = 0.42, p < 0.01). Controlling for the latter, RVMi correlated with RVEDVi (r = 0.3, p < 0.01), RVESVi (r = 0.35, p < 0.01), and peripheral pulmonary stenosis (r = 0.3, p < 0.01). The most significant independent predictors of RVMi were peripheral pulmonary stenosis and RVEDVi.
No direct relation existed between RVEF and PRF in patients as a whole, nor was there such a relation in the subgroups with and without RVOT aneurysm/akinesia. The RVEF correlated with RVEDVi (r = −0.3, p = 0.01), RVESVi (r = −0.7, p < 0.01), RVMi (r = −0.3, p = 0.007), and RVOT aneurysm/akinesia (r = −0.3, p = 0.002). Both RVMi and RVOT aneurysm/akinesia were the only significant independent predictors of RVEF. Left ventricular ejection fraction (LVEF) was significantly lower in patients than in controls and correlated with RVEF (r = 0.67, p < 0.01, Fig. 4). Length of time that patients remained palliated, aortic regurgitant fraction, and RVEF were independent predictors of LVEF.
This study provides data on the spectrum of PR and on biventricular volumes, mass, and function from a large cohort of adults with rTOF. A concept addressed and examined here is the relationship of RV volumes and function with RVOT aneurysmal or akinetic regions. The latter, in combination with chronic PR, emerge as the main predisposing factors for RV dysfunction late after rTOF. Furthermore, this study demonstrated the presence of LV dysfunction in adults with rTOF and provides insights into causative mechanisms.
Pulmonary regurgitation was present in the vast majority—but not all—of our adult patients. Reconstruction of the RVOT with a transannular patch was associated with higher PRF in accord with previous reports (17). There was a tendency toward increased PRF in patients with peripheral pulmonary stenosis (18,19); these patients had significantly higher RV mass.
The RV mass was increased in patients compared to controls. We hypothesize that RV dilation—secondary to chronic PR—and RV hypertrophy were adaptive mechanisms for preservation of RVEF and maintenance of a low ventricular systolic wall stress, as in aortic regurgitation (20). Peripheral pulmonary stenosis augments this hypertrophic response. The inverse relationship between RVEF and RVMi observed suggests that RV contractility may be adversely affected by this hypertrophy. Demand ischemia, accompanying fibrosis, altered ventricular geometry, and/or affected electromechanical coupling could be the underlying mechanisms of this adverse interaction.
RV function, RVOT aneurysm/akinesia, and patch repair
The RVEF did not relate directly to PR. Although PR was a significant predictor of RVEDVi, and this in turn of RVESVi, there was an additional independent predictor of RV dilation and systolic dysfunction, namely RVOT aneurysm/akinesia. Aneurysm and akinesia of the RVOT in rTOF have received limited attention in the literature (21,22). Two studies reported a high incidence of RVOT aneurysms in similar patient cohorts, which related to sustained ventricular tachycardia underscoring their proarrhythmic role (12,23). Our study extends this negative effect of RVOT aneurysms and demonstrates a strong relation between the latter and/or RVOT akinesia with RV dysfunction. Redington et al. (24)showed angiographically a lack of direct relation between degree of PR and RVEF. The researchers speculated that impaired RV systolic function may be secondary to a noncontractile RVOT patch rather than the direct effect of PR. Our study confirms this hypothesis that a noncontractile RVOT region contributes to decreased RVEF. Furthermore, we have shown that RVOT contractile dysfunction is not necessarily related to the usage of a patch; RVOT akinesia or aneurysm was present in a significant number of patients (n = 17/35, 48.5%) who did not undergo a patch type of repair, begging the question whether other factors, such as extreme myectomy (infundibular resection) and/or ischemic insult (perhaps due to conal branch interruption), are also responsible for the genesis of RVOT aneurysm/akinesia. Further support for this comes from Atallah-Yunes et al. (25), who reported less RV dilation and preserved RV systolic function late after rTOF with a modified approach for relieving RVOT obstruction, employing a short infundibular resection and avoiding extensive myectomy. Furthermore, Miura et al. (26)showed reduced wall motion of the upper part and other areas of the RV at rest and during isoproterenol infusion. These wall motion abnormalities were thought to be responsible for a reduction in angiographic RVEF and were more common following transventricular versus transatrial repair. Only two patients from our study underwent a transatrial repair (neither of them manifested RVOT aneurysm/akinesia); hence, we cannot answer this question.
Recent studies have shown RV patching (RVOT or transannular) to be a significant predictor of late adverse events after rTOF (4,27). In keeping with the observations of d’Udekem et al. (27), we found increased RVEDVi, RVESVi, and PRF in patients with transannular or RVOT patching compared to patients without patch repair, whereas no such differences existed between the two patch subgroups. Our data suggest that RVOT reconstruction with either transannular or RVOT patching may have a detrimental long-term effect because a) this predisposes to more severe PR and b) patch repair—perhaps in conjunction with other factors like RVOT myectomy—leads to RVOT aneurysm/akinesia. We would submit that preservation of pulmonary valve function and avoidance of or limiting RVOT or transannular patching and perhaps avoidance of extensive RVOT myectomy are likely to preserve long-term RV systolic function.
Furthermore, ventricular function may not fully recover following late, elective pulmonary valve implantation unless RVOT aneurysmal or akinetic regions are specifically addressed. Therrien et al. (14)recently reported no improvement in RVEF, assessed with radionuclide angiography, following late PVR in adults with previous rTOF. However, radionuclide angiography is not sensitive in detecting RVOT aneurysmal or akinetic regions. Moreover, even in patients without RVOT aneurysm/akinesia, RVEF may fail to identify early reversible RV systolic dysfunction, and serial assessment of RV diastolic and particularly systolic volumes may be required. We concur with Therrien et al. (14), therefore, on the need for prospective assessment of the effects of pulmonary valve replacement on biventricular function in these patients with CMR.
LV function and RV-to-LV interaction
The LVEF was significantly lower in patients, compared to controls. Relative increase of LVESVi in relation to LVEDVi in patients may represent an unfavorable LV remodeling process associated with decreased LVEF. Our study showed three independent predictors of LVEF: length of time that patients remained palliated, aortic regurgitant fraction, and RVEF; the first two correlated inversely with LVEF. Both volume loading from arterial shunts in palliated patients and relative hypoxia may have affected LV contractility (28). Aortic regurgitation, in turn, its pathogenic mechanisms and its negative effect on LV function, which were suggested in our study, need further investigation. The most significant predictor of LVEF was, nevertheless, RVEF. This suggests a ventricular-ventricular interaction, whereby RV dilation and dysfunction are both associated with LV dysfunction. Indeed, an independent positive association existed between RVEDVi and LVEDVi. The latter was an independent predictor of LVESVi. Furthermore, an independent negative relation existed between RVEF and LVESVi. Kondo et al. (29)previously suggested RV volume overload as the underlying mechanism for LV dysfunction during exercise. We have shown that this adverse interaction between RV and LV volumes and function is also present at rest. Altered patterns of septal systolic motion, patching of the septum with resultant akinesia, septal fibrosis, and/or demand ischemia with globally reduced LV contractility may all be responsible. Moreover, myocardial injury at the time of repair could contribute to long-term LV (and RV) dysfunction. It is unlikely, however, that the latter was the main predisposing factor for LV dysfunction as there were only 6/59 (10%) patients with normal RVEF and reduced LVEF (with a tendency toward longer palliation among them).
Two previous studies by Niezen et al. (6,30)have assessed LV hemodynamics employing CMR in rTOF. In their second study (30), examining a group of patients similar to ours, the investigators reported a lower LVEDVi in patients and no difference in LVEF between patients and controls. Furthermore, no correlation existed between RV and LV volumes in their study. Shorter length of follow-up from repair, overall preserved RVEF in the study by Niezen et al. (30), and the larger number of patients enrolled in our study may explain these differences.
Our data clearly demonstrate that a longer period of palliation with an arterial shunt and late aortic regurgitation irrespective of cause, predispose to LV dysfunction. Repair of tetralogy in infancy and avoidance of long-term palliation with arterial shunts may, therefore, preserve late LV function. Furthermore, there was a direct RV and LV interaction underscoring the importance of preserving RV function for multiple long-term benefits (9,11,14), including maintenance of LV function.
Our study is limited by its cross-sectional design. Prospective studies are required to define further the causative mechanisms of RVOT aneurysm/akinesia and record their longitudinal course together with the course of PR after rTOF. Additional predictors of biventricular dysfunction may exist and be identified with a larger patient sample and longer periods of observation, in both older cohorts—like ours—and contemporary ones, where different surgical strategies are employed. Furthermore, future studies need to address the effects of drug therapy and late catheter and surgical intervention on biventricular function and the right-to-left interaction reported here.
There is a spectrum of PR—quantified by CMR—in adults late after rTOF, associated with RV dilation and RV hypertrophy. RVOT aneurysm/akinesia is common and contributes to increased RV systolic volumes and decreased RVEF, irrespective of the degree of PR. These aneurysmal or akinetic regions relate only in part to RVOT or transannular patching. Left ventricular systolic dysfunction exists in adults with rTOF relating to the length of palliation with arterial shunts, aortic regurgitation, and RV dilation and dysfunction. Measures to maintain or restore pulmonary valve function and to avoid RVOT aneurysm/akinesia are mandatory for preserving RV and LV function late after tetralogy repair.
☆ Dr. Davlouros was supported by the Clinical Research Committee of the Royal Bromptom Hospital and by the Greek Cardiology Society. Dr. Hornung was supported by the Waring Trust, Royal Bromptom Hospital. The Cardiac Magnetic Resonance Unit is supported by the British Heart Foundation.
- cardiovascular magnetic resonance
- left ventricle/ventricular
- left ventricular end-diastolic volume index (ml/m2)
- left ventricular ejection fraction (%)
- left ventricular end-systolic volume index (ml/m2)
- total left ventricular mass index (g/m2)
- pulmonary regurgitation
- pulmonary regurgitant fraction (%)
- pulmonary valve replacement
- repair of tetralogy of Fallot
- right ventricle/ventricular
- right ventricular end-diastolic volume index (ml/m2)
- right ventricular ejection fraction (%)
- right ventricular end-systolic volume index (ml/m2)
- right ventricular mass index (g/m2)
- right ventricular outflow tract
- Received April 4, 2002.
- Revision received July 17, 2002.
- Accepted August 19, 2002.
- American College of Cardiology Foundation
- Norgaard M.A.,
- Lauridsen P.,
- Helvind M.,
- Pettersson G.
- Nollert G.,
- Fischlein T.,
- Bouterwek S.,
- Bohmer C.,
- Klinner W.,
- Reichart B.
- Roos-Hesselink J.,
- Perlroth M.G.,
- McGhie J.,
- Spitaels S.
- Therrien J.,
- Siu S.C.,
- Harris L.,
- et al.
- Therrien J.,
- Siu S.C.,
- McLaughlin P.R.,
- Liu P.P.,
- Williams W.G.,
- Webb G.D.
- Rebergen S.A.,
- Chin J.G.,
- Ottenkamp J.,
- van der Wall E.E.,
- de Roos A.
- Chaturvedi R.R.,
- Kilner P.J.,
- White P.A.,
- Bishop A.,
- Szwarc R.,
- Redington A.N.
- Bonow R.O.,
- Carabello B.,
- de Leon A.C.,
- et al.
- Harrison D.A.,
- Harris L.,
- Siu S.C.,
- et al.
- Redington A.N.,
- Oldeshaw P.J.,
- Shinebourne E.A.,
- Rigby M.L.
- Niezen R.A.,
- Helbing W.A.,
- van Der Wall E.E.,
- van Der Geest R.J.,
- Vliegen H.W.,
- de Roos A.