Author + information
- Received January 12, 1996
- Revision received August 3, 1996
- Accepted August 13, 1996
- Published online December 1, 1996.
- Willem A. Helbing, MD*,⁎,
- R. Andre Niezen, MD*,†,
- Saskia Le Cessie, MSc*,
- Rob J. Van Der Geest, MSc*,
- Jaap Ottenkamp, MD* and
- Albert De Roos, MD*
- ↵⁎From the Address for correspondence: Dr. Willem A. Helbing, Department of Pediatric Cardiology, University Hospital Leiden, J-6-S, Albinusdreef 2, P.O. Box 9600, 2333 AA Leiden, The Netherlands.
Objectives We sought to assess right ventricular diastolic function in young patients with corrected tetralogy of Fallot and pulmonary regurgitation.
Background Pulmonary regurgitation is an important problem in repair of tetralogy of Fallot. Its effects on right ventricular diastolic function in children are unknown.
Methods Nineteen children with repair of tetralogy of Fallot (mean age [±SD] 12 ± 3 years, mean age at operation 1.5 ± 1) and 12 healthy children were studied. Summation of magnetic resonance velocity mapping pulmonary and tricuspid volume flow curves provided right ventricular time-volume curves. Ventricular size was assessed with tomographic magnetic resonance imaging (MRI). Graded exercise testing was performed.
Results Systematic and random differences (mean ± SD) of velocity mapping and Doppler tricuspid time to peak velocities (peak E: 1 ± 26 ms, r = 0.43; peak A: 2 ± 11 ms, r = 0.76), E/A ratios (0.04 ± 0.5, r = 0.63) and duration of pulmonary regurgitation (20 ± 35 ms, r = 0.74) were satisfactory. In 6 patients (group I), late diastolic forward pulmonary artery flow was absent; in 13 patients (group II), this flow contributed 1% to 14% to right ventricular stroke volume. Significant differences were increased deceleration time (315 ± 91 vs. 168 ± 28 ms, p < 0.001), decreased filling fraction (44 ± 11 vs. 55 ± 16%, p = 0.02) and increased peak early filling rate (378 ± 124 vs. 286 ± 112 ml/s, p = 0.018) between control subjects and group I, and increased deceleration time (230 ± 40, p = 0.03) between control subjects and group II. Pulmonary regurgitation, ventricular size and ejection fraction did not differ significantly between patient groups. Exercise function was diminished with restrictive right ventricular physiology (p < 0.001, group II vs. control subjects).
Conclusions Impaired relaxation and restriction to filling affect diastolic right ventricular function in children with repair of tetralogy of Fallot and pulmonary regurgitation. Restrictive right ventricular physiology is associated with decreased exercise function.
Surgical repair of tetralogy of Fallot can be performed with low mortality and acceptable long-term results (1). Current surgical policy, aimed at early complete surgical correction, results in considerable transannular patch rates, which induces residual pulmonary regurgitation (2). The negative effects of pulmonary regurgitation on exercise capacity and right ventricular systolic function have been well documented, and in some patients restoration of pulmonary valve competence may eventually be required (3,4). Recently it has been suggested that restriction to late diastolic filling of the right ventricle, indicated by diastolic forward flow in the pulmonary artery after atrial contraction, may reduce pulmonary regurgitation and increase exercise capacity in adult patients with repair of tetralogy of Fallot (5). This observation underlines the need for adequate assessment of diastolic right ventricular function in such patients. In clinical practice, transtricuspid Doppler echocardiography is used for this purpose. However, if ventricular filling occurs from more than a single source, as with pulmonary regurgitation, right ventricular early diastolic filling cannot be quantified adequately with this method (5). Furthermore, data on right ventricular diastolic function of patients with repair of tetralogy of Fallot operated on with recent surgical techniques in infancy and early childhood are lacking. Magnetic resonance velocity mapping has been validated extensively as a practical method to assess flow volume, including tricuspid flow and pulmonary artery flow in postoperative patients with repair of tetralogy of Fallot with pulmonary regurgitation (6–9).
The present study aimed to assess characteristics of right ventricular diastolic function in young patients with tetralogy of Fallot, in the presence of pulmonary regurgitation, and to compare these characteristics with those in healthy children.
- Abbreviations and Acronymns
- = electrocardiogram, electrocardiographic
- = magnetic resonance imaging
Study subjects. Twelve healthy volunteers (mean [±SD] age 12 ± 3 years, range 7 to 15) and 19 children with repair of tetralogy of Fallot, all with residual pulmonary regurgitation (mean age 12 ± 3 years, range 6 to 16 years) were studied. Patients were selected randomly from the institutional data base, excluding those with known residual intracardiac shunting or pulmonary stenosis with a Doppler echocardiography gradient >30 mm Hg, but without prior knowledge of the examiners with regard to other demographic, procedural or hemodynamic data. Characteristics of all children are given in Table 1. The patients were classified into two subgroups on the basis of the occurrence of late diastolic forward flow in the pulmonary artery as assessed with magnetic resonance imaging (MRI): group I included those without and group II those with late diastolic forward flow (5). Surgical correction had been performed at a mean age of 1.5 ± 1 years. Patients were studied at a mean age of 10 ± 3 years postoperatively. Details of the operative procedures are given in Table 1. All operations were performed with a transatrial-transpulmonary approach, using cardiopulmonary bypass and profound hypothermia. In 10 patients (9 from group II), a transannular patch was used. All patients were in New York Heart Association functional class I, except for one patient in group I and two in group II, who were in class II. All subjects were in sinus rhythm during the examinations. All but two patients (one from each patient group) had complete right bundle branch block. All children were studied without sedation. The study protocol was approved by the committee on medical ethics of our institution, and children were entered in the study after informed consent had been obtained.
Echocardiography. Transthoracic two-dimensional and Doppler echocardiography were performed within2hofthe MRI examination, using a Hewlett-Packard Sonos 2500 ultrasound machine with 5.5- and 3.5-MHz transducers. Pulmonary artery and tricuspid flow velocities were recorded using pulsed wave Doppler echocardiography. If flow velocities exceeded 2 m/s, continuous wave Doppler was used. Pulmonary recordings were obtained from the parasternal short-axis view and tricuspid measurements from the apical four-chamber view. All measurements were performed during normal quiet respiration. An electrocardiogram and respiratory signal were recorded simultaneously. Spectral recordings were made with minimal filtering at a speed of 100 cm/s and stored on magnetooptical disk. Doppler echocardiography measurements of all heartbeats in two complete consecutive respiratory cycles were analyzed and averaged to minimize the influence of respiration on outcome.
Magnetic resonance imaging. The MRI studies were performed using a Gyroscan NT15 system (Philips Medical Systems, Best, The Netherlands) operating at 1.5-T. Spin echo localizing views were used to determine the position of the ventricles and the orientation of the planes for velocity mapping, perpendicular to the direction of measured flows. Volume of flow into or out of the right ventricle was assessed with magnetic resonance velocity mapping of flow across the tricuspid orifice and the main pulmonary artery. Magnetic resonance velocity mapping is a modified gradient-echo sequence that uses a velocity-encoding magnetic field gradient in the direction of flow (10). For the tricuspid valve, the imaging plane was positioned halfway between the end-diastolic and end-systolic position of the valve annulus, and for the pulmonary artery halfway between the pulmonary valve “annulus” and the bifurcation of the main pulmonary artery. Velocity maps were acquired with a flip angle of 20°, an echo time of 12 ms, a section thickness of 8 mm and a field of view of 300 × 300 mm. A 128 × 128 scan matrix was reconstructed to a display matrix of 256 × 256. Through-plane flow was encoded at a maximal velocity of 1 to 3 m/s for the pulmonary artery and 0.8 to 2.5 m/s for the tricuspid valve, anticipated from the flow velocity from previous Doppler echocardiography measurements and cine-loop displays of gradient-echo images. The cardiac cycle was divided in 30 time frames, with temporal resolution ranging from 21 to 36 ms, depending on the individual child's heart rate.
Furthermore, ventricular volumes were measured from a multisection image set. Details of this technique have been previously reported (11). In brief, a transverse stack of 10 to 12 of multiphase gradient-echo magnetic resonance images was acquired with a flip angle of 50°, an echo time of 4 to 10 ms and a section thickness of 8 to 10 mm with an 0.8- to 1-mm slice gap. Temporal resolution ranged from 22 to 35 ms.
Data acquisition of tomographic MRI was triggered to the R wave of the ECG. Retrospective ECG gating was used for the velocity mapping measurements (10). Because most of the patients had right bundle branch block, particular attention was given to adequate ECG registration during the studies, to ascertain that triggering occurred on the first deflection of the QRS complex.
Magnetic resonance image analysis. Analyses of the magnetic resonance images were performed using the Magnetic Resonance Analytical Software System (MASS), implemented on a SUN IPC workstation (12). Window and level settings were standardized during the analysis.
Volumetric flow measurements. Analysis of the magnetic resonance flow velocity studies was performed using the FLOW software package (13). Volume of flow was calculated by separately tracing a region of interest along the inner borders of the tricuspid orifice or pulmonary artery wall in each time frame of a velocity map series. For every time frame instantaneous volumetric flow was calculated by a computer algorithm by multiplying spatial average flow velocity with the area of the region of interest. Integration of all instantaneous volumetric flow data yielded total volume flow per cardiac cycle.
Ventricular volumes and mass. Ventricular volume was calculated by summation of ventricular cavity areas, assessed by manual tracing of the endocardial border on a stack of gradient-echo image sections of a specific time frame, and multiplied by section thickness, with correction for the interslice gap. Papillary muscles and the moderator band were not included in the ventricular area. On end-systolic frames, epicardial contours were drawn to determine right ventricular free wall mass, excluding the interventricular septum. Ventricular wall volume was calculated as myocardial area (epicardial minus endocardial area) multiplied by the sum of slice and interslice gap thickness. A specific gravity of 1.05 g/ml was used for calculation of ventricular mass.
Calculations. Assuming that changes in right ventricular volume are equal to the sum of the volume of flow entering or leaving through the tricuspid valve and pulmonary artery, right ventricular time-volume change curves were reconstructed by summation of the volume flow data from the pulmonary artery and tricuspid valve. Trigger delay from time 0 (in ms) was plotted on the x-axis and summated volume flow (ml/s) from the pulmonary artery and tricuspid valve for a specific time frame on the y-axis. If trigger delays of the same time frame number on both flow curves were not identical, the summated flow volume was assigned to a trigger delay time calculated by averaging the respective trigger delays. Thus, time on the x-axis represented duration of a complete reconstructed heartbeat, averaged from 128 to 256 actual beats. Furthermore, timevolume curves were obtained by integration of the timevolume change curves (Fig. 1). From these curves the following indexes of diastolic function were derived: 1) peak early filling rate (ml/s); 2) time to peak filling rate (ms) = time from end-systole to maximal rate of increase of right ventricular volume before atrial contraction; 3) filling fraction (%) = volume increase (ml) during the first one-third of diastole, normalized to right ventricular stroke volume (=systolic flow volume derived from the summated volume flow curves); 4) deceleration time, calculated, analogous to Doppler echocardiography practice, on the time-volume change curve by extrapolation of the deceleration of flow to the baseline and measuring the time from peak early filling to the extrapolation of deceleration to the baseline (14); 5) peak atrial (A) filling rate (ml/s); 6) time to peak A filling rate = time from onset of acceleration of increase of right ventricular volume after atrial contraction to the maximal increase in volume after atrial contraction; 7) atrial filling fraction (%) = volume increase after atrial contraction, normalized to right ventricular stroke volume; and 8) the ratio of peak early and peak atrial filling rate (Fig. 1).
Stroke volume was defined as end-diastolic (maximal ventricular) volume minus end-systolic (minimal) volume. Ejection fraction was calculated as stroke volume divided by end-diastolic volume. Ventricular volume indexes were calculated as volume divided by the body surface area.
Validation of volume flow curves. The accuracy of volumetric flow measurements with magnetic resonance velocity mapping has been validated extensively (6–9). The volumetric accuracy of the combined right ventricular time-volume curves was assessed by comparing stroke volume, derived from this combined curve, to stroke volume calculated from tomographic right ventricular volumes. Furthermore, the accuracy of the tricuspid and pulmonary artery velocity mapping curves was validated by comparison with Doppler echocardiography tracings (Table 2).
Exercise testing. Seventeen patients (5 from group I, 12 from group II) underwent graded exercise testing. Those <12 years old performed treadmill exercise according to a modified Bruce protocol; those ≥12 years old performed bicycle ergometer tests, increasing work load with 20 W/min. Patients exercised until exhaustion. Maximal heart rates were recorded from the ECG.
Statistical analysis. Results are expressed as mean value ± SD. Results were compared between groups using analysis of variance and analysis of covariance, correcting for heart rate and body surface area and also for right ventricular end-diastolic volume if appropriate. To avoid the problem of multiple comparison testing, pairwise comparisons, using the Tukey least significant difference procedure, were made in case only the overall group effect was significant. Agreement of tomographic and velocity mapping magnetic resonance results and of velocity mapping and Doppler echocardiography was assessed by linear regression analysis. Furthermore, the systematic and random differences between corresponding measurements were calculated as average difference ± SD of the (paired) differences (15). Results of the exercise tests in patients were compared to percentiles of the distribution of exercise function in a healthy Dutch pediatric population, using a one-sample chi-square test (16). Percentages were compared between groups using Fisher exact tests. Exercise percentiles were correlated with magnetic resonance data using Kendall's tau. Correlation of hemodynamic measurements was assessed using linear regression analysis. All analyses were performed using the SPSS-PC statistical software package. A p value <0.05 was considered to indicate statistical significance.
Echocardiography. Intracardiac shunting was excluded in all children by echocardiography. Nonphysiologic pulmonary regurgitation and late diastolic forward flow in the pulmonary artery throughout respiration were absent in healthy children. In 11 patients from group II, late diastolic forward flow in the pulmonary artery throughout the entire respiratory cycle was noted, with a maximal flow velocity of 40 ± 12 cm/s (range 18 to 60) and velocity time integral of 3.2 ± 1.07 cm (range 1.3 to 5.2) (pulsed wave Doppler measurements). In two group II patients, this flow pattern was present only during inspiration. Minimal residual pulmonary stenosis was present in 12 patients (gradients of 10 to 20 mm Hg) and moderate pulmonary stenosis in 2 (gradients of 25 and 30 mm Hg). Minimal tricuspid regurgitation was documented in 14 patients and moderate tricuspid regurgitation in 1 (flow velocities 160 to 300 cm/s), with no differences between the two patient groups. Transtricuspid peak E velocities and E deceleration times were similar for patients and control subjects (Table 3). In patients, peak A velocity was significantly higher than that in control subjects, and the E/A ratio was significantly lower (Table 3). Differences between group I and II patients were not statistically significant.
Magnetic resonance imaging. Because of motion artifacts, one tomographic study (group I) was of insufficient quality. Therefore, the tomographic data reported are from 5 group I patients and 13 group II patients.
Ventricular volumes and mass. Right ventricular end-diastolic volume index was significantly larger in patients than control subjects (Table 4) (control subjects: 71 ± 13 ml/m2; patients: 106 ± 19 ml/m2 for group I [p = 0.007], 129 ± 40 ml/m2 for group II [p < 0.001]). Right ventricular ejection fraction was significantly lower in both patient groups than healthy children (control subjects: 66 ± 6%; patients: 54 ± 7% for group I [p = 0.01], 51 ± 11 for group II [p < 0.001]). Right ventricular free wall mass in healthy children was significantly less than that in both patient groups (controls: 17 ± 2 g/m2; patients: 25 ± 7 gm/m2 for group I [p = 0.025], 26 ± 6 g/m2 for group II [p 5 0.004]). Right ventricular size, mass and ejection fraction did not differ significantly between groups I and II.
Left ventricular end-diastolic volume in group II patients was significantly larger than that in healthy children. For group I patients, left ventricular end-diastolic volume did not differ significantly from that in control subjects (Table 4). Left ventricular ejection fraction was significantly smaller in both patient groups than control subjects, but differences between groups I and II were not statistically significant (Table 4).
Validation of magnetic resonance flow measurements. Right ventricular stroke volume derived from the combined tricuspid and pulmonary artery magnetic resonance velocity mapping volume of flow curves and from tomographic MRI ventricular volumes showed excellent agreement (regression equation: Velocity mapping = 5.3 + 0.89 × Tomographic stroke volume, r = 0.93, p < 0.001; systematic difference: 22.7 ± 8.5 ml). Comparison of magnetic resonance velocity mapping and Doppler echocardiography variables showed acceptable correlation and limited systematic differences (Table 2). Agreement for time to peak velocities ranged from + ± 26 ms (r = 0.43, p = 0.035) for time to peak E wave velocity to 9 ± 22 ms (r = 0.82, p < 0.001) for time to peak systolic pulmonary artery velocity (Table 2). Peak velocity measurements also correlated well but with clear underestimation by magnetic resonance velocity mapping: Systematic differences ranged from 9 ± 9 cm/s (r = 0.74, p < 0.001) for peak A wave velocity to 12 ± 11 cm/s (r = 0.48, p = 0.042) for peak E wave velocity.
Doppler echocardiography and velocity mapping agreed with regard to the presence of late diastolic anterograde pulmonary artery flow in the 13 patients from group II. This flow pattern occurred directly after the a wave of the ECG in all Doppler measurements. With velocity mapping, late diastolic anterograde pulmonary flow was noted to begin at time frame 27 or 28 (depending on heart rate 66 to 140 ms before the R wave [time 0]) in 10 patients, and in time frame 30 (22 to 35 ms before time 0) in 3. These three patients all had minimal late diastolic forward flow, and in two of them, Doppler studies showed this flow pattern only during inspiration. In the 10 patients in whom velocity mapping showed late diastolic forward flow in more than one time frame, the maximal flow velocity (ml/s) was reached at time frame 28 or 29. A rapid decline of forward pulmonary artery flow at end-diastole, as noted with Doppler measurements, clearly separating late diastolic and early systolic flow, was not observed with velocity mapping.
When the time of maximal trigger delay (time frame 30) of the pulmonary and tricuspid velocity mapping curves of each child were compared, the overall mean difference was 7 ± 32 ms (p = NS). With 30 time frames, the difference between 2 subsequent time frame numbers on either curve was 0.4 ± + ms.
Pulmonary flow. In the 13 patients from group II, the volume of late diastolic anterograde pulmonary artery flow ranged from 0.5% to 14% (mean 3.6 ± 3.5%) of right ventricular stroke volume. Contribution of late diastolic flow to right ventricular output was <4% in 11 patients, 7% in + and 14% in another. Pulmonary regurgitation volume in group I was 33 ± 21% (range 8% to 51%) of right ventricular stroke volume and 42 ± 14% (range 17% to 67%) in group II (p = NS). Duration of pulmonary regurgitation did not differ significantly between groups (292 ± 178 ms in group II, 359 ± 83 in group II, p = NS). The extent of pulmonary regurgitation (in ml/m2) correlated significantly with right ventricular enddiastolic volume index (r = 0.78, p < 0.001) and right ventricular ejection fraction (r = 20.53, p = 0.023) but not with right ventricular mass.
In the patients with repair of tetralogy of Fallot, the distribution of flow velocities across the imaged cross section of the pulmonary artery was inhomogeneous (nonlaminar), particularly at end-systole, early diastole and after atrial contraction.
Tricuspid regurgitation. Tricuspid regurgitant volume was 1% and 5% of right ventricular stroke volume in two group I patients and from 2% to 9% in four group II patients (p = NS).
Indexes of right ventricular diastolic function. Indexes of right ventricular diastolic function are summarized in Table 5. In both patient groups, the filling fraction during the first one-third of diastole was lower than that in the normal group (control subjects: 55.1 ± 15.7%; patients: 44 ± 10.5% in group 1,48.4 ± 10.8% in group II), but the difference was statistically significant only between control subjects and group I (p = 0.02). For group I, significant differences with control subjects were observed for deceleration time (control subjects: 168 ± 28 ms; group I: 315 ± 91 ms, p < 0.001). By definition, late diastolic forward flow in the pulmonary artery was found only in group II patients. For this patient group, a significant increase compared with control subjects was observed for peak early filling rate (control subjects: 286 ± 112 ml/s; group II: 378 ± 124 ml/s, p = 0.018) and deceleration time (control subjects: 168 ± 28 ms; group II: 230 ± 40 ms, p = 0.004). For group I versus group II, only deceleration time differed significantly (p = 0.03). No differences between the three groups of children were observed for time to peak filling rate, peak A filling rate, atrial filling fraction or the ratio of peak early and peak A filling rate (Table 5).
Late diastolic pulmonary forward flow did not correlate significantly with right ventricular or left ventricular size or ejection fraction, right ventricular mass or the amount of pulmonary regurgitation. A significant correlation was observed between peak filling rate and pulmonary regurgitation (r = 0.69, p = 0.001). No other variable of diastolic function correlated significantly with the amount of pulmonary regurgitation or with diastolic forward flow in the pulmonary artery or the right ventricular wall mass. The amount of late diastolic pulmonary flow, as expressed in the Doppler velocity time integral, showed poor correlation with volumetric assessment of late diastolic pulmonary flow (r = 0.08, p = NS).
Exercise function. Exercise tolerance in group II patients was diminished compared to that in healthy children (p < 0.001) and group I patients (p = 0.04) (Table 6). Exercise function in group I did not differ significantly from control subjects. Exercise performance was weakly associated with the amount of pulmonary regurgitation (tau-0.5, p = 0.01) but not with right or left ventricular volume or ejection fraction. Maximal heart rates (173 ± 16 beats/min for group 1,176 ± 18 beats/min for group II) did not differ significantly from heart rates in the normal population (p = 0.065).
Right ventricular diastolic function in patients with repair of tetralogy of Fallot. In the present study, we demonstrated abnormalities in right ventricular diastolic function in young patients with residual pulmonary regurgitation after surgical correction of tetralogy of Fallot. The filling pattern in group I, characterized by prolonged early filling, as indicated by longer deceleration times, and low early filling rates, is consistent with impaired relaxation (14,17–19). When restriction to filling occurs, early filling rates increase and deceleration time decreases, as was observed in group II (17,19). Differences in diastolic function between the patient groups were limited and most likely reflect a continuum of changes, with signs of impaired relaxation and restriction to filling present in both groups (14,17). In patients with predominantly restrictive right ventricular physiology, exercise performance was clearly diminished. This result could only be explained by differences in diastolic function, not by differences in ventricular size, ejection fraction or amount of pulmonary regurgitation. No indexes of diastolic function could directly be related to right ventricular wall mass.
Use of magnetic resonance velocity mapping time-volume curves. Assuming that changes in right ventricular volume can be calculated as the sum of the volume of flow entering or leaving the tricuspid valve and pulmonary artery, right ventricular time-volume curves were obtained by summation of magnetic resonance velocity mapping pulmonary and tricuspid volume flow data. With velocity mapping, flow velocity is measured at numerous pixels along the entire cross section of a vessel or valve orifice. Area changes of this region of interest can easily be accounted for in analysis. Thus, accurate assessment of the actual volume of flow throughout the cardiac cycle is possible (6,7,9). Although not obtained simultaneously, with steady heart rates, net pulmonary and tricuspid flow volumes during the 128 to 256 heart beat acquisition period are equal, which allows summation of flow data. This was confirmed by the high agreement between right ventricular stroke volume from tomographic data and the reconstructed time-volume curves.
Agreement for timing of systolic and diastolic events measured with Doppler echocardiography and velocity mapping MRI was acceptable (Table 2). With velocity mapping, the onset of late diastolic anterograde pulmonary flow was somewhat delayed with respect to the ECG a wave compared with Doppler measurements, and correlation between velocitymap-ping volume of flow and Doppler velocity time integrals was poor. This finding may be explained by two factors: 1) With nonlaminar flow patterns, as occurred at end-diastole in the pulmonary artery of the patients with repair of tetralogy of Fallot, Doppler echocardiography is not accurate enough to assess volume of flow. If forward flow in the center of a vessel is detected with this technique, this does not mean that there is simultaneous net forward displacement of a certain volume of flow across the entire vessel area (10). 2) Retrospective ECG gating is required for velocity mapping, and temporal resolution differs from that of Doppler echocardiography, which results in relatively large random differences when comparing timing of events (Table 2). Furthermore, small errors in retrospective ECG gating are difficult to rule out and may partly explain why anterograde pulmonary flow was noted only in the last time frame before the R wave in three patients (with minimal amounts of this flow). Nevertheless, Doppler echocardiography and velocity mapping were in agreement with regard to the presence of this flow pattern in all group II patients.
Assuming that time-flow velocity curves adequately reflect time-volume change curves, Doppler echocardiography is widely used for assessment of ventricular filling dynamics (14). Doppler flow patterns have been validated as indexes of diastolic function against invasive data (14,17). For the right ventricle, tricuspid flow patterns are generally used. However, with pulmonary regurgitation, the transtricuspid flow pattern alone does not adequately reflect right ventricular filling (5), as was confirmed in our study. Caval vein and pulmonary artery flow patterns provide additional but only semiquantitative information (5,14).
In previous reports (7,10), differences in sampling site, temporal resolution, spatial orientation and the effects of throughplane motion have consistently reduced magnetic resonance maximal velocity levels compared with those of Doppler echocardiography, particularly for transtricuspid flow. More important, the ratios of peak velocities (E/A ratio) have not been affected (7,9). This finding suggests that magnetic resonance velocity mapping volume flow patterns may be interpreted comparable to Doppler echocardiographic results.
Reconstruction of time-volume curves can be achieved with tomographic MRI (20). However, this process requires extremely time-consuming manual tracing of the endocardial outline in 200 to 250 images (20). We have observed considerable interobserver and intraobserver variation for ventricular volumes of children (5% to 12%) (10). This variation makes accuracy of measurements of small volume changes questionable. Furthermore, most tomographic techniques do not provide data during late diastole. Magnetic resonance velocity mapping has distinct advantages: Observer variability, in our experience and that of others, is less (~2% to 3%); with retrospective gating, data are also provided in late diastole; and time required for analysis is considerably less than with tomographic techniques (7,9,10).
Right ventricular mass measurements. At present, MRI is the only available method to assess right ventricular wall mass in patients, and its accuracy has been validated (21,22). Reference data in children with this technique are lacking, but results in healthy children in the current study were similar to those from autopsy data (23).
Comparison with previous studies. The increase in right ventricular wall mass observed in our patients is in accordance with histologic data from similar patients (24). Ventricular filling patterns are influenced by complex interactions between myocardial relaxation, ventricular compliance, contractile state, heart rate, age and loading conditions (14). Noninvasive imaging methods, including the method used in the present study, do not allow clear differentiation between the effects of ventricular dynamics and loading conditions. Ventricular hypertrophy, increased ventricular size and decreased ejection fraction, residual pulmonary stenosis and intraventricular conduction disturbances have been shown (14,18,19,20,25,26) to negatively influence diastolic filling by altering both ventricular compliance and relaxation. These factors may in part explain the differences observed in diastolic properties between patients and control subjects but not between patient groups. With pulmonary regurgitation, loading conditions are affected by residual pulmonary valve function and pulmonary vascular resistance. Differences in residual pulmonary valve function, as present in group I and II patients, were accounted for in our calculations of the amount of pulmonary regurgitation but may influence the pattern of right ventricular filling. This influence has to be taken into account when comparing diastolic right ventricular function in patients with pulmonary regurgitation.
Consistent with our results, restriction to filling has been associated with diminished functional state in left ventricular disease states (14,17,19). Remarkably, the contrary was noted with right ventricular restrictive physiology in a recent Doppler echocardiography study of adult patients with repair of tetralogy of Fallot with pulmonary regurgitation who underwent operation 15 to 35 years ago (5). In that study, improved exercise performance (percent predicted peak oxygen consumption) was attributed to reduction of pulmonary regurgitation by the “restrictive” right ventricle and by the contribution of diastolic forward pulmonary flow to cardiac output (5). Our results confirm the applicability of this flow pattern as a marker of restrictive physiology, but the volumetric contribution of this flow to cardiac output was limited in most of our patients. Although restrictive physiology appeared to be unfavorable in young patients with repair of tetralogy of Fallot, it is noteworthy that in the “restrictive” group (group II), pulmonary regurgitation was not increased relative to that in group I, despite the higher percentage of transannular patches in group II. Therefore the hypothesis of Gatzoulis et al. (5) of a myocardial restrictive process that limits pulmonary regurgitation cannot be rejected. Whether this restrictive process is a stationary, intrinsic factor in some patients, or a dynamic adaptation to chronic volume overload that eventually may reduce pulmonary regurgitation and improve exercise function, is not clear from available data. Distinction between these possible mechanisms is hampered by the lack of data on postoperative changes in myocardial architecture and how these are affected by preoperative factors, including age at operation, extent of hypoxemia, surgical techniques and chronic loading abnormalities (24,27).
Clinical implications. The direct association with exercise function and the potential of diastolic function to limit pulmonary regurgitation call for adequate quantification of diastolic function, ventricular size and the amount of pulmonary regurgitation in the follow-up of patients with repair of tetralogy of Fallot. None of these factors can be fully appreciated with (Doppler) echocardiography. The combination of velocity mapping and tomographic MRI provides a practical method to serially assess these data in clinical practice.
Limitations of the study. Ideally, assessment of relaxation should include isovolumetric relaxation time. However, in many patients with repair of tetralogy of Fallot, actual pulmonary valve closure is absent, preventing measurement of this factor. Increased filling pressures augment the effects of restriction to filling and may mask signs of impaired relaxation (14,25). Apart from pulmonary valve function, pulmonary regurgitation may be influenced by pulmonary vascular resistance. We did not obtain invasive data in the present study. However, because right ventricular end-diastolic pressure is usually normal or only slightly elevated in postoperative patients with repair of tetralogy of Fallot (28), and pulmonary vascular resistance is usually normal in the absence of (residual) pulmonary stenosis, it is unlikely these factors would explain the differences observed. With current MRI techniques, beat to beat information cannot be obtained. Therefore, the effects of respiration, an important modulator of right ventricular function, could not be assessed.
Conclusions. Right ventricular diastolic function in young patients with repair of tetralogy of Fallot is characterized by impaired relaxation and restriction to filling. Restrictive right ventricular filling is associated with diminished exercise tolerance. With pulmonary regurgitation, right ventricular diastolic function can be quantified with magnetic resonance velocity mapping.
The valuable help in MRI analysis of Sietske Meinesz, BSc, is gratefully acknowledged.
* Financial support to Dr. Niezen was provided by the Interuniversity Cardiology Institute of The Netherlands (ICIN), Utrecht, The Netherlands.
- Received January 12, 1996.
- Revision received August 3, 1996.
- Accepted August 13, 1996.
- American College of Cardiology
- Gatzoulis MA,
- Clark AL,
- Cullen S,
- Newman CGH,
- Redington AN
- Rebergen SA,
- Chin JGJ,
- Ottenkamp J,
- Van der Wall EE,
- de Roos A.
- van der Geest RJ,
- Jansen E,
- Buller VGM,
- Reiber JHC.
- van der Geest RJ,
- Buller VGM,
- Reiber JHC.
- Binkhorst RA,
- van’t Hof MA,
- Saris WHM.
- Nishimura RA,
- Holmes DR,
- Reeder GS,
- Tajik AJ,
- Hatle LK
- Vanoverschelde J-LJ,
- Raphael DA,
- Robert AR,
- Cosyns JR
- Suzuki J-I,
- Chang J-M,
- Caputo GR,
- Higgins CB
- Lentner C
- Hegerty A,
- Anderson RH,
- Deanfield JE.