Background
Assessing right ventricular (RV) performance is essential for patients with tetralogy of Fallot (TOF). The aim of this study was to investigate the reliability and validity of tricuspid annular plane systolic excursion (TAPSE) against cardiac magnetic resonance imaging measures and cardiopulmonary exercise testing.
Methods
A retrospective study was performed in 125 outpatients with repaired TOF with available protocol-driven echocardiography, cardiac magnetic resonance imaging, and exercise stress testing obtained as part of a cross-sectional study. TAPSE was measured on the two-dimensional apical four-chamber view on echocardiography by two readers. Multivariate linear regression was used to examine the association between TAPSE and measures of RV function and exercise capacity.
Results
The mean age was 12.6 ± 3.3 years, 41 patients (33%) were female, and 104 (83%) were white. TAPSE averaged 1.6 ± 0.37 cm, with an interreader intraclass correlation coefficient of 0.78 ( n = 18). TAPSE was significantly associated with cardiac magnetic resonance–based RV stroke volume after adjustment for gender and body surface area (β = 13.8; 95% confidence interval, 2.25–25.30; P = .02). TAPSE was not associated with cardiac magnetic resonance–based RV ejection fraction ( P = .77). On exercise testing, TAPSE was not associated with peak oxygen consumption, percentage of predicted oxygen consumption, oxygen pulse, or the ventilatory equivalent for carbon dioxide in patients with maximal exercise stress testing ( n = 73 [58%]).
Conclusions
TAPSE is reproducibly measured by echocardiography in patients with TOF. It is not associated with RV ejection fraction or exercise performance, and its association with RV stroke volume may be confounded by body size. On the basis of these results, TAPSE is not representative of global RV performance in patients with TOF.
Quantification of right ventricular systolic performance is vital given the growing number of survivors with congenital heart disease, in particular those with involvement of the right heart, such as tetralogy of Fallot (TOF). Patients operated on for TOF experience pulmonary insufficiency, which results in RV dilation and ultimately RV failure. Therefore, assessment of RV function is of paramount importance in the longitudinal follow-up of this patient population. The assessment of RV function by echocardiography in patients with TOF is challenging and currently is largely qualitative.
Tricuspid annular plane systolic excursion (TAPSE) is a well-described echocardiographic measure of RV function in the adult population without congenital heart diseases. Representing the distance of the tricuspid valve annular descent toward the apex of the heart during systole, this measurement reflects the longitudinal component of RV contraction. In normal subjects as well as in patients with pulmonary hypertension and other cardiac diagnoses, TAPSE is associated with global measures of RV systolic function, including RV ejection fraction (RVEF). This likely reflects the fact that longitudinal RV shortening accounts for the majority of total RV contraction in normal subjects. Studies in pulmonary hypertension and heart failure have shown that lower TAPSE is associated with a higher risk for death. Therefore, this measure could similarly be of use in congenital heart disease, in particular because TAPSE depends neither on endocardial border definition nor on geometric assumptions. Studies have examined the use of TAPSE in TOF and other congenital heart diseases. Of note, those studies have produced mixed results regarding TAPSE and RVEF; in addition, none have examined its association with maximal exercise capacity. Therefore, further investigation is needed to determine the usefulness of TAPSE to assess RV function in congenital heart disease.
We sought to examine the reliability of the interpretation of TAPSE and to determine its validity against measures of RV function by cardiac magnetic resonance (CMR) and indices of exercise performance in a large cohort with TOF.
Methods
Subjects
This was a retrospective study of participants in a cross-sectional protocol conducted at the Children’s Hospital of Philadelphia (RO1-HL74731) that recruited 177 patients and performed echocardiography, CMR, and exercise stress testing (EST). The study sample for this analysis included subjects with surgically treated TOF aged 8 to 19 years who underwent transthoracic echocardiography, CMR, and EST within a maximum of 3 months of one another. TAPSE measurements were retrospectively obtained in this study sample. The study was approved by the Children’s Hospital of Philadelphia Institutional Review Board.
Echocardiography
Echocardiography was performed using a Phillips iE33 machine (Phillips Medical Systems, Andover, MA) using a standard protocol. Briefly, images were acquired with 3-MHz to 8-MHz transducers, suited for patient size and acoustic windows and digitally stored using syngo Dynamics version 3.0 (Siemens Healthcare, Ann Arbor, MI). Although TAPSE was originally measured using two-dimensional (2D) tricuspid annular displacement, TAPSE is currently more commonly measured using M-mode echocardiography, by placing the cursor through the tricuspid annulus and measuring the amount of longitudinal motion of the annulus at peak systole. Because M-mode imaging was not available in this study population, we performed offline TAPSE measurements using 2D apical four-chamber views. We identified the lateral annulus of the tricuspid valve and measured the distance at end-diastole and end-systole to a point on the screen that was the same for systole and diastole. We measured the maximal excursion and used the electrocardiographic tracing as a guide to ensure that all measurements were obtained at the same time in the cardiac cycle. The distance the annulus travels vertically should be the same whether the point measured is the apex of the heart or the apex of the sector. To corroborate the technique we used in this study, we repeated measurements in 12 patients, comparing the measurement of the annulus with a point on the screen to measurement of the annulus to the apex. We found that the measurements are essentially the same (correlation coefficient = 0.98; P < .0001; intraclass correlation coefficient [ICC] = 0.99; 95% confidence interval [CI], 0.96–1.02).
Measurements were performed by two trained readers blinded to clinical information and to CMR and EST results. The tricuspid valve annulus was identified at end-systole and end-diastole using electrocardiographic guidance, and the distance was measured as has been previously reported ( Figure 1 ). The average of three TAPSE measurements was used for each patient. We randomly selected 10% of studies for blind reinterpretation by the same reader and 10% for interpretation by the second reader.
To assess the agreement of 2D TAPSE with M-mode measurements, we compared measurements performed using both techniques performed in a blinded fashion in 20 children without congenital heart disease.
CMR
CMR studies were performed using a 1.5-T Avanto magnetic resonance imaging scanner (Siemens Healthcare, Erlangen, Germany) with a six-channel phased-array body coil using a standard imaging protocol. The magnetic resonance imaging sequences included a steady-state free precession sequence through the cardiac apex. Sedation was used when appropriate according to patient age and ability to lie still for the scan. To assess RV end-systolic volume (RVESV) and RV end-diastolic volume (RVEDV), a cine magnetic resonance sequence in a short-axis view was used (echo time, 2.0 ms; repetition time, 4.5ms; flip angle, 75°–90°; matrix size, 196 × 196). RV stroke volume was calculated by subtracting RVESV from RVEDV. RVESV and RVEDV were indexed to body surface area (BSA), and ventricular size was compared with published normative data. RVEF was calculated from RV stroke volume divided by RVEDV. RVEF < 50% was defined as abnormal on the basis of prior studies in normal children. Pulmonary insufficiency was graded as mild if the regurgitant fraction was ≤20%, moderate if it was 20% to 40%, and severe if it was ≥40%. RV regional wall motion abnormalities were assessed qualitatively by examining four-chamber and short-axis views on cine CMR and recorded as present or absent. The RV outflow tract was assessed on the sagittal RV outflow tract view.
EST
Patients exercised to maximal ability using an electronically braked cycle ergometer (SensorMedics, Yorba Linda, CA). Eighteen subjects who were <130 cm tall exercised on a treadmill (Series 2000; Marquette, Milwaukee, WI). Metabolic data were obtained throughout the study and for the first 2 min of recovery on a breath-by-breath basis using a metabolic cart (V29; SensorMedics), including oxygen consumption (V o 2 ), carbon dioxide production (V co 2 ), maximum work (physical working capacity), oxygen pulse, respiratory exchange ratio, and the ventilatory equivalent of carbon dioxide (VE/V co 2 ). Anaerobic threshold was measured using the V-slope method. The predicted percentage of maximum V o 2 (V o 2 max%) was calculated for each patient, according to normative values for age, gender, and body size. Exercise performance was defined by peak V o 2 and V o 2 max%. Abnormal aerobic performance was defined as V o 2 max% < 80. Ventilatory efficiency was assessed by VE/V co 2 at the anaerobic threshold and by the slope of the VE/V co 2 relationship from initiation to peak exercise (VE/V co 2 slope). Maximal EST was defined as attaining a respiratory exchange ratio > 1.1.
Statistical Analysis
Continuous variables are presented as mean ± SD or as medians with interquartile ranges, as appropriate. Categorical variables are described using counts and percentages. Reproducibility of TAPSE interpretations and the agreement of 2D TAPSE measurements with M-mode measurements are expressed using ICCs with 95% CIs. Student’s t or Mann-Whitney U tests were used to compare continuous variables according to their distribution.
Multivariate linear regression was used to assess the association of TAPSE (the independent variable) with CMR and EST parameters of interest (the dependent variables). The final models included clinically relevant covariates that were significant ( P < .15) on univariate analysis or thought to be confounders, defined as a covariate that resulted in a >15% change in the TAPSE β coefficient. Body size was accounted for as BSA in the multivariate model. Receiver operating characteristic curve analysis was used to test the ability of TAPSE to predict RV dysfunction (RVEF < 50%) by CMR.
Analyses of exercise covariates included only those with maximal EST (respiratory exchange ratio > 1.1). All analyses were performed using Stata version 11.0 (StataCorp LP, College Station, TX). Statistical significance was defined as P < .05.
Results
Description of the Study Sample
A total of 177 individuals were recruited for the parent study. One hundred forty-one subjects completed all three studies of interest (echocardiography, CMR, and EST). Of these 141, 125 had interpretable TAPSE plus the other CMR and EST covariates of interest and therefore constituted the study sample. Most echocardiographic (70%) and CMR studies were performed within 2 weeks of each other, and all three studies were performed within 3 months of one another. There were no differences in age, gender, and race between those in the study sample ( n = 125) and those excluded ( n = 52). The time between echocardiography and CMR averaged 14 ± 25 days (range, 0–96 days). Half of the subjects underwent echocardiography and CMR on the same day ( n = 62). Those with echocardiography and CMR performed within 2 weeks of each other were comparable in terms of RV volumes, RVEF, TAPSE, and peak V o 2 with those who had studies performed >2 weeks and <3 months apart.
The mean age was 12.5 ± 3.3 years (range, 8–19 years), and 41 patients (33%) were female. All age groups were represented ( Figure 2 ). One hundred four (83%) were non-Hispanic white. Most subjects (76%) were repaired before 1 year of age and presented with pulmonary valve stenosis (79%), followed by atresia and absent pulmonary valve leaflets (15% and 6%, respectively). Surgical repair included a transannular patch in most subjects (67%) ( Table 1 ). No subjects had undergone pulmonary valve replacement at the time of the study.
| Variable | Value |
|---|---|
| Age (y) | 12.5 ± 3.2 |
| Male | 84 (67.2%) |
| Race/ethnicity | |
| Non-Hispanic white | 104 (83%) |
| African American | 15 (12%) |
| Asian | 6 (5%) |
| Weight (kg) | 42.8 ± 15.7 |
| Height (cm) | 147 ± 17.5 |
| Ideal body weight (kg) | 51.3 ± 11 |
| Body mass index (kg/m 2 ) | 19.8 ± 4.2 |
| BSA (m 2 ) | 1.31 ± 0.3 |
| Original pulmonary valve | |
| Stenosis | 99 (79%) |
| Atresia | 19 (15%) |
| Absent leaflets | 7 (6%) |
| Age at surgical repair (mo) | 4.4 (4.1–4.7) |
| Time elapsed since surgical repair (y) | 7.5 (5.2–10.7) |
| Use of transannular patch | 82 (66%) |
| TAPSE (cm) | 1.6 ± 0.38 |
On CMR, there was significant pulmonary insufficiency and RV dilation. Pulmonary insufficiency was severe in 40% of the subjects (regurgitant fraction > 40%) ( Table 2 ). RV volumes were increased but RV function was preserved, with a mean RVEF of 61 ± 8%. RV systolic dysfunction (RVEF < 50%) was present in 10% of patients. Despite overall preserved RVEF, we found average TAPSE to be 1.6 ± 0.38 cm (range, 0.75–2.75 cm), lower than that of children without congenital heart disease. Moreover, TAPSE was decreased in each age group in our study compared with published normal values ( Table 3 ). Regional wall motion abnormalities of the right ventricle were noted in most subjects (89%). Most subjects had mild or less than mild tricuspid regurgitation (mean regurgitant fraction, 5.6 ± 5%).
| Variable | Value |
|---|---|
| CMR | |
| RVEF (%) | 60.6 ± 8.3 |
| RVEF | |
| <50% ( n = 13) | 46 ± 4 |
| ≥50% ( n = 112) | 62 ± 7 |
| RV cardiac output (L/min) | 7.2 ± 2.5 |
| RV cardiac index (L/min/m 2 ) | 5.5 ± 1.5 |
| RV stroke volume (mL) | 92.8 ± 37 |
| RV stroke volume index (mL/m 2 ) | 70.3 ± 19.3 |
| RVEDV (mL) | 155.2 ± 63 |
| Indexed RVEDV (mL/m 2 ) | 117.7 ± 34.4 |
| RVESV (mL) | 62.4 ± 31.2 |
| Indexed RVESV (mL/m 2 ) | 47 ± 20 |
| Pulmonary regurgitant fraction (%) | 34.1 ± 16.73 |
| RV mass (g) | 91 (67–127) |
| RV mass (g/m 2 ) | 76 (58–93) |
| Pulmonary insufficiency ∗ | |
| Mild or less | 20 (16%) |
| Moderate | 49 (39%) |
| Severe | 50 (40%) |
| Regional wall motion abnormalities † | |
| Yes | 111 (89%) |
| No | 13 (10%) |
| Left ventricular ejection fraction (%) | 68.4 ± 7.1 |
| EST | Overall ( n = 125) | RER > 1.1 ( n = 73) | RER ≤ 1.1 ( n = 52) |
|---|---|---|---|
| Peak V o 2 (mL/kg/min) | 31.7 ± 8.41 | 33. 5 ± 7.8 | 29.2 ± 8.4 |
| V o 2 (L/min) | 1.34 ± 0.56 | 1.56 ± 0.53 | 0.96 ± 0.38 |
| V o 2 max% | 76 ± 18.4 | 80 ± 18 | 70 ± 18 |
| Maximum work (W) | 114 ± 48.6 | 128.1 ± 50 | 84.3 ± 29 |
| Oxygen pulse (mL/beat) | 7.4 ± 3 | 8.58 ± 2.9 | 5.63 ± 2.1 |
| RER | 1.14 ± 0.13 | 1.22 ± 0.9 | 1.02 ± 0.07 |
| VE/V co 2 at anaerobic threshold | 39 ± 6.9 | 37.8 ± 6.61 | 42.3 ± 6.73 |
| VE/V co 2 slope | 37.9 ± 8.23 | 35.7 ± 6 | 41.6 ± 10.1 |
| Measured anaerobic threshold | 89 (68%) | 65 (86%) | 24 (48%) |
∗ Not assessed in six subjects.
| Patients with TOF | Normal controls | ||||||
|---|---|---|---|---|---|---|---|
| Age (y) | n | Mean | SD | n | Mean | SD | P |
| 8 | 15 | 1.45 | 0.2720 | 23 | 1.97 | 0.1525 | <.0001 |
| 9 | 17 | 1.48 | 0.3456 | 20 | 2.01 | 0.1425 | <.0001 |
| 10 | 20 | 1.51 | 0.3518 | 27 | 2.05 | 0.13 | <.0001 |
| 11 | 8 | 1.79 | 0.4137 | 25 | 2.1 | 0.1325 | <.0001 |
| 12 | 10 | 1.45 | 0.2424 | 18 | 2.14 | 0.1475 | <.0001 |
| 13 | 15 | 1.58 | 0.2554 | 20 | 2.2 | 0.1725 | <.0001 |
| 14 | 9 | 1.60 | 0.4491 | 35 | 2.26 | 0.195 | <.0001 |
| 15 | 9 | 1.74 | 0.3479 | 25 | 2.33 | 0.205 | <.0001 |
| 16 | 11 | 1.72 | 0.5198 | 34 | 2.39 | 0.2 | <.0001 |
| 17 | 12 | 1.93 | 0.4984 | 27 | 2.45 | 0.21 | <.0001 |
| 18 | 7 | 1.62 | 0.1502 | 21 | 2.47 | 0.215 | <.0001 |
On EST, the overall cohort had decreased aerobic performance (mean V o 2 max%, 76 ± 18.4%). Seventy-three subjects (58%) achieved maximum effort. In those, the peak V o 2 was 33.7 ± 7.9 mL/kg/min (V o 2 max%, 80 ± 17%) compared with 29.3 ± 8.4 mL/kg/min (V o 2 max%, 70 ± 18%) in subjects who achieved submaximal EST ( P < .001).
Comparisons among TAPSE, CMR, and EST
TAPSE measurements showed excellent intrareader reliability and interreader reliability (intrareader 1 ICC = 0.85 [95% CI, 0.68–1]; intrareader 2 ICC = 0.66 [95% CI, 0.27–1]; interreader ICC = 0.78 [95% CI, 0.52–1]). In addition, we found excellent agreement between 2D and M-mode TAPSE measured in a sample of 20 patients without congenital heart disease (ICC = 0.90, P < .001), confirming that measurements by these two different methods are essentially identical.
TAPSE was associated with stroke volume after adjustment for sex and BSA (β = 13.8; 95% CI, 2.25–25.30; P = .02; Table 4 ). There also appeared to be direct associations between TAPSE and RVEDV and RVESV. There was no significant association between TAPSE and RVEF ( Figure 3 ). Similarly, subanalysis limited to patients with pulmonary valve stenosis demonstrated no association of TAPSE and RVEF (β = −0.78, P = .72).
