Background
Transthoracic echocardiographic estimates of peak systolic pulmonary artery pressure are conventionally calculated from the maximal velocity of the tricuspid regurgitation (TR) jet. Unfortunately, there is insufficient TR to determine estimated peak systolic pulmonary artery pressure (EPSPAP) in a significant number of patients. To date, in the absence of TR, no noninvasive method of deriving EPSPAP has been developed.
Methods
Five hundred clinically indicated transthoracic echocardiograms were reviewed over a period of 6 months. Patients with pulmonic stenosis were excluded. Pulsed-wave Doppler was used to measure pulmonary artery acceleration time (PAAT) and right ventricular ejection time. Continuous-wave Doppler was used to measure the peak velocity of TR (TR Vmax ), and EPSPAP was calculated as 4 × TR Vmax 2 + 10 mm Hg (to account for right atrial pressure). The relationship between PAAT and EPSPAP was then assessed.
Results
Adequate imaging to measure PAAT was available in 99.6% of patients (498 of 500), but 25.3% (126 of 498) had insufficient TR to determine EPSPAP, and 1 patient had significant pulmonic stenosis. Therefore, 371 were included in the final analysis. Interobserver variability for PAAT was 0.97. There were strong inverse correlations between PAAT and TR Vmax ( r = −0.96), the right atrial/right ventricular pressure gradient ( r = −0.95), and EPSPAP ( r = −0.95). The regression equation describing the relationship between PAAT and EPSPAP was log 10 (EPSPAP) = −0.004 (PAAT) + 2.1 ( P < .001).
Conclusions
PAAT is routinely obtainable and correlates strongly with both TR Vmax and EPSPAP in a large population of randomly selected patients undergoing transthoracic echocardiography. Characterization of the relationship between PAAT and EPSPAP permits PAAT to be used to estimate peak systolic pulmonary artery pressure independent of TR, thereby increasing the percentage of patients in whom transthoracic echocardiography can be used to quantify pulmonary artery pressure.
Estimates of pulmonary artery pressure obtained using transthoracic echocardiography (TTE) are most commonly performed using continuous-wave Doppler to measure the maximum velocity of tricuspid regurgitation (TR Vmax ). This technique of pulmonary artery pressure quantification has been widely adopted for clinical and research purposes because it provides a noninvasive, direct estimate of right ventricular (RV) systolic pressure (RVSP) that correlates closely with invasive hemodynamic measurement. In the absence of significant RV outflow tract obstruction, this measurement provides an accurate means of obtaining estimated peak systolic pulmonary artery pressure (EPSPAP). However, this method of pulmonary artery pressure assessment is not feasible when TR is absent or trivial.
Alternative transthoracic echocardiographic methods for pulmonary artery pressure estimation have been proposed. These include the measurement of blood flow through an anatomic defect (ventricular septal defect, patent ductus arteriosus, or aortopulmonary shunt ), measurement of the peak systolic and end-diastolic pulmonic valve regurgitant velocity, and measurement of the pulmonary artery acceleration time (PAAT). PAAT measurement, previously shown to correlate with invasively measured mean pulmonary artery pressure (MPAP), is a particularly attractive alternative to the TR Vmax -dependent method, because it does not rely on the presence of an anatomic defect or valvular regurgitation and is therefore measurable in the vast majority of individuals. Currently, measurement of PAAT cannot be used as an alternative to TR Vmax for the derivation of EPSPAP, because the relationship between these two parameters has not been described.
We hypothesized that PAAT would correlate strongly with TR Vmax and thus could be used to quantify EPSPAP. To address this hypothesis, we examined the relationship between PAAT and both TR Vmax and EPSPAP in a large unselected group of patients undergoing clinically indicated TTE.
Methods
All aspects of this study were approved by our institutional review board (Partner’s Human Research Committee). Five hundred (of a total of 9,200) clinically indicated transthoracic echocardiographic studies performed in the Massachusetts General Hospital echocardiography laboratory over a 6-month period (March to September 2009) were randomly selected for retrospective review. To arrive at this total number of participants, a research staff member who was blinded to patient age, gender, and study indication randomly selected 84 studies from each month of the study period for formal review. Once an a priori established enrollment total of 500 patients was reached, study enrollment ended. As detailed below, TR Vmax , PAAT, and RV ejection time (RVET) were measured on each study. Studies that did not contain adequate imaging for the measurements of each of these variables were excluded from the final analysis. In addition, patients with significant pulmonic valvular stenosis, as defined by a continuous-wave peak jet velocity ≥2 m/sec across the pulmonic valve, were excluded from analysis. Studies were performed using commercially available echocardiographic systems (iE33, Philips Medical Systems, Andover, MA; Vivid 7, GE Healthcare, Milwaukee, WI).
To assess the reproducibility of our findings, a separate validation cohort was then constructed. Specifically, 100 consecutive patients referred for clinically indicated TTE who had measurable TR signals (50 from the Massachusetts General Hospital echocardiography laboratory [Boston, MA] and 50 from the Scarborough Hospital echocardiography laboratory [Toronto, ON, Canada]) were prospectively assessed in a manner identical to that used during creation of the derivation cohort.
All measurements were made offline (Xcelera; Philips Medical Systems). Heart rate and rhythm were assessed on the final image of each study. TR Vmax was defined as the maximal velocity of the TR jet measured using continuous-wave Doppler in the echocardiographic view that yielded the most complete TR Doppler profile. EPSPAP was calculated using the modified Bernoulli equation: 4 × TR Vmax 2 + 10 mm Hg (to account for right atrial pressure ). The use of 10 mm Hg as an estimate of right atrial pressure yields similar correlates with catheter-based estimates of RVSP as a method using the Doppler gradient plus mean jugular venous pressure. Pulsed-wave and continuous-wave Doppler interrogation of the proximal pulmonary artery was then performed in the parasternal short-axis view with the sample volume placed at the annulus of the pulmonary valve ( Figures 1 and 2 ). We chose to place the pulsed-wave sample volume at the pulmonary valve annulus and not more proximally in the RV outflow tract to maximally align blood flow and Doppler interrogation. The same view was used to exclude pulmonic stenosis. PAAT, RVET, and heart rate were measured using the pulse-wave Doppler profile. PAAT was defined as the interval between the onset of systolic pulmonary arterial flow and peak flow velocity ( Figure 2 ). RVET was defined as the interval between the onset of RV ejection to the point of systolic pulmonary arterial flow cessation. All values used for analysis represent the average of three consecutive cardiac cycles, with the exception of patients with atrial fibrillation, in whom five-beat averages were obtained. All measurements were performed by three study cardiologists (A.L.B., P.N., and K.Y.).
Statistical Analysis
Echocardiographic data are reported as mean ± SD with the corresponding range of values. The correlation between PAAT and EPSPAP was examined using Pearson’s or Spearman’s method as appropriate for data distribution. Histograms of PAAT and EPSPAP were constructed and tested for normality using the Shapiro-Wilks test. Although both variables were found to have significant skew, transforming RVSP to its log 10 value fulfilled assumptions for linear regression. PAAT was then plotted against log 10 (RVSP), and linear regression was performed. Interobserver variability for PAAT was assessed using linear regression and Bland-Altman analysis for 50 random individuals from the cohort.
Results
Among this cohort of randomly selected clinical studies, pulsed-wave Doppler imaging of the main pulmonary artery was sufficient to measure PAAT in 99.6% of patients (498 of 500). In contrast, 25.2% (126 of 500) did not have sufficient TR to measure TR Vmax . One individual was excluded from analysis because of the presence of significant pulmonary valve stenosis. As such, 371 individuals were retained in the final analysis.
In the final cohort ( n = 317), the mean age was 63 ± 17 years (range, 17–96 years), and 51% were men (189 of 371). The mean heart rate during TTE was 70 ± 10 beats/min (range, 42–112). Atrial fibrillation or flutter was present in 8% of patients (30 of 371), while the remaining 92% were in normal sinus rhythm. Indications for TTE included assessment of left ventricular function, congestive heart failure, or cardiomyopathy ( n = 108); assessment of suspected or established valvular heart disease ( n = 89); dyspnea of uncertain etiology ( n = 53); chest pain of uncertain etiology ( n = 46); fever of uncertain etiology ( n = 23); syncope ( n = 13); and other ( n = 39). In 11% of studies (41 of 371), ordering providers specifically requested assessments of RVSP or pulmonary artery pressure.
Group mean values and ranges for all TR-derived and pulmonary artery–derived flow parameters are shown in Table 1 . The interobserver variability for PAAT measurement was 0.97. By Bland-Altman analysis, the interobserver variability was no more than ±7.5 msec ( Figure 3 ). The group mean TR Vmax of 282 ± 65 cm/sec (range, 150–480 cm/sec) yielded a mean EPSPAP of 44 ± 16 mm Hg (range, 19–102 mm Hg). The mean PAAT was 124 ± 34 msec (range, 43–186 msec), and the mean RVET was 362 ± 51 msec (range, 216–580 msec). There were strong inverse correlations between PAAT and TR Vmax ( r = −0.96), the right atrial/RV pressure gradient (4 × TR Vmax 2 ; r = −0.95), and EPSPAP ( r = −0.95) ( Figure 4 ). Adjustment of PAAT for RVET ( r = −0.89) and heart rate ( r = −0.91) did not lead to stronger correlations with EPSPAP. The relationship between PAAT and log 10 (EPSPAP) is shown in Figure 5 . The highly significant ( P < .001) regression equation describing this relationship was log 10 (EPSPAP) = −0.004(PAAT) + 2.1.
| Variable | Mean ± SD (range) |
|---|---|
| TR-derived values | |
| TR Vmax (cm/sec) | 282 ± 65 (150–480) |
| RA/RV gradient (mm Hg) | 34 ± 16 (9–92) |
| EPSPAP (mm Hg) | 44 ± 16 (19–102) |
| Pulmonary artery–derived values | |
| PAAT (ms) | 124 ± 34 (43–186) |
| Log 10 (PAAT) | 2.07 ± 0.14 (1.64–2.27) |
| PA Vmax (cm/sec) | 89 ± 20 (44–175) |
| RVET (msec) | 362 ± 51 (216–580) |
| PAAT/RVET (m/sec 2 ) | 0.35 ± 0.11 (0.08–0.64) |
