The ratio of tricuspid regurgitation velocity (TRV) to the time-velocity integral of the right ventricular outflow tract (TVI RVOT ) has been studied as a reliable measure to distinguish elevated from normal pulmonary vascular resistance (PVR). The equation TRV/TVI RVOT × 10 + 0.16 (PVR echo ) has been shown to provide a good noninvasive estimate of PVR. However, its role in patients with significantly elevated PVR (> 6 Wood units [WU]) has not been conclusively evaluated. The aim of this study was to establish the validity of the TRV/TVI RVOT ratio as a correlate of PVR. The role of TRV/TVI RVOT was also compared with that of a new ratio, TRV 2 /TVI RVOT , in patients with markedly elevated PVR (>6 WU).
Data from five validation studies using TRV/TVI RVOT as an estimate of PVR were compared with invasive PVR measurements (PVR cath ). Multiple linear regression analyses were generated between PVR cath and both TRV/TVI RVOT and TRV 2 /TVI RVOT . Both PVR echo and a new derived regression equation based on TRV 2 /TVI RVOT : 5.19 × TRV 2 /TVI RVOT – 0.4 (PVR echo2 ) were compared with PVR cath using Bland-Altman analysis. Logistic models were generated, and cutoff values for both TRV/TVI RVOT and TRV 2 /TVI RVOT were obtained to predict PVR > 6 WU.
One hundred fifty patients remained in the final analysis. Linear regression analysis between PVR cath and TRV/TVI RVOT revealed a good correlation ( r = 0.76, P < .0001, Z = 0.92). There was a better correlation between PVR cath and TRV 2 /TVI RVOT ( r = 0.79, P < .0001, Z = −0.01) in the entire cohort as well as in patients with PVR > 6 WU. Moreover, PVR echo2 compared better with PVR cath than PVR echo using Bland-Altman analysis in the entire cohort and in patients with PVR > 6 WU. TRV 2 /TVI RVOT and TRV/TVI RVOT both predicted PVR > 6 WU with good sensitivity and specificity.
TRV/TVI RVOT is a reliable method to identify patients with elevated PVR. In patients with TRV/TVI RVOT > 0.275, PVR is likely > 6 WU, and PVR echo2 derived from TRV 2 /TVI RVOT provides an improved noninvasive estimate of PVR compared with PVR echo .
The noninvasive determination of pulmonary vascular resistance (PVR) has been the subject of multiple reports. In 2003, we reported a novel index obtained by Doppler echocardiography, which reliably distinguished between patients with elevated and normal PVR. Invasively, PVR is derived from the ratio of transpulmonary pressure gradient to transpulmonary flow. The noninvasive index presented was the ratio of the peak tricuspid regurgitation velocity (TRV) (m/sec) to the time-velocity integral of the right ventricular outflow tract (TVI RVOT ) (cm), where TRV represents a surrogate for transpulmonary pressure and TVI RVOT a surrogate for transpulmonary flow.
In our original study, all patients had PVR ≤ 6 Wood units (WU), and the equation TRV/TVI RVOT × 10 + 0.16 (PVR echo ) was shown to provide a good estimate of invasively derived PVR (PVR cath ). The ratio was further simplified to TRV/TVI RVOT × 10 and was subsequently evaluated in three different groups of studies. The first group consisted of validation studies, confirming the validity of the ratio as a noninvasive correlate of PVR and included patients with varying pathologies. The second group included prognostic studies, which used the ratio as a prognostic indicator in various clinical scenarios. Finally, the last group consisted of modification studies with adjustments to the original ratio and equation in an attempt to provide improved accuracy in exchange for complexity. This was, in part, because some of the above studies noted that PVR echo became less accurate in patients with significantly elevated PVR (>6 WU).
The goals of our present study were to confirm the validity of TRV/TVI RVOT in a larger group of patients collected from multiple centers and to include patients with markedly elevated PVR (>6 WU). In addition, we compared the TRV/TVI RVOT ratio with a proposed modified ratio, TRV 2 /TVI RVOT , as an improved noninvasive correlate of PVR. As has been studied, TRV/TVI RVOT discriminates normal from elevated PVR, provides a noninvasive estimate when PVR is ≤6 WU, and provides a prognostic marker in patients with myriad pathologies. We hypothesized that the modified ratio, TRV 2 /TVI RVOT , might perform better in situations of markedly elevated PVR (>6 WU). The latter ratio was introduced because, according to the convective acceleration component of the Bernoulli equation, a quadratic relationship between pressure and velocity exists.
In an effort to create a larger database of noninvasive Doppler correlates of PVR cath , we corresponded with authors who had published validation studies using the TRV/TVI RVOT ratio, or a derivative thereof, and compared these values with PVR cath . We were able to obtain the anonymous raw data from five centers, including our original data set. Analysis was performed by combining preexisting data from the original databases. Original echocardiographic images were not reanalyzed, but an independent review of each data worksheet was performed. We included only patients with complete sets of Doppler and hemodynamic data. Of note, not all patients in the studies included had simultaneous right heart catheterization and Doppler interrogation as outlined in the original publication. The institutional review boards of the respective institutions approved the published studies. Doppler variables including TRV and TVI RVOT were obtained as previously described. Briefly, Doppler echocardiography was performed using the following ultrasound systems: Vivid 5 and Vivid 7 (GE Medical Systems, Milwaukee, WI), Acuson Sequoia (Acuson, Mountain View, CA; Siemens Healthcare, Erlangen, Germany), and iE33 (Philips Healthcare, North Ryde, Australia; Philips Medical Systems, Andover, MA).
TVI RVOT (cm) was obtained by placing a 1-mm to 2-mm pulsed-wave Doppler sample volume in the proximal right ventricular outflow tract when imaged from the parasternal short-axis view. The sample volume was placed so that the closing but not opening click of the pulmonary valve was visualized. Continuous-wave Doppler was used to determine the peak TRV (m/sec). The highest velocity obtained from multiple views was used. Agitated saline was used to enhance suboptimal Doppler signals. In patients with atrial fibrillation, the average of five measurements were used. The TRV/TVI RVOT and TRV 2 /TVI RVOT ratios were then calculated.
Invasive PVR measures were also obtained as previously described and recorded from their respective studies. Briefly, a flow-directed pulmonary artery catheter was used for hemodynamic measurements. Pulmonary capillary wedge pressure (PCWP), pulmonary artery systolic pressure, pulmonary artery diastolic pressure, and mean pulmonary artery pressure were measured. Cardiac output was calculated by thermodilution as a mean of three consecutive measurements not varying by >10%. PVR (WU) was calculated using the equation PVR = mean pulmonary artery pressure − PCWP/cardiac output. Individuals interpreting invasive and Doppler variables were blinded to each other’s results.
Multiple linear regression analyses were generated, the first between the entire cohort of PVR cath and TRV/TVI RVOT and the second between the same PVR cath cohort and TRV 2 /TVI RVOT . Spearman’s correlation coefficient was obtained for each analysis. The patients were then divided into two groups, the first with PVR cath ≤ 6 WU and the second with PVR cath > 6 WU. Linear regression analysis was then performed for each of the two groups with both TRV/TVI RVOT and subsequently TRV 2 /TVI RVOT . Spearman’s correlation coefficient was obtained for each analysis.
A plot of PVR echo (TRV/TVI RVOT × 10 + 0.16) compared with PVR cath was generated using Bland-Altman analysis. A new regression equation was derived on the basis of the modified ratio TRV 2 /TVI RVOT (PVR echo2 = 5.19 × TRV 2 /TVI RVOT − 0.4). A second plot for PVR echo2 compared with PVR cath was then generated using Bland-Altman analysis.
Using receiver operating characteristic curves, dichotomized PVR was analyzed based first on TRV/TVI RVOT and then TRV 2 /TVI RVOT . A logistic model was generated, and cutoff values for both TRV/TVI RVOT and TRV 2 /TVI RVOT , with balanced sensitivity and specificity, were obtained to predict elevated PVR > 6 WU. Confidence intervals (CIs) were calculated for the sensitivity and specificity values using the binomial method.
In an attempt to determine the effect of left atrial pressure, patients were divided into two subgroups, those with elevated and normal PCWP (>15 and ≤15 mm Hg, respectively). Moreover, both ratios were also correlated with invasively derived PVR and with total pulmonary resistance (TPR). The latter is an invasively obtained variable that excludes left atrial pressure in determining pulmonary resistance (TPR = mean pulmonary artery pressure/cardiac output). Bland-Altman analysis was performed for both ratios (TRV/TVI RVOT and TRV 2 /TVI RVOT ) in patients with PCWP ≤ 15 and > 15 mm Hg.
Similarly, in an attempt to determine the effect of flow, patients were divided into two subgroups, those with normal or diminished TVI RVOT (>15 or ≤ 15 cm, respectively). TVI RVOT was used as a surrogate of stroke volume. A second Bland-Altman analysis was performed for patients with TVI RVOT ≤ 15 and > 15 cm.
Thirty patients were excluded because of incomplete data. A total of 150 patients remained in the final analysis. The baseline demographics are listed in Table 1 . A total of 126 patients had PVR cath ≤ 6 WU, while 24 patients had PVR cath > 6 WU. The mean, median, and standard deviation of invasive PVR and noninvasive correlates are shown in Table 2 .
|Age (y)||62 ± 13|
|Left ventricular ejection fraction (%)||58 ± 13|
|Mean pulmonary artery pressure (mm Hg)||25 ± 11|
|PCWP (mm Hg)||13 ± 6|
|Right atrial pressure (mm Hg)||8 ± 6|
|PVR (WU)||3.47 ± 3.63|
|Valvular heart disease||12|
|Renal and liver transplantation||6|
|Acute respiratory failure||4|
|Primary pulmonary hypertension||20|
|Congestive heart failure||15|
|PVR cath ≤ 6 WU||0.19||0.17||0.06||0.08–0.48|
|PVR cath > 6 WU||0.4||0.41||0.14||0.13–0.71|
|TRV 2 /TVI RVOT|
|PVR cath ≤ 6 WU||0.55||0.45||0.29||0.1–1.86|
|PVR cath > 6 WU||1.72||2.02||0.69||0.45–2.84|
|PVR cath (WU) (all patients)||3.47||2||3.63||0.31–21|
|PVR echo (WU) (all patients)||2.4||2||1.16||1.04–7.26|
|PVR echo2 (WU) (all patients)||3.44||2.16||2.97||0.11–14.33|
|PVR cath ≤ 6 WU||2.11||1.8||1.09||0.31–6|
|PVR echo (WU) (PVR cath ≤ 6)||2||1.89||0.69||1.04–4.99|
|PVR echo2 (WU) (PVR cath ≤ 6)||2.47||1.96||1.5||0.11–9.29|
|PVR cath > 6 WU||10.59||9.83||4||6.07–21|
|PVR echo (WU) (PVR cath > 6)||4.25||4.32||1.42||1.45–7.26|
|PVR echo2 (WU) (PVR cath > 6)||8.53||10||3.59||1.95–14.33|
The linear regression analysis between PVR cath and TRV/TVI RVOT revealed a good correlation ( r = 0.76, P < .0001, Z = 0.92; Figure 1 A). There was a better correlation between PVR cath and TRV 2 /TVI RVOT ( r = 0.79, P < .0001, Z = −0.01; Figure 1 B). PVR echo was derived as previously described ([TRV/TVI RVOT × 10] + 0.16). PVR echo2 was derived from the regression analysis involving TRV 2 /TVI RVOT :
PVR echo 2 = ( 5.19 × TRV 2 / TVI RVOT ) – 0.4