Abnormal Left Ventricular Diastolic Filling in Chronic Thromboembolic Pulmonary Hypertension

True Diastolic Dysfunction or Left Ventricular Underfilling?

Methods

Results

All 61 patients had normal LV systolic function as well as moderate-to-severe pulmonary hypertension with evidence of RV pressure overload on echocardiography. All patients were in sinus rhythm during both the echocardiogram and cardiac catheterization. The demographic, echocardiographic, and hemodynamic data are shown in Table 1.Preoperatively, all patients had Doppler evidence of impaired relaxation, with a mean E-wave velocity of 54 ± 16 cm/s, mean A-wave velocity of 77 ± 22 cm/s, and a mean E/A ratio of 0.72 ± 0.16. After PTE, there was a significant increase in E velocity (81 ± 20 cm/s, p < 0.0001 compared with preoperative values), whereas A velocity decreased slightly to 71 ± 20 cm/s (p = 0.001). This translated to a normalization of the E/A ratio at 1.22 ± 0.40 (p < 0.0001).

Pulmonary venous interrogation revealed systolic-predominant flow signals preoperatively, with a mean systolic (S) velocity of 57 ± 13 cm/s and a mean diastolic (D) velocity of 39 ± 15 cm/s. Both S and D velocities increased significantly after PTE, with the mean S velocity increasing to 68 ± 16 cm/s (p < 0.0001 compared with preoperative values) and the mean D velocity increasing to 70 ± 21 cm/s (p < 0.0001). Figure 1shows an example of a typical patient’s pre- and postoperative mitral inflow and pulmonary venous Doppler pattern.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Example of Transmitral and Pulmonary Venous Doppler Interrogation Before and After Pulmonary Thromboendarterectomy

Transmitral pulsed-wave Doppler interrogation (left)and pulmonary venous Doppler interrogation (right)in a patient with chronic thromboembolic pulmonary hypertension before (above)and after (below)pulmonary thromboendarterectomy.

Despite the abnormal preoperative E/A ratios, tissue Doppler interrogation of the septal portion of the mitral annulus revealed normal mean early diastolic (E_m) tissue velocity of 8.0 ± 3.0 cm/s whereas the lateral portion of the mitral annulus had a similar mean E_mvelocity of 9.4 ± 3.2 cm/s. Tissue velocities remained normal after surgery and were unchanged in the septal region (mean E_mvelocity of 8.1 ± 2.0 cm/s, p = 0.66 compared with preoperative values). Lateral E_mvelocity increased to 11.8 ± 3.1 cm/s (p < 0.0001). The resulting E/E_mratios also changed significantly after surgery (septal: 7.8 ± 3.8 vs. 10.9 ± 4.6, p < 0.001; lateral: 6.2 ± 2.2 vs. 7.4 ± 3.2, p = 0.004). Although marked RV hypertrophy was present in this population, M-mode measurements of LV septal and posterior wall thickness were normal before PTE and did not change after surgery (septal: 10 ± 2 mm vs. 10 ± 1 mm; posterior: 10 ± 2 mm vs. 10 ± 1 mm, p = NS for both comparisons). Figure 2summarizes the Doppler echocardiographic findings.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Preoperative and Postoperative Parameters in Patients Undergoing PTE

The p values are significant for all comparisons between preoperative and postoperative parameters except E_m(septal). All variables are presented as mean ± standard deviation. Solid bars= postoperative; open bars= preoperative. A = A velocity; CTEPH = chronic thromboembolic pulmonary hypertension; D = diastolic pulmonary venous flow velocity; E = E velocity; E_m= early mitral annular tissue Doppler velocity; PTE = pulmonary thromboendarterectomy; S = systolic pulmonary venous flow velocity.

The changes in echocardiographic measurements were accompanied by substantial reductions in mean PA pressure (from 47 ± 9 mm Hg to 30 ± 8 mm Hg, p < 0.0001) and mean PVR (from 893 ± 390 dyne s/cm⁵to 335 ± 146 dyne s/cm⁵, p < 0.0001). Cardiac output significantly improved from 3.7 ± 1.1 l/min to 5.3 ± 1.5 l/min (p < 0.0001). Mean PCWP increased after PTE from 9.3 ± 3.2 mm Hg to 10.6 ± 3.9 mm Hg (p = 0.035). The transmitral E velocity correlated inversely with PVR (r = 0.54, p < 0.0001), with a nonlinear logarithmic regression analysis showing the best fit. These results are shown in Fig. 3.Inter- and intraobserver variability in Doppler velocity measurements were tested in a random sample of 6 patients and were both <5%.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Comparison of E and PVR in Patients With CTEPH Undergoing PTE

Solid circles= preoperative patients; open circles= postoperative patients. Logarithmic regression analysis with 95% confidence intervals is also shown. PVR = pulmonary vascular resistance; other abbreviations as in Figure 2.

Discussion

In this study of patients with CTEPH, diastolic abnormalities of mitral inflow resolved rapidly after successful PTE. This resolution was manifested primarily by a substantial increase in E velocity: as PVR decreased after successful surgery, there was a predictable increase in E velocity. Furthermore, both S and D velocities increased significantly after PTE, suggesting a higher preload state for the LV after a decrease in PVR. The increase in E/E_mratio is also consistent with an increase in LV filling pressure and, indeed, we found a significant increase in mean PCWP after PTE (consistent with a previous report from our laboratory) (3). Importantly, although an E<A transmitral Doppler pattern was present, early diastolic mitral annular (E_m) velocities were normal before PTE. Although septal E_mvelocity did not change after surgery, lateral E_mvelocity increased significantly. Finally, M-mode measurements of LV septal and posterior wall thickness were normal before PTE and did not change after surgery. These findings support the notion that the LV inflow diastolic abnormalities seen in patients with CTEPH are more a reflection of LV underfilling than a marker of true diastolic dysfunction from septal hypertrophy and LV compression.

We suspect that the rise in lateral annular E_mvelocity observed after PTE is a result of increased LV preload. Although it is well established that transmitral inflow Doppler measurements are load-dependent (8,9), there initially was some controversy regarding the load-dependence of tissue Doppler measurements. Sohn et al. (10) demonstrated that E velocity and deceleration time were preload dependent but that E_mvelocity in the same patients was unchanged under various loading conditions. However, Firstenberg et al. (11) found that E_mvelocities were indeed load-dependent, decreasing concordantly with LV preload. Similarly, Oguzhan et al. (12) found that acute hemodialysis caused a significant decline in Doppler tissue velocities in both the septal and lateral mitral annular positions. Although one might expect septal E_mvelocity to increase as well after PTE, we recently have found that cardiac surgery with cardiopulmonary bypass leads to a decrease in septal E_mvelocity (but no effect on lateral E_m) (13).

Interdependence between the right and left ventricles was described by the French physiologist P.I. Bernheim in 1910 (14). He hypothesized that the geometric changes from progressive LV enlargement and hypertrophy could directly compress the RV and thereby reduce RV filling. This so-called “Bernheim effect” could lead to elevated central venous pressure and peripheral edema in cases of LV overload. The “reverse” Bernheim phenomenon (i.e., enlargement of the RV leading to compression and underfilling of the LV) also has been described in both animals and humans with pulmonary hypertension and RV failure (5,15). For example, Marcus et al. (16) studied 12 patients with primary pulmonary hypertension using magnetic resonance imaging (MRI) and found a reduction in LV peak filling rate and LV stroke volume compared with control patients. Similarly, several echocardiographic studies (3,17) have shown a high incidence of decreased early LV filling in pulmonary hypertension and RV enlargement.

The point of some contention is not if this phenomenon occurs, but why. On the one hand, direct ventricular interaction could limit LV diastolic filling. The LV does appear distorted and compressed in the setting of severe RV pressure overload; the interventricular septum appears hypertrophied and functions more as part of the RV than the left. It is certainly plausible that these changes could lead to LV diastolic dysfunction and abnormal early diastolic relaxation. For example, Gan et al. (18) recently reported a study using MRI, right heart catheterization, and transesophageal echocardiography in patients with pulmonary arterial hypertension (PAH, not CTEPH). With cardiac MRI, they found impaired early diastolic filling in patients with PAH compared with control patients. They also found that leftward curvature of the interventricular septum correlated with LV filling rate and so concluded that direct ventricular interaction impairs ventricular filling in pulmonary hypertension. Of note, they also reported an E<A transmitral pattern, a lower stroke volume index, and a lower cardiac index in PAH versus control patients. Tissue Doppler imaging was not performed. Their reported mean systolic and diastolic pulmonary venous flow velocities (59 ± 16 cm/s and 42 ± 10 cm/s, respectively) in patients with PAH (18) are remarkably similar to the preoperative velocities in our study. In addition, PCWP in the PAH group was normal at 7 ± 5 mm Hg. However, transesophageal echocardiography and right heart catheterization were not performed in the control group and, thus, the differences in Doppler filling patterns and LV filling pressure could not be compared.

On the other hand, the high pulmonary vascular resistance and low cardiac output observed in severe PAH and CTEPH could lead to underfilling of the left heart. Indeed, LV underfilling in normal individuals can cause reversal of the transmitral E/A ratio and simulate impaired LV relaxation (19). Although previous work in our laboratory has shown a markedly abnormal LV eccentricity index in CTEPH (20), our current data show that the interventricular septum is not particularly hypertrophied, and its thickness matches that of the LV posterior wall. Also, even though the LV in CTEPH appears compressed within the pericardium by the RV, pericardiectomy leads to no change in LV shape or diastolic pressure/volume relationships (21). Although the present study does not exclude a direct effect on LV diastolic function from chamber distortion in CTEPH, it does demonstrate the importance of diastolic LV underfilling. We would also submit that the results of Gan et al. (18), described previously, are not inconsistent with the hypothesis that LV underfilling contributes to altered diastolic inflow.

The dramatic increases in transmitral E and pulmonary venous flow velocities after PTE are consistent with a marked improvement in LV preload. The rapid normalization of the E/A ratio after PTE and the presence of normal preoperative interventricular septal thickness by echo would also argue against any significant lingering effects from septal hypertrophy. Finally, the coexistence of an E<A transmitral filling pattern and an essentially normal mitral annular E_mvelocity before PTE is perhaps the most persuasive evidence of LV underfilling with relatively preserved diastolic function.

Conclusions

In patients with CTEPH, successful reduction of RV afterload after pulmonary thromboendarterectomy predictably improves LV preload and transmitral diastolic flow patterns. Pulmonary venous Doppler velocities also increase substantially and shift from a systolic-dominant pattern to a diastolic-dominant pattern. Despite the abnormal (E<A) preoperative transmitral Doppler flow pattern, mitral annular (E_m) velocity is normal. After successful thromboendarterectomy, transmitral Doppler E velocity increases significantly and the E/A ratio becomes normal. Mitral E_mvelocity increases somewhat but remains in the normal range. Thickness of the interventricular septum and LV posterior wall are normal preoperatively and do not change after surgery. The finding of an abnormal transmitral filling pattern (E<A) together with a normal mitral E_mvelocity before thromboendarterectomy supports the concept that the “abnormal relaxation” pattern of transmitral LV filling in CTEPH is due in large part to low LV preload and underfilling rather than RV hypertrophy, enlargement, and LV compression.