Background
Assessments of right ventricular (RV) function using myocardial velocities in patients with pulmonary embolism (PE) may add vital information.
Methods
Thirty-four patients with PE were studied in the acute stage and 3 months afterward. Tricuspid annular velocity was recorded using pulsed-wave Doppler tissue imaging.
Results
At the time of diagnosis, tricuspid annular velocities were significantly decreased in patients compared with controls in systole (12.9 vs 14.8 cm/s, P < .05) and early diastole (11.9 vs 15.3 cm/s, P < .01) and normalized during follow-up. Decreases in tricuspid annular velocity were most pronounced in patients with increased RV pressure. The myocardial performance index was prolonged and pulmonary vascular resistance was higher in patients with increased RV pressure. The ratio of tricuspid flow to myocardial velocity (E/Em) was also increased compared with controls (4.5 vs 3.5, P < .05).
Conclusion
RV dysfunction in patients with PE was common in the acute phase but normalized within 3 months. Patients presenting with normal RV pressure had normal systolic but disturbed diastolic function.
Patients with pulmonary embolism (PE) and right ventricular (RV) dysfunction detected by echocardiography are known to be at risk for in-hospital clinical worsening and PE-related death. At present, thrombolysis is a well-established treatment for patients who are hemodynamically unstable or in shock. It has also been shown that patients with PE who are clinically stable but show echocardiographic RV dysfunction may benefit from thrombolysis. Therefore, it may be of importance to detect early, and correctly, RV dysfunction in this patient group.
Pulsed-wave Doppler tissue imaging (DTI) has been used to detect RV dysfunction in different clinical conditions. One advantage of measuring tricuspid annular velocity with pulsed DTI is that it is independent of geometric assumptions and endocardial border detection. Also, DTI has excellent temporal resolution and a good signal-to-noise ratio, and tricuspid systolic annular velocity has been shown to have a good correlation with global systolic RV function. Diastolic tricuspid annular velocities have also been used to determine RV relaxation and to predict RV filling pressure. However, to our knowledge, estimation of RV filling pressure in patients with PE has not been fully investigated previously using DTI. Therefore, the aims of this study were to determine the feasibility of using pulsed DTI to detect RV dysfunction and to estimate RV filling pressure in the acute phase and also to assess whether the values change over time in relation to clinical status.
Methods
Patients
Thirty-four consecutive patients (19 women) attending the emergency departments of Karolinska University Hospital Solna or Södersjukhuset Stockholm because of dyspnea and diagnosed with PE were included in the study. The diagnosis of PE was made by means of diagnostic pulmonary angiography, spiral computed tomography, and/or high-probability ventilation-perfusion scanning. The total number of screened patients was 148. Twelve of these patients were clinically assessed as having massive PE requiring immediate thrombolytic therapy and could not be subjected to more extensive and time-consuming echocardiographic investigation prior to therapy and were therefore excluded. Patients with histories of coronary artery bypass grafting or other thoracic surgery and having pacemakers, left bundle branch blocks, or atrial fibrillation at the time of arrival were excluded. Also, patients with histories of ≥2 previous PEs or deep-vein thromboses were excluded, because these patients are often under ongoing treatment with anticoagulants and may have chronic pulmonary hypertension.
The duration of symptoms at inclusion ranged from 3 hours to 12 days (average, 2.9 days; median, 46 hours). Two of the studied patients had clinically nonsignificant chronic obstructive pulmonary disease, and 4 had histories of myocardial infarction. None of the patients showed evidence of acute coronary syndromes on the basis of electrocardiography and serial cardiac enzymes or electrocardiographic signs of previous Q-wave inferior myocardial infarctions. Twenty percent of the patients had histories of hypertension, defined as blood pressure ≥ 140/90 mm Hg. Nine patients were smokers.
All patients were treated initially with low–molecular weight heparin and warfarin sodium. When therapeutic level was reached, only warfarin sodium was continued for ≥6 months according to the clinical indication. Twenty-three age-matched healthy subjects without histories of cardiopulmonary disease or hypertension and with normal electrocardiographic and echocardiographic findings served as controls. The local ethics committee of Karolinska Institute approved the study.
Echocardiography
A commercially available echocardiographic apparatus was used (Sequoia, Siemens Medical Solutions USA, Inc, Mountain View, CA; or System V, GE Vingmed Ultrasound AS, Horten, Norway). The subjects were studied <24 hours after arrival at the emergency department, and follow-up studies were conducted 3 months afterward. The different cardiac dimensions and the ejection fraction were calculated according to the recommendations of the American Society of Echocardiography. RV end-diastolic dimension was measured from an apical 4-chamber view orthogonal to the long axis, one third of the distance from the base. The M-mode tricuspid annular plane systolic excursion (TAPSE) was recorded from the apical 4-chamber view, as described previously. Interventricular septal motion was assessed from the parasternal short-axis view and from the apical 4-chamber view and determined to be abnormal if a flattened septum or paradoxical septal motion was present.
Tricuspid inflow velocity patterns were recorded from the 4-chamber view with the pulsed-wave Doppler sample volume positioned at the tip of the tricuspid leaflets. The early diastolic inflow velocity (E), obtained from the tricuspid inflow velocities, was measured for further analysis. The ratio of tricuspid regurgitation velocity (TRV) to the RV outflow tract time-velocity integral (TVI RVOT ) was multiplied by 10 (ie, 10 × [TRV/TVI RVOT ]) and used to calculate pulmonary vascular resistance (PVR) in Wood units (WU), as described elsewhere. A qualitative visual segmental wall motion analysis of the right ventricle from the 4-chamber view was performed by two experienced echocardiographers, as described previously. McConnell’s sign was considered positive in the presence of abnormal motion of the mid free wall of the right ventricle but with normal motion at the apex. In the presence of discrepancies in assessing McConnell’s sign between the two observers, a consensus was established.
Myocardial Velocities Using DTI
Recordings were made by spectral pulsed-wave Doppler. Filters were set to exclude high-frequency signals, and the Nyquist limit was adjusted to <25 cm/s. Minimal optimal gain was used to ensure the best signal-to-noise ratio. A small sample volume size was used, adjusted proportionally to annular motion. Myocardial velocities were obtained at the RV free wall near the tricuspid annulus and at the septal and lateral sites of the mitral annulus from the apical 4-chamber view, as described previously. The pulsed-wave cursor was aligned in such a way that the annulus moved along the sample volume line, always keeping the angle of insonation <20°. The peak annular tricuspid systolic (tricuspid Sm) and early (tricuspid Em) and late (tricuspid Am) diastolic velocities were measured from the myocardial velocity profiles. Tricuspid Sm was measured with exclusion of the velocities recorded during isovolumic contraction ( Figure 1 ). DTI tricuspid velocities were also recorded at the mid and apical portions of the RV free wall. The resulting velocities were recorded for 5 consecutive cardiac cycles at a sweep speed of 100 mm/s and were stored digitally as well as on a VHS videotape for later analysis. From the RV free wall near the tricuspid annulus, the isovolumic contraction time, isovolumic relaxation time, and ejection time were measured using DTI and used to calculate the myocardial performance index (MPI) of the right ventricle. The MPI was calculated as (isovolumic contraction time + isovolumic relaxation time)/ejection time, as described elsewhere. The sample volume was repositioned in the same location when required. Recordings were made whenever possible with patients holding their breath at end-expiration, otherwise during as shallow respiration as possible.
Estimation of RV Pressure and Right Atrial Pressure
The degree of tricuspid regurgitation was assessed and classified as absent to severe according to methods described previously. Systolic RV pressure was calculated from the summation of the estimated right atrial pressure and the peak difference in pressure between the right atrium and right ventricle. Patients with estimated systolic RV pressures ≥ 40 mm Hg were considered to have increased RV pressure. The ratio between tricuspid early diastolic inflow velocity and tricuspid annular early diastolic velocity (tricuspid E/Em) was calculated. The value for right atrial pressure was also estimated by measuring the end-expiratory diameter of the inferior vena cava and the change in diameter during inspiration. If the diameter of the inferior vena cava was <2.5 cm and the inspiratory collapse was >50%, right atrial pressure was estimated to be 5 mm Hg, and if the inspiratory collapse was <50%, estimated right atrial pressure was 10 mm Hg. If the inferior vena cava diameter was >2.5 cm and there was inspiratory collapse of <50%, right atrial pressure was estimated to be 15 mm Hg, and if there was no inspiratory change in the caval diameter, right atrial pressure was estimated to be 20 mm Hg. Invasive mean right atrial pressures were obtained in patients undergoing pulmonary angiography, prior to contrast injection, using a 7Fr pigtail catheter (Cook, Inc, Indianapolis, IN) connected to an electrically calibrated fluid-filled transducer (Navilyst Medical, Inc, Marlborough, MA), positioned at the midchest level of the patient.
Follow-Up
All patients underwent follow-up clinical assessment by the attending physician at the outpatient department 3 months later. They were asked about their physical well-being and if they could perform normal daily activities without dyspnea or fatigue.
Statistical Analysis
Results are expressed as mean ± SD. The significance of differences between patients on day 1 and controls was tested using Mann-Whitney’s nonparametric test for unpaired comparisons. Follow-up results from day 1 to day 90 in the patient group were assessed using Wilcoxon’s test for paired comparisons. The correlation between invasive right atrial pressure and the tricuspid E/Em ratio and between invasive right atrial pressure and noninvasive right atrial pressure was determined by computing Pearson’s correlation coefficient. Using receiver operating characteristic curves, dichotomized right atrial pressure was analyzed on the basis of the tricuspid E/Em ratio, and a cut-off value for tricuspid E/Em with sensitivity and specificity values was obtained to predict right atrial pressure values ≥ 10 mm Hg. A P value < .05 was regarded as statistically significant. Reproducibility was assessed in 20 randomly selected patients. The recorded DTI parameters were analyzed by an investigator on two different occasions and by another investigator on one occasion without prior knowledge of the clinical data or the results of the previous analysis. Variability is expressed as the mean percentage error, derived as the absolute difference between the two sets of observations divided by the mean of the observations.
Results
The mean age of the patients was similar to that of the healthy subjects (58 ± 16 vs 57 ± 11 years, P = NS). The mean transcutaneous oxygen saturation for the patient group was 94 ± 3%. The systolic and diastolic blood pressures in patients were 138 ± 17 and 82 ± 11 mm Hg, respectively. Abnormal septal motion was present in 16 patients: in 11 patients with systolic RV pressures ≥ 40 mm Hg and in 5 patients with systolic RV pressures < 40 mm Hg. None of the patients had severe tricuspid regurgitation. Five patients had moderate and 29 patients mild or physiologic tricuspid regurgitation.
Tricuspid Annular Velocities or Excursion During Admission
The DTI data acquired from the tricuspid annular region were of acceptable quality for analysis in all patients studied. Nearly all the measurements of tricuspid velocities from the mid portion of the right ventricle could be analyzed (n = 32). The velocities from the apical portion were more difficult to interpret and therefore were not used in the present study. Tricuspid Sm at the mid portion of the RV free wall was similar to that of the annular portion (11.1 ± 3.3 vs 11.9 ± 3.6, P = NS). The same was true for diastolic velocities between the two sites (tricuspid Em, 11.9 ± 3.7 vs 11.9 ± 3.6; tricuspid Am, 18.2 ± 4.3 vs 17.8 ± 5.1; P = NS for both). Because of the attainment of similar velocities at the mid or annular site of the right ventricle, only the annular velocities of the right ventricle were used for further analysis. The intraobserver variabilities for tricuspid annular Sm, Em, and Am velocities were low: 5.2 ± 2.1%, 5.3 ± 3.7%, and 5.8 ± 1.0%, respectively. Similar low results were obtained when assessing the interobserver variabilities: 6.3 ± 2.1%, 5.0 ± 0.8%, and 2.0 ± 1.6%, respectively.
The clinical and echocardiographic parameters of the patients at the time of diagnosis are shown in comparison with those of healthy controls in Table 1 . Tricuspid Sm and tricuspid Em were decreased in patients compared with controls. Tricuspid Am was similar in patients and controls. The tricuspid E/Em ratio, an expression of RV filling pressure, was increased in patients compared with controls. There was no correlation between tricuspid Am and RV filling pressure. Similar to tricuspid annular velocities, TAPSE was decreased in patients compared with controls. The MPI for the whole patient group was 0.64 ± 0.29. Patients with increased systolic RV pressures had higher MPIs than those with normal systolic RV pressures (0.76 ± 0.31 vs 0.40 ± 0.16, P < .001). Patients with normal systolic RV pressures had lower tricuspid Em compared with healthy controls (12.8 ± 3.9 vs 15.3 ± 3.6, P < .05) but did not differ regarding tricuspid Sm or tricuspid Am (Sm, 13.7 ± 3.1 vs 14.8 ± 1.8 cm/s; Am, 17.2 ± 4.9 vs 16.4 ± 3.6 cm/s; P = NS for both).
| Variable | Controls | Patients |
|---|---|---|
| Heart rate (beats/min) | 65 ± 8 | 77 ± 16 † |
| Tricuspid annular velocity (cm/s) | ||
| Systolic (Sm) | 14.8 ± 1.8 | 12.9 ± 3.1 ∗ |
| Early diastolic (Em) | 15.3 ± 3.6 | 11.9 ± 3.6 † |
| Late diastolic (Am) | 16.4 ± 3.6 | 17.8 ± 5.1 |
| Tricuspid E/Em ratio | 3.5 ± 0.7 | 4.5 ± 2.0 ∗ |
| TAPSE (mm) | 26 ± 4 | 19 ± 5 ‡ |
| RV end-diastolic dimension (mm) | 24 ± 3 | 34 ± 5 ‡ |
| Mitral annular velocity (cm/s) | ||
| Septal systolic (Sm) | 9.0 ± 1.5 | 8.6 ± 2.5 |
| Septal early diastolic (Em) | 12.0 ± 2.5 | 9.2 ± 3 ‡ |
| Lateral systolic (Sm) | 10.4 ± 1.8 | 10.7 ± 3.1 |
| Lateral early diastolic (Em) | 15.0 ± 2.8 | 12.0 ± 4.5 ‡ |
| LV ejection fraction (%) | 62 ± 5 | 58 ± 5 † |
| Transmitral E/A ratio | 1.2 ± 0.3 | 0.9 ± 0.2 ‡ |
| LV end-diastolic dimension (mm) | 47 ± 4 | 45 ± 7 |
Noninvasive Right Atrial Pressure and Annular Velocity
Right atrial pressure derived from the inferior vena cava was compared with the tricuspid E/Em ratio at the first investigation in 29 of 34 patients. The lack of tricuspid inflow E waves in the other 5 patients was due to high heart rates, which made it difficult to record adequate Doppler curves. The mean value for the tricuspid E/Em ratio for the whole patient group was 4.5 (range, 2.1-10.4). Using a receiver operating characteristic curve, a tricuspid E/Em cutoff value of ≥4 provided the best balanced sensitivity (71%) and specificity (83%) to detect patients with right atrial pressures ≥ 10 mm Hg (area under the curve, 0.774; 95% confidence interval, 59%-96%; P = .018). In the group of patients with tricuspid E/Em ratios ≥ 4, 3 of 14 patients had right atrial pressures < 10 mm Hg, and in the group of patients with tricuspid E/Em ratios < 4, 4 of 15 patients had right atrial pressures ≥ 10 mm Hg. Tricuspid E/Em ratio had sensitivity of 75% and specificity of 73% for identifying patients with increased right atrial pressures.
Invasive Right Atrial Pressure
Right atrial pressure was assessed in 10 patients using an invasive method. There was a moderate correlation between invasive right atrial pressure and right atrial pressure derived from the inferior vena cava ( r = 0.6, P < .05). Similar results were also obtained for the correlation between invasive right atrial pressure and the tricuspid E/Em ratio ( r = 0.6, P < .05).
PVR
A noninvasive measure of PVR was derived using the formula 10 × (TRV/TVI RVOT ) at the first investigation. The calculated PVR for the whole patient group was 1.8 ± 0.03 WU. When separated according to RV systolic pressure, patients with increased RV pressures had PVR of 2.8 ± 1.1 WU, and those with normal RV pressures had PVR of 1.7 ± 0.3 WU ( P < .01).
McConnell’s Sign and Tricuspid Annular Velocity
McConnell’s sign could be assessed in 31 patients (91%). Nine of these 31 patients showed positive McConnell’s sign (29%). Patients with positive McConnell’s sign showed decreased tricuspid Sm (12.9 ± 2.9 vs 15.5 ± 4.0 cm/s, P < .05) and tricuspid Em (10.9 ± 2.6 vs 13.5 ± 5.0 cm/s, P < .01) compared with those with negative McConnell’s sign. The tricuspid Am was similar in both groups. Furthermore, 7 of 9 patients (78%) had increased systolic RV pressures in the presence of positive McConnell’s sign. In contrast, most of the patients with negative McConnell’s sign had normal RV pressures (77%). Also, the tricuspid E/Em ratio was borderline significant in patients with positive McConnell’s sign compared with negative McConnell’s sign (5.7 ± 2.8 vs 4.0 ± 1.4, P = .054).
Mitral Annular Velocities and Left Ventricular Parameters
Mitral annular early diastolic velocity (mitral Em) was found to be reduced at the time of diagnosis compared with controls, while peak systolic mitral annular velocity (mitral Sm) did not differ from that of controls. There was no difference in the left ventricular end-diastolic dimension in patients compared with controls, nor did the left atrial dimension differ between the two groups (patients, 35.7 ± 6.4 mm; controls 34.4 ± 3.2 mm; P = NS). The left ventricular ejection fraction and the transmitral E/A ratio were reduced in patients compared with controls ( Table 1 ).
Follow-Up
There were no in-hospital deaths in our study. One patient died during follow-up. At 3-month follow-up, most of the patients were totally recovered clinically, and only 2 patients reported mild dyspnea. Heart rate was decreased at follow-up (77 ± 16 vs 70 ± 13 beats/min, P < .01). Tricuspid Sm and tricuspid Em had increased significantly ( Figure 2 ). RV end-diastolic dimension and RV systolic pressure also decreased significantly. The tricuspid E/Em ratio normalized during follow-up ( Table 2 ). Table 3 shows a comparison between patients with or without increased RV systolic pressures on arrival and at follow-up. In both groups, tricuspid Em improved during follow-up. Patients with increased RV pressures also showed a significant improvement of tricuspid Sm and the tricuspid E/Em ratio at follow-up.
