We describe a protocol to measure vortex formation time, an index of left ventricular filling efficiency, using standard transesophageal echocardiography techniques in patients undergoing cardiac surgery. We apply this technique to analyze vortex formation time in several groups of patients with differing cardiac pathologies.
Trans-mitral blood flow produces a three-dimensional rotational body of fluid, known as a vortex ring, that enhances the efficiency of left ventricular (LV) filling compared with a continuous linear jet. Vortex ring development is most often quantified with vortex formation time (VFT), a dimensionless parameter based on fluid ejection from a rigid tube. Our group is interested in factors that affect LV filling efficiency during cardiac surgery. In this report, we describe how to use standard two-dimensional (2D) and Doppler transesophageal echocardiography (TEE) to noninvasively derive the variables needed to calculate VFT. We calculate atrial filling fraction (β) from velocity-time integrals of trans-mitral early LV filling and atrial systole blood flow velocity waveforms measured in the mid-esophageal four-chamber TEE view. Stroke volume (SV) is calculated as the product of the diameter of the LV outflow track measured in the mid-esophageal long axis TEE view and the velocity-time integral of blood flow through the outflow track determined in the deep transgastric view using pulse-wave Doppler. Finally, mitral valve diameter (D) is determined as the average of major and minor axis lengths measured in orthogonal mid-esophageal bicommissural and long axis imaging planes, respectively. VFT is then calculated as 4 × (1-β) × SV/(πD3). We have used this technique to analyze VFT in several groups of patients with differing cardiac abnormalities. We discuss our application of this technique and its potential limitations and also review our results to date. Noninvasive measurement of VFT using TEE is straightforward in anesthetized patients undergoing cardiac surgery. The technique may allow cardiac anesthesiologists and surgeons to assess the impact of pathological conditions and surgical interventions on LV filling efficiency in real time.
Fluid mechanics is a critical yet often underappreciated determinant of left ventricular (LV) filling. A three-dimensional rotational body of fluid, known as a vortex ring, is generated whenever a fluid traverses an orifice1,2,3. This vortex ring improves the efficiency of fluid transport compared with a continuous linear jet4. Movement of blood through the mitral valve during early LV filling causes a vortex ring to form5,6,7,8 and facilitates its propagation into the chamber by preserving fluid momentum and kinetic energy9. These actions enhance LV filling efficiency4,10,11,12,13. The ring not only inhibits blood flow stasis in the LV apex14,15,16,17 but also directs flow preferentially beneath the anterior mitral leaflet7,18, effects that decrease the risk of apical thrombus formation and facilitate filling of the LV outflow track19, respectively. Contrast echocardiography17, Doppler vector flow mapping6,20,21, magnetic resonance imaging7, and particle imaging velocimetry9,22,23,24 have been used to demonstrate the appearance and behavior of trans-mitral vortex rings under normal and pathological conditions. The left atrial-LV pressure gradient, the degree of diastolic mitral annular excursion, the minimum LV pressure achieved during diastole, and the rate and extent of LV relaxation are the four major determinants of the duration, size, flow intensity, and position of the trans-mitral ring2,12,25,26,27,28,29.
Vortex ring development is most often quantified with a dimensionless parameter (vortex formation time; VFT) based on fluid ejection from a rigid tube3, where VFT is defined as the product of the time-averaged fluid velocity and the duration of ejection divided by the orifice diameter. The optimal size of a vortex ring is achieved when VFT is 4 in vitro because trailing jets and energetic limitations prevent it from attaining a larger size3,4. Mitral valve VFT has been approximated clinically using transthoracic echocardiography8,30,31. Based on analysis of trans-mitral blood flow velocity and mitral valve diameter (D), it can be easily shown8 that VFT = 4 × (1-β) × EF ×α3, where β = atrial filling fraction, EF = LV ejection fraction, and α = EDV1/3/D, where EDV = end-diastolic volume. Ejection fraction is the ratio of stroke volume (SV) and EDV, allowing this equation to be simplified to VFT = 4 × (1-β) × SV/(πD3). Because VFT is dimensionless (volume/volume), this index allows direct comparison between patients of varying size without adjustment for weight or body surface area8. Optimal VFT ranges between 3.3 and 5.5 in healthy subjects8, and results are consistent with those obtained in fluid dynamics models3,32. VFT was shown to be ≤ 2.0 in patients with depressed LV systolic function, findings that are also supported by theoretical predictions8. Reductions in VFT independently predicted morbidity and mortality in patients with heart failure30. Elevated LV afterload33, Alzheimer's disease34, abnormal diastolic function19, and replacement of the native mitral valve with a prosthesis35 have also been shown to decrease VFT. Measurement of VFT may also be useful to identify blood flow stasis or thrombosis in patients with acute myocardial infarction36,37.
Our group is interested in factors that affect LV filling efficiency during cardiac surgery38,39,40,41. We use standard two-dimensional and Doppler transesophageal echocardiography (TEE) to noninvasively derive the variables required to calculate VFT. In this report, we describe this methodology in detail and review our findings to date.
The Institutional Review Board of the Clement J. Zablocki Veterans Affairs Medical Center approved the protocols. Written informed consent was waived because invasive cardiac monitoring and TEE are routinely used in all patients undergoing cardiac surgery in our institution. Patients with relative or absolute contraindications for TEE, those undergoing repeat median sternotomy or emergency surgery, and those with atrial or ventricular tachyarrhythmias were excluded from participation.
1. Anesthesia
2. Transesophageal Echocardiography
3. Experimental Design
4. Statistics
The current technique allowed us to reliably measure VFT during cardiac surgery under a variety of clinical conditions by obtaining each determinant from blood flow and dimensional recordings in standard TEE imaging planes. A pulse-wave Doppler sample volume was placed at the tips of the mitral leaflets in the mid-esophageal four-chamber view to obtain the trans-mitral blood flow velocity profile necessary to calculate atrial filling fraction (β; Figure 1). Stroke volume was determined using the continuity equation (velocity-time integral of the LV outflow track blood flow velocity waveform multiplied by the area of the outflow track) and LV outflow track diameter was measured in the mid-esophageal LV long-axis view (Figure 2A), whereas blood flow through the outflow tract was determined in the deep transgastric short axis imaging plane (Figure 2B). Finally, average mitral valve diameter was calculated as the average of major and minor axis diameters measured in the mid-esophageal bicommissural and LV long-axis planes (Figure 3A and 3B, respectively). Measurement of VFT was associated with intra- and interobserver variability of 5% and 7%, respectively, similar to other indices of dimension and blood flow measured using TEE (data not shown). Using this technique, we first showed that exposure to CPB reduced VFT (5.3 ± 1.8 before vs. 4.0 ± 1.5 15 minutes after bypass, p < 0.05; Figure 4) in patients undergoing coronary artery surgery. VFT recovered to baseline values within 60 minutes after CPB. An increase in β (0.33 ± 0.04 before vs. 0.41 ± 0.07 15 minutes after CPB, p < 0.05) consistent with greater atrial contribution to LV filling was primarily responsible for the decline in VFT because SV and mitral valve diameter remained unchanged.
We also showed that a decrease in VFT occurs in patients with severe aortic valve stenosis and LV pressure-overload hypertrophy compared with those with normal LV wall thickness (3.0 ± 0.6 vs. 4.3 ± 0.5, respectively; p < 0.05; Figure 5). Early LV filling was attenuated (e.g., E/A, 0.77 ± 0.11 compared with 1.23 ± 0.13; β, 0.43 ± 0.09 compared with 0.35 ± 0.02; p < 0.05 for each), and SV was reduced (72 ± 12 mL compared with 95 ± 10 mL; p < 0.05) in patients with vs. without LV hypertrophy; however, mitral valve diameter was similar between groups. A significant inverse correlation between VFT and posterior wall thickness (PWT) was shown with linear regression analysis (VFT = -2.57 × PWT + 6.81; r = 0.408; p = 0.017). In addition, our results using this technique demonstrated that the presence compared to absence of moderate aortic insufficiency in patients with severe aortic valve stenosis increased VFT (5.7 ± 1.7 vs. 3.0 ± 0.6, respectively; p < 0.05; Figure 5) concomitant with a decrease in mitral valve diameter (2.2 ± 0.2 vs. 2.6 ± 0.1 cm, respectively; p < 0.05), whereas indices of LV diastolic dysfunction and SV were similar between groups. Finally, we were able to use our technique of measuring VFT to show that VFT was lower in octogenarians compared with younger patients (3.0 ± 0.9 vs. 4.5 ± 1.2; p < 0.05) concomitant with an impaired relaxation pattern of LV diastolic dysfunction (e.g., E/A of 0.81 ± 0.16 vs.1.29 ± 0.19; β of 0.44 ± 0.05 vs.0.35 ± 0.03, p < 0.05 for each). A significant inverse correlation between VFT and age was also demonstrated (VFT = -0.0627 × age + 8.24; r = 0.639; p = 0.0139; Figure 6).
Figure 1: Trans-mitral blood flow velocity waveforms. Trans-mitral blood flow velocity waveforms during early LV filling (E) and atrial systole (A) obtained in the mid-esophageal four-chamber TEE view (left side of image); the area of each envelope was integrated using the equipment's software to obtain velocity-time integrals (right side of image) and the atrial filling fraction (β) was calculated. In this example, β = 4.28 cm/(4.28 cm + 6.73 cm) = 0.39 (see text). Please click here to view a larger version of this figure.
Figure 2: Measurement of LV outflow track diameter. Measurement of LV outflow track diameter during mid-systole in the aortic valve long axis TEE view (A) (diameter = 2.23 cm); (B) blood flow velocity was measured in the in the distal LV outflow track using the deep transgastric long axis TEE view and the area of the resulting envelope (left side of panel B) integrated using the equipment's software to obtain a velocity-time integral (white arrow, right side of panel B). In this example, stroke volume = π/4 × (2.23 cm)2 × 19.8 cm = 77 mL (see text). Please click here to view a larger version of this figure.
Figure 3: Average mitral valve diameter was calculated as the average of major and minor axis diameters measured in the mid-esophageal bicommissural and LV long-axis planes. Mid-esophageal bicommissural (A) and LV outflow tract (B) TEE images were used to determine major (transcommissural anterior-lateral-posterior-medial) and minor (anterior-posterior) axis diameters, respectively. In this example, mitral valve diameter = (3.04 cm + 2.18 cm)/2 = 2.61 cm. This figure is reproduced with permission from Elsevier38. Please click here to view a larger version of this figure.
Figure 4: Temporal changes in VFT. Temporal changes in VFT before and 15, 30, and 60 minutes after cardiopulmonary bypass (CPB) in patients undergoing coronary artery surgery; *indicates significant (p < 0.05) difference from the "before CPB" measurement. Please click here to view a larger version of this figure.
Figure 5: Effects of LV pressure-overload hypertrophy resulting from severe aortic valve stenosis in the absence (-) or presence (+) of moderate aortic insufficiency (AI) in patients undergoing aortic valve replacement. Patients with normal LV wall thickness undergoing coronary artery surgery served as controls (normal). *Significantly (p < 0.05) different from normal; †Significantly (p < 0.05) different from both normal and hypertrophy-AI. Please click here to view a larger version of this figure.
Figure 6: Correlation between age and VFT in 14 patients undergoing coronary artery surgery. VFT = -0.0627 × age + 8.24; r = 0.639; p = 0.0139. Please click here to view a larger version of this figure.
The current results illustrate that VFT can be reliably measured during cardiac surgery using the TEE techniques described here. Previous descriptions of VFT used transthoracic echocardiography in conscious subjects, but this approach cannot be utilized when the chest is open. We used intraoperative TEE to determine VFT in the anesthetized patients undergoing cardiac surgery during which changes in LV filling dynamics are often encountered as a result of ischemia-reperfusion injury or surgical interventions. Our findings indicate that VFT measurements reflect changes in LV filling efficiency produced by transient CPB-induced impaired relaxation pattern diastolic dysfunction, aortic valve disease, and aging. The current technique for calculating VFT during cardiac surgery requires high-quality TEE images and video clips during steady-state hemodynamic conditions to assure precise measurements of mitral valve and LV outflow tract dimension and blood flow (Figure 1, Figure 2, and Figure 3). Not all patients will have optimal imaging windows because of off-axis rotation of the heart or pathological changes in cardiac geometry. Despite these potential limitations, experienced intraoperative echocardiographers should be able to easily obtain the necessary mid-esophageal four-chamber, mid-esophageal bicommissural, mid-esophageal LV long axis, and deep transgastric long axis views during the comprehensive TEE examination42. The technique may also be unreliable when rapidly changing hemodynamic conditions are present. It does not provide direct visualization of blood flow movement within the LV associated with the vortex, as previously characterized using Doppler vector flow mapping6,20,21 or particle imaging velocimetry9,22,23,24. Accurate measurement of LV outflow track diameter using two-dimensional echocardiography is especially important because this variable is squared in the calculation of area and errors are magnified as a result. Similarly, accurate measurements of mitral valve minor and major axis length are essential because the cube of the average of these two dimensions appears in the denominator of the VFT formula. Two-dimensional echocardiography consistently underestimates aortic and mitral valve areas compared with three-dimensional reconstruction techniques43,44. The impact of these differences between two- and three-dimensional TEE on VFT is an area of current research by our group.
Additionally, isoflurane was used for maintenance of anesthesia in our studies. This volatile anesthetic is a vasodilating negative inotrope that reduces LV preload and afterload, decreases myocardial contractility, and affects LV diastolic function in a dose-related manner45,46. These cardiovascular changes may have influenced atrial filling fraction and stroke volume in our studies. Nevertheless, the values of VFT obtained in anesthetized patients with normal preoperative LV ejection fraction undergoing coronary artery surgery before CPB were similar to those described in healthy conscious subjects8. These data suggest that baseline anesthesia does not substantially alter LV filling efficiency, but we are currently examining this hypothesis. VFT has been previously shown to be an independent predictor of mortality in patients with congestive heart failure30, but it is unknown whether intraoperative changes in VFT are predictive of perioperative morbidity or mortality in cardiac surgery patients. This topic is also an area of interest that we are actively pursuing.
We first used this technique of noninvasively calculating VFT in a study examining the impact of CPB on VFT in isoflurane-fentanyl-anesthetized patients with normal preoperative LV ejection fraction undergoing coronary artery surgery39. LV diastolic dysfunction occurs after cardiopulmonary bypass as a result of global ischemia-reperfusion injury and a profound systemic inflammatory response47,48,49. This diastolic dysfunction eventually recovers within minutes to hours based on the efficacy of myocardial protection during and the duration of CPB50. Indeed, our findings confirmed that LV diastolic dysfunction occurs after CPB. This effect was accompanied by transient reductions in VFT that recovered within one hour after separation from CPB. The declines in VFT resulted from an increase in β and a modest decrease in SV because mitral valve diameter was unchanged. The recoveries of VFT, β, E/A, and SV after CPB were similar. Notably, VFT observed here did not fall below the normal range of VFT (3.3 to 5.5) in healthy individuals. Our patients had normal preoperative LV systolic and diastolic function, were exposed to relatively short CPB times (93 ± 27 min), and were treated with regular doses of antegrade and retrograde cardioplegia. These factors probably combined to reduce ischemia-reperfusion injury during aortic cross-clamp application39. CPB has also been shown to cause transient declines in trans-mitral blood flow propagation velocity (Vp) consistent with attenuated early LV filling in patients undergoing coronary artery surgery49 as a result of decreases in LV compliance51 and reductions in early diastolic intraventricular pressure gradients52. A relationship between vortex ring formation and Vp was previously demonstrated53, and our findings supported those of other investigators49 in similar patient populations.
We subsequently examined the effects of pressure-overload LV hypertrophy produced by severe calcific degenerative aortic valve stenosis in patients with preserved LV systolic function undergoing aortic valve replacement40. A second group of patients with normal LV wall thickness undergoing coronary artery surgery served as controls. Chronically elevated LV end-systolic wall stress causes LV pressure-overload hypertrophy as a compensatory response in the presence of aortic valve stenosis54. LV wall thickening without dilatation occurs as a consequence of an increase in the diameter of individual myocytes. This LV remodeling is associated with interstitial fibrosis55,56. Delays in apical recoil57,58 also occur that further attenuate early LV filling58,59, which causes LV diastolic dysfunction by delaying LV relaxation and reducing LV compliance55,60. Thus, VFT is reduced in the presence of delayed relaxation in patients with LV pressure-overload hypertrophy vs. those with normal LV wall thickness. Our findings were attributed to an increase in β and decline in SV at similar filling pressures consistent with a decrease in LV compliance. A significant correlation between decreases in VFT and the severity of hypertrophy was shown using linear regression analysis. This observation suggests that the degree of pressure-overload hypertrophy is inversely related to LV filling efficiency quantified using vortex formation time.
Valvular insufficiency often occurs in conjunction with severe calcific degenerative aortic valve stenosis because prominent leaflet calcification prevents complete coaptation. We conducted another investigation to ascertain whether regurgitant blood flow into the LV through an incompetent aortic valve affects LV filling efficiency by interfering with trans-mitral blood flow38. We compared patients with severe aortic valve stenosis undergoing valve replacement who had moderate centrally-directed aortic insufficiency with a second group of patients who did not have regurgitation. We quantified aortic insufficiency using the regurgitant jet width LV outflow track diameter ratio measure with color Doppler M-mode echocardiography61. Our results showed that moderate aortic insufficiency increases VFT in patients with aortic valve stenosis. However, this increase in VFT does not suggest an improvement in LV filling efficiency has occurred because of abnormal regurgitant flow into the LV through the aortic valve. LV diastolic pressure rapidly increases in moderate to severe aortic insufficiency62, attenuating trans-mitral LV filling and reducing mitral valve area63,64,65. The results indicate that mitral valve diameter and area were reduced in patients with moderate aortic insufficiency versus those without regurgitation. These observations were most likely due to a decrease in minor axis length, resulting from attenuated anterior mitral leaflet opening caused by aortic regurgitant during LV filling, thereby, falsely elevated VFT. Indeed, VFT reported in our study (5.7 ± 1.7) was greater than the upper limit of normal VFT (5.5) in healthy conscious individuals8 and patients with normal LV geometry during anesthesia (4.3 ± 0.5)40. Therefore, it is highly likely that abnormal diastolic flow into the LV invalidates VFT as an index of LV filling efficiency.
We recently studied the influence of advanced age on VTF in elderly patients undergoing coronary artery surgery41. Progressive LV diastolic stiffening66, decreased intraventricular diastolic kinetic energy67, and attenuated diastolic suction68 cause LV diastolic function in the elderly69,70,71,72. Octogenarians with normal preoperative LV ejection were compared with a younger cohort of patients (≤ 62 years of age). We found that VFT was lower in octogenarians compared with younger patients. These observations were expected and occurred in conjunction with an impaired relaxation pattern of LV diastolic dysfunction and a modest reduction in SV at similar LV filling pressures. Mitral valve diameter was similar in octogenarians versus younger patients and did not contribute to the differences in VFT between groups. It is noteworthy that VFT was similar in octogenarians compared with patients with severe aortic valve stenosis that we previously reported38,40. Indeed, aortic stenosis is another condition characterized by impaired relaxation LV diastolic dysfunction and reductions in LV compliance. A significant inverse correlation between VFT and age was also demonstrated despite the small sample size (n = 7 per group; Figure 6). The decline in VFT that occurs with age that may eventually become indistinguishable from heart failure produced by pathological processes such as restrictive diastolic dysfunction19 or dilated cardiomyopathy8. Our results were consistent with reductions in early peak diastolic intraventricular kinetic energy in elderly subjects with depressed LV function67.
In summary, noninvasive measurement of VFT using standard two-dimensional and Doppler TEE is straightforward in anesthetized patients undergoing cardiac surgery. This technique may allow cardiac anesthesiologists and surgeons to assess the impact of pathological conditions and surgical interventions on LV filling efficiency in real time.
The authors have nothing to disclose.
This material is the result of work supported with resources and the use of the facilities at the Clement J. Zablocki Veterans Affairs Medical Center in Milwaukee, Wisconsin.
Echocardiography Machine | Philips Ultrasound, Bothall, WA | iE33 | |
Transesophageal Echocardiography Probe | Philips Ultrasound, Bothall, WA | X7-2t | |
Statistical Software | AnalystSoft, Walnut, CA | StatPlus:mac Pro |