Relationship of Left Atrial Pressure and Pulmonary Venous Flow Velocities: Importance of Baseline Mitral and Pulmonary Venous Flow Velocity Patterns Studied in Lightly Sedated Dogs

Relationship of Left Atrial Pressure and Pulmonary Venous Flow Velocities: Importance of Baseline Mitral and Pulmonary Venous Flow Velocity Patterns Studied in Lightly Sedated Dogs Christopher P. Appleton, MD, MarkS. Gonzalez, MD, and Michael A. Basnight, MD, with the technical assistance of Alice McArthur and Tammi Zine, BA, Tucson, Arizona

Prior clinical and animal studies have shown a markedly different relationship between left atrial pressure and the systolic fraction of pulmonary venous flow but have not discussed possible reasons for this discrepancy. To examine the possibility that these disparate results are due to differences in baseline mitral and pulmonary venous flow velocities, we recorded both velocities with left atrial and left ventricular pressure under different loading conditions in eight lightly sedated normal dogs. With constant atrial pacing at 85 beats/min, mean left atrial pressure was increased from 5.3 ± 1.1 mm Hg at baseline to 16.1 ± 1.7 mm Hg with volume and methoxamine infusion (p < 0.05). As left atrial pressure increased, the operating compliance of the left atrium decreased, whereas left atrial volumes and ejection fraction increased. Baseline pulmonary venous diastolic flow velocity was larger than systolic velocity (66 ± 9 versus 36 ± II em/sec), with the systolic fraction of pulmonary venous flow 31% ± 8%. With increasing left atrial pressure, pulmonary venous diastolic velocity did not change, but peak systolic velocity (57 ± 16 em/sec) and the systolic fraction (48% ± 9%) both increased (p < 0.05). Changes in pulmonary venous diastolic flow velocity closely followed changes in early diastolic mitral flow velocity (r = 0.85, p < 0.05). Mean left atrial pressure, or change in mean left atrial pressure, was related to the ratio of pulmonary venous systolic to diastolic velocity time integral (r = 0.59 to 0.62, p < 0.01) and the pulmonary venous systolic fraction (r = 0.58 to 0.60; p < 0.01). When expressed as change from baseline, these variables showed even stronger correlations with left atrial pressure (r = 0.72 to 0.76,p < 0.001). These results are consistent with previous animal and clinical results that indicate pulmonary venous diastolic flow is closely related to early mitral flow velocity, whereas systolic flow is determined primarily by left atrial systolic function. The markedly different relationships observed between left atrial pressure and pulmonary venous systolic flow in animal and clinical studies are most likely due to different baseline flow velocity patterns and differences in left atrial systolic reserve. Future studies investigating these relationships should include data on mitral and pulmonary venous flow velocities as well as left atrial size and systolic function. (JAM Soc EcHocARDIOGR 1994;7:264-75.)

Mitral and pulmonary venous flow velocities obtained with pulsed wave Doppler echocardiography

From the Section of Cardiology, Veterans Administration Medical Center, and the University Heart Center, Tucson, Arizona. Supported by the Flinn Foundation and the Arizona Affiliate of the American Heart Association, Phoenix, Arizona. Reprint requests: Christopher Appleton, MD, Cardiovascular Diseases, Mayo Clinic Scortsdale, 13400 E. Shea Blvd., Scortsdale, AZ 85259. 27/l/51282

264

are being used for the indirect evaluation of left ventricular (LV) diastolic function and filling pressures.1·ll Mitral flow velocity and related variables, such as the LV isovolumic relaxation time, have been correlated with LV end-diastolic pressure in cardiac patients. 2 •6 •10. 12 Pulmonary venous flow velocity variables have also been related to LV filling pressures in patients undergoing cardiac surgery7· 9 or cardiac catheterization. 10·12 In most of these studies/·9 •10·12 the proportion of forward pulmonary venous flow that occurred in systole, the systolic fraction, has

Journal of the American Society of Echocardiography Volume 7 Number 3 Part l

shown a strong negative correlation with mean left atrial, pulmonary wedge, or LV pre-A pressure (the LV pressure just prior to atrial contraction). This inverse relationship has been strong enough (r = >0.80) so that the systolic fraction of pulmonary venous flow has been proposed even as a rapid and reliable method to estimate LV filling pressures in patients undergoing cardiac bypass surgery. 7 More recently, a study in anesthetized, open-chest dogs 13 has shown opposite results, with an acute increase in left atrial pressure associated with an increase in pulmonary venous systolic to diastolic velocity-time integral (r = 0.75). Because pulmonary venous and mitral diastolic flow are closely related7 •8 •1+ 16 and pulmonary venous systolic flow most closely follows atrial systolic function, 10•13•16 we hypothesized that the marked disparity between clinical and experimental results might be explained by differences in baseline left atrial and LV filling patterns and the response of the left atrium to an acute increase in volume and pressure. 17 To test this hypothesis, we recorded changes in mitral and pulmonary venous flow velocities with a transthoracic Doppler technique over a physiologic range of filling pressures in normal dogs. This experimental model was used because of the impracticality of left atrial and LV pressure recording in healthy normal adults and because canine pulmonary venous flow shows a diastolic predominance 13 that is similar to young adults. 7-10 The animals had intact pericardia and were spontaneously breathing with only light sedation.

MATERIALS AND MEmODS

Surgical Preparation

The surgical preparation has been described previously. 18 Briefly, eight mongrel dogs (25-35 kg) were anesthetized with halothane gas, then intubated and given ventilation with a Harvard respirator (Harvard Apparatus, Natick, Mass.) and supplemental oxygen. With a sterile technique, we performed a left thoracotomy in the fifth intercostal space. An 8F silastic catheter was placed l to 2 em into the body of the left atrium via the left superior pulmonary vein and was secured with a purse-string suture. A second 8F silastic catheter was placed into the left lateral pleural space. The silastic catheters were filled with heparin and were placed through the chest wall with tunneling subcutaneously to an area between the scapulae. The muscle, subcutaneous tissue, and skin were closed in layers to provide an airtight seal. Analgesics

Appleton et al. 265

and antibiotics were administered postoperatively. The animals were allowed to recover for a minimum of 7 days before undergoing study. Protocol Instrumentation and Calibration

On the day of the experiment, the animals were sedated with diazepam (0.5 mg/kg) and hydromorphone (0.3 mg/kg). After administration of halothane and nitrous oxide, they were intubated and given mechanical ventilation with a Harvard respirator. With fluoroscopic guidance, a high-fidelity, multisensor (two sensors, 8 em apart) 7F micromanometer catheter with pigtail tip (Millar Instruments, Houston, Tex.) was placed in the LV apex via the previously implanted left pulmonary vein silastic catheter. The catheter was positioned parallel to the long axis of the left ventricle. Intrathoracic pressure was recorded from a SF micromanometertipped catheter placed into the left pleural space. A 7F micromanometer-tipped catheter with fluid-filled reference lumen was inserted percutaneously into the ascending aorta via the left femoral artery. A SF bipolar pacing wire was placed in the right atrial appendage via the right external jugular vein. A flowdirected pulmonary artery catheter was placed via the right external jugular vein for cardiac output measurement. A l2F inflatable, intravascular balloon was placed in the high inferior vena cava via the left femoral vein. After instrumentation, anesthesia was stopped, and the animals were extubated and allowed to recover for 30 minutes before the start of the protocol. An arterial blood gas was checked before data was gathered to check oxygenation and acidbase status. Further sedation was provided as necessary by diazepam in doses ofO.Ol to 0.05 mg/kg. Zero reference for the high-fidelity catheters was obtained by immersing them in body temperature saline for 60 minutes before insertion at a height equal to 50% of the transthoracic diameter of the animal's chest. After insertion, zero reference for the multisensor catheter was checked before each stage by advancing the aortic catheter into the left ventricle and using the fluid-filled lumen. The late diastolic (preatrial contraction) pressure waveforms from the micromanometer transducers were adjusted to match the comparable fluid-filled pressure recording on cardiac cycles with R-R intervals of greater than l 000 msec. 18 The phase of respiration was decided by changes in intrapleural pressure. Data Gathering and Analysis

The hemodynamic and Doppler variables were averaged for three consecutive cardiac cycles obtained

Journal of the American Society of Echocardiography May-June 1994

266 Appleton et a!.

during apnea. Heart rate, LV end-diastolic pressure, and mean aortic and mean left atrial pressure were obtained from the hard copy of the physiologic recorder. The time constant of LV isovolumic relaxation (T 1/2) was measured on-line. 18 At each loading condition left atrial "operating" compliance was estimated by dividing maximum left atrial volume at end systole by left atrial pressure at mitral valve opening. Cardiac output was determined in triplicate by thermodilution technique. Two-dimensional and Doppler echocardiograms were obtained from an apical transducer position with a commercially available ultrasound machine (CFM 700, Vingmed Sound, Salt Lake City, Utah). Left atrial and LV maximum and minimum volumes and ejection fractions were calculated from apical two- and four-chamber views. 10•29 Mitral flow velocity was obtained with pulsed wave Doppler technique, with the sample volume placed between the tips of the mitral valve leaflets. Pulmonary venous flow velocity was obtained with a similar technique, with the sample volume placed approximately 1 em into the right upper pulmonary vein. 10 The mitral and pulmonary venous flow velocity variables measured are shown in Figure l. These variables included peak mitral flow velocity and velocity time integral in early diastole, flow velocity and velocity time integral at atrial contraction, the ratio of velocities and velocity time integrals in early and late diastole, and the mitral deceleration time. Pulmonary venous flow velocity variables measured included peak systolic flow velocity and velocity time integral, peak diastolic flow velocity and velocity time integral, and the ratio of these systolic and diastolic velocities. An attempt was made to record reverse pulmonary venous flow at atrial contraction in all animals. When pulmonary venous systolic flow was biphasic, the highest velocity recorded was considered peak velocity. The systolic fraction of pulmonary venous flow was calculated as the percentage of the total forward time velocity integral that occurred during systole. 7 Protocol

Hemodynamic and echocardiographic data was obtained at four sequential stages in each animal at an atrial pacing rate of 85 beats/min. A standardized rate was used because heart rate can affect both mitral 18 and pulmonary venous flow velocity variables independently. 10 The stages were designed to give a physiologic range of LV end-diastolic pressures and included baseline, volume expansion, increase in afterload together with volume expansion, and increase in afterload with left atrial pressure held at baseline levels by inferior vena cava (IVC) balloon inflation. Preload was increased from baseline by rap-

idly infusing normal saline solution (200 ml/min) until the native sinus rate began to override the atrial pacing. After volume loading, LV end-diastolic pressure was increased further by infusing methoxamine hydrochloride (15-40 J..Lg/min) until mean aortic pressure was increased 40% or until reflex-associated atrioventricular heart block developed. Finally, with methoxamine infusion continuing, left atrial pressure was returned to baseline by inflating the IVC balloon and diminishing venous return. This protocol was approved by the animal research committees of the Tucson Veterans Affairs Hospital and the University of Ariwna. Specific attention was given to the appropriateness and welfare of the animal model, the adequacy of anesthesia, and the methods of instrumentation. This protocol was also in accordance with the "Position of the American Heart Association on Research Animal Use." Statistical Analysis

All results are expressed as mean ± 1 SD. Differences in mean values for each variable were compared between stages with an AN OVA for repeated measures. When differences were found between groups, a Student-Newman-Keuls procedure was used to determine which groups differed. Correlation of hemodynamic and echocardiographic variables was performed with Pearson's correlation coefficients. Significance of all statistical tests was considered p < 0.05. RESULTS Hemodynamics

Heart rate, cardiac output, and select hemodynamics are shown for all stages in Table l. Data from an individual animal is shown in Figure 2. LV enddiastolic pressure ranged from 7.4 ± 1.9 to 21.6 ± 2.9 mm Hg (p < 0.05). Cardiac output was 4.06 ± 0.6 L/min at baseline and increased to 5.79 ± 1.15 L/min after volume infusion (p < 0.05). Two-dimensional and Doppler Variables

Two-dimensional and Doppler variables are shown in Tables 1 and 2. Representative changes in mitral and pulmonary venous flow velocities are shown in an individual animal in Figure 3. With increasing filling pressures, left atrial operating compliance decreased, but left atrial volume and ejection fraction increased. Changes in pulmonary venous diastolic flow velocity closely followed changes in early mitral flow velocity (r = 0.85,p < 0.05). At baseline conditions pulmonary venous diastolic velocity was larger than the systolic velocity (66 ± 9 versus

Journal of the American Society of Echocardiography Volume 7 Number 3 Part l

Appleton et a!.

pulmonary venous flow velocity

e·cG m/s

o.e

PYa

.PVd

B *$

- -

Figure 1 Mitral and pulmonary venous flow velocities obtained with transthoracic Doppler echocardiography illustrating the variables measured in this study. A, The mitral variables included peak mitral flow velocity (Ml) and flow velocity integral (Ml VII) in early diastole, peak mitral flow velocity (.M2) and velocity time integral (.M2 VII) in late diastole, and mitral deceleration time (Mdt). B, Pulmonary venous flow velocity variables included peak flow velocity (PVs) and velocity time integral (PVs VII) during ventricular systole and peak flow velocity (PVd) and velocity time integral during ventricular diastole (PVd VII).

267

Journal of the American Society of Echocardiography May-June 1994

268 Appleton et al.

Table 1 Hemodynamics and left atrial and LV ejection fractions in eight dogs at baseline and different loading conditions Basellne

Heart rate (beats/min) Ao mean (mm Hg) co (Limin) Mean LAP (mm Hg) LVEDP (mm Hg) T V2 (msec) Left Atrium Maximum volume (cm3 ) Minimum volume (cm3 ) Compliance (cm3 /mm Hg) LAEF (%) Left ventricle Maximum volume (cm3 ) Minimum volume (cm 3 ) LV EF (%)

Volume expansion

Volume plus afterload

Afterload/NI preload

1.9 5.1

89.6 94.3 5.79 11.4 16.2 58.3

± ± ± ± ± ±

5.8 8.3 1.15* 1.3* 2.4* 5.8

87.0 118.9 4.99 16.1 21.6 64.2

± ± ± ± ± ±

3.9 7.4*,t 1.09 l.7*,t 2.9*,t 4.3*,t

85.3 97.6 3.92 5.7 8.6 57.8

± ± ± ± ± ±

5.5 9.7:j: l.Olt l.Ot:f: 2.3t:f: 5.3*,:j:

9.3 9.7 1.3 9.3

57.7 31.2 4.2 42.8

± ± ± ±

14.9* 10.0 1.7* 11.6

64.1 31.4 3.4 45.8

± ± ± ±

15.1* 8.3* 1.3* 8.5*

38.5 21.4 5.3 38.7

± ± ± ±

12.4t:f: 8.4t l.5t 1l.Ot

85.8 87.6 4.06 5.3 7.4 51.2

± ± ± ± ± ±

4.6 8.4 0.6

38.7 22.9 6.1 36.7

± ± ± ±

l.l

64.2 ± 10.2 27.1 ± 7.9 57.4 ± 9.9

88.9 ± 15.2* 39.7 ± 8.6 55.3 ± 7.1

90.6 ± 19.7* 43.6 ± 8.7* 51.7 ± 8.8

64.8 ± 15.4t:f: 28.8 ± 6.1 t:f: 53.6 ± 10.4

All values are mean ± SD. Left atrial and LV volumes are averages of apical rwo- and four-chamber two-dimensional imaging views. Afterload, increase in systemic blood pressure with methoxamine infusion; Ao mean, mean aortic pressure; CO, cardiac ourput; EF, ejection fraction; L VEDP, left ventricular enddiastolic pressure; mean LAP, mean left atrial pressure; Nl preload, normal preload; T 1!2, the time constant of LV isovolumic relaxation; LA, left atrial. *p < 0.05 compared with baseline. tp < 0.05 compared with volume expansion. "4:p < 0.05 compared with volume plus afterload.

36 ± 11 em/sec), with a systolic fraction of 31% ± 8%. The pulmonary venous systolic fraction increased to 44% ± 7% after volume loading and then to 48% ± 9% after methoxamine infusion (both p < 0.05). Correlation of Doppler Variables with Mean Left Atrial Pressure

The correlations between Doppler variables and left atrial pressure (Tables 3 and 4) were calculated with the left atrial pressure expressed in three ways: as an absolute value, as the change from baseline, 7 and as the quotient of the intervention value divided by the control value. 8 The best correlation with mean left atrial pressure, or change in mean left atrial pressure, was with the ratio of pulmonary venous systolic to diastolic velocity time integral (r = 0.59 to 0.62, p < 0.01) and the pulmonary venous systolic fraction (r = 0.58 to 0.60, p < 0.01, Figure 4, A and B). If the pulmonary venous variables were expressed as change from baseline, even stronger correlations couldbeobtained(r = 0.72to0.76,p < 0.001,Figure 4, C).

DISCUSSION

Previous clinical studies have shown a strong inverse relationship between LV filling pressures and pul-

monary venous systolic fraction. 7 •9 •10•12 In contrast, an opposite, positive relationship has been demonstrated in anesthetized, open-chest dogs. 13 In an attempt to explain this difference, this study in conscious dogs was designed to reexamine these relationships while simultaneously analyzing changes in mitral flow velocity and left atrial ejection fraction. The results are similar to those in open-chest dogs 13 and show that with acute increases in left atrial pressure, both the systolic fraction of pulmonary venous flow and left atrial ejection fraction increase. The marked disparity in results compared with clinical studies appear to be the result of differences in baseline mitral and pulmonary venous flow velocity patterns and to differences in atrial systolic reserve. The clinical implication is that pulmonary venous flow velocities should be used to estimate left atrial pressures only after considering baseline atrial and ventricular filling characteristics and left atrial systolic function. Comparison with Prior Studies

Blood flow in the pulmonary veins is biphasic, with peaks of forward flow occurring in both systole and diastole. 14•20 In normal individuals pulmonary venous flow changes with age; diastolic flow predominates before the age of 40 before gradually diminishing in later life as the rate of LV relaxation and LV filling in early diastole slows. 21 •22 By the sixth decade LV filling in early and late diastole is approximately

Journal of the American Society of Echocardiography Volume 7 Number 3 Part l

Appleton et al.

269

Ao

mmHg 20

I

·:.tJ I

0·

~...:....-.~t~

ECG-•_...,[

baseline

1-JL volume

!

.l ____,----f\-

vol & methox

v&m/nl preload

Figure 2 Simultaneous left atrial (LA) and left ventricular (LV) pressure in an individual animal at baseline and during the different loading conditions studied. Heart rate is constant at 85 to 90 beats/min because of right atrial pacing. Note the increase in LA, LV end-diastolic and aortic (Ao) pressure after volume loading (volume) and after volume loading and methoxamine infusion (vol & methox). Last panel, LA pressure has been lowered to near baseline level after volume and methoxamine infusion (v&mlnl preload) by inflation of a balloon located in the inferior vena cava. IP, Intrapleural pressure.

equal, and peak pulmonary venous systolic flow velocity usually exceeds diastolic flow velocity. 21 ·22 This changeover to pulmonary venous systolic flow predominance is accelerated in patients with LV hypertrophy or coronary artery disease, where left atrial and LV filling in early diastole is reduced because of a slowing of LV relaxation. 4•7. 10 In patients with advanced cardiac disease, pulmonary venous diastolic flow may again predominate if high filling pressures increase LV filling in early diastole 5•7•9•10 and atrial systolic dysfunction occurs. 10 With transesophageal echocardiography, Kuercherer et al. 7•9 measured mitral and pulmonary venous flow velocities with left atrial pressure in patients undergoing cardiac surgery. Atrial systolic function was not reported. Patients with a mean left atrial pressure of less than 15 mm Hg had approximately equal early and late diastolic mitral flow velocity and a predominance of pulmonary venous systolic flow. Patients with a left atrial pressure of 15 mm Hg or more had an increase in early diastolic mitral flow velocity and a predominance of pulmonary diastolic flow. The proportion of total flow velocity that occurred in systole (the systolic fraction) correlated most strongly with mean left atrial pressure

(r = - 0.88). Similarly, changes in the systolic fraction also showed a strong inverse relationship with changes in mean left atrial pressure (r = - 0. 78), suggesting this relationship could be used intraoperatively to estimate left atrial pressure. More recently, Rossvoll and Hatle 12 reported that in patients undergoing cardiac catheterization a pulmonary venous systolic fraction of less than 40% identifies all patients with a LV pre-A wave pressure of more than 18 mm Hg, with the correlation between the two variables -0.70. We have reported a somewhat weaker inverse relationship (r = - 0.47) between peak pulmonary venous systolic flow velocity and pulmonary wedge pressure in a similar patient population.10 Nishimura et al. 8 also studied the effect of different loading conditions on mitral and pulmonary venous flow velocities in patients undergoing cardiac surgery. A direct correlation existed between changes in early diastolic mitral velocity and peak pulmonary diastolic velocity (r = 0.61) and between the deceleration time of mitral and pulmonary venous diastolic flows (r = 0.84). Changes in pulmonary venous systolic flow velocity were most strongly related to changes in cardiac output (r = 0.60). Pulmonary

Journal of the American Society of Echocardiography May-June 1994

270 Appleton et al.

A

baaellne

volume

vol & methox

pulmonary venous flow velocity

ECG··

m/s

PVd.

0.6

B

baseline

volume

vol & methox

Figure 3 Pulsed Doppler mitral and pulmonary venous flow velocities together with left atrial (LA) and left ventricular (LV) pressure in an individual animal under different loading conditions. A, After volume infusion (volume) and volume combined with methoxamine (vol & metho.x) peak mitral flow velocity in early diastole and at atrial contraction both increase; the ratio of early to late diastole decreasing slightly. B, At baseline pulmonary venous diastolic flow velocity (PVd) predominates, and pulmonary venous systolic flow velocity (PVs) is small. After volume infusion LA pressure (LAP) increases, and there is an increase in pulmonary venous systolic flow velocity with a minimal increase in diastolic flow velocity. When methoxamine is added in addition to the volume LA pressure and peak systolic flow velocity and velocity time integral increase further. Note that peak diastolic pulmonary flow velocity also increases but that less change occurs in the flow velocity integral because of a more rapid deceleration of flow.

Journal of the American Society of Echocardiography Volume 7 Number 3 Part l

Appleton et al.

271

20

o; I

E

15

v

~

10

c 0 v

5

_s

0

0

0

0: y = -1.45 + 0.25x

0

r-o.se.n-32

0

"'

0

20

30

A

SEE = 3.69 mm Hg

0

50

40

15

15

I

E

Y = 1.13 + O.J6x r - 0.73, n - 32 SEE= 3.13 mm Hg

o;

"'

_s

60

Systolic Fraction (%)

I

E

_s

10

0

10 0

0 0

0

0

0

0

0

c 0

v

"'

0

0

0

G>

0

0

y = -5.99 + 0.25x 32 0.58, n r SEE .. 3.74 mm Hg

=

=

0

c

0

v

0

"'

-5L-~------L-----~----~------L---

20

8

0

0



30

40

50

60

Systolic Fraction (%)

c

- 5 ~---1~0------~------~1~0------~20~----

6

Systolic Fraction (%)

Figure 4 Scatterplots of the correlation between the left atrial (LA) pressure variables and the systolic fraction of pulmonary venous flow. A, Mean left atrial pressure versus pulmonary venous systolic fraction; B, change (A) in mean left atrial pressure compared with baseline at each loading condition versus pulmonary venous systolic fraction; and C, change (A) in mean left atrial pressure compared with baseline at each loading condition versus the change compared with baseline in pulmonary venous systolic fraction. SEE, Standard error of the estimate.

wedge pressure correlated best with flow reversal in the pulmonary vein at atrial contraction (r = 0.81); the correlation between wedge pressure and pulmonary venous systolic flow velocity variables was not reported. Our study also showed a correlation between changes in pulmonary venous systolic flow velocity and cardiac output (r = 0.54, p < 0.01). In anesthetized, open-chest dogs, Hoit et al. 13 studied the influence of three loading conditions on atrial function and pulmonary venous flow. Changes in mitral flow velocities were not analyzed. With volume loading mean left atrial pressure varied from 7 to 19 mm Hg. Over this range mean heart rates varied between 95 and 116 beats/min. Compared with the current study, pulmonary venous peak systolic flow velocity was higher (43 versus 36 em/sec) and peak diastolic flow velocity was lower (43 versus 66 em/sec). The results showed that left atrial pressure was correlated with the ratio of pulmonary venous systolic to diastolic velocities (r = 0.64) and

velocity time integrals (r = 0.75). However, when the same protocol was repeated during halothane infusion, little change was seen in the peak velocity ratio, and an opposite change was observed in the ratio of flow velocity integrals. At the same time atrial systolic shortening, as assessed by sonomicrometer crystals, was seen to decrease markedly after halothane infusion, where it had increased previously. Stepwise multiple regression analysis indicated that atrial systolic shortening was the most important determinant of the ratio of pulmonary venous systolic to diastolic flow. Although the pulmonary venous flow velocities and integrals are somewhat different, the results of the current study in conscious dogs are similar to those studied by Hoit et al. 13 that did not receive halothane. The positive correlations between left atrial pressure and pulmonary venous variables are similar (Table 3), and changes in the ratio of pulmonary systolic to diastolic velocities paralleled

Journal of the American Society of Echocardiography May-June 1994

272 Appleton eta!.

Table 2 Doppler mitral and pulmonary venous flow velocity variables in eight dogs at baseline and at different loading conditions Variable

Doppler mitral inflow Peak early diastolic flow velocity (em/sec) Peak late diastolic flow velocity (em/sec) Ratio peak early I late diastolic velocity Early diastolic VTI (em) Late diastolic VTI (em) Early I late diastolic VTI Mitral deceleration time (msec) Doppler pulmonary venous Peak systolic flow velocity (em/sec) Peak diastolic flow velocity (em/sec) Ratio peak systolic/ diastolic velocity Systolic VTI (em) Diastolic VTI (em) Ratio systolic I diastolic VTI Systolic fraction of VTI (%)

Volume expansion

Volume plus afterload

Afterload/Nl preload

79 ± 7

88 ± 7*

81 ± 10

72 ± 9t

46 ± 9

58± 10

62 ± 12*

52 ± 9

1.60 ± 0.31

1.33 ± 0.21*

1.41 ± 0.28*

8.26 4.82 1.84 103

8.41 4.24 2.1'2 119

Baseline

1.82 ± 0.27 9.22 3.32 2.93 121

± ± ± ±

0.77 0.56 0.40 18

10.23 4.73 2.31 114

± ± ± ±

1.18 1.22* 0.44* 21

± ± ± ±

l.OO*t 1.15* 0.50* 15

± ± ± ±

l.20t l.08t 0.61* 19

36 ± 11

61 ± 15*

57± 16*

38 ± 16t:J:

66 ± 9

81 ± 14

71 ± 13

59± 11t

0.54 ± 0.16

0.75 ± 0.14*

0.81 ± 0.18*

0.63 ± 0.20

6.95 15.21 0.48 31

± ± ± ±

2.25 3.05 0.20 8

11.06 14.39 0.80 44

± ± ± ±

2.51* 3.51 0.23* 7*

10.19 11.27 0.96 48

± ± ± ±

2.49 3.07* 0.36* 9*

6.80 11.76 0.58 36

± ± ± ±

2. 73t:J: 3.04 0.20:j: 8:j:

All values are mean ± SD. V77, velocity time integral; see Table l for explanation of other abbreviations.

*p < 0.05 compared with baseline. tp < 0.05 compared with volume expansion. t-p < 0.05 compared with volume plus overload.

changes in left atrial ejection fraction (Tables l and 2). Left atrial minimum volume and ejection fraction are also the most important determinants of pulmonary venous systolic flow in patients with coronary artery disease undergoing cardiac catheterization.10 Resolution of Results That Differed from Experimental and Clinical Studies Resolution of differences can be achieved by considering the determinants of mitral and pulmonary venous flow velocities and viewing both as an interrelated, dynamic continuumP As shown in this and prior clinical studies, 8 - 10·14-16 pulmonary venous diastolic flow closely follows early diastolic mitral flow velocity and presumably the factors (rate of LV relaxation, left atrial pressure, and transmitral pressure gradient) that determine LV filling. Pulmonary venous systolic flow is related to atrial systolic function 10·13·16 and, to a lesser degree, LV systolic function. 9 Clinical studies investigating changes in pulmonary venous flow with altered loading conditions 7•8 have been composed of patients with reduced LV filling in early diastole, normal filling

pressures, and an increased proportion of pulmonary venous systolic flow. In these cases an acute increase in left atrial pressure increases the early diastolic transmitral pressure gradient and early diastolic mitral and pulmonary venous flow. If no change occurs in cardiac output, systolic flow would have to decrease, perhaps markedly, if a decrease in atrial systolic function occurred. In contrast, young adults have more LV filling in early diastole and a predominance of pulmonary venous diastolic flowY- 22 As seen in normal dogs, an acute elevation in left atrial pressure in normal subjects is associated with an increase in left atrial ejection fraction, an atrial FrankStarling response/ 2 and an increase in pulmonary venous systolic flow. Pulmonary venous diastolic flow may increase or decrease depending on what happens to the determinants of the transmitral pressure gradient and early mitral flow. Therefore analysis of pulmonary venous flow is aided by considering mitral flow velocity and its determinants as well as left atrial systolic function. Clinical Implications The experimental results discussed do not invalidate previous clinical observations/·9•10·12 but they do im-

Journal of the American Society of Echocardiography Volume 7 Number 3 Part l

Appleton et al.

273

Table 3 Correlation of Doppler variables with mean left atrial pressure, the change (•) in mean left atrial pressure from control, and mean left atrial pressure expressed as quotient of intervention divided by control (I + C) in eight dogs at all loading conditions studied Mean LA pressure

pressure

Mean lA pressure (I+C)

0.35*

0.32

0.24

0.36*

0.43*

0.48**

~eaniA

Variable

Doppler mitral inflow Peak early diastolic velocity (em/sec) Peak late diastolic velocity (em/sec) Peak early /late diastolic velocity (em/sec) Early diastolic VTI (em) Late diastolic VTI (em) Ratio early /late diastolic VTI Mitral deceleration time (msec) Doppler pulmonary venous Peak systolic velocity (em/sec) Peak diastolic velocity (em/sec) Ratio peak systolic/ diastolic velocity Systolic VTI (em) Diastolic VTI (em) Ratio systolic/diastolic VTI Systolic fraction ofVTI (%)

-0.15

-0.22

-0.29

-0.06 0.27 -0.30

0.03 0.34 -0.32

0.06 0.41* -0.33

-0.44*

-0.38*

-0.35*

0.61***

0.50**

0.35*

0.34

0.21

0.18

0.60***

0.54**

0.48**

0.44* -0.28 0.60*** 0.58***

0.41* -0.32 0.62*** 0.60***

0.46** -0.25 0.59*** 0.58***

VII, velocity time integral; LA, left atrial. Statistics: * = p < 0.05; ** = p < 0.01; *** = p < 0.001.

ply that estimating LV filling pressures with pulmonary venous variables should be performed only after consideration of the patient's age, clinical history, mitral flow velocity pattern, and atrial systolic function. Although an inverse relationship between the systolic fraction of pulmonary venous flow and mean left atrial pressure appears to work well in patients older than 50 years of age who have coronary artery disease, this relationship is unlikely to be valid in younger individuals who have markedly different baseline mitral flow velocity patterns. Similarly, interpreting hemodynamics from acute changes in pulmonary venous flow is aided by assessing corresponding changes in mitral flow velocity and left atrial size and function, and this data should be included in future studies. Study Limitations

Several limitations to this study exist. First, results from normal dogs have been compared with results from clinical studies. A human study group of normal individuals would have been ideal, but such a study was impractical because of the hemodynamic instru-

mentation required. Normal dogs have less pulmonary venous systolic flow than normal young adults (unpublished data), perhaps reflecting some species variability. However, because of similarities in myocardial structure, hemodynamics, and mitral flow velocity/8 we believe the results reported would be similar to healthy young adults. Pulmonary venous flow velocities in this study were significantly different from the other experimental study discussed. 13 Several possible reasons exist. These reasons include different types of experimental design (anesthetized, open chest versus lightly sedated, spontaneously breathing), intact versus open pericardium, different heart rates, different LV filling pressures, different sample volume locations, and different methods (transesophageal versus transthoracic) of recording pulmonary venous flow velocities. Unfortunately, mitral flow velocities were not reported in the prior study, which would have helped determine whether the observed differences were physiologic or methodologic. Anesthesia may slow the rate of LV relaxation and result in a reduced proportion of early diastolic mitral and pulmonary

Journal of the American Society of Echocardiography May-June 1994

274 Appleton et al.

Table 4 Correlation of changes (A) in Doppler variables with mean left atrial pressure, the change (A) in mean left atrial pressure from control, and mean left atrial pressure expressed as quotient of intervention divided by control (I-;- C) in eight dogs at all loading conditions studied Variable

A Mean LA

Mean LA pressure

Mean LA pressure

pressure

(I+C)

0.38* 0.53** -0.31 -0.07 0.34 -0.29 -0.45**

0.39* 0.50** -0.25 -0.03 0.31 -0.18 -0.49**

0.29 0.50** -0.24 -0.02 0.32 -0.11 -0.51**

0.63** 0.37* 0.63*** 0.67*** -0.20 0.72*** 0.72***

0.59*** 0.35* 0.60*** 0.67*** -0.16 0.75*** 0.73***

0.47** 0.27 0.54** 0.63*** -0.15 0.76*** 0.72***

Doppler mitral inflow A Peak early diastolic flow velocity (em/sec) A Peak late diastolic velocity (em/sec) A Peak early/late diastolic velocity (em/sec) A Early diastolic VTI (em) A Late diastolic VTI (em) A Ratio early I late diastolic VTI A Mitral deceleration time (msec) Doppler pulmonary venous A Peak systolic velocity (em/sec) A Peak diastolic velocity (em/sec) A Ratio peak systolic/ diastolic velocity A Systolic VTI (em) A Diastolic VTI (em) A Ratio systolic/ diastolic VTI A Systolic fraction ofVTI (%) VTI, velocity time integral; LA, left atrial. Statistics: * = p < 0.05; ** = p < 0.01; *** = p < 0.001.

venous flow velocity. Alternately, pulmonary venous systolic velocities are about lO em/ sec lower when recorded by transthoracic as compared with transesophageal technique. 24 We also used a sample volume placement l em into the pulmonary vein rather than at the venoatrial orifice. Although this method facilitates the recording of reverse pulmonary venous flow at atrial contraction, systolic pulmonary venous flow may be slightly lower at this location. Whatever the reason for the differences, the two studies remain complementary by showing that normal dogs, whether conscious or anesthetized, respond to acute increases in filling pressures by increasing atrial ejection fraction and the proportion of pulmonary venous systolic flow. Mean left atrial pressures in this study were lower than in other studies examining acute changes in pulmonary venous flow. 7•8 •13 However, maximum LV end-diastolic pressures did exceed 20 mm Hg. Higher acute filling pressures could not be obtained without altering heart rate, which can independently alter mitral and pulmonary venous flow velocity variables 4 •10•18 and make serial interventions difficult to compare. We believe that alterations in heart rate have made previous results difficult to interpret, especially in cases where mitral flow velocity was not reported. PR interval changes, which were not controlled in this or any previous studies, can also affect peak pulmonary venous systolic velocity by altering the velocity at the start of the later systolic compo-

nent. As shown by Hoit et al., 13 it is likely that very high filling pressures might have resulted in atrial systolic dysfunction and decreasing pulmonary venous systolic flow. If true, a parabolic relationship between pulmonary flow variables with left atrial pressure would have been seen, again indicating the need to consider baseline atrial filling and systolic function. Atrial contractility was estimated by use of twodimensional imaging ejection fractions. Although these methods make certain geometric assumptions, 19 the changes in atrial size and function observed with increasing pressures are similar to those observed in open-chest animals studied with left atrial sonomicrometer crystals. 13 Conclusion The results of this study and previous experimental and clinical studies show that pulmonary venous diastolic flow closely follows early diastolic mitral flow velocity, whereas pulmonary venous systolic flow is determined mainly by left atrial systolic function. Like mitral flow velocity, pulmonary venous flow represents a dynamic continuum, which varies with age, cardiac disease, and loading conditions. Therefore pulmonary venous flow variables should be used to estimate filling pressures only after considering the clinical history, baseline mitral and pulmonary venous flow velocities, and left atrial size and function. Future studies investigating pulmonary venous flow

Journal of the American Society of Echocardiography Volume 7 Number 3 Pan l

Appleton et al.

and hemodynamics should include this data to aid in interpreting the results. 12.

REFERENCES l. Kitabatake A, Inoue M, Asao M, et al. Transmitral blood flow reflecting diastolic behavior of the left ventricle in health and disease. A study by pulsed Doppler technique. Jpn Circ J 1982;46:92-102. 2. Appleton CP, Harle LK, Popp RL. The relationship of transmitral flow velocity patterns to left ventricular diastolic function: new insights from a combined hemodynamic and Doppler echocardiographic study. JAm Coli Cardiol1988;12:42640. 3. Stoddard MF, Pearson AC, Kern MJ, Ratcliff], Mrosek DG, Labovitz AJ. Left ventricular diastolic function: comparison of pulsed Doppler echocardiographic and hemodynamic indexes in subjects with and without coronary artery disease. JAm Coli Cardiol 1989;13:327-36. 4. Nishimura RA, Abel MD, Harle LK, Tajik AJ. Assessment of diastolic function of the heart: background and current applications of Doppler echocardiography. Part II. Clinical studies. Mayo Clin Proc 1989;64:181-204. 5. Klein AL, Harle LK, Taliercio CP, et al. Serial Doppler echocardiographic follow-up of LV diastolic function in cardiac amyloidosis. JAm Coli Cardiol 1990;16: 1135-41. 6. Vanoverschelde JJ, Raphael DA, Robert AR, Cosyns JR. Left ventricular filling in dilated cardiomyopathy: relation to functional class and hemodynamics. J Am Coli Cardiol 1990;15:1288-95. 7. Kuercherer HF, Muhiudeen IA, Kusumoto FM, et al. Estimation of mean left atrial pressure from transesophageal pulsed Doppler echocardiography of pulmonary venous flow. Circulation 1990;82: 1127-39. 8. Nishimura RA, Abel MD, Hade LK, Tajik AJ. Relation of pulmonary vein to mitral flow velocities by transesophageal echocardiography. Effect of different loading conditions. Circulation 1990;81:1488-97. 9. Kuercherer HF, Kusumoto FM, Muhiudeen IA, Cahalan MK, Schiller NB. Pulmonary venous flow patterns by transesophageal pulsed Doppler echocardiography: relation to parameters ofleft ventricular systolic and diastolic function. Am Heart J 1991;122:1683-92. 10. Basnight MA, Gonzalez MS, Kershenovich SC, Appleton CP. Pulmonary venous flow velocity: relation to hemodynamics, mitral flow velocity and left atrial volume and ejection fraction. JAM Soc EcHOCARDIOGR 1991;4:547-58. 11. Mulvagh S, Quinones MA, Kleiman NS, Jorge CB, Zoghbi W A. Estimation ofleft ventricular end-diastolic pressure from

13.

14.

15.

16. 17.

18. 19.

20. 21.

22.

23. 24.

275

Doppler transmitral flow velocity in cardiac patients independent of systolic performance. J Am Coli Cardiol 1992; 20:112-9. Rossvoll 0, Harle LK. Pulmonary venous flow velocities recorded by transthoracic Doppler ultrasound: relations to left ventricular diastolic pressures. J Am Coli Cardiol 1993; 21:1687-96. Hoit BD, Shao Y, Gabel M, Walsh RA. Influence ofloading conditions and contractile state on pulmonary venous flow. Validation by Doppler velocimetry. Circulation 1992;86: 651-9. Keren G, Sherez J, Megedish R, Levitt B, Laniado S. Pulmonary venous flow pattern-its relationship to cardiac dynamics. A pulsed Doppler echocardiographic study. Circulation 1985;71:1105-12. Keren G, Sonnenblick EH, LeJemtel TH. Mitral annulus motion. Relation to pulmonary venous and transmitral flows in normal subjects and in patients with dilated cardiomyopathy. Circulation 1988;78:621-9. Keren G, Bier A, Sherez J, Miura D, Keefe D, LeJemtel T. Atrial contraction is an important determinant of pulmonary venous flow. J Am Coli Cardiol 1986;7:693-5. Appleton CP, Harle LK. The natural history ofleft ventricular filling abnormalities: assessment by two-dimensional and Doppler echocardiography. Echocardiography 1992;9:43757. Appleton CP. Influence of incremental changes in heart rate on mitral flow velocity: assessment in lightly sedated, conscious dogs. JAm Coli Cardiol 1991;17:227-36. Ren JF, Kottler MN, DePace NL, et al. Two-dimensional echocardiographic determination of left atrial emptying volume: a non-invasive index in quantifYing the degree of nonrheumatic mitral regurgitation. J Am Coli Cardiol 1983; 2:729-36. Skagseth E. Pulmonary vein flow pattern in man during thoracotomy. Scand J Thorac Cardiovasc Surg 1976;10:36-42. Masuyama T, Lee J-M, Tarnai M, Tanouchi J, Kitabatake A, Kamada T. Pulmonary venous flow velocity pattern as assessed with transthoracic pulsed Doppler echocardiography in subjects without cardiac disease. Am J Cardiol 1991; 67:1396-1404. Klein AL, Tajik AJ. Doppler assessment of pulmonary venous flow in healthy subjects and patients with heart disease. J AM Soc ECHOCARDIOGR 1991;4:379-92. Payne RM, Stone HL, Engelken EJ. Atrial function during volume loading. J Appl Physiol 1971;31:326-31. Castello R, Pearson AC, Lenzen P, Labowitz AJ. Evaluation of pulmonary venous flow by transesophageal echocardiography in subjects with a normal heart: comparison with transthoracic echocardiography. J Am Coli Cardiol 1991; 18:65-71.