Ring annuloplasty prevents delayed leaflet coaptation and mitral regurgitation during acute left ventricular ischemia.

OBJECTIVE: Incomplete mitral leaflet coaptation during acute left ventricular ischemia is associated with end-diastolic mitral annular dilatation and ischemic mitral regurgitation. Annular rings were implanted in sheep to investigate whether annular reduction alone is sufficient to prevent mitral regurgitation during acute posterolateral left ventricular ischemia. METHODS: Radiopaque markers were inserted around the mitral anulus, on papillary muscle tips, and on the central meridian of both mitral leaflets in three groups of sheep: control (n = 5), Physio ring (n = 5) (Baxter Cardiovascular Div, Santa Ana, Calif), and Duran ring (n = 6) (Medtronic Heart Valve Div, Minneapolis, Minn). After 8 +/- 1 days, animals were studied with biplane videofluoroscopy before and during left ventricular ischemia. Annular area was calculated from 3-dimensional marker coordinates and coaptation defined as minimal distance between leaflet edge markers. RESULTS: Before ischemia, leaflet coaptation occurred just after end-diastole in all groups (control 17 +/- 41, Duran 33 +/- 30, Physio 33 +/- 24 ms, mean +/- SD, P >.2 by analysis of variance). During ischemia, regurgitation was detected in all control animals, and leaflet coaptation was delayed to 88 +/- 8 ms after end-diastole (P =.02 vs preischemia). This was associated with increased end-diastolic annular area (8.0 +/- 0.9 vs 6.7 +/- 0.6 cm(2), P =.004) and septal-lateral annular diameter (2.9 +/- 0.1 vs 2.5 +/- 0.1 cm, P =.02). Mitral regurgitation did not develop in Duran or Physio sheep, time to coaptation was unchanged (Duran 25 +/- 25 ms, Physio 30 +/- 48 ms [both P >.2 vs preischemia]), and annular area remained fixed. CONCLUSION: Mitral annular area reduction and fixation with an annuloplasty ring eliminated delayed leaflet coaptation and prevented mitral regurgitation during acute left ventricular ischemia after ring implantation.

Deformational dynamics of the aortic root: modes and physiologic determinants.

BACKGROUND: Current surgical methods for treating aortic valve and aortic root pathology vary widely, and the basis for selecting one repair or replacement alternative over another continues to evolve. More precise knowledge of the interaction between normal aortic root dynamics and aortic valve mechanics may clarify the implications of various surgical procedures on long-term valve function and durability. METHODS AND RESULTS: To investigate the role of aortic root dynamics on valve function, we studied the deformation modes of the left, right, and noncoronary aortic root regions during isovolumic contraction, ejection, isovolumic relaxation, and diastole. Radiopaque markers were implanted at the top of the 3 commissures (sinotubular ridge) and at the annular base of the 3 sinuses in 6 adult sheep. After a 1-week recovery, ECG and left ventricular and aortic pressures were recorded in conscious, sedated animals, and the 3D marker coordinates were computed from biplane videofluorograms (60 Hz). Left ventricular preload, contractility, and afterload were independently manipulated to assess the effects of changing hemodynamics on aortic root 3D dynamic deformation. The ovine aortic root undergoes complex, asymmetric deformations during the various phases of the cardiac cycle, including aortoventricular and sinotubular junction strain and aortic root elongation, compression, shear, and torsional deformation. These deformations were not homogeneous among the left, right, and noncoronary regions. Furthermore, changes in left ventricular volume, pressure, and contractility affected the degree of deformation in a nonuniform manner in the 3 regions studied, and these effects varied during isovolumic contraction, ejection, isovolumic relaxation, and diastole. CONCLUSIONS: These complex 3D aortic root deformations probably minimize aortic cusp stresses by creating optimal cusp loading conditions and minimizing transvalvular turbulence. Aortic valve repair techniques or methods of replacement using unstented autograft, allograft, or xenograft tissue valves that best preserve this normal pattern of aortic root dynamics should translate into a lower risk of long-term cusp deterioration.

Functional evaluation of the medtronic stentless porcine xenograft mitral valve in sheep.

BACKGROUND: Recently, renewed interest in allograft and stentless "freehand" bileaflet xenograft mitral valve replacement has arisen. The variability of human papillary tip anatomy and scarcity of donors limit allograft availability, making xenograft mitral valves an attractive alternative; however, these valves require new surgical implantation techniques, and assessment of their hemodynamics and functional geometry is lacking. METHODS: Seven sheep underwent implantation of a new stentless, glutaraldehyde-preserved porcine mitral valve (Physiological Mitral Valve [PMV], Medtronic) and were studied acutely under open-chest conditions. A new method of retrograde cardioplegia was developed. Hemodynamic valve function was assessed by epicardial Doppler echocardiography. 3D motion of miniature radiopaque markers sutured to the valve leaflets, annulus, and papillary tips was measured. Six other sheep with implanted markers served as controls. RESULTS: Both papillary muscle tips avulsed in the first animal, leaving 6 other animals. Mitral regurgitation was not observed in any xenograft valve. The peak and mean transvalvular gradients were 4.6+/-1.8 mm Hg and 2.6+/-1.5 mm Hg, respectively. The average mitral valve area was 5.7+/-1.6 cm(2). Valve closure in the xenograft group occurred later (30+/-11 ms, P<0. 015) and at higher left-ventricular pressure (61+/-9 mm Hg, P<0.001) than in the control group; furthermore, leaflet coaptation was displaced more apically (5.6+/-2.2 mm, P<0.001) and septally (5. 8+/-1.5 mm, P<0.001), and the anterolateral papillary tip underwent greater septal-lateral displacement (2.7+/-1.5 mm, P<0.001). Annular contraction during the cardiac cycle was similar in the 2 groups (xenograft 9.2+/-4.5% versus control 10.6+/-4.5% [mean+/-SD; 2-factor ANOVA model]). CONCLUSIONS: Successful freehand stentless porcine mitral valve implantation is feasible in sheep and was associated with excellent early postoperative hemodynamics. Physiological mitral valve annular contraction and functional leaflet closure mechanics were preserved. Long-term valve durability, calcification, and hemodynamic performance remain to be determined in models.

Mitral annular dilatation and papillary muscle dislocation without mitral regurgitation in sheep.

BACKGROUND: Asymmetrical mitral annular (MA) dilatation and papillary muscle dislocation are implicated in the pathogenesis of functional mitral regurgitation (MR). METHODS AND RESULTS: To determine the mechanism by which annular and papillary muscle geometric alterations result in MR, we implanted radiopaque markers in the left ventricle, mitral annulus, anterior and posterior mitral leaflets, and papillary muscle tips and bases in 2 groups of sheep. One group served as controls (CTL, n=7); an experimental group (EXP, n=9) underwent topical phenol application to obliterate anterior annular and leaflet muscle (confirmed histologically ex vivo). After 1 week of recovery, markers were imaged with biplane videofluoroscopy, and hemodynamic data were recorded. MA area (computed from 3-dimensional marker coordinates) was 11% to 13% larger in the EXP group than in the CTL group (P<0.05 by ANOVA). This area increase resulted exclusively from intercommissural axis increase except in 1 heart with large (>1 cm) increases in both the intercommissural and septolateral annular axes. The anterior papillary muscle tip in EXP was displaced from CTL by 2.9+/-0.23 mm toward the anterolateral left ventricle and 2.5+/-0.12 mm toward the mitral annulus at end systole; the posterior papillary muscle geometry was unchanged. Transthoracic echocardiography revealed MR only in the heart exhibiting biaxial annular enlargement. CONCLUSIONS: MA dilatation in the intercommissural dimension with anterior papillary muscle tip displacement toward the annulus is insufficient to produce MR in sheep. Functional MR may require MA dilatation in the septolateral axis, as observed with proximal circumflex coronary occlusion.

Potential mechanism of left ventricular outflow tract obstruction after mitral ring annuloplasty.

OBJECTIVES: The purpose of this study was to explore whether geometric changes that predispose to left ventricular outflow tract obstruction after mitral ring annuloplasty are coupled to subvalvular apparatus disturbances. METHODS: Radiopaque markers were implanted in sheep: 9 in the ventricle, 1 in the high interventricular septum, 1 on each papillary muscle tip, 8 around the mitral anulus, 4 on the anterior mitral leaflet, and 2 on the posterior leaflet. One group served as control (n = 5); the others were randomized to undergo annuloplasty with the Duran ring (n = 6; Medtronic, Inc, Minneapolis, Minn) or Carpentier-Edwards Physio ring (n = 6; Baxter Healthcare Corp, Irvine, Calif). After a 7- to 10-day recovery period, 3-dimensional marker coordinates were measured with biplane videofluoroscopy. RESULTS: At the beginning of ejection, (1) the anterior leaflet was displaced toward the left ventricular outflow tract; (2) the normal atrially flexed anterior anulus was flattened into the left ventricular outflow tract; (3) the posterior anulus was displaced toward the left ventricular outflow tract; (4) the anterior papillary muscle was displaced septally; and (5) the posterior papillary muscle was dislocated inwardly toward the anterior papillary muscle in the Physio ring group compared with the control group. During ejection, all these structures moved septally, encroaching further on the left ventricular outflow tract. In the Duran ring group, only the posterior anulus was displaced toward the left ventricular outflow tract; the anterior leaflet was not displaced toward the left ventricular outflow tract, and it did not move septally during ejection. CONCLUSIONS: The semirigid Physio ring was associated with perturbations in annular dynamics that caused changes in papillary muscle geometry. We propose an integrated valvular-subvalvular mechanism to explain displacement of the anterior leaflet into the left ventricular outflow tract after mitral ring annuloplasty.

Mitral annular size and shape in sheep with annuloplasty rings.

BACKGROUND: Mitral annuloplasty is an important element of most mitral repairs, yet the effects of various types of annuloplasty rings on mitral annular dynamics are still debated. Recent studies suggest that flexible rings preserve physiologic mitral annular area change during the cardiac cycle, while rigid rings do not. METHODS: To clarify the effects of mitral ring annuloplasty on mitral annular dynamic geometry, we sutured 8 radiopaque markers equidistantly around the mitral anulus in 3 groups of sheep (n = 7 each: no ring, Carpentier-Edwards semi-rigid Physio-Ring [Baxter Healthcare Corp, Edwards Division, Santa Ana, Calif], and Duran flexible ring [Medtronic, Inc, Minneapolis, Minn]). Ring sizes were selected according to anterior leaflet area and inter-trigonal distance (Physio-Ring 28 mm, n = 7; Duran ring 31 mm, n = 5, and 29 mm, n = 2). After 8 +/- 1 days of recovery, the sheep were sedated and studied by means of biplane videofluoroscopy. Mitral annular area was calculated from 3-dimensional marker coordinates without assuming circular or planar geometry. RESULTS: In the no ring group, mitral annular area varied during the cardiac cycle by 11% +/- 2% (mean +/- SEM; maximum = 7.6 +/- 0.2, minimum = 6.8 +/- 0.2 cm2; P

Semirigid or flexible mitral annuloplasty rings do not affect global or basal regional left ventricular systolic function.

BACKGROUND: Previous studies have revealed that rigid mitral annuloplasty rings may be associated with left ventricular (LV) systolic dysfunction, but whether ring type affects regional systolic function at the base of the LV, in the region near the mitral annulus, is unclear. We tested the hypothesis that rigid fixation of the mitral annulus results in significant regional systolic dysfunction at the base of the LV. METHODS AND RESULTS: Twenty-six adult male sheep underwent placement of 13 miniature tantalum markers into the LV epicardium and around the mitral annulus to allow calculation of LV volume and regional epicardial area. Group I (n = 7) sheep served as controls; animals randomized to groups II (n = 11) and III (n = 8) underwent mitral annuloplasty with either a semirigid or flexible ring, respectively. After a 7- to 10-day recovery period, animals were studied in a closed-chest, sedated, autonomically blocked state. Global LV systolic function (end-systolic elastance and preload recruitable stroke work) were not significantly different among the 3 groups (P = 1.0, ANOVA). Regional systolic function at the base of the LV (fractional area shrinkage [FAS] of 4 epicardial areas) at comparable LV preload and afterload was similar in the 4 basal areas (P = 0.223, MANOVA). With the use of load-insensitive indexes (slope and area intercept of the end-systolic pressure-regional area relationship and regional stroke work-end-diastolic area relationship), regional systolic function also was not different between groups at baseline or with inotropic stimulation in any basal region (P > 0.05, MANOVA). Furthermore, neither annuloplasty ring perturbed the regional pattern of basal LV systolic function. CONCLUSIONS: Postoperative LV systolic function, both globally and in the region of the base of the LV (near the mitral annulus), was not altered with either semirigid or flexible ring fixation of the mitral annulus.

Early systolic mitral leaflet "loitering" during acute ischemic mitral regurgitation.

BACKGROUND: The mechanism by which incomplete mitral leaflet coaptation develops during ischemic mitral regurgitation is debated, with recent studies suggesting that incomplete mitral leaflet coaptation may be due to apically displaced papillary muscle tips. Yet quantitative in vivo three-dimensional mitral leaflet motion during ischemic mitral regurgitation has never been described. METHODS: Radiopaque markers (sutured around the mitral anulus, to the central free mitral leaflet edges, and to both papillary muscle tips and bases) were imaged with the use of biplane videofluoroscopy in six closed-chest, sedated sheep before (control) and during induction of acute ischemic mitral regurgitation. Leaflet coaptation was defined as the minimum distance measured between edge markers during control conditions. RESULTS: During control, leaflet coaptation occurred 23 +/- 7 msec (mean +/- standard error of the mean) after end-diastole, when left ventricular pressure was 27 +/- 6 mm Hg. During ischemic mitral regurgitation, coaptation was delayed to 115 +/- 19 msec after end-diastole (p < or = 0.01 vs control [n = 4]) when left ventricular pressure was 88 +/- 4 mm Hg. At end-diastole during ischemic mitral regurgitation, the mitral anulus area was 14% +/- 2% larger than control (7.4 +/- 0.3 cm2 vs 6.5 +/- 0.2 cm2, p < or = 0.005) as the result of the lengthening of muscular annular regions (76.0 +/- 2.5 mm vs 70.5 +/- 1.4 mm, p < or = 0.01). Mitral anulus shape (ratio of two diameters) at end-diastole was more circular during ischemic mitral regurgitation (0.79 +/- 0.01 vs 0.71 +/- 0.02, p < 0.01). At end-diastole during ischemic mitral regurgitation, the posterior papillary muscle tip was displaced 1.5 +/- 0.5 mm laterally and 2.0 +/- 0.6 mm posteriorly (p < or = 0.02 vs control), but there was no apical displacement of either papillary muscle tip. CONCLUSIONS: Incomplete mitral leaflet coaptation during acute ischemic mitral regurgitation occurred early in systole, not at end-systole, and was due to "loitering" of the leaflets associated with posterior mitral anulus enlargement and circularization, as well as some posterolateral, but not apical, posterior papillary muscle tip displacement. These data suggest that early systolic mitral anulus dilatation and shape change and altered posterior papillary muscle motion are the primary mechanisms by which incomplete mitral leaflet coaptation occurs during acute ischemic mitral regurgitation.

Mitral valve opening in the ovine heart.

To study the three-dimensional size, shape, and motion of the mitral leaflets and annulus, we surgically attached radiopaque markers to sites on the mitral annulus and leaflets in seven sheep. After 8 days of recovery, the animals were sedated, and three-dimensional marker positions were measured by computer analysis of biplane videofluorograms (60/s). We found that the oval mitral annulus became most elliptical in middiastole. Both leaflets began to descend into the left ventricle (LV) during the rapid fall of LV pressure (LVP), before leaflet edge separation. The anterior leaflet exhibited a compound curvature in systole and maintained this shape during opening. The central cusp of the posterior leaflet was curved slightly concave to the LV during opening. Markers at the border of the "rough zone" were separated by 10 mm during systole. We conclude that coaptation occurs very near the leaflet edges, that the annulus and leaflets move toward their open positions during the rapid fall of LVP, and that leaflet edge separation, the last event in the opening sequence, occurs near the time of minimum LVP.

Most ovine mitral annular three-dimensional size reduction occurs before ventricular systole and is abolished with ventricular pacing.

BACKGROUND: Conventional surgical thinking indicates that mitral annular (MA) size reduction plays a key role in mitral valve closure, and most MA size and shape changes are thought to occur during left ventricular (LV) systole. The influences of left atrial (LA) and LV systole on MA size and shape, however, remain debated. METHODS AND RESULTS: Eight radiopaque markers were placed equidistantly around the MA and imaged using high-speed simultaneous biplane videofluoroscopy in seven closed-chest, sedated sheep before and during asynchronous LV pacing. Marker images were used to compute the three-dimensional coordinates of each marker every 16.7 ms throughout the cardiac cycle, allowing calculation of three-dimensional MA area, septal-lateral (SL) dimension, and commissure-commissure (CC) dimension under control and LV pacing conditions. Maximum MA area occurred in early diastole, and minimum MA area near end-diastole; maximum area reduction was 12+/-1% (P< or =.001). Interestingly, 89+/-3% of area reduction occurred before LV systole. During this "presystolic" period, SL decreased by 8+/-1% and CC by 2+/-1%; the SL/CC ratio fell from 0.73+/-0.02 to 0.69+/-0.01 (P< or =.005), indicating a less circular shape at end-diastole. With LV pacing, total MA area reduction was similar (13+/-2 versus 12+/-1%, P=NS versus control); however, all MA area reduction occurred during LV systole with minimum MA area occurring at end-systole. Presystolic shortening in both SL and CC dimensions was lost, and presystolic ellipticalization disappeared. CONCLUSIONS: Changes in MA size and shape coincident with LA systole included area reduction and shape change prior to the onset of LV contraction. These presystolic changes vanished when LA systole was absent (LV pacing). Thus, LA systole plays a pivotal role in MA size reduction and shape alteration. The unexpected timing of these MA dynamics should be taken into account during mitral valve reparative procedures.

Segmental wall motion abnormalities in patients undergoing total hip replacement: correlations with intraoperative events.

We examined the effect of methylmethacrylate cement on venous embolization and cardiac function in 20 patients having total hip arthroplasty under general anesthesia. Segmental wall motion abnormalities and intracardiac targets (presumably emboli) were investigated by making videotaped recordings of the transgastric short axis and longitudinal 4-chamber views of the heart with transesophageal echocardiography at different points during surgery. The incidence of segmental wall motion abnormalities was the most frequent during insertion of cemented femoral prostheses (8 of 14 patients had wall motion abnormalities). This was significantly different from baseline measurements taken at the beginning of surgery (P < 0.05). In addition, there were also significantly more segmental wall motion abnormalities in patients having a cemented femoral component compared to those having an uncemented femoral prosthesis (P < 0.05). The incidence of wall motion abnormalities during acetabular and femoral reaming and during wound closure was not significantly different from baseline. Intracardiac targets (emboli) were seen in all 20 patients during surgery. The largest number of emboli occurred during reaming of the femur and during insertion of the femoral prosthesis. Significantly more emboli were seen with cemented components (P < 0.02). Most emboli were small (< 2 mm) and appeared similar to the microbubbles produced by agitating saline with a small amount of air. Six patients also had larger (> 5 mm) emboli that appeared to be solid material. One patent foramen ovale was detected (5% incidence). There were no adverse cardiac or neurologic events, and heart rate and arterial blood pressure remained within normal limits throughout surgery.(ABSTRACT TRUNCATED AT 250 WORDS)

Regional contractile performance during acute ischemia in porcine right ventricle.

Studies of regional ischemia in the canine left ventricle (LV) have shown augmented systolic segmental shortening in nonischemic regions. To characterize these responses in the right ventricle (RV), acute occlusions of the right coronary artery (RCAO) were produced in anesthetized pigs. Orthogonal pairs of ultrasonic segment-length crystals were implanted in the mid-RV free wall (ischemic region) and in the outflow tract (nonischemic region). RV pressure-area work loops were generated during inferior vena caval occlusions, and both slope (M) of the preload recruitable stroke work relationship and elastance (Ees) of the end-systolic pressure-area relationship were determined by linear regression. Under control conditions, there was no significant difference in M or Ees of the two regions. After RCAO, fractional area shortening decreased by 84% in the ischemic region but increased by 21% in the nonischemic region. Although RV end-diastolic pressure and areas in both regions increased, M was significantly reduced by 78% only in the ischemic region. There was no change in contractile function in the nonischemic region as described by M and Ees despite the increase in systolic fractional area shortening. Therefore the nonischemic region of the porcine RV exhibits an apparent compensatory hyperkinesis in response to regional ischemia similar to that shown for the canine LV, consistent with the Frank-Starling mechanism.

Effects of acute right ventricular ischemia on ventricular interactions during prosthetic left ventricular support.

Systolic ventricular interactions may be partially responsible for right ventricular failure that sometimes occurs during clinical use of prosthetic left ventricular assist devices. In this hypothesis, it is proposed that the left ventricular assist device reduces left ventricular pressure and its contribution to right ventricular performance, thus impairing right ventricular output. On the other hand, these effects may be small compared with other causes of right ventricular failure such as ischemia. To test the systolic interaction hypothesis in the normal and ischemic right ventricle, we used a left ventricular assist device to pressure unload the left ventricle of anesthetized pigs, and we compared its effect on right heart function before and after 2 minutes of acute right coronary artery occlusion as a model of right heart failure. Pigs were instrumented for measurements of septal to left ventricular and right ventricular free wall dimensions with ultrasonic crystals, ventricular chamber pressures, and cardiac output with a pulmonary artery blood flow probe. Without right ventricular ischemia, the left ventricular assist device produced an 80% +/- 6% reduction in left ventricular pressure-time integral while maintaining aortic pressure. This resulted in a leftward septal shift with an 11.6% +/- 1.8% decrease in left ventricular septal-to-free wall dimension and a 12.5% +/- 2.4% increase in right ventricular septal-to-free wall dimension, with no changes in right ventricular cardiac output or stroke work. In contrast, right coronary artery occlusion alone produced right heart failure, with a 50% +/- 6% reduction in right ventricular global stroke work and 26% +/- 6% and 27% +/- 3% reductions in cardiac output and right ventricular peak systolic pressure, respectively. This right heart failure persisted during left ventricular unloading with the left ventricular assist device, which resulted in further leftward septal shifting and unchanged but still depressed stroke work and flow output. These findings support the hypothesis that a preexisting pathologic condition is the dominant factor in determining right ventricular function during prosthetic left ventricular support and that direct anatomic interactions play a minor role.

Complications of general anesthesia for Nd:YAG laser resection of endobronchial tumors.

We studied the incidence and mechanisms of cardiovascular complications and postoperative respiratory insufficiency associated with GA and Nd:YAG laser endobronchial tumor resection. The records of 73 patients undergoing 87 procedures were reviewed. Preoperative status, anesthetic methods and perioperative complications were analyzed by multiple regression to determine predictors of outcome. Twenty-three percent of patients had greater than 90 percent mainstem bronchus obstruction. Longer serum elimination half-life of relaxant drug was significantly correlated with longer duration of mechanical ventilation after neuromuscular blockade reversal. Cardiovascular complications were noted in 24 procedures and often required therapeutic intervention. Variables predicting cardiovascular complications included longer duration of GA and increasing age. Perioperative respiratory and cardiovascular complications are common after GA for Nd: YAG laser resection. Short-acting neuromuscular relaxants, careful assessment prior to postoperative extubation, limiting duration of GA and cardiovascular monitoring are recommended when implementing GA for Nd: YAG laser resection of endobronchial tumors.

Elasticity of small pulmonary veins in the cat.

The distensibility of pulmonary veins of cats, in the diametric range of 100-1200 micrometer, was studied as a function of the venous pressure p upsilon and pleural pressure p PL, while the alveolar air pressure was maintained at zero (atmospheric). The resulting percentage changes in diameter normalized with respect to the diameter at delta P of 10 cm H20 (D10) are expressed as function of delta = p upsilon - p PL. It was found that in most cases the diameter varies linearly with delta P: D/D10 = 1 + alpha (p upsilon - p PL) where alpha is the compliance coefficient. The results show that smaller veins of the cat are more compliant than larger veins. For example, when pleural pressure is -10 cm H2O, the values of alpha for vessels in the ranges of diameters of 100-200 micrometer, 200-400 micrometer, 400-800 micrometer and 800-1200 micrometer are, respectively, 2.05, 1.44, 1.08 and 0.71 percent per cm H2O or Pa-1. The effects of lung inflation on the distensibility of pulmonary veins are also studied. Our results show that for vessel sizes in the range of 400-800 micrometer and 800-1200 micrometer the compliance constant alpha is not affected by the inflation of the lungs (changes in pleural pressure to more negative values). For smaller veins in the size ranges 100-200 micrometer and 200-400 micrometer our results show an increase in compliance from 2.05 to 2.79 and from 1.44 to 2.01 percent per cm H20 or Pa-1, respectively, when pleural pressure is changed from -10 to -15 cm H20. When the pleural pressure is more negative than -15 cm H20, however, the compliance of the vessels in the foregoing two size ranges is observed to decrease.