Ann Thorac Surg 1997;63:1227-1234
© 1997 The Society of Thoracic Surgeons
Original Article: Cardiovascular
Impact of Size Mismatch and Left Ventricular Function on Performance of the St. Jude Disc Valve After Aortic Valve Replacement
Ole Lund, MD, PhD,
Kristian Emmertsen, MD, PhD,
Torsten T. Nielsen, MD, PhD,
Finn T. Jensen, MSc,
Christian Flø, MSc,
Hans K. Pilegaard, MD,
Bodil S. Rasmussen, MD,
Ole K. Hansen, MD,
Liselotte H. Kristensen, MD
Departments of Thoracic and Cardiovascular Surgery, Cardiology, and Clinical Physiology and Nuclear Medicine, Skejby Sygehus, Aarhus University Hospital, Aarhus, Denmark
 |
Abstract
|
|---|
Background. The hemodynamic function of the St. Jude valve may change relative to changes in left ventricular function after aortic valve replacement for aortic stenosis. From theoretical reasons one may hypothesize that prosthetic valve hemodynamic function is related to left ventricular failure and mismatch between valve size and patient/ventricular chamber size.
Methods. Forty patients aged 24 to 82 years who survived aortic valve replacement for aortic stenosis with a standard St. Jude disc valve (mean size, 23.5 mm; range, 19 to 29 mm) were followed up prospectively with Doppler echocardiography and radionuclide left ventriculography preoperatively and 9 days, 3 months, and 18 months after the operation with assessment of intravascular hemolysis at 18 months. Follow-up to a maximum of 7.4 years (mean, 6.3 years) was 100% complete.
Results. Left ventricular muscle mass index decreased from 198 ± 62 g•m-2 preoperatively to 153 ± 53 g•m-2 at 18 months (p < 0.001), paralleled by a significant increase in left ventricular ejection fraction, peak ejection rate, and peak filling rate; only 18% of the patients had normal left ventricular muscle mass index and only 32% normal ventricular function (normal left ventricular ejection fraction, peak ejection rate, peak filling rate, early filling fraction, and late filling fraction during atrial contraction) at 18 months. Prosthetic valve peak Doppler gradient dropped from 20 ± 6 mm Hg at 9 days to 17 ± 5 mm Hg at 18 months (p < 0.05). Reduction of left ventricular muscle mass index was unrelated to peak gradient and size of the valve. Peak gradient at 18 months rose with valve orifice diameter of 17 mm or less (by 6 mm Hg), orifice diameter/body surface area of 9 mm•m-2 or less (by 5 mm Hg), left ventricular end-diastolic dimension (by 23 mm Hg per 10 mm increase), and impaired ventricular function (by 3 mm Hg). All but 2 patients (5%) had intravascular hemolysis; none had anemia. Two patients with moderate paravalvular leak had the highest serum lactic dehydrogenase levels; 4 patients with trivial leak had higher serum lactic dehydrogenase levels than those without leak. Serum lactic dehydrogenase levels rose with moderate paravalvular leak, impaired ventricular function, and valve orifice diameter. Six patients with trivial or moderate paravalvular leak had a cumulative 7-year freedom from bleeding and thromboembolism of 44% ± 22% compared with 87% ± 5% for those without leak (p < 0.05).
Conclusions. The peak gradient of the St. Jude aortic valve dropped marginally over the first 18 postoperative months in association with incomplete left ventricular hypertrophy regression and marginal improvement of ventricular function. Mismatch between valve size and ventricular cavity size or patient size and impaired function of a dilated ventricle significantly compromised the performance of the St. Jude valve. Probably explained by platelet destruction or activation, paravalvular leak was related to bleeding and thromboembolic complications.
|
Introduction
|
|---|
The St. Jude standard bileaflet disc valve has been in clinical use for 18 years [1], and is established as the reference valve for newer valve designs. In their metaanalysis, Cannegieter and associates [2] recently concluded that the bileaflet design is less thrombogenic than the single-disc design, and as regards hemodynamic function newer prosthetic valves have failed to show an advantage. It is well known that the pressure gradient across a prosthetic valve increases during exercise [3] and with decreasing valve size [4]. The narrow aortic root accommodating only the smallest prosthetic valve is accordingly a serious problem. Size mismatch in a broader sense, prosthetic aortic valve size relative to patient size and left ventricular cavity size, may also have an influence. Impaired left ventricular function has, furthermore, been related to increased intravascular hemolysis in patients with a mechanical aortic valve prosthesis [5].
The aim of the present study was to examine the influence of size mismatch and left ventricular function on hemodynamic performance of aortic St. Jude valves using serial follow-up examinations after aortic valve replacement (AVR). For the sake of homogeneity as regards changes in left ventricular function and mass, we only included patients operated on for aortic stenosis (AS).
|
Material and Methods
|
|---|
The present study included 40 patients with AS who underwent AVR with a standard St. Jude disc valve during 1989 and who were alive 18 months after the operation. No one had significant aortic regurgitation. A further 9 patients operated on for AS during 1989 were excluded: 4 who received another type of prosthetic valve, 3 who died early postoperatively, 1 who died 40 days after the operation of an autopsy-verified cerebral thrombosis, and 1 who underwent repeat AVR due to endocarditis 50 days after the first operation. Patients with associated cardiac disease other than coronary artery disease and patients referred to us for operation after diagnostic investigation at other centers were primarily excluded.
|
Preoperative and Serial Postoperative Investigations
|
|---|
All patients underwent left-sided heart catheterization with aortic root and coronary angiography. Transthoracic Doppler echocardiography using standard methods and equipment [6, 7] and electrocardiogram-gated radionuclide ventriculography as previously described [8–10] were performed preoperatively, as well as 9 days (range, 6 to 15 days), 3 months (range, 2.8 to 3.5 months), and 18 months (range, 17.5 to 18.6 months) after the operation in all patients except 2 who did not want to attend the 18-month investigation (both were women, age 77 and 81 years). For the 9-day echocardiogram a poor acoustic window precluded a sufficient study in 9 patients. All echocardiographies were performed by one of two experienced observers.
|
Operation
|
|---|
Previously described standard methods [11, 12] involving complete extracorporeal circulation, general hypothermia (30°C), topical cardiac cooling, and cardioplegic cardiac arrest were used in all patients. The St. Jude valves were implanted using tightly placed simple interrupted sutures in 25 patients and Teflon-plegeted U-sutures placed supraannularly in 15. Two patients with a narrow annulus had an annuloplasty with a pericardial patch in the acoronary part of the annulus (Nick's method [13]) to accommodate a 21-mm valve. Fourteen patients had concomitant coronary artery bypass grafting performed. Thirty-nine patients were started on life-long warfarin treatment aiming at maintaining Owren's prothrombin-proconvertin index in the 0.1 to 0.3 range (normal, 0.8 to 1.2). One patient with preoperative bleeding from a cecal angioma received low-dose acetylsalicylic acid and dipyridamole treatment. One patient with bleeding from colon diverticulosis 1 year after the operation had his warfarin administration discontinued and was given low-dose acetylsalicylic acid and dipyridamole instead.
|
Definitions and Calculation of Variables
|
|---|
Coronary artery disease was defined as luminal diameter reduction of at least 50% of a major epicardial artery or a first branch. From the echocardiographic M-mode tracings the following were recorded according to internationally accepted criteria [7]: left ventricular end-diastolic diameter (EDD) and wall thickness (EDWTh; average of posterior free wall and interventricular septum). Left ventricular muscle mass was calculated as 1.04 [(EDD + 2EDWTh)3 - EDD3] - 14 (g) [14]. Left ventricular muscle mass index (g•m-2; LVMMi) is left ventricular muscle mass divided by body surface area. A value of 108 g•m-2 [7, 14] was used as upper normal level for LVMMi. The peak instantaneous gradient across the aortic valve was calculated from the peak continuous Doppler velocity in the ascending aorta using the modified Bernoulli equation [6]. The peak gradient in the left ventricular outflow tract proximal to the valve was measured in 16 patients at the 9-day investigation, in 26 at 3 months, and in 28 at 18 months using pulsed Doppler echocardiography [6]. Using a standard semiquantitative color Doppler method [6], regurgitation to the left ventricle peripherally to the prosthetic valve (paravalvular leak) was quantified as mild (trivial), moderate, or severe (no one had severe leak). The perivalvular leak, the regurgitation through the diastasis between the discs and the housing ring of the St. Jude valve, was hemodynamically insignificant and trivial in all examinations.
The left ventricular function indices of the radionuclide ventriculographies were calculated as previously described [8]: ejection fraction (LVEF; stroke volume in percent of end-diastolic volume), fast filling fraction (volume filled during first half of diastole in percent of total filling volume), PQ filling fraction (volume filled during PQ interval of electrocardiogram, atrial contraction, in percent of total filling volume), peak ejection rate and peak filling rate in first half of diastole (end-diastolic volume•s-1; normalized for end-diastolic volume), time from ejection start to peak ejection rate/duration of systole ratio (an index of prolonged contraction), and time from end-systole to peak filling rate/duration of diastole ratio (an index of reduced and delayed early fast filling). Left ventricular end-diastolic volume index (mL•m-2) was calculated using the monoplane area-length method as previously described [8]. The lower or upper 95% confidence limits of a reference population of healthy individuals [8] were used to identify patients with impaired systolic function (subnormal LVEF, subnormal peak ejection rate, or supranormal time from ejection start to peak ejection rate/duration of systole ratio), impaired diastolic function (subnormal peak filling rate, subnormal fast filling fraction, supranormal time from end-systole to peak filling rate/duration of diastole ratio, or supranormal PQ filling fraction), and impaired left ventricular function (impaired systolic or diastolic function or both). Upper 95% confidence limit for end-diastolic volume index was 89 mL•m-2 [8].
|
Investigation of Hemolysis at 18 Months
|
|---|
Venous blood samples were collected and analyzed for the following (with the normal reference range for our laboratory in parenthesis): blood hemoglobin (male, 8.0 to 11.0 mmol•L-1; female, 7.0 to 10.0 mmol•L-1), hematocrit (male, 0.40 to 0.50; female, 0.35 to 0.46), red blood cell count (male, 4.1 to 5.8 x 1012 female, 3.8 to 5.2 x 1012•L-1), red blood cell mean corpuscular volume (85 to 100 x 10-15 L) and hemoglobin (18 to 22 mmol•L-1), serum iron level (10 to 34 µmol•L-1), and serum transferrin level (45 to 70 µmol•L-1) to check the red blood cell and iron status; serum lactic dehydrogenase level (LDH; 100 to 400 U•L-1), plasma hemoglobin level (<3 µmol•L-1), blood reticulocyte fraction (<0.02), and serum haptoglobin level (0.40 to 3.50 U•L-1) to detect and quantitate intravascular hemolysis; serum creatine kinase level (male < 270 U•L-1; female < 150 U•L-1), serum creatine kinase B (myocardial) subunit level (<12 U•L-1), and serum aspartate aminotransferase level (10 to 40 U•L-1) to check for other possible organ sources of elevated serum LDH; and serum 
1-antitrypsin level (1.7 to 4.1 g•L-1), which is a sensible acute phase reactant, with the same as the preceeding aim and to check for possibly elevated serum haptoglobin level.
|
Follow-up
|
|---|
In addition to the serial follow-up investigations of the present study, the patients attended annual out-patient clinic controls at their local hospital. A complete follow-up was performed from March 25 to April 3, 1996, involving telephone contact to the patients who were still alive and to the general practitioner of the patients who had died. All hospital admissions after the 18-month investigation were identified (nationally covering computerized registration) and the relevant hospital notes checked. Prosthetic valve-related complications were recorded as previously described [12] abiding by internationally accepted guidelines [15]. Emboli were recorded as minor if symptoms subsided within 48 hours or as major if they did not. The observation time ranged from 6.4 to 7.4 years for 32 patients who were alive at the end of the study and 21 months to 6.9 years for 8 who died, with death certificates available in all and autopsy reports in 2. A total of 254 patient-years at risk had been accumulated.
|
Statistical Analysis
|
|---|
All tests were computerized using the BMDP Dynamic Release 7.0 software package [16]. Simple comparisons were performed using a standard Pearson 
2-test, a nonpaired t test, or a one-way analysis of variance as appropriate. Paired comparisons were done with a paired t test or a paired 2-test (McNemar). Linear relations were checked with a standard least squares linear regression analysis. Identification of independent determinants of Doppler aortic valve gradients and of serum LDH levels was provided by a stepwise multiple linear regression analysis employing a comprehensive formalized test sequence [11]. Cumulative event freedom curves were constructed using Kaplan and Meier's product-limit method and compared using a log-rank and a Gehan test. Cumulative event freedom and linearized incidence rates are given with ± one standard error, whereas quantitative data are given as mean ± one standard deviation (range). The level of statistical significance was chosen as 0.05.
|
Results
|
|---|
Preoperative general and invasive data are given in Table 1
. All but one 24-year-old patient had a calcified aortic valve, and judged by the peak systolic pressure all had significant pressure load of the left ventricle with a normal (
12 mm Hg) end-diastolic pressure in only 1 patient. Size characteristics of the patients in relation to prosthetic valve size are shown in Table 2. Body surface area and EDD rose with increasing valve size, whereas EDD corrected for body surface area did not differ between the valve sizes; neither did LVEF. However, valve orifice diameter-to-body surface area ratio and valve orifice diameter/EDD ratio indicated a possible size mismatch associated with the smaller valves. Seventy percent of the patients were in New York Heart Association classes III and IV (Table 3) and all had significant left ventricular hypertrophy with LVMMi greater than 108 g•m-2 preoperatively. Subnormal LVEF (<0.61) preoperatively was present in 38% (n = 15) of the patients, whereas impaired systolic function was noted in 45% and impaired diastolic function in 68% (see Table 3). Patients with impaired diastolic function (n = 27) had higher LVMMi (212 ± 60 g•m-2) than those without (166 ± 48 g•m-2; n = 13; p < 0.05) and there was a similar tendency for patients with reduced systolic function. Patients with reduced LVEF at the 18-month investigation (n = 12) had higher preoperative LVMMi (231 ± 76 g•m-2) than those without (184 ± 49 g•m-2; n = 26; p < 0.05), with a similar tendency for 18-month LVMMi.
View this table:
[in this window]
[in a new window]
|
Table 3. . Clinical, Echocardiographic, and Radionuclide Ventriculographic Data at Preoperative and Serial Postoperative Investigationsa
|
|
|
Serial Changes in Left Ventricular and Prosthetic Valve Function
|
|---|
Table 3
shows selected clinical, echocardiographic, and radionuclide ventriculography data at the preoperative and at the 9-day, 3-month, and 18-month postoperative investigations. The patients experienced a pronounced improvement in New York Heart Association classification. Cardiothoracic index and EDD increased early after the operation but decreased subsequently, having reached the preoperative level at the 18-month investigation. The LVMMi and wall thickness decreased significantly after the 9-day investigation; only 18% of patients (n = 7), however, had normal LVMMi at the 18-month investigation. The LVEF increased slightly after the operation, whereas a pronounced early increase was noted for peak ejection rate. Although the latter dropped again after 9 days, the 18-month level was still significantly higher than the preoperative value. Diastolic function deteriorated somewhat up to the 3-month investigation but had returned to preoperative level at 18-months.
The Doppler aortic valve gradient was 20 ± 6 mm Hg (range, 3 to 36 mm Hg) at the 9-day investigation but dropped significantly to 17 ± 5 mm Hg (range 8 to 30 mm Hg) at the 18-month investigation (Table 3, Fig 1). The peak gradient in the left ventricular outflow tract did not change from the 9-day examination (3.7 ± 1.9 mm Hg; range, 2 to 9 mm Hg; n = 16) to the 3-month analysis (3.5 ± 1.5 mm Hg; range, 2 to 7 mm Hg; n = 26) to the 18-month study (3.7 ± 1.6 mm Hg; range, 2 to 7 mm Hg; n = 28). The Doppler gradient corrected for the left outflow tract gradient, calculated using the Bernoulli equation [7], was 15 ± 5 mm Hg (range, 7 to 26 mm Hg; n = 28) at 18 months (compare with Table 3), with the highest values (17 ± 5 mm Hg) for nine 19-mm and 21-mm valves. One patient had moderate paravalvular leak at the 9-day investigation. A patient with trivial paravalvular leak at 9 days was regarded to have moderate leak at 3 months, whereas a patient without previous leak had developed trivial leak at 18 months (see Table 3). No leak developed in the 2 patients who had an annuloplasty, and no one with a paravalvular leak had had a febrile episode or other signs of endocarditis after discharge.
View larger version (43K):
[in this window]
[in a new window]
|
Fig 1. . Peak instantaneous Doppler gradient across the St. Jude aortic valve measured 9 days, 3 months, and 18 months after the operation in relation to size (outer diameter) of the prosthesis. The gradient dropped significantly from the 9-day to the 18-month investigation (see Table 3).
|
|
|
Intravascular Hemolysis at 18 Months
|
|---|
Levels of serum LDH, plasma hemoglobin, and serum haptoglobin varied with the absence and presence of paravalvular leak (Table 4). All patients had normal reticulocyte fraction. Only 3 patients had normal serum LDH levels, 1 of whom had an elevated plasma hemoglobin level and subnormal serum haptoglobin level. The remaining 2 patients with no hemolysis both had a 21-mm valve. Elevated serum LDH level stemming from other organs than the red blood cells were unlikely because all patients had normal serum levels of creatine kinase, creatine kinase B-subunit, and 1-antitrypsin, whereas only marginally elevated serum levels of aspartate aminotransferase (45 to 60 U/L) were observed in 4 patients. No one had anemia (subnormal values of at least two of the following: blood hemoglobin, hematocrit, and red blood cell count).
|
Determinants of Prosthetic Valve Function
|
|---|
The variables of Table 2; age; coronary artery disease; simultaneously recorded heart rate, arrhythmia, and systolic and diastolic blood pressure; the clinical variables of Table 3; and all echocardiographic and radionuclide ventriculography data including those of Table 3 were examined for their relation with paravalvular leak and Doppler aortic valve gradient at the 9-day, 3-month, and 18-month investigations, and with serum LDH level at the 18-month investigation.
No variable including suturing technique was related to development of trivial or moderate paravalvular leak. The 2 patients with moderate leak had higher end-diastolic volume index at the 18-month investigation (95 and 115 mL•m-2) than preoperatively (60 and 98 mL•m-2). Only valve orifice diameter correlated with the Doppler gradient of the 9-day investigation (r = -0.47; p < 0.01). The 3-month gradient was 19.6 ± 5.8 mm Hg for orifice diameter of 19 mm or less (n = 25) and 15.6 ± 4.8 for larger valves (n = 15; p < 0.05). At the 18-month investigation the Doppler gradient did not correlate with valve orifice diameter (r = -0.30; not significant) but did correlate with valve orifice diameter/EDD ratio (r = -0.43; p < 0.01), EDD (r = 0.32; p < 0.05), and peak ejection rate (r = -0.32; p < 0.05). The gradient was 20.6 ± 5.3 mm Hg for valve orifice diameter of 17 mm or less (n = 10) and 16.1 ± 4.4 mm Hg for larger valves (n = 28; p < 0.05); it was 22.6 ± 7.5 mm Hg for a valve orifice diameter-to-body surface area ratio of 9 mm•m-2 or less (n = 5) and 16.5 ± 4.1 mm Hg for a larger ratio (n = 33; p < 0.01). There was no significant correlation between 18-month gradient and LVMMi or reduction in LVMMi (preoperative to 18-month level). There was no relation between the latter and valve orifice diameter. The final regression model with the independent determinants of 18-month Doppler gradient is shown in Table 5.
At the 18-month investigation, the serum LDH level was significantly related to paravalvular leak as shown in Table 4. The serum LDH level correlated with valve orifice diameter (r = 0.38; p < 0.05), LVMM (r = 0.36; p < 0.05), and EDD (r = 0.32; p < 0.05). The serum LDH level was 652 ± 133 U•L-1 for patients with impaired left ventricular function (n = 25) and 504 ± 96 U•L-1 in those without (n = 11; 2 patients with moderate paravalvular leak excluded). The final regression model is shown in Table 6. A model excluding the 2 patients with moderate paravalvular leak identified impaired left ventricular function (p < 0.001) and valve orifice diameter (p < 0.05) as independent determinants of serum LDH level.
|
Embolism and Bleeding Complications
|
|---|
There were two minor and three major (one fatal) embolic episodes (1.97 ± 0.06%•patient-year-1) and 3 bleeding episodes (gastrointestinal; no sequelae; 1.18 ± 0.04%•patient-year-1) during follow-up. One of the bleeding episodes occurred in the patient whose warfarin treatment was discontinued due to bleeding from colon diverticulosis; he and the other patient on low-dose aspirin and dipyridamole treatment had no embolic complications. Patients in whom paravalvular leak did not develop (n = 34) had a 5-year and 7-year freedom from embolism and bleeding combined of 94% ± 4% and 87% ± 6%, respectively, compared with 67% ± 19% and 44% ± 22%, respectively, for the 6 who did (p < 0.05). No one had an embolic or bleeding episode before development of the leak.
|
Comment
|
|---|
Left ventricular function improved over the first 18 postoperative months in the present study, but only to a minor degree. The explanation probably is that the underlying pathologic factor, concentric hypertrophy, was still present. Although left ventricular hypertrophy regression was highly significant, it was incomplete in the majority of patients. Rediker and associates [17] observed that subnormal LVEF improved early after AVR for AS in half of the patients with no further change at restudy 41 months after the operation. In the other half of the patients the early improved LVEF had returned to the subnormal preoperative starting point at the 41-month restudy [17]. Reduction of afterload and the associated reduction of myocardial oxygen demand [18] probably explain the early improvement, with irreversible pathologic influence of nondwindling myocardium taking its toll later in some patients. Also in accordance with the present results, Diver and colleagues [19] noted that afterload reduction in AS alone did not result in improved diastolic function because of unchanged hypertrophy. The possibility that incomplete hypertrophy regression was caused by some degree of maintenance of left ventricular pressure load caused by relatively obstructive mechanical valves was partially contradicted by our study: there were no relations between peak Doppler gradient or orifice diameter of the St. Jude valves and LVMMi or reduction in LVMMi at the 18-month investigation. Others have shown that insignificant LVMMi reduction was only associated with 19-mm valves (only 1 in the present study) [20].
The peak Doppler gradient of our St. Jude valves dropped significantly but on average only 3 mm Hg from the 9-day to the 18-month study, in parallel with the minor improvement in left ventricular function. The reduction in prosthetic valve gradient was not explained by hypertrophy regression in the left ventricular outflow tract because the peak pressure drop just proximal to the valve, in the outflow tract, did not change. A similar drop in peak gradients of a stentless xenograft aortic valve from 7 days to 6 months after AVR along with regression of left ventricular hypertrophy was reported by Walther and associates [21]. Others have reported peak Doppler gradients (calculated as in the present study) of the individual valve sizes of aortic St. Jude valves comparable with or slightly higher than the present ones (see Fig 1) [3, 4].
In the early phase after AVR, when the left ventricle tended to dilate compared with the preoperative level, the peak Doppler gradients of our St. Jude valves were only related to valve size. At the 18-month investigation, on the other hand, the gradient was related to several patient-related factors as well. According to the regression model shown in Table 5, a valve orifice diameter of 17 mm or less (19- and 21-mm valve) independently indicates a gradient that is 6 mm Hg greater (1 x regression coefficient) than with larger valves. An orifice diameter-to-body surface area ratio of 9 mm•m-2 or less, being prevalent in all valve sizes except the 27- or 29-mm valves (Table 2), would similarly increase the gradient by about 5 mm Hg. The influence of left ventricular EDD was more pronounced; an increase or dilatation by 10 mm would signify a 23 mm Hg gradient increase (10 x regression coefficient). The strong inverse relation between gradient and orifice diameter/EDD ratio (r = -0.43) tells a similar tale: a patient of average size but with a dilated left ventricle may not achieve acceptable hemodynamic function with, for instance, a 23-mm valve. Together with the influence of more direct valve-patient size mismatch these findings point toward a more liberal attitude toward annular enlargement. This may be performed without related increase in mortality or complications [13]. An assisting option is to use newer valve designs with an improved ratio between valve (aortic annulus) size and valve orifice diameter [22].
The quantitative influence of size mismatch and of a large, dilated left ventricle indicated above seems to be clinically relevant. The present findings were related to parameters measured with the patients at rest, but the peak Doppler gradient increases by a factor of 2 to 3 relative to valve size during exercise [3]. The clinical significance may also be underlined by a direct correlation between valve size and exercise capacity of the patients [3]. Roithinger and colleagues [23] were able to show, furthermore, that the reduction in pulmonary pressure after AVR for AS in patients with pulmonary hypertension was correlated directly with valve size.
All but 2 (5%) of the present patients had intravascular hemolysis. As in numerous other articles, paravalvular leak was the predominant determinant of the degree of hemolysis. The clinical significance was limited, though, because none of our patients had frank anemia. Valve orifice diameter was an independent determinant of serum LDH level, with a direct correlation. The explanation most probably lies in the perivalvular diastolic leak through the circular gap between the leaflets and inner housing ring of the St. Jude valve. Diastolic shear stress sufficient to cause damage to blood corpuscles has been measured in the perivalvular regurgitant flow of disc valves [24]. Quantitatively, perivalvular regurgitation is greater in the St. Jude valve compared with most other mechanical valves although not of a magnitude to cause volume load of the left ventricule [25]. Clinically, this seems to be insignificant: according to our regression model a 1-mm increase in orifice diameter should result in a serum LDH level increase of 25 U•L-1 (see Table 6). The influence of impaired left ventricular function was more pronounced; in statistical terms it was as strong as that of moderate paravalvular leak. Increased hemolysis associated with impaired ventricular function has similarly been shown in patients with Starr-Edwards ball valves examined 12 years after AVR for AS [5]. It has, furthermore, previously been shown that variables underlying deranged heart status before AVR were predictors of increased prosthesis-related complication rates [12]. The same mechanism relating impaired left ventricular function to increased hemolysis may well cause lethal platelet damage with increased bleeding tendency or sublethal damage, platelet activation, and thrombus formation and embolism.
How impaired ventricular function and mismatch between a large ventricle and valve size cause more hemolysis and higher aortic prosthetic valve gradients is not clear. Interestingly, peak ejection rate of the left ventricle was inversely related to peak Doppler gradient of the present study. A relation between regional left ventricular dysfunction and increased serum LDH level has been observed previously [5]. A tentative explanation is that a regionally malfunctioning or universally impaired pump gives rise to a more sluggish and dyscoordinated flow profile of the prosthetic valve with more turbulent random flow where high shear stress is located.
With regard to a characteristic flow profile as a potential determinant of hemolysis and hemodynamic function, the St. Jude valve has a more favorable design than the ball valve [26]. However, Nygaard and colleagues [27] have shown that shear stresses downstream of aortic St. Jude, CarboMedics, and Starr-Edwards ball valves in humans were of the same magnitude and with exposure times sufficient to cause sublethal damage to blood corpuscles. The explanation of this apparent controversy probably is that the highest shear forces act in the immediate vicinity of the valve, where measurements so far have not been feasible [27].
Paravalvular leak of hemodynamic significance, being related to uncontrollable hemolysis and to volume load and wear of the left ventricle, may be regarded as a major complication. However, we were able to show that even hemodynamically insignificant paravalvular leak may have deleterious consequences: an increased rate of bleeding and embolism. This is not surprising, accepting that the same mechanisms that cause increased red blood cell destruction probably also result in increased tendency to platelet destruction or activation. In the absence of endocarditis, paravalvular leak usually develops early after AVR in relation to a massively calcified aortic valve and annulus [28]. No variable predicted the development of paravalvular leak in the present study, which underscores the importance of meticulous surgical technique.
We conclude that the hemodynamic function of the standard St. Jude valve improved slightly over the first 18 months after AVR for AS in association with hypertrophy regression and improvement of left ventricular function. The performance of the St. Jude valve may, however, be significantly compromised subject to two factors: (1) mismatch between valve size and patient size in general and left ventricular cavity size in particular, and (2) impairment of left ventricular function. Apart from causing a slight and clinically insignificant increase in degree of intravascular hemolysis, even hemodynamically trivial paravalvular leak may be related to increased rates of bleeding and embolism, probably due to increased platelet destruction or activation.
|
Acknowledgments
|
|---|
This study was supported by grants from the Danish Heart Foundation and St. Jude Medical Inc.
|
Footnotes
|
|---|
Presented at the Poster Session of the Thirty-third Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Feb 3–5, 1997.
Address reprint requests to Dr Lund, Department of Cardio-Thoracic Surgery, Aalborg Sygehus Syd, PO Box 365, 9100 Aalborg, Denmark.
|
References
|
|---|
- Lillehei CW, Nakib A, Kaster RL, Kalke BR, Rees JR. The origin and development of three new mechanical valve designs: toroidal disc, pivoting disc, and rigid bileaflet cardiac prostheses. Ann Thorac Surg 1989;48:S35–7.
- Cannegieter SC, Rosendaal FR, Briet E. Thromboembolic and bleeding complications in patients with mechanical heart valve prostheses. Circulation 1994;89:635–41.[Abstract/Free Full Text]
- Tatineni S, Barner HB, Pearson AC, Halbe D, Woodruff R, Labovitz AJ. Rest and exercise evaluation of St. Jude Medical and Medtronic Hall prostheses. Influence of primary lesion, valvular type, valvular size and left ventricular function. Circulation 1989;80 (Suppl 1):16–23.
- Panidis IP, Ross J Mintz GS. Normal and abnormal prosthetic valve function as assessed by Doppler echocardiography. J Am Coll Cardiol 1986;8:317–26.[Abstract]
- Lund O. Late chronic hemolysis after valve replacement for aortic stenosis. Relation to residual hypertrophy and impaired left ventricular function. Angiology 1990;41:836–47.
- Hatle L, Angelsen BA, Tromsdal A. Non-invasive assessment of aortic stenosis by Doppler ultrasound. Br Heart J 1980;43:284–92.[Abstract/Free Full Text]
- Sahn DJ, DeMaia A, Kisslo J, Weyman A. Recommendations regarding quantitation in M-mode echocardiography: results of a survey of echocardiographic measurements. Circulation 1978;58:1072–83.[Abstract/Free Full Text]
- Lund O, Flø C, Rasmussen BS, Jensen FT. Factors with modifying influence on normal right and left ventricular function at rest and during submaximal exercise in healthy volunteers. Int J Angiol 1997;6:80–8.
- Lund O, Rasmussen K, Jensen FT, Hansen HH. Left ventricular ejection fraction determined with the nuclear stethoscope, gamma camera and contrast ventriculography. Nucl Med Commun 1986;7:337–48.[Medline]
- Jensen FT, Lund O, Erlandsen M. Reliability of three computer methods in the analysis of ECG-gated radionuclide left ventriculography: interrecording, interobserver and intraobserver variability. Angiology 1991;42:866–77.
- Lund O. Valve replacement for aortic stenosis: the curative potential of early operation. Scand J Thorac Cardiovasc Surg 1993;27 (Suppl 40):1–137.
- Lund O, Pilegaard HK, Magnussen K, Knudsen MA, Nielsen TT, Albrechtsen OK. Long-term prosthesis-related and sudden cardiac-related complications after valve replacement for aortic stenosis. Ann Thorac Surg 1990;50:396–406.[Abstract]
- Nakano S, Matsuda H, Shimazaki Y, et al. An appraisal of patch enlargement of the small aortic annulus in 33 patients undergoing aortic valve replacement. Eur J Cardiothorac Surg 1992;6:347–9.[Abstract]
- Devereux RD, Reichek N. Echocardiographic determination of left ventriuclar mass in man: anatomic validation of the method. Circulation 1977;55:613–8.[Abstract/Free Full Text]
- Edmunds LH Jr, Clark RE, Cohn LH, Grunkemeier GL, Miller DC, Weisel RD. Guidelines for reporting morbidity and mortality after cardiac valvular operations. Ann Thorac Surg 1996;62:932–5.[Abstract/Free Full Text]
- Dixon NJ, ed. BMDP statistical software manual. Version 7.0. Vols 1 and 2. Berkeley: University of California Press, 1992:1–1500.
- Rediker DE, Boucher CA, Block PC, Akins CW, Buckley MJ, Fifer MA. Degree of reversibility of left ventricular systolic dysfunction after aortic valve replacement for isolated aortic valve stenosis. Am J Cardiol 1987;60:112–8.[Medline]
- Strauer B-E. Myocardial oxygen consumption in chronic heart disease: role of wall stress, hypertrophy and coronary reserve. Am J Cardiol 1979;44:73–4.
- Diver DJ, Royal HD, Aroesty JM, et al. Diastolic function in patients with aortic stenosis: influence of left ventricular load reduction. J Am Coll Cardiol 1988;12:642–8.[Abstract]
- Gonzales-Juanatey JR, Garcia-Acuna JM, Fernandez MV, et al. Influence of the size of aortic valve prostheses on hemodynamics and change in left ventricular mass: implications for the surgical management of aortic stenosis. J Thorac Cardiovasc Surg 1996;112:273–80.[Abstract/Free Full Text]
- Walther T, Falk V, Autschbach R, et al. Hemodynamic assessment of the stentless Toronto SPV bioprosthesis by echocardiography. J Heart Valve Dis 1994;3:657–65.[Medline]
- Fisher J. Comparative study of the hydrodynamic function of six size 19 mm bileaflet heart valves. Eur J Cardiothorac Surg 1995;9:692–6.[Abstract]
- Roithinger FX, Krennmaier G, Deutsch M, Pachinger O. The influence of aortic prosthesis diameter on the reversibility of pulmonary hypertension in isolated aortic stenosis. J Heart Valve Dis 1994;3:185–9.[Medline]
- Tillmann W, Reul H, Herold M, Bruss K-H, van Gilse J. In-vitro wall shear measurements at aortic valve prostheses. J Biomech 1984;17:263–79.[Medline]
- Knott E, Reul H, Knoch M, Steinseifer U, Rau G. In vitro comparison of aortic heart valve prostheses. Part I: Mechanical valves. J Thorac Cardiovasc Surg 1988;96:952–61.[Abstract]
- Paulsen PK, Hasenkam JM, Stødkilde-Jørgensen H, Albrechtsen O. Three-dimensional visualization of velocity profiles in the ascending aorta in humans. A comparative study among normal aortic valves, St. Jude Medical and Starr-Edwards Silastic ball valves. Int J Artif Organs 1988;11:277–92.[Medline]
- Nygaard H, Paulsen PK, Hasenkam JM, Pedersen EM, Rovsing PE. Turbulent stresses downstream of three mechanical aortic valve prostheses in human beings. J Thorac Cardiovasc Surg 1994;107:438–46.[Abstract/Free Full Text]
- Rizzoli G, Russo R, Valente S, et al. Dehiscence of aortic valve prostheses: analysis of a ten-year experience. Int J Cardiol 1984;6:207–18.[Medline]
Top
Abstract
Introduction
