Author + information
- Received July 8, 1996
- Revision received November 4, 1996
- Accepted December 19, 1996
- Published online April 1, 1997.
- Ian Adatia, FRCP(C)A,* (, )
- Phillip M Moore, MDA,
- Richard A Jonas, MDA,
- Steven D Colan, MDA,
- James E Lock, MD, FACCA and
- John F Keane, MDA
- ↵*Dr. Ian Adatia, Cardiology and Critical Care Medicine, Hospital for Sick Children, 555 University Avenue, Toronto, Ontario, Canada M5G 1X8.
Objectives. We report the clinical course and unique hemodynamic findings after placement of a supraannular mitral valve prosthesis.
Background. Children with symptomatic mitral valve disease whose annulus is too small for the smallest prosthesis are difficult to manage. One option is valve replacement with a prosthesis positioned entirely within the left atrium (LA).
Methods. We reviewed 17 patients (median age 10 months) with symptomatic mitral valve disease who underwent placement of a supraannular valve prosthesis between 1980 and 1994.
Results. The actuarial survival rates were 88% at 1 month and 71%, 62% and 53% at 1, 2 and 10 years, respectively. Preoperative hemodynamic data (mean ± SD) compared with those after placement of the supraannular mitral prosthesis were as follows: “a” wave to left ventricular end-diastolic pressure gradient 17 ± 5 versus 4 ± 4 mm Hg (p = 0.003), mean LA pressure 25 ± 6 versus 20 ± 6 mm Hg (p = 0.07), “a” wave 30 ± 6 versus 19 ± 5 mm Hg (p = 0.006), “v” wave 28 ± 5 versus 30 ± 9 mm Hg (p = 0.31), mean pulmonary artery pressure 54 ± 19 versus 42 ± 15 mm Hg (p = 0.07) and left ventricular end-diastolic pressure 14 ± 5 versus 16 ± 4 mm Hg (p = 0.12).
Conclusions. Supraannular mitral valve replacement provides relief of mitral stenosis or mitral regurgitation. However, LA to left ventricular early diastolic gradients with large atrial “v” waves contribute to elevated mean LA pressures in the absence of prosthetic valve obstruction or regurgitation. As a result of this unexpected finding, associated left heart obstructive lesions and pulmonary and left ventricular end-diastolic hypertension, the outlook remains poor.
(J Am Coll Cardiol 1997;29:1089–94)
© 1997 by the American College of Cardiology
Congenital mitral valve disease is uncommon and severe congenital mitral stenosis (MS) requiring valve replacement in infancy is even rarer (). Nevertheless, valve replacement may be the only option for small children with congenital MS, particularly if balloon dilation has failed or resulted in intolerable regurgitation (). Therefore, children requiring mitral valve replacement whose native annulus is too small to accept the smallest prosthesis available present a difficult management problem. One approach, which is feasible in only a subset of patients, is surgical repair of the mitral valve (). An alternative option is replacement of the mitral valve with a prosthesis positioned above the native annulus entirely within the left atrium (LA). As the operative mortality after supra-annular mitral valve replacement (SMVR) has improved ([3, 4]), these children have required late hemodynamic prosthesis assessment to aid future management. The purposes of this study are to review our experience with SMVR with the emphasis on subsequent hemodynamic observations and to attempt an explanation of these findings and their use as indicators for further intervention. Furthermore, this analysis may be helpful when deciding between the treatment options of a Norwood-type single-ventricle palliation or heart transplantation versus a two-ventricle repair for neonates who have multiple left heart obstructive lesions.
Between 1980 and 1994, 17 patients (10 females) underwent mitral valve replacement with a prosthesis in the supra-annular position. They ranged in age from 3 months to 3.5 years (median 10 months) at the time of this procedure. All available data, including echocardiographic, cardiac catheterization and surgical information, were reviewed in detail and form the basis of this report. The data were analyzed and presented as follows: 1) preoperative evaluation (i.e., before SMVR) of associated lesions, procedures and mitral valve echocardiographic features (Table 1); 2) indications for and observations at the time of SMVR (Table 2); 3) comparison of cardiac catheterization data before and after SMVR; and 4) clinical course after SMVR (Table 2).
1.2 Statistical analysis.
Preoperative hemodynamic data were compared with postoperative data using the paired Student ttest. A p value <0.05 was considered significant. Actuarial survival was calculated using the Kaplan-Meier method.
2.1 Associated lesions and previous procedures (Table 1).
In addition to the mitral valve lesion, all but one patient (no. 7) had associated congenital heart disease. Previous surgical repair of aortic coarctations had been performed in eight patients, resection of supramitral stenosing rings in three and repair of ventricular septal defects in four (two of these had atrioventricular canal defects). Stenotic mitral valves were balloon dilated in nine patients, with resultant regurgitation, and a patent ductus arteriosus was closed with a coil in one patient. Thus, 10 of 17 patients had had a previous unsuccessful attempt to relieve MS.
2.2 Supraannular mitral valve replacement (Table 2).
The indications for SMVR were MS in nine patients (due to prior prosthesis stenosis in one of these, Patient 7), both MS and mitral regurgitation (MR) in six and pure MR in two. Mitral valve apparatus abnormalities were evident in all 17 with available echocardiographic data. The mitral annulus mean (±SD) diameter was 12 ± 2.5 mm (with a median Z score −0.8; range −3.8 to 1.9), requiring placement of the prosthesis in the supra-annular position.
2.3 Preoperative hemodynamic data.
At cardiac catheterization the mean (±SD) LA “a” wave was 30 ± 6 mm Hg, LA “v” wave 28 ± 5 mm Hg, LA mean pressure 25 ± 6 mm Hg, left ventricular end-diastolic (LVED) pressure 14 ± 5 mm Hg and LA “a” wave to LVED pressure gradient 17 ± 5 mm Hg. All patients had pulmonary hypertension with a mean pulmonary artery (PA) pressure 54 ± 19 mm Hg. In three patients, severe MR was the primary lesion and indication for mitral valve replacement. Mitral regurgitation followed repair of a complete atrioventricular canal in two patients (Patients 2 and 16) (complicated by hypertrophic cardiomyopathy and left ventricular outflow tract obstruction in one patient [Patient 2]). One child (Patient 7) without congenital heart disease developed severe MR after acute bacterial endocarditis. This child initially underwent mitral valve replacement with a no. 17 Björk-Shiley prosthesis placed in the annular position, but 1 year later, because of prosthetic valve stenosis, underwent SMVR with a no. 21 St. Jude prosthesis.
Mitral valve replacement was performed using the largest prosthesis available that could be placed in the supra-annular position. The following prostheses were used: no. 17 Björk-Shiley (n = 6), no. 19 St. Jude (n = 4), no. 21 St. Jude (n = 1) and no. 16 Carbomedics (n = 6). All patients received long-term anticoagulation postoperatively.
2.4 Hemodynamic data after SMVR.
Of the 17 patients, postoperative hemodynamic data were available in 12. Nine patients have undergone postoperative cardiac catheterization, three patients have been catheterized on three occasions and two twice. In seven patients, direct LA with simultaneous LVED pressure was used and in two PA wedge and simultaneous LVED pressure was used. In three patients, not yet catheterized after SMVR, postoperative traces of LA, PA and systemic arterial pressures, obtained on the intensive care unit, were available for analysis.
Postoperatively there was a reduction of the mean (±SD) LA “a” wave to LVED pressure gradient from 17 ± 5 to 4.5 ± 4 mm Hg (p = 0.006), with a significant fall in the LA “a” wave from 30 ± 6 to 19 ± 5 (p = 0.003). However, there was no significant change in LA “v” wave (30 ± 9 mm Hg, p = 0.31), mean LA pressure (20 ± 6 mm Hg, p = 0.07), LVED pressure (16 ± 4 mm Hg, p = 0.12) or mean PA pressure (42 ± 15 mm Hg, p = 0.07).
Characteristically all patients had very tall “v” waves, which occurred in the absence of LA “a” wave to LVED pressure gradients or MR (neither angiographic nor echocardiographic) (Fig. 1). The peak of the “v” wave occurred close to the downstroke of ventricular diastole and the rate of fall or Ry/v of 1.12 ± 0.15 (±SD) was not typical of either the Ry/v ratio of MS (range 0.6 to 1.0) or MR (range 2.0 to 6.0) (). There were two exceptions to the finding of large “v” waves (the only two patients without direct LA pressure measurements), and in these two patients the “v” waves were only between 2 and 4 mm Hg greater than the mean PA wedge pressure. Both of these patients died after SMVR. In one, pulmonary vein obstruction by the prosthesis was evident on autopsy. In all but one patient, pulmonary hypertension persisted, again in the absence of MR or MS. The exception (Patient 7) underwent valve replacement for isolated MR due to bacterial endocarditis in an otherwise normal heart.
2.5 Postoperative angiography.
Angiography of the left ventricle produced a characteristic picture (Fig. 2). The prosthesis was very mobile. During diastole the valve moved toward the LA. At the beginning of ventricular systole, just before the opening of the aortic valve, the “ventricularized” portion of the atrium beneath the valve expanded, and the prosthesis moved transiently and abruptly further backward to compress the remnant of the LA. With continued ventricular contraction the prosthesis moved farther away from the apex.
No child had significant MR by either angiographic assessment or echocardiography. All valve leaflets appeared to be functioning well by fluoroscopy in those who underwent cardiac catheterization.
2.6 Echocardiography after SMVR.
The echocardiographic features mirrored in many respects the angiographic findings. The prosthesis appeared mobile and moved paradoxically away from the ventricular apex during systole.
Analysis of the diastolic Doppler wave form in 13 patients who did not have regurgitation or stenosis of the prosthesis yielded a peak early diastolic gradient of 14.0 ± 6.1 mm Hg, a mean gradient of 5.2 ± 2.4 mm Hg and a pressure half-time of 61 ± 34.0 ms, all values are mean ± SD. The prosthetic valve flow typically demonstrated an early diastolic peak gradient that decreased rapidly and was absent at end-diastole. A Doppler “a” wave was usually not discernible (Fig. 3). In one patient who was not catheterized malfunction of the anterior prosthetic leaflet, which failed to close in ventricular systole, was seen. In another patient, a vegetation on the prosthesis was noted and confirmed during surgical replacement.
2.7 Clinical course.
Actuarial survival rates were 88% at 30 days, 71% at 1 year, 62% at 2 years and 53% at 5 and 10 years. There have been seven deaths, and five patients have required reoperation. Two children died in the immediate postoperative period with pulmonary hypertension and low cardiac output. Autopsy in Patient 13 demonstrated obstruction of the left pulmonary veins by the prosthesis and a small hypertrophied LA proximal to the prosthesis with endocardial fibroelastosis, also present in the left ventricle.
Five children died late at 2 months, 4 months, 6 months, 2 years and 2.5 years after the operation. Postmortem examinations were carried out in three children. In Patient 4 the mitral prosthesis was judged to be obstructing the lower pulmonary veins, and in both Patients 4 and 5 the supraannular prosthesis was clean and functioning well. Autopsy findings common to both patients included a small hypertrophied LA proximal to the prosthesis with dilation of the appendage and marked endocardial fibrosis diffusely involving the left ventricle and atrium. The pulmonary vasculature demonstrated histologic changes of severe pulmonary vascular disease. Patients 3 (confirmed at autopsy) and 16 died after acute obstruction of the mitral prosthesis.
Four children were reoperated on because of concern about the prosthesis. In two patients the prosthesis was functioning normally despite high LA pressures and tall “v” waves. However, in Patients 8 and 2, who had an “a” wave to LVED gradient, there was pannus ingrowth obstructing the prosthesis. Both patients underwent successful prosthesis replacement, one in the annular position.
3.1 General findings.
The management of children with variable degrees of symptomatic left heart inflow and outflow obstruction who do not fit the criteria for single-ventricle palliation in the neonatal period remains extremely challenging. We report replacement of the mitral valve with a prosthesis entirely within the LA in infants as young as 3 months and weighing as little as 3.8 kg. The early surgical mortality rate was 12%. However, alleviation of the mitral valve gradient alone was not sufficient to return LA pressure to normal, probably because of a persistent elevation in LVED pressure and the hemodynamic findings suggestive of a noncompliant LA.
Follow-up cardiac catheterization of these complex patients documented an interesting hemodynamic profile. Despite a well-functioning prosthesis, we observed the triad of elevated LA “v” waves and left ventricular end-diastolic and pulmonary artery hypertension. The outcome of the children who underwent reoperation suggests that an early diastolic gradient and pulmonary hypertension do not imply prosthesis stenosis in the absence of abnormal valve movement. Failure to appreciate the significance of the elevated LA “v” wave and thus mean LA pressure resulted in an unnecessary surgical procedure in two of our patients.
3.2 Left atrial hemodynamic data.
The hemodynamic profile in this cohort of children might be related to a combination of factors. The presence of a rigid prosthetic ring in a normally compliant part of an already small and hypertrophied LA whose volume has been further reduced drastically alters the volume–pressure relation of the LA. The “ventricularized” LA contracts before the left ventricle, and with ventricular systole there is regurgitation through the native annulus with posterior motion of the prosthesis. The normal LA “v” wave occurs after complete atrial relaxation as pulmonary venous blood fills the LA against a closed mitral valve and depends on five factors: 1) volume of blood entering the atrium (either from the pulmonary veins or the left ventricle if there is MR) during ventricular systole; 2) rate of forward flow into the atrium; 3) compliance of the LA/pulmonary vein chamber; 4) systemic afterload; and 5) left ventricular contractile force ().
Most studies investigating LA hemodynamic data associated with abnormal mitral valves have been in adults with acquired disease. In contrast, our study involves infants with congenital MS. Tall LA “v” waves may occur in MR, but with poor sensitivity and specificity ([7–13]). The highest “v” waves are seen in acute MR when the LA is small and less distensible, whereas in chronic MR with slow dilation over time, the systolic atrial filling is better absorbed by a more compliant LA with concomitantly lower “v” waves ([8, 14]). Indeed, very prominent “v” waves have been reported after mitral valve replacement when there is extensive fibrosis and scarring of the LA (“stiff left atrial syndrome”) (). Thus we suggest that after SMVR the “v” wave is contributed to by a small, stiff LA, which quickly fills and is unable to fulfill its requirement as an easily distensible reservoir for pulmonary venous blood during ventricular systole () by upward movement of the prosthesis in systole () and elevated LVED pressure (), these last two mechanisms being evident in almost all of our patients.
Most investigators suggest that extraparenchymal pulmonary venous flow is an inverted record of LA pressure changes and minimal pulmonary vein flow coincides with the peak of the LA “v” wave ([17–22]). However, in patients with SMVR with the onset of diastole and high LA pressures, there will be rapid pulmonary vein and LA emptying accounting for a rapid decline in the early diastolic gradient.
3.3 Pulmonary hypertension.
It is a source of concern that there has been little reduction in the PA pressure. Indeed, the only patient in whom this decreased to normal had acquired rather than congenital left heart disease. This is in contrast to improvement in pulmonary vascular pressures in adults with acquired mitral valve disease ([23–26]). Children with obstructed total anomalous pulmonary venous connection and hypoplastic left heart syndrome have increased muscularity and intimal lesions of the pulmonary veins ([27, 28]). Thus there are some differences between pulmonary vascular changes due to acquired and congenital left ventricular inflow obstruction. The extraparenchymal pulmonary veins are highly compliant structures that isolate the lung capillaries from LA pressure events ([20, 29]). Presumably the pulmonary veins in children with congenital MS lose this capacity and this, coupled with the continued elevation in LVED pressure, accounts for the persistence of pulmonary hypertension. Interestingly, in two of our patients the pulmonary vascular bed remains reactive to vasodilator drugs ().
Supraannular mitral valve replacement provides effective relief of MS or MR in children whose native annulus is too small to accommodate a prosthesis. However, their outlook remains poor because of associated left heart obstruction and persistent LVED and PA hypertension. Left atrial “v” wave to left ventricular diastolic gradients do not indicate prosthetic valve obstruction, as “a” wave to left ventricular end-diastolic gradients are uncommon. Similarly, postoperative high LA “v” waves reflect the underlying LA pathology rather than prosthetic valve incompetence.
We gratefully acknowledge Kimberlee Gauvreau, ScD, Harvard School of Public Health, for her help and advice with the statistical analysis.
- left atrium, left atrial
- left ventricular end-diastolic
- mitral regurgitation
- mitral stenosis
- pulmonary artery
- supraannular mitral valve replacement
- Received July 8, 1996.
- Revision received November 4, 1996.
- Accepted December 19, 1996.
- The American College of Cardiology
- Moore P,
- Adatia I,
- Spevak PJ,
- et al.
- Schaffer MS,
- Clarke DR,
- Campbell DN,
- Madigan CK,
- Wiggins JW,
- Wolfe RR
- Millar GAH
- Braunwald E,
- Brockenbrough EC,
- Frahm CJ,
- Ross J
- Grossman W
- Braunwald E,
- Awe WC
- Wood P
- Böök K,
- Holmgren A,
- Szamosi A
- Morkin E,
- Collins JA,
- Goldman HS,
- Fishman AP
- Dixon SH,
- Nolan SP,
- Morrow AG
- Rajagopalan B,
- Friend JA,
- Stallard T,
- Lee G de J
- Rajagopalan B,
- Friend JA,
- Stallard T,
- Lee G de J
- Keren G,
- Sherez J,
- Megidish R,
- Levitt B,
- Laniado S
- Smallhorn JF,
- Freedom RM,
- Olley PM
- Braunwald E,
- Braunwald N,
- Ross J,
- Morrow AG
- Dalen JE,
- Matloff JM,
- Evans GL,
- et al.
- Kaul TK,
- Bain WH,
- Jones JV,
- et al.
- Haworth SG,
- Reid L
- Rajagopalan B,
- Bertram CD,
- Stallard T,
- Lee GdeJ