Author + information
- Received January 31, 1996
- Revision received June 27, 1996
- Accepted September 17, 1996
- Published online January 1, 1997.
- Charles R Cannan, MB, ChBA,
- Rick A Nishimura, MD, FACCA,*,
- Guy S Reeder, MD, FACCA,
- Duane R Ilstrup, MAA,
- Dirk R Larson, MSA,
- David R Holmes, MD, FACCA and
- A.Jamil Tajik, MD, FACCA
- ↵*Dr. Rick A. Nishimura, Division of Cardiovascular Diseases and Internal Medicine, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905.
Objectives. This study was undertaken to determine whether the presence of calcium in the mitral valve commissures, as demonstrated echocardiographically, could predict outcome and to compare this with an established echocardiographic scoring system.
Background. Percutaneous mitral balloon valvotomy is an effective form of treatment for mitral valve stenosis. It is important to identify patients who would benefit from this procedure. Commissural splitting is the dominant mechanism by which mitral valve stenosis is relieved by this technique, and thus commissural morphology may predict outcome.
Methods. One hundred forty-nine consecutive patients who underwent percutaneous mitral balloon valvotomy at the Mayo Clinic were evaluated retrospectively. The morphology of the mitral valve apparatus on the baseline echocardiograms was scored in blinded manner using a semiquantitative grading system of leaflet thickening, mobility, calcification and subvalvular thickening (Abascal score). Additionally, each of the medial and lateral commissures was graded for the presence or absence of calcification. End points were death, New York Heart Association functional class, repeat percutaneous mitral balloon valvotomy and mitral valve replacement at follow-up.
Results. The mean follow-up period was 1.8 years (maximum 7.9 years). Univariate predictors of death and all events combined included age, the use of a double-balloon technique, the presence of calcium in a commissure and the Abascal score, as continuous variables. Patients with an Abascal score ≤8 showed a trend toward improved survival at 36 months free of death, repeat percutaneous mitral balloon valvotomy or mitral valve replacement (78 ± 6% vs. 67 ± 8%, p = 0.07) and free of all events combined (75 ± 6% vs. 64 ± 8%, p = 0.07) versus those patients with a score >8. However, survival at 36 months free of death, repeat percutaneous mitral balloon valvotomy or mitral valve replacement (86 ± 4% vs. 40 ± 4%) and free of all events combined (82 ± 5% vs. 38 ± 10%) at follow-up was significantly different between patients without commissural calcium and those with commissural calcium (p < 0.001). In a Cox regression model with Abascal score and commissural calcium and their interaction, calcification emerged as the only significant variable (p < 0.01).
Conclusions. The presence of commissural calcium is a strong predictor of outcome after percutaneous mitral balloon valvotomy. Patients with evidence of calcium in a commissure have a lower survival rate and a higher incidence of mitral valve replacement and all end points combined. Thus, the simple presence or absence of commissural calcification assessed by two-dimensional echocardiography can be used to predict outcome.
(J Am Coll Cardiol 1997;29:175–80)>
Percutaneous mitral balloon valvotomy has become an effective and established form of treatment for selected patients with mitral valve stenosis ([1–3]). Randomized trials have shown percutaneous mitral balloon valvotomy to be as effective as closed or open commissurotomy in short-term follow-up for patients with symptomatic mitral valve stenosis ([4, 5]). The beneficial differences in cost, length of hospital stay and return to work from percutaneous mitral balloon valvotomy as compared with surgical intervention make this catheter-based therapy an attractive first-line procedure for patients with severely symptomatic mitral stenosis. However, the immediate success rate and long-term outcome of this procedure are dependent on the underlying mitral valve morphology. Patients with pliable, noncalcified mitral leaflets and no subvalvular fusion will have a higher success rate and better long-term results than those with thickened, immobile, calcified leaflets with concomitant subvalvular fusion ([1, 3, 6, 7]).
The mitral valve morphology in patients with mitral stenosis is currently assessed in many institutions by the appearance of the mitral valve apparatus on two-dimensional echocardiography and is based on a mitral valve “score” in which leaflet thickening, calcification and mobility as well as the degree of subvalvular fusion are determined ([1, 3, 6, 7]). A low score (≤8) is indicative of a pliable, noncalcified valve with little subvalvular fusion and has been shown to be predictive of immediate success as well as a low rate of restenosis. Higher scores, which are indicative of more calcified, immobile and thickened valve leaflets with subvalvular fusion, have been shown to be associated with a higher complication rate, a lower immediate success rate and a higher rate of restenosis.
Despite the overall value of the mitral valve score, there remain a number of patients with a relatively high mitral valve score who will do well after mitral valvotomy (). In addition, there are patients with low mitral valve scores who may not do well after mitral valvotomy. As commissural splitting is the dominant mechanism by which mitral valve stenosis is relieved by the balloon technique ([9, 10]), assessment of the commissure morphology should logically be an important part of the evaluation of patients with mitral stenosis considered for percutaneous mitral balloon valvotomy, and the commissural appearance is a factor that is not directly assessed by the Abascal score. Direct assessment of the mitral commissural morphology by two-dimensional echocardiography was shown in 30 patients to be a better predictor of immediate outcome than the mitral echocardiographic score (). However, there have not been studies examining the effect of commissural morphology on intermediate-term outcome in a larger group of patients.
The purpose of this study was to determine the usefulness of examining the morphology of the mitral valve commissures for predicting outcome after percutaneous mitral balloon valvotomy specifically assessing the presence or absence of calcium in the commissures by two-dimensional echocardiography. This new approach is compared to the currently established echocardiographic scoring system ([6, 7]).
1.1 Patient group.
The study consisted of the first 149 consecutive patients who underwent percutaneous mitral balloon valvotomy at the Mayo Clinic between September 1987 and June 1995. All patients had symptomatic mitral stenosis and underwent a comprehensive two-dimensional and Doppler echocardiogram before the percutaneous mitral balloon valvotomy. The decision to proceed with the valvotomy procedure was made by one of three experienced operators based on the clinical situation and the morphology of the mitral valve on two-dimensional echocardiography. It has been previously reported from our institution that all patients with severe mitral stenosis were initially considered candidates for valvotomy irrespective of the mitral valve morphology ([1, 12]). However, with experience, the procedure has been limited to patients with pliable, minimally calcified valves or to those not thought to be suitable candidates for mitral valve surgery. The procedure of percutaneous mitral balloon valvotomy was approved by the Institutional Review Board of the Mayo Clinic.
1.2 Echocardiographic scoring.
For this study, the morphology of the mitral valve apparatus visualized on preprocedure two-dimensional echocardiograms was retrospectively reviewed by two observers who had no knowledge of the outcome of the procedure. A mitral valve score, as described by Wilkins et al. () and Abascal et al. (), was determined. This score involves a semiquantitative grading of mitral valve leaflet thickening, mobility, calcification and subvalvular thickening, each on a scale of 1 to 4 with 1 being the least involvement and 4 being the most severe involvement. The total mitral valve score is the sum of these four individual numbers. In addition, echocardiographic assessment of the presence or absence of commissural calcification was performed by directly examining the appearance of the medial and lateral commissures on the two-dimensional echocardiographic parasternal short-axis view, as has been previously described (). Echocardiographic calcium was defined as a bright echocardiographic density () with acoustic shadowing. The aortic root was used as a point of reference, and calcification was said to be present if there was a brighter echocardiographic density in the commissures than was seen in the adjacent aortic root. Specifically, the presence or absence of calcification in each commissure was evaluated (Fig. 1).
Patients were classified into the following groups for the purpose of outcome analysis: they were dichotomized into two groups based on the Abascal score (score ≤8 vs. >8) and were also dichotomized based on the presence or absence of commissural calcium.
The double-balloon technique and the Inoue single-balloon technique were both used as previously described ([1, 14]). The procedure was performed through an anterograde approach across the mitral valve by performing a transseptal puncture. For the double-balloon technique, the balloon size was chosen using a nomogram, which incorporated the patient’s body surface area. The Inoue single-balloon technique () included sequential increases in balloon size, assessing the valve gradient and the degree of mitral regurgitation after each inflation. Echocardiographic guidance was employed in the majority of cases after December 1991 to determine balloon position before and during inflation. Mitral regurgitation was assessed by either a left ventriculogram or color Doppler echocardiography after the procedure and graded semiquantitatively on a scale from 1 to 4. Major periprocedural complications were defined as in-hospital death, cardiac tamponade, systemic embolism and an increase in mitral regurgitation by two grades.
1.4 Follow-up analysis.
Patients were contacted by telephone every 6 months, and the majority were seen in clinical follow-up at yearly intervals. End points at follow-up were death, New York Heart Association functional class, repeat percutaneous mitral balloon valvotomy and mitral valve replacement. Freedom from the combined end points of 1) death, mitral valve replacement or repeat percutaneous mitral balloon valvotomy, or 2) death, mitral valve replacement or repeat percutaneous mitral balloon valvotomy and functional class III or IV were also evaluated.
Several preoperative variables were compared with short-term results to assess the statistical significance of differences and associations. For continuous variables, comparisons were made using two-tailed ttests if the data were approximately gaussian. If the data were determined to be non-gaussian, rank-sum tests were used. For categorical variables, tests for association were performed using the chi-square test or the Fisher exact test, and the exact Wilcoxon rank-sum test was used to analyze ordered categorical data. Kaplan-Meier survival curves and Cox regression models were computed for combined end points to identify predictors of survival.
For the first analysis the end point was taken to be mitral valve replacement, repeat percutaneous mitral balloon valvotomy or death; otherwise the patients were censored at the time of last known follow-up. For the second analysis the end point was taken to be mitral valve replacement, repeat percutaneous mitral balloon valvotomy, death or functional class III or IV; again, the patients not experiencing any of these end points were censored at the time of last known follow-up.
Univariate analyses using the Cox model with Abascal score (coded as 0 = ≤8, and 1 = >8) as the single covariate, and another Cox model with calcification (coded as 0 = no and 1 = yes) as the single covariate, were performed initially.
Subsequently, two multivariate Cox models were evaluated for each end point. First a main effects’ Cox model was run with Abascal score and calcification as covariates. For both end points, Abascal score was no longer significant and calcification remained highly associated with the end point. In addition, another Cox model was run for each end point with the two main effect covariates and an interaction term. For both end points there was no evidence of an interaction effect involving Abascal score and calcification.
All Cox model analyses were performed using PROC PHREG in SAS version 6.09.
2.1 Study patients.
One hundred and forty-nine patients underwent percutaneous mitral balloon valvotomy. The average patient age was 54.6 ± 14 years (range 24 to 86). Before the procedure, 5 patients (3%) were in functional class I, 48 (32%) in class II, 87 (58%) in class III and 9 (6%) in class IV. Eight-five patients were in normal sinus rhythm, 60 were in atrial fibrillation and 4 had other rhythms (3 low atrial rhythms and 1 paced rhythm). Twenty-three patients had undergone prior mitral commissurotomy procedures, 16 of which had been performed under open visualization.
2.2 Echocardiographic findings.
Baseline echocardiograms were available in 141 (95%) of the 149 patients. Ten studies for commissural scoring and six studies for Abascal scoring were assessed as having suboptimal image quality and were not used. The average Abascal score was 8 ± 2.1 (range 4 to 13). Eighty-two patients (61%) had a score ≤8, whereas 32 (24%) had an Abascal score ≥10. Twenty-nine patients (21%) had calcification in either the medial or lateral mitral valve commissure (3 of these patients had bilateral commissural calcification). Thirty-three patients (24%) had an Abascal grading ≥3 for calcification, 16 of whom (48%) had the presence of calcium in a commissure. Of the 32 patients with an Abascal score ≥10, 14 (44%) had calcium present in a commissure.
2.3 Procedural results.
Echocardiographic guidance was used in 51 patients (34%). The mean mitral valve area of 1.1 ± 0.36 cm2before the procedure increased to 1.95 ± 0.65 cm2after the procedure (p < 0.01). The mean mitral valve gradient of 13 ± 5.8 mm Hg before the procedure was reduced to 6.1 ± 3.5 mm Hg after the procedure (p < 0.1). In-hospital complications included death (2%), cardiac tamponade (3%) and severe mitral regurgitation (6%). Four patients (3%) had a technically unsuccessful procedure due to the inability to cross the mitral valve with the balloon. Operative reports were available in six patients who underwent mitral valve replacement due to severe mitral regurgitation. The mechanism of mitral regurgitation was identified to be secondary to torn valve leaflets in four patients, a torn leaflet and chordae in one patient and avulsion of the papillary muscle and left ventricular rupture in one patient. All four patients with torn leaflets alone had unilateral commissural calcification identified on echocardiography. Leaflet tears were related to the calcified commissure if splitting occurred in that commissure (two patients) and to the contralateral commissure if the calcified commissure did not split (two patients). No echocardiographic commissural calcification was seen in the patient with leaflet and chordal tearing and dense calcification was noted in the lateral commissure of the last patient during the procedure.
The single-balloon technique (Inoue) was used in 99 patients (66%) and the double-balloon technique in 48 patients (32%). A combination of both techniques was used in two patients (1%). When comparing the type of balloon used (single vs. double) there was no difference in final mitral valve area (<1.5 vs. ≥1.5 cm2; p = 0.8) or the combined end point of mitral valve area, complications and procedural death (p = 0.6). However, there was a significant difference in freedom from death, mitral valve replacement and percutaneous mitral balloon valvotomy (p = 0.03) and all end points combined (p = 0.005) at last follow-up when comparing the single- versus the double-balloon technique.
The mean follow-up period was 1.84 years (range 30 days to 7.95 years). At last follow-up, 14 patients (9%) were dead, 2 (1.5%) underwent a repeat percutaneous mitral balloon valvotomy, 24 (16%) underwent mitral valve replacement and 32 (21%) were functional class III or higher.
2.5 Echocardiographic predictors. Short-term results.
When Abascal score is taken as a discrete variable (≤8, >8) and compared with the final mitral valve area (<1.5, ≥1.5), low values of Abascal score are significantly associated with high values of the final mitral valve area (Fisher exact test, p = 0.043). When Abascal score (≤8, >8) is compared with the multiple event end point of final mitral valve area (<1.5, ≥1.5), complication (yes, no) or procedural death (yes, no), low values of Abascal score are significantly associated with high values of mitral valve area and no complications or deaths (Fisher exact test, p < 0.01). The absence of commissural calcification is also significantly associated with high values of mitral valve area and no complications or deaths (Fisher exact test, p < 0.01), but not with the final mitral valve area (<1.5, ≥1.5) alone (p = NS). When the extent of commissural calcification (unilateral vs. bilateral) is compared with the multiple event end point just described, there is no significant difference between the two subgroups (Fisher exact test, p = NS); however, patients with bilateral commissural calcification were less likely to achieve a final mitral valve area ≥1.5 cm2(Fisher exact test, p < 0.05).
2.5.1 Survival free of death, mitral valve replacement or repeat percutaneous mitral balloon valvotomy.
The 3-year Kaplan-Meier estimate for overall survivorship free of death, mitral valve replacement and repeat valvotomy is 69.9% (95% confidence interval 60.4 to 80.2). The actuarial survival curves with freedom from death, mitral valve replacement or repeat percutaneous mitral balloon valvotomy for patients with Abascal scores ≤8 and >8 are shown in Fig. 2. The 1-, 2- and 3-year Kaplan-Meier estimated actuarial event-free survival rates were 93 ± 3% versus 77 ± 6%, 82 ± 5% versus 71 ± 7% and 79 ± 6% versus 67 ± 8%, respectively (p = NS). In contrast to these results, Fig. 2shows the estimated actuarial event-free survival curves for those patients without commissural calcium versus those patients with calcium in a commissure. The 1-, 2- and 3-year Kaplan-Meier estimated event-free actuarial survival rates were 93 ± 3% versus 63 ± 10%, 86 ± 4% versus 53 ± 10% and 86 ± 4% versus 40 ± 11%, respectively (p < 0.001). The results of the Cox regression support these Kaplan-Meier estimates. In a model with Abascal score and commissural calcification and their interaction, calcification is the only significant variable (p < 0.01).
2.6 Event-free survival.
The 1-, 2- and 3-year Kaplan-Meier estimated actuarial survival rates free of all events (death, mitral valve replacement or repeat percutaneous mitral balloon valvotomy and functional class III or IV) were 91 ± 4% versus 73 ± 7%, 78 ± 6% versus 67 ± 7% and 75 ± 6% versus 64 ± 8%, respectively, for patients with Abascal scores ≤8 and >8 (p = 0.07). Similarly, these results are in sharp contrast with the event-free survival rates of patients without calcium in a commissure versus those with commissural calcium (90 ± 3% vs. 59 ± 10%, 82 ± 5% vs. 51 ± 10% and 82 ± 5% vs. 38 ± 11%, respectively [p < 0.001]). Once again, in a model with Abascal score and commissural calcification and their interaction, calcification is the only significant variable (p < 0.01).
Since the introduction of percutaneous mitral balloon valvotomy by Inoue et al. () in 1984, the procedure has been widely accepted as an effective therapy for mitral stenosis. In two randomized studies percutaneous mitral balloon valvotomy has been shown to be as effective as open or closed mitral commissurotomy ([4, 5]). A number of previous studies have shown that percutaneous mitral balloon valvotomy results in satisfactory immediate- and long-term outcome ([1–3, 7, 15, 16]) with both the single- and double-balloon techniques. These studies also highlighted subgroups of patients according to clinical criteria that fared poorly with percutaneous mitral balloon valvotomy, including the elderly, those with a history of previous surgical commissurotomy and those in atrial fibrillation. Although these clinical variables are helpful, by far the strongest independent predictor of outcome in these studies has been the appearance of the mitral valve on two-dimensional echocardiographic evaluation ([1, 3, 6, 7, 16]). Most centers offering percutaneous mitral balloon valvotomy now use the echocardiographic scoring system of Abascal et al. (), which involves a semiquantitative grading of mitral valve leaflet thickening, mobility, calcification and subvalvular thickening. Patients with a score ≤8 are considered good candidates for the procedure and have been shown to have better immediate- and long-term outcomes ([3, 6, 7]).
The dominant mechanism by which mitral valve stenosis is relieved by the balloon technique is by splitting of the commissures ([9, 10]). This study demonstrates the importance of the assessment of mitral commissural morphology by two-dimensional echocardiography before percutaneous mitral balloon valvotomy, in addition to the more established methods of mitral echocardiographic scoring. We specifically used the presence or absence of commissural calcium as a simple reproducible method for describing a valve with less likelihood of responding to balloon valvotomy. Herein, there was a trend but no significant difference in intermediate-term outcomes between patients with a score ≤8 and those with a score >8. This was in sharp contrast to a large difference in intermediate-term outcomes between those patients with calcification in commissures and those without calcification. These results extend the observations of Fatkin et al. (), who demonstrated the influence of commissural calcification on the short-term outcome in a series of 30 patients. Thus, a simple evaluation of the presence or absence of commissural calcium on two-dimensional echocardiography is a better predictor of intermediate-term outcome than the Abascal mitral echocardiographic score ([7, 8]).
Previous studies on long-term outcome and mitral calcification are limited. Tuzcu et al. () reviewed the association between mitral valve calcification on fluoroscopy and long-term outcome and found that survival rates became progressively worse when severity of calcification increased. This series did not take into account whether the calcification was in the commissure, given the limitations of fluoroscopy. Post et al. () reviewed the Inoue balloon registry and studied the outcome of patients with Abascal scores ≥10, but also evaluated patients for the presence of commissural calcification as proposed by the University of Southern California (USC) scoring system (). The immediate-term outcome success was inversely proportional to the number of commissures with calcium; however, the USC commissural score was not predictive of long-term outcome. Finally, a study of patients developing severe mitral regurgitation after percutaneous mitral balloon valvotomy showed that of six patients requiring early mitral valve replacement, five had evidence of commissural calcium on pathologic examination ().
Our study highlights the importance of examining the presence or absence of calcification directly in the commissures for selecting patients for percutaneous mitral balloon valvotomy. The present Abascal scoring system uses calcification as one of four variables but does not differentiate between calcium in the leaflets versus calcium directly in the commissures. Of those patients in this study with significant calcification recognized by the Abascal criteria (calcification subscore ≥3), only 48% had calcium present in a commissure, placing them in a more favorable risk group. Furthermore, of 32 patients with a high total Abascal score (≥10), such as in the group analyzed by Post et al. (), only 14 (44%) had calcium present in a commissure. Based on these observations it is our recommendation to proceed with percutaneous mitral balloon valvotomy if there is no commissural calcification, despite a high Abascal score. In those patients with calcification in the commissures, mitral valve replacement should be considered because the risk of complications in follow-up is increased. However, if those patients with commissural calcification are at high risk for open heart surgery, percutaneous mitral balloon valvotomy could be considered if the physician and patient are made aware of the higher risk and lower success rate. The procedure should be performed beginning with smaller balloon inflations and accepting final mitral valve areas than those seen in patients without commissural calcification.
3.1 Study limitations.
This was a retrospective study of a group of referred patients and was subject to the drawbacks attributed to such a study. Sick, elderly and nonsurgical candidates were prevalent among the earlier patients treated at our institution before widespread acceptance of the procedure. These patients who represent a higher risk group were also on the steep portion of the learning curve that accompanies all new procedures (). Furthermore, the double-balloon technique was used in the majority of these earlier patients, and although there was no difference in short-term outcome, the significant difference in long-term outcome seen in these patients compared with those in whom the single-balloon technique was used should be interpreted with caution.
Routine follow-up echocardiography was not performed on patients before hospital discharge after percutaneous mitral balloon valvotomy. Thus, the impact of commissural calcium on the immediate echocardiographic outcome, as reported previously by Fatkin et al. (), could not be performed.
An inherent bias of the study is the fact that all patients who were being considered for percutaneous mitral balloon valvotomy at this institution underwent scoring with the conventional Abascal system first. Some patients with high scores may have been eliminated, despite the appearance of the commissures. Thus, the presence of commissural calcification as a lone predictor of outcome needs prospective testing in a large group of patients before this can be deemed applicable to all patients with mitral stenosis.
The present study demonstrates that the presence of commissural calcification identified by two-dimensional echocardiography is a useful predictor of outcome after percutaneous mitral balloon valvotomy. Patients with evidence of calcium in a commissure have a lower survival rate and a higher incidence of mitral valve replacement at intermediate follow-up. This simple observation may allow better selection of patients for percutaneous mitral balloon valvotomy and provide a powerful adjunct to the present echocardiographic evaluation of mitral valve morphology.
- Received January 31, 1996.
- Revision received June 27, 1996.
- Accepted September 17, 1996.
- The American College of Cardiology
- (1992) Complications and mortality of percutaneous balloon mitral commissurotomy: a report from the National Heart, Lung, and Blood Institute Balloon Valvotomy Registry. Circulation 85:2014–2024.
- Palacios IF,
- Tuzcu ME,
- Weyman AE,
- Newell JB,
- Block PC
- Patel JJ,
- Sharma D,
- Mitha AS,
- et al.
- Wilkins GT,
- Weyman AE,
- Abascal VM,
- Block PC,
- Palacios IF
- Post JR,
- Feldman T,
- Isner J,
- Hermann HC
- Reid CL,
- MacKay CR,
- Chandraratna PAN,
- Kawanishi DT,
- Rahimtoola SH
- Fatkin D,
- Roy P,
- Morgan JJ,
- Feneley MP
- Weyman AE
- Tuzcu EM,
- Block PC,
- Griffin BP,
- Newell JB,
- Palacios IF
- Tuzcu EM,
- Block PC,
- Griffin B,
- Dinsmore R,
- Newell JB,
- Palacios IF