2017 Volume 81 Issue 11 Pages 1730-1735
Background: As mitral valve (MV) repair for Barlow’s disease remains surgically challenging, it is important to distinguish Barlow’s disease from fibroelastic deficiency (FED) preoperatively. We hypothesized that the prolapse volume to prolapse height ratio (PV-PH ratio) may be useful to differentiate Barlow’s disease and FED.
Methods and Results: In 76 patients with MV prolapse who underwent presurgical transesophageal echocardiography, the 3D MV morphology was quantified: 19 patients were diagnosed with Barlow’s disease and 57 with FED. The patients with Barlow’s disease had greater prolapse volume and height than the patients with FED, as well as greater PV-PH ratio (0.61±0.35 vs. 0.17±0.10, P<0.001). Receiver-operating characteristic analysis revealed that with a cutoff value of 0.27, the PV-PH ratio differentiated Barlow’s disease from FED with 84.2% sensitivity and 84.2% specificity. Net reclassification improvement showed that the differentiating ability of the PV-PH ratio was significantly superior to prolapse volume (1.30, P<0.001). After being adjusted by each of prolapse volume and height, annular area and shape, and the number of prolapsed segments, the PV-PH ratio had an independent association with Barlow’s disease.
Conclusions: The PV-PH ratio was able to differentiate Barlow’s disease from FED with high accuracy. 3D quantification including this value should be performed before MV repair.
Mitral regurgitation (MR) caused by mitral valve (MV) prolapse is one of the most common valvular heart diseases requiring surgical treatment in developed countries.1 Over the past 3 decades, the efficacy of MV repair over MV replacement has been continuously reported,2–5 and it has been established as the first therapeutic choice for MV prolapse.
There is a wide spectrum of degenerative MV diseases that result in MV prolapse, ranging from Barlow’s disease to fibroelastic deficiency (FED).6 Compared with FED, which is generally characterized by localized involvement of the leaflet with healthy adjacent segments,7,8 Barlow’ disease is a clinical entity characterized by diffuse excess tissue and significantly enlarged leaflets and annulus.8–12 In Barlow’s disease, the whole mitral apparatus is different from a normal valve; multiple segmental involvement of both leaflets with myxomatous change, diffuse chordal elongation in addition to chordal rupture, severely enlarged annulus, and various degrees of annular and subvalvular calcification are observed.6,10–14 Because of its complex valve morphology, MV repair for Barlow’s disease still remains a challenge to surgeons compared with FED.
Quantification of MV morphology using 3-dimensional transesophageal echocardiography (3DTEE) is consistently reported to be superior to 2D echocardiography for providing a precise diagnosis.15–17 A previous study reported that prolapse volume (i.e., the volume under the prolapsed leaflet of the MV above mitral annulus) could differentiate Barlow’s disease from FED completely (with 100% sensitivity and specificity) when using a cutoff value 1.15 mL.11 In clinical practice, however, we often encounter patients with a prolapse volume >1.15 mL who are diagnosed with FED rather than Barlow’s disease intraoperatively. Because Barlow’s disease is characterized by a more widespread prolapse lesion compared with FED, in which prolapse is localized but often associated with significant prolapse height by a flail leaflet, we hypothesized that the prolapse volume to prolapse height ratio (PV-PH ratio) may be useful to differentiate Barlow’s disease from FED.
From a pooled preoperative TEE database of patients who underwent MV surgery during November 2013 and September 2015 in the Sakakibara Heart Institute of Okayama, we analyzed the 3DTEE data of 19 consecutive patients with Barlow’s disease and 57 age- and sex-matched patients with FED (1 Barlow’s disease patient per 3 FED patients). Barlow’s disease and FED were diagnosed by experienced surgeons (T.T. and T.S.) based on the reported criteria of surgical findings. The findings of typical Barlow’s disease are: severely dilated annulus, multisegmental prolapse of leaflet(s) thickened with excess tissue, thickened and elongated chordae with occasional calcification, and multisegmental billowing. On the other hand, in typical FED patients, the leaflet thickened with excess tissue, the elongation of the chords, and the billowing and prolapse are localized in a few segments and the other segments remain normal.11,18 Pathological findings were also investigated to validate the diagnosis.
We obtained approval from the Investigational Review Board Committee of the Sakakibara Heart Institute of Okayama for the collection and publication of data.
Image Acquisition and AnalysisTransthoracic 2D echocardiographic images were acquired in accordance with published guidelines.19 In all patients, we acquired 3DTEE images under mild sedation using an iE33 ultrasound system and X7-2t transducer (Philips Medical Systems, Andover, MA, USA). The probe was positioned at the mid-esophageal level, taking care to include the entire MV; the mid and apical left ventricular segments were excluded to maximize the volume rate. We essentially used multibeat (2–6 beats) acquisitions because of the high volume rate. To avoid stitch artifacts, we used the 1-beat real-time 3D zoomed mode for patients with atrial fibrillation rhythm and those who could not hold their breath during the examination.
Acquired datasets were analyzed by 2 experienced observers (N.K. and M.T.) blinded to the clinical diagnosis. The 3D analysis was performed off-line using commercially available semi-automated software, MVN (Philips Medical Systems). We previously reported the detailed method of 3D quantification using MVN and showed that it is accurate and has high reproducibility, comparable to standard manual tracing software; the interobserver agreements of annular and leaflet parameters assessed using Cronbach’s α were high (0.82–0.97).20 In detail, prolapse volume and height were defined as the volume between the annulus surface and prolapsed leaflet and the height above the annulus surface to the most distant point of the MV leaflet, respectively. These prolapse parameters were automatically calculated by the software. The PV-PH ratio was defined as prolapse volume divided by prolapse height. The 3D parameters were analyzed in the mid-systolic phase. MR grade was evaluated by a comprehensive approach using the proximal isovelocity surface area method and/or vena contracta width, as recommended in the guidelines.19,21 Effective regurgitant orifice measured by the proximal isovelocity surface area method >0.4 cm2 and/or vena contracta width >7 mm were considered as severe MR.
The echocardiographic equipment was managed according to the guidelines of the Japanese Society of Echocardiography.22
Statistical AnalysisData are presented as mean±standard deviation for continuous variables and as frequency (%) for categorical variables. Student’s t-test and the Mann-Whitney U test were used to compare continuous variables with normal and non-normal distributions, respectively. The χ2 test was used to compare categorical variables. The receiver-operating characteristics (ROC) curves were constructed to assess the potential for prolapse volume, prolapse height, and the PV-PH ratio to identify Barlow’s disease. The area under the curve (AUC) of each of the ROC curves was compared according to the method of DeLong et al.23 An appropriate cutoff value to discriminate Barlow’s disease from FED was then decided according to the Youden index. Continuous net reclassification improvement (NRI) was also calculated.24 Multivariable logistic regression models were constructed to assess the independency of the PV-PH ratio. In this study, we did not use matched analysis instead of standard analysis, based on a previous report.25 All statistical analyses were performed with R (The R Foundation for Statistical Computing, Vienna, Austria) and its graphical user interface, EZR (version 1.27; Saitama Medical Center, Jichi Medical University, Saitama, Japan).26 In all analyses, P<0.05 was taken to indicate statistical significance.
Pathological reports on the MV leaflets of 43 patients (75.4%; 12 with Barlow’s disease and 31 with FED) were available. All the reports of the patients with Barlow’s disease stated there was (1) significant myxomatous change with excess tissue and (2) the change involved the majority of the leaflet, whereas only 1 report for the patients with FED included the presence of both findings.
Table 1 shows the patients’ characteristics. Age and sex were successfully matched between the 2 groups. There were no significant differences in body surface area, prevalence of sinus rhythm, blood pressure, and heart rate. As expected, Barlow’s disease had a higher prevalence of bileaflet prolapse (100% vs. 17.5%, P<0.001) and lower prevalence of torn chordae (15.8% vs. 66.7%, P<0.001) compared with FED. All patients were successfully treated by MV repair and there was no conversion to MV replacement. Operative cardiopulmonary bypass time was significantly longer in the patients with Barlow’s disease (140±37 vs. 106±32 min, P=0.001).
| Barlow’s disease (n=19) |
FED (n=57) |
P value | |
|---|---|---|---|
| Age, years | 51.7±12.0 | 52.7±11.2 | 0.74 |
| Female, n (%) | 6 (31.6) | 18 (31.6) | >0.99 |
| Body surface area, m2 | 1.56±0.19 | 1.60±0.19 | 0.44 |
| Cardiac rhythm, n (%) | >0.99 | ||
| Sinus | 18 (94.7) | 53 (93.0) | |
| Atrial fibrillation | 1 (5.3) | 4 (7.0) | |
| SBP, mmHg | 118±15 | 120±24 | 0.70 |
| DBP, mmHg | 66±12 | 70±12 | 0.17 |
| Heart rate, beats/min | 80±18 | 76±14 | 0.26 |
| Prolapse site, n (%) | <0.001 | ||
| Only anterior leaflet | 0 (0.0) | 10 (17.5) | |
| Only posterior leaflet | 0 (0.0) | 37 (64.9) | |
| Bileaflet | 19 (100) | 10 (17.5) | |
| Torn chordae, n (%) | 3 (15.8) | 38 (66.7) | <0.001 |
| Cardiopulmonary bypass time, min | 140±37 | 106±32 | 0.001 |
DBP, diastolic blood pressure; FED, fibroelastic deficiency; SBP, systolic blood pressure.
The volume rate was 18.5±11.1 Hz and there was no significant difference between the 2 groups. Figure 1 shows typical TEE images of the MV obtained in a patient with Barlow’s disease and FED, respectively.
Typical images of Barlow’s disease and fibroelastic deficiency (FED). (A) Barlow’s disease usually has a significantly enlarged mitral annulus and multisegmental prolapse. The prolapse has a wide and low shape. (B) FED typically has localized prolapse with healthy adjacent segments. Although the area of prolapse is small, the leaflet is typically flail with a high prolapse height.
The MV parameters quantified using TEE and dedicated software are summarized in Table 2. Although there was no significant difference in annulus height, the annulus area and circumference were significantly greater in Barlow’s disease compared with FED. Anterior and posterior leaflet areas were also significantly larger in Barlow’s disease. Nonplanar angle, anterior and posterior leaflet angles, and tenting height and volume were significantly smaller in Barlow’s disease, suggesting the MV was flattened and deviated away from the left ventricle.
| Barlow’s disease (n=19) |
FED (n=57) |
P value | |
|---|---|---|---|
| Annulus area, mm2 | 1,730±427 | 1,073±250 | <0.001 |
| Annulus circumference, mm | 155±19 | 123±14 | <0.001 |
| Annulus height, mm | 6.4±2.1 | 5.7±1.5 | 0.14 |
| Nonplanar angle, ° | 160±29 | 130±22 | <0.001 |
| Aorto-mitral angle, ° | 116±10 | 119±11 | 0.40 |
| Mitral leaflet area, mm2 | 2,336±654 | 1,359±356 | <0.001 |
| Anterior leaflet area, mm2 | 1,145±308 | 684±185 | <0.001 |
| Anterior leaflet length, mm | 27.1±5.0 | 22.9±4.5 | 0.001 |
| Anterior leaflet angle, ° | 6.8±10.9 | 18.9±8.1 | <0.001 |
| Posterior leaflet area, mm2 | 1,191±408 | 662±240 | <0.001 |
| Posterior leaflet length, mm | 20.1±6.1 | 15.2±5.8 | 0.002 |
| Posterior leaflet angle, ° | 13.4±18.0 | 32.4±13.4 | <0.001 |
| Tenting volume, mL | 0.24±0.39 | 1.06±1.05 | 0.001 |
| Tenting height, mm | 2.1±1.9 | 5.5±3.2 | <0.001 |
| Prolapse volume, mL | 6.88±5.80 | 1.37±1.48 | <0.001 |
| Prolapse height, mm | 9.9±3.4 | 6.9±3.7 | 0.002 |
| PV-PH ratio | 0.61±0.35 | 0.17±0.10 | <0.001 |
| MR ERO, cm2 | 0.45±0.18 | 0.55±0.22 | 0.15 |
FED, fibroelastic deficiency.
Prolapse volume was 6.88±5.80 mL in Barlow’s disease and 1.37±1.48 mL in FED (P<0.001); prolapse height was 9.9±3.4 mm in Barlow’s disease and 6.9±3.7 mm in FED (P=0.002). As a result, the PV-PH ratio in Barlow’s disease was 0.61±0.35, significantly greater than in FED (0.17±0.10, P<0.001).
Scatterplots shown in Figure 2 demonstrate that although there tended to be a greater prolapse volume in Barlow’s disease than in FED, there was significant overlap between the patient groups for a prolapse volume between 1.0 and 6.5 mL.
Distribution of prolapse volume and height. Although Barlow’s disease tended to have a greater prolapse volume than fibroelastic deficiency (FED), there is significant overlap between the 2 patient groups with prolapse volumes between 1.0 and 6.5 mL. Black circles and red triangles indicate Barlow’s disease and FED, respectively.
ROC curve analysis showed that the AUCs of prolapse volume, prolapse height, and the PV-PH ratio were 0.89 (95% confidence interval (CI) 0.81–0.96), 0.76 (95% CI 0.65–0.87), and 0.92 (95% CI 0.86–0.99), respectively (Figure 3). The AUCs of both prolapse volume and the PV-PH ratio were significantly greater than that of prolapse height. The superiority of the AUC for the PV-PH ratio over prolapse volume was not significant (P=0.065). However, the calculated NRI between prolapse volume and the PV-PH ratio was significant (NRI=1.30, P<0.001), suggesting that the PV-PH ratio was able to more precisely differentiate Barlow’s disease from FED. The optimal cutoff value of the PV-PH ratio assessed by ROC curves was 0.27, with 84.2% sensitivity and 84.2% specificity.
Receiver-operating characteristic (ROC) curves analysis. The area under the curve (AUC) of the PV-PH ratio (red) is slightly higher than the AUC of the prolapse volume (blue: P=0.065). Calculated net reclassification improvement (NRI) between prolapse volume and the PV-PH ratio was significant (NRI=1.30, P<0.001). The optimal cutoff value of the PV-PH ratio assessed by ROC curves is 0.27 (arrow), with 84.2% sensitivity and 84.2% specificity.
Table 3 summarizes the results from the multivariable logistic regression models. Because there were only 19 patients with Barlow’s disease, a detailed multivariate model including many independent variables was not appropriate. Accordingly, we adjusted the PV-PH ratio by only 1 factor in each model. In all models, the PV-PH ratio remained significant even after adjustment for well-known parameters of Barlow’s disease.
| OR [95% CI] | P value | |
|---|---|---|
| Model 1 | ||
| PV-PH ratio/% | 1.31 [1.09–1.57] | 0.0048 |
| Prolapse volume/mL | 0.33 [0.11–0.98] | 0.046 |
| Model 2 | ||
| PV-PH ratio/% | 1.16 [1.06–1.27] | 0.0017 |
| Prolapse height/mm | 0.82 [0.61–1.11] | 0.21 |
| Model 3 | ||
| PV-PH ratio/% | 1.09 [1.02–1.17] | 0.017 |
| Annular area/cm2 | 1.36 [1.00–1.84] | 0.049 |
| Model 4 | ||
| PV-PH ratio/% | 1.12 [1.06–1.20] | <0.001 |
| Ellipsoid index/% | 0.99 [0.93–1.05] | 0.80 |
| Model 5 | ||
| PV-PH ratio/% | 1.12 [1.04–1.21] | 0.0041 |
| No. of prolapsed segments | 2.23 [1.29–3.87] | 0.0043 |
In the present study, we successfully demonstrated that the PV-PH ratio, a novel parameter of MV prolapse, differentiated Barlow’s disease from FED more precisely than crude prolapse volume.
Although there has not been a randomized controlled trial comparing MV repair and MV replacement for MR caused by MV prolapse, many observational studies have consistently shown the superiority of MV repair over MV replacement. MV repair has been established as a standard therapy for MV prolapse.2–5 Guidelines only recommend MV surgery for asymptomatic severe MR if there is a high likelihood of a successful and durable repair.21 Therefore, preoperative assessment of reparability, which depends primarily on the complexity of the prolapse, is one of the most important roles of sonographers and cardiologists.
Barlow’s disease is characterized by its complex pathology and morphology, such as a dilated and floppy annulus, elongated leaflets and chordae with severe myxomatous change, and multisegmental involvement. These patients tend to have bileaflet prolapse and a multi-regurgitant orifice; therefore reparability is generally lower than with FED, in which the pathology and morphology is less complex, with mild annular dilatation, thin and mildly elongated leaflets and chordae, with usually 1- or 2-segment involvement and normal adjacent segments.7,8
A previous report from the University of Chicago demonstrated there was no overlap between Barlow’s disease and FED in prolapse volume quantified with 3DTEE.11 However, in our current study, there were both FED patients and Barlow’s disease patients with a prolapse volume between 1.15 and 6.5 mL, as shown in Figure 2. This difference may be related to the severity of disease, because in our population the prolapse volume and height were greater and higher compared with the previous report (in Barlow’s disease 6.88±5.80 mL and 9.9±3.4 mm vs. 4.71±2.52 mL and 8.4±2.9 mm; in FED 1.37±1.48 mL and 6.9±3.7 mm vs. 0.34±0.34 mL and 4.1±2.4 mm, respectively). In addition, volume rate, one of the most important qualities in analysis of 3D images, was quite low (mean 9±3 Hz) in the previous study. Because the prolapse volume and height change throughout systole, this low frame rate might cause inaccurate quantification.
As shown in Figure 1, typical Barlow’s disease has a widespread, low prolapse, as a “plateau”, whereas FED has a localized high prolapse, as a “tower”. The PV-PH ratio, a novel parameter of prolapse calculated by dividing prolapse volume by prolapse height, reflects the mean area of prolapse. Accordingly, it is reasonable that the PV-PH ratio can differentiate Barlow’s disease from FED precisely; the present results supported this hypothesis.
A possible question based on our results may be whether measurement of the crude prolapse area is adequate. However, to the best of our knowledge, there is no commercially available software that can measure prolapse area, and quantification of the prolapse area is difficult. On the other hand, one of the most valuable features of the PV-PH ratio is its easiness to calculate. With most of the manual and automatic software available for 3D analysis, prolapse volume and height are calculated simultaneously. Therefore, the PV-PH ratio can be calculated without any additional time or effort.
Study LimitationsFirst, this was a single-center retrospective observational study including a relatively small number of patients. Because this was a type of pilot study proposing a hypothesis, the usefulness of the PV-PH ratio should be confirmed in the prospective study including a larger population. Second, because of the small number of patients, subanalysis based on the location of prolapse was inappropriate. If a larger sample size is available, further study, such as a comparison between bileaflet FED and Barlow’s disease, might be important. Third, we used operative findings as the gold standard reference of the differentiation between Barlow’s disease and FED. It is true that the operative findings are observed mainly in the cardiac arrest state and influenced by the experience of the surgeon. However, morphological changes in cardiac rhythm are rather small considering the individual variance, and operative findings were used as the gold standard reference in many previous studies.10,11 In addition, both surgeons in this study were highly experienced, having performed more than 500 valve surgeries. Furthermore, in almost all the patients with pathological reports, pathological findings were concordant with the operative diagnosis. Fourth, because all the subjects were Japanese, the results possibly differ in Western populations. Finally, as mentioned before, we did not measure the prolapse area because it is not possible with commercially available software. Further study to compare the predictability of the PV-PH ratio and prolapse area may be needed when novel software capable of calculating prolapse area becomes available.
PV-PH ratio, a novel parameter of prolapse, was able to differentiate Barlow’s disease from FED more precisely than crude prolapse volume or height. Because this parameter is very easy to calculate when the prolapse volume is available, it may be useful in clinical practice. In preoperative planning for repair of MV prolapse, the PV-PH ratio should be taken into consideration for differentiating Barlow’s disease from FED.
