Detecting bioprosthetic mitral valve dysfunction on transthoracic echocardiography can be challenging because of acoustic shadowing of regurgitant jets and a wide normal range of transvalvular gradients. Several studies in mechanical mitral valves have demonstrated the utility of the transthoracically derived parameters E (peak early mitral inflow velocity), pressure half-time, and the ratio of mitral inflow velocity-time integral (VTI MV ) to left ventricular outflow tract velocity-time integral (VTI LVOT ) in detecting significant prosthetic dysfunction. Uncertainty exists as to their applicability and appropriate cutoff levels in bioprosthetic valves. This study was designed to establish the accuracy and appropriate normal limits of routinely collected transthoracic Doppler parameters when used to assess bioprosthetic mitral valve function.
A total of 135 clinically stable patients with bioprosthetic mitral valves who had undergone both transthoracic and transesophageal echocardiography within a 6-month period were retrospectively identified from the past 11 years of the echocardiography database. Transthoracic findings were labeled as normal ( n = 81), regurgitant ( n = 44), or stenotic ( n = 10) according to the patient’s transesophageal echocardiographic findings. Univariate and multivariate analyses were performed to identify Doppler parameters that detected dysfunction; then receiver operating characteristic curves were created to establish appropriate normal cutoff levels.
The VTI MV /VTI LVOT ratio was the most accurate Doppler parameter at detecting valvular dysfunction, with a ratio of >2.5 providing sensitivity of 100% and specificity of 95%. E > 1.9 m/sec was slightly less accurate (93% sensitivity, 72% specificity), while a pressure half-time of >170 msec had both 100% specificity and sensitivity for detecting significant bioprosthetic mitral valve stenosis, (although it did not differentiate between regurgitant and normal).
This study demonstrates that Doppler parameters derived from transthoracic echocardiography can accurately detect bioprosthetic mitral valve dysfunction. These parameters, particularly a VTI MV /VTI LVOT ratio of >2.5, are a sensitive way of selecting patients to undergo more invasive examination with transesophageal echocardiography.
Transthoracic Doppler parameters identify dysfunctional mechanical mitral valves.
The authors examined these parameters in normal and abnormal bioprosthetic mitral valves.
The VTI MV /VTI LVOT ratio was best at identifying valvular dysfunction.
VTI MV /VTI LVOT > 2.5 was 100% sensitive for bioprosthetic mitral valve dysfunction.
VTI MV /VTI LVOT > 2.5 was 95% specific for bioprosthetic mitral valve dysfunction.
Detecting bioprosthetic mitral valve dysfunction (significant stenosis or regurgitation) on transthoracic echocardiography (TTE) can be challenging yet vital to appropriately select patients for more invasive evaluation by transesophageal echocardiography (TEE). The large normal range of transmitral gradients published for each individual prosthesis complicates the assessment of pathologic stenosis, and acoustic shadowing of the regurgitant jet by the prosthesis can render significant regurgitation invisible to standard color Doppler imaging. Transmitral gradients are also significantly affected by heart rate and cardiac output, and normal mean gradients have been reported to be as high as 15 mm Hg. Transthoracically derived Doppler parameters that do not rely on visualization of the regurgitant jet and are relatively independent of heart rate and stroke volume (SV) have been shown to be effective in detecting significant mechanical mitral prosthetic dysfunction. Few studies, however, have evaluated these parameters in bioprosthetic valves. Current American Society of Echocardiography guidelines recommend assessing various Doppler parameters in prosthetic mitral valve function but note uncertainty in cutoff values in bioprosthetic valves. Several studies have reported the full range of Doppler values for normally functioning, in vivo bioprosthetic mitral valves, but none have investigated these parameters in dysfunctional valves.
Early Doppler studies on normal bioprosthetic valves focused on peak gradient and valve area, calculated via the pressure half-time (PHT). More recently, E (the peak velocity of early left ventricular filling), the PHT, and the ratio of mitral inflow velocity-time integral (VTI MV ) to left ventricular outflow tract velocity-time integral (VTI LVOT ) have been shown to be the most accurate Doppler parameters when assessing mechanical mitral prosthetic dysfunction. In mechanical valves, these studies have shown that either an E of ≥1.9 m/sec or a VTI MV /VTI LVOT ratio of ≥2.2 detect significant prosthetic valve dysfunction (stenosis or regurgitation) with an accuracy of >90%. No studies have evaluated whether these cutoffs apply to modern bioprosthetic valves. The few studies that have reported these Doppler parameters for normally functioning bioprosthetic valves have suggested that higher cutoffs may be needed. The present study was therefore designed to determine the accuracy and best cutoff levels of routinely collected Doppler values when used to detect significant bioprosthetic mitral valve dysfunction, using both normal and abnormally functioning bioprosthetic mitral valves.
All patients with bioprosthetic mitral valves who had undergone both TTE and TEE within a period of 6 months were retrospectively identified in our echocardiography database. Patients were included if they had undergone comprehensive transthoracic studies within 180 days of their transesophageal studies and had no interval clinical evidence of a change in valvular function. All studies between January 2004 and May 2015 were eligible for inclusion. Exclusion criteria were subvalvular left ventricular outflow tract (LVOT) obstruction or moderate or greater aortic regurgitation (both of which affect VTI LVOT ) or image quality precluding accurate measurement on transthoracic imaging. Of the approximately 220,000 studies in the database, 146 transthoracic studies met the inclusion criteria, with 135 included in the final analysis after the exclusion of three patients because of subvalvular LVOT obstruction, four patients because of insufficient image quality, and four patients because of moderate or greater aortic regurgitation ( Figure 1 ).
Patients’ transthoracic studies were divided into three categories on the basis of prosthetic mitral function on TEE (which served as the gold standard in the majority of cases): normal, regurgitant (defined as those with moderate or severe regurgitation), or stenotic (defined as two-dimensional evidence of abnormal valve structure on TEE [thickened leaflet with decreased mobility, thrombus, or pannus] with a mean gradient of >5 mm Hg on TTE and no more than mild regurgitation). Because stenotic values were required to have two-dimensional evidence of abnormal function on TEE, a mean gradient of >5 mm Hg was chosen, because this level is given as the upper limit of definitely normal valves in the 2009 American Society of Echocardiography guidelines on prosthetic valvular assessment. Additionally, we prospectively elected to include in the regurgitant group any transthoracic studies judged to have greater than moderate prosthetic mitral regurgitation by color Doppler (as judged by two independent reviewers). This enabled inclusion of those patients with clearly significant regurgitation who did not proceed to TEE or surgery, thus reducing selection bias. Multiple transthoracic studies from a single patient were included only in cases of redo mitral valve surgery in which preoperative TEE demonstrated bioprosthetic dysfunction, and intraoperative images obtained after valve replacement demonstrated normal valve function. In this case, preoperative TTE was allocated to the dysfunctional group (regurgitant or stenotic) and postoperative TTE to the normal group. Ethical approval for the study was obtained from the University of British Columbia.
Studies were performed using commercially available ultrasound systems (iE33 [Philips Medical Imaging, Andover, MA], Vivid I [GE Healthcare, Milwaukee, WI], or HP Sonos 5500 [Hewlett-Packard, Andover, MA]) and recorded on an image management system (Xcelera; Philips Medical Imaging).
During TEE, leaflet mobility was examined with two-dimensional imaging, and the maximum mean gradient through the bioprosthetic mitral valve was sought using continuous-wave Doppler from multiple views. The presence of mitral regurgitation on TEE was evaluated with color Doppler with adjustment of the transducer to maximize the regurgitant jet. The source of the jet was established, and the maximal area of the high-velocity regurgitant flow was traced to assist in classifying severity. Central jets were classified as mild (jet area < 4 cm 2 ), moderate (4–7 cm 2 ), or severe (>7 cm 2 ). Eccentric jets were considered moderate if the jet impacted the left atrial wall and changed direction and severe if this was associated with systolic pulmonary vein flow reversal. These criteria are in accordance with the latest American Society of Echocardiography guidelines on prosthetic valve assessment and were used to allow inclusion of studies from the entire study period, as quantitative assessment was not standard during the early years of the study.
A single investigator, who was blinded to the transesophageal findings, made all transthoracic measurements using an offline workstation (Xcelera). Transthoracic imaging included continuous-wave Doppler of mitral inflow performed via the apical window, with guidance by color Doppler to reduce angle-dependent error. From this, E (peak early diastolic flow in meters per second), mean gradient (millimeters of mercury) and prosthetic valve velocity-time integral (VTI; centimeters) were measured, ensuring that overgained parts of the Doppler signal were not included (see Figure 2 ). Evidence of mitral regurgitation was pursued with both color and continuous-wave Doppler from multiple windows. Measurement of VTI LVOT was performed from the apical window using pulsed-wave Doppler 0.5 to 1 cm proximal to the aortic valve. In patients in atrial fibrillation, an average of at least five beats was used. Left ventricular systolic function was defined as normal (ejection fraction [EF] > 50%) or mildly (EF 40%–50%), moderately (EF 30%–40%), or severely (EF < 30%) reduced. The VTI MV /VTI LVOT ratio was calculated by dividing VTI MV by VTI LVOT , and the effective prosthetic valve area was calculated using the continuity equation (this is subsequently referred to as the SV/VTI MV ratio because it does not represent the true mitral valve area in significantly regurgitant valves). SV was calculated using the LVOT diameter and VTI and is subsequently referred to as forward SV.
Descriptive statistics were used to describe patients’ baseline characteristics and transthoracic echocardiographic parameters. Continuous variables are summarized as mean ± SD or as medians and interquartile ranges. Categorical variables are summarized as frequencies and percentages, and associations were established using the χ 2 test. Univariate testing was then carried out (using one-way, between-groups analysis of variance to test for significance) to determine whether these parameters predicted bioprosthetic valve dysfunction. Receiver operating curves were constructed for each of the predictive parameters to select appropriate cutoff values, slightly favoring sensitivity over specificity, as would be appropriate in clinical practice. Specificity, sensitivity, positive predictive values, and negative predictive values of the selected receiver operating characteristic values are quoted to better explain these results but should be viewed with caution until validated in independent cohorts. The areas under the receiver operating characteristic curves for predictors were compared to assess their predictive ability (using the method described by Hanley and McNeil ). Stepwise and multinominal logistic regression was used to establish multivariate predictors of prosthetic valve dysfunction, which were used to create models using combinations of parameters that would best predict valve dysfunction. The six strongest univariate predictors of a finding of dysfunction were chosen for multivariate analysis (VTI MV , E, VTI MV /VTI LVOT ratio, mean gradient, PHT, and effective valve area [SV/VTI MV ratio]).
A total of 135 transthoracic studies (81 normal, 44 regurgitant, 10 stenotic) were available for analysis after application of the exclusion criteria. Sixteen transthoracic studies were included in the regurgitant group without transesophageal confirmation, as per the inclusion criteria. Two of these patients had evidence on TEE of severe regurgitation within 12 months; the remainder did not undergo TEE, as the patients declined redo surgery. The median number of days between TTE and TEE was 14 (interquartile range, 4–38) in the regurgitant group, 22 (interquartile range, 6–88) in the normal group, and 10 (interquartile range, 1–28) in the stenotic group.
Fifty-five of the patients (68%) in the normal group had this identified on intraoperative TEE, immediately after mitral valve surgery. Indications for TEE in the remainder of the normal group were hemodynamic instability (9%), exclusion of endocarditis (7%), assessment of bioprosthetic mitral valve function (7%), and other unrelated indications (9%), such as exclusion of thrombus before cardioversion. Transesophageal echocardiographic indications in the abnormally functioning group were evaluation of suspected prosthetic mitral valve dysfunction (62%), exclusion of endocarditis (22%), and other unrelated indications (16%), such as assessment of suitability for percutaneous left atrial appendage closure. In the stenotic group, the original TTE report correctly suggested the diagnosis 100% of the time, whereas transthoracic imaging missed significant regurgitation in 20% of the regurgitant group and incorrectly suggested dysfunction in 7% of the normal group.
Thirty-one of the patients (57%) with abnormally functioning mitral prostheses were considered for redo surgery (with 22 proceeding) at an average of 10.5 years after their initial operations. In all 22 cases, direct inspection of the explanted valve confirmed anatomic abnormalities consistent with prosthetic malfunction. None of the patients with normally functioning prostheses have undergone redo surgery during a median follow-up period of 5.3 years (range, 0–18 years).
The mean age of 77 years and mean body surface area (BSA) of 1.8 m 2 did not differ significantly between the groups. Atrial fibrillation was present in 46% of patients with normal bioprosthetic function, in 23% of patients with regurgitant valves, and in 10% of patients with stenotic valves ( P < .001). All patients’ heart rates were between 50 and 100 beats/min during their transthoracic studies. Left ventricular systolic function was not significantly different between the groups, although there was a trend toward higher function in the regurgitant group ( Table 1 ). Valve sizes ranged from 25 to 33 mm (median, 29 mm), with valve size information available in 93% of patients (see Supplemental Table 1 , available at www.onlinejase.com ). Seven bioprosthetic valve types made up the majority (93%) of valves studied, with the St Jude Medical (St Paul, MN) Epic (porcine) being the most common (41%). The Medtronic (Minneapolis, MN) Mosaic (porcine) made up 26%, the Carpentier-Edwards (Edward Lifesciences, Irvine, CA) Magna (pericardial, models 7000 FTX and 7300 FTX) 10%, the Carpentier-Edwards 6650 (porcine, supra-annular valve) 8%, and the Carpentier-Edwards PERIMOUNT Plus (pericardial, model no. 6900P) 7%.
|Variable||Normal ( n = 81)||Regurgitant ( n = 44)||Stenotic ( n = 10)||P|
|Age (y)||75 ± 12||79 ± 12.9||79.3 ± 6.8||.17|
|BSA (m 2 )||1.8 ± 0.2||1.8 ± 0.2||1.8 ± 0.1||.88|
|AF||37 (46)||14 (32)||1 (10)||<.05 †|
|LVOT diameter (mm)||20.0 ± 1.4||20.0 ± 1.3||20.6 ± 1.7||.39|
|Normal LV systolic function||46 (57)||31 (70)||4 (40)||.13|
|Mild LV systolic dysfunction||17 (21)||8 (18)||4 (40)||.31|
|Moderate LV systolic dysfunction||12 (15)||3 (7)||2 (20)||.33|
|Severe LV systolic dysfunction||6 (7)||2 (5)||0 (0)||.58|
|VTI LVOT (cm)||19.4 ± 4.7||16.8 ± 3.4||17.8 ± 3.6||<.01 ∗|
|SV LVOT (mL)||61 ± 17||53 ± 13||60 ± 17||<.03 ∗|
|VTI MV (cm)||39.6 ± 10.3||60.4 ± 11.4||79.0 ± 25.9||<.0001 †|
|E velocity (m/sec)||1.7 ± 0.3||2.3 ± 0.2||2.3 ± 0.2||<.0001 ‡|
|VTI MV /VTI LVOT||2.1 (1.8–2.3)||3.4 (3–4.2)||4.3 (3.9–4.6)||<.0001 †|
|Mean gradient (mm Hg)||4.8 (3.7–6.2)||8 (5.9–10.5)||12.3 (10–16.5)||<.0001 †|
|PHT (msec)||81 (64–105)||109 (91–131)||252 (223–284)||<.0001 †|
|Effective valve area (cm 2 )||1.6 (0.3–1.5)||0.9 ± 0.2||0.8 ± 0.1||<.0001 ‡|
Univariate Analysis of Echocardiographic Parameters
All Doppler-derived echocardiographic parameters predicted the finding of valve dysfunction on linked TEE using univariate analysis ( Table 1 ). Mean gradient, VTI MV , and PHT were all significantly higher in the stenotic group than the regurgitant group, which were also significantly higher than the normal group. Figure 2 shows representative Doppler images of prosthetic mitral flow in stenotic, regurgitant, and normal valves. The longer PHT in the regurgitant group is likely secondary to concomitant stenosis in many of the regurgitant valves. Forward SV and VTI LVOT were significantly lower in regurgitant valves compared with normal valves, but there was no difference between normal and stenotic valves. Peak early inflow velocity (E) was higher in both stenotic and regurgitant valves but was not different between the two groups.
Calculated Doppler variables also detected valve dysfunction, with the ratio of VTI MV /VTI LVOT being significantly different among the three groups and highest in stenotic valves (see Figure 3 ). Effective valve area (SV/VTI MV ) was smaller in both regurgitant and stenotic groups but was not different between the two.