The long-term outcome of mechanical aortic and mitral prosthetic valve (A-PV, M-PV) dysfunction (PVD) remains a serious complication associated with high morbidity and mortality. We sought to evaluate the incremental diagnostic value of combined transthoracic echocardiography (TTE) and fluoroscopy (F) in patients with suspected PVD. A total of 354 patients (178 A-PV, 176 M-PV) were imaged by TTE and F within 5 days of hospital admission. PVD was confirmed by transesophageal echocardiography, computed tomography, effective thrombolysis, or surgical inspection. PVD was confirmed in 101 patients (57%) with M-PV and 99 (55%) with A-PV. Regardless of the mechanism of PVD, TTE shows good sensitivity and specificity, with accuracy of 80% for M-PV and 91% for A-PV. F shows high specificity, but low sensitivity with accuracy of 68% for M-PV and 78% for A-PV. The integration of TTE + F significantly improved accuracy both for M-PV (83%) and A-PV (96%). At ROC analysis, the combined model of TTE + F showed the highest area under the curve for the detection of PVD compared with TTE and F alone (p < 0.001). In conclusion, in patients with a clinical suspicion of PVD, the combined model of TTE + F offers incremental value over TTE or F alone. This multimodality imaging approach overcomes limitations of TTE or F alone and provides prompt identification of patients who may require further imaging assessment and/or closer follow up.
The long-term outcome of mechanical aortic and mitral prosthetic valves (A-PV and M-PV) has greatly improved. Nevertheless, PV dysfunction (PVD) remains a very serious and challenging complication, associated with high mortality and morbidity. , Several studies have investigated the etiology of PVD, showing that the development of thrombosis, pannus or paraprosthetic regurgitation being closely related to PV failure, , The clinical presentation may vary from no symptoms to acute heart failure and sudden cardiac arrest, and the time between valve replacement and PVD may be very broad. Currently, transthoracic echocardiography (TTE) is the first-line imaging modality for the assessment of patients with PV. Transesophageal echocardiography (TEE) can better define the cause of PVD and help in guiding therapy, risk stratification and follow-up. , Computed tomography (CT) has recently emerged as a complementary approach offering excellent spatial resolution to identify a range of PV complications, including pannus formation. Although TEE and CT are considered the gold-standard in the diagnosis of PVD, an integrated multimodality imaging approach comprising several parameters by TTE and fluoroscopy (F) is essential to appropriately refer suspected PVD to TEE or CT and to devise the optimal therapeutic pathway. The aims of this study were: 1) to evaluate the incremental diagnostic value of combined TTE derived parameters and F over each imaging modality alone in symptomatic patients with high clinical suspicion of PVD, 2) to provide the basis for developing an algorithm allowing for rapid PVD diagnosis.
Methods
We retrospectively enrolled 354 patients with suspected PVD, admitted between 2000 and 2019 to Centro Cardiologico Monzino IRCCS (Milan, Italy). Symptoms of suspected PVD were dyspnea, cerebral and/or peripheral embolism, infective-like disease or hemolytic anemia. All patients underwent TTE, F and TEE within 5 days after hospital admission. Since 2010, 3D TEE was included in the standard diagnostic work-up and since 2013, patients with suspected PVD underwent also CT imaging. The gold standard for establishing PVD was the surgical inspection or effective thrombolysis and, when those were not available, the 2D/3D TEE imaging combined to CT data were used. The Ethical Committee of our institution approved the study (R1264/20-CCM 1328).
Normal PV function (N-PVF) was defined at TEE as the absence of obstructive or non-obstructive thrombus, vegetation or pannus ingrowth with normal leaflet(s) motion, and trivial PV regurgitation. Obstructive PVD (O-PVD) could be due to the presence of thrombus or pannus. Thrombus was defined by the evidence of irregular shaped mobile masses with low echogenicity at 2D/3D TEE and by low-attenuation lesions ranging between 90 and 145 HU, starting in the valve ring area and then spreading to the leaflet(s) at CT. Pannus ingrowth was defined as a mass with lower area and lower density than thrombus at TEE, and as a low-attenuation lesion or a calcified lesion ≥145 HU, protruding into the valvular strut, beneath the PV at CT. At surgical inspection thrombus was confirmed as a mobile mass more commonly located on the atrial side of M-PV and on the aortic root side of A-PV, and pannus was confirmed as a fibrotic ingrowth frequently seen on the ventricular side of M-PV and A-PV. In case of O-PVD, thrombolysis was carried out when thrombotic prosthetic mass size was <0.8 cm 2 at TEE. The results of the thrombolysis were defined as “full success” (leaflet motion normalization), “partial success” (improvement without normalization of leaflet motion), and “ineffective” (absence of improvement of the leaflet motion) using F. Paraprosthetic valve dysfunction (P-PVD) was defined as paraprosthetic regurgitation with an echo-free space close to the sewing ring at TEE, leak by the contrast medium at CT or by surgical inspection.
Comprehensive TTE evaluation was performed using commercially available equipment (Sonos 7500, iE33 or EPIQ system, Philips Medical Systems, Andover, MA, USA; and Vivid 7 or E9, GE Healthcare, Horten, Norway). Multiple cross-sectional and off-axis views were performed to visualize the mobility of PV leaflet(s). Color Doppler was used for screening and evaluating the degree of intra and/or paraprosthetic regurgitation. Doppler-derived parameters in M-PV evaluation included: early peak mitral velocity (E wave), mean transprosthetic gradient (∆P mean ), pressure half time, and Doppler velocity index (DVI). , , Doppler-derived parameters in A-PV evaluation included: peak transprosthetic velocity (V max ), peak and mean transprosthetic gradients (∆P max and ∆P mean ), DVI, effective prosthetic orifice area, acceleration time (AT), ejection time (ET), and the ratio AT/ET. , ,
F was carried out with the patient in the supine position. Due to surgical PV orientation, multiple projections (0-90°) were performed, including cranial-caudal angulations. The examination was considered successful when the tilting disc projection (with the radiographic beam parallel to both the valve ring plane and the tilting axis of discs) was obtained and allowed calculation of opening and closing angles . Short F film (3–10 beats) at 50 frames/s was recorded. Frames of interest were selected to mesure disc(s) motion. For bileaflet prostheses, opening and closing angles were calculated as the distance between the 2 leaflets in the fully open and closed position. For the single disc prostheses, the opening angle was defined as the distance between the housing and the disc at its full open position. Normal reference values for opening and closing angles were obtained from the manufacturer and used to evaluate leaflet motion, specifically for each prosthesis size and type.
The TEE studies were performed using standard, commercially available systems (Sonos 7500, iE33, or EPIQ system, Philips Medical System, Andover, MA) with a 5 MHz multiplane probe or X7-2t probe. TEE examination included 2D and 3D conventional and off-axis views.
All CT examinations were performed using a 256-slice wide volume coverage scanner (Revolution CT, GE Healtcare, Milwaukee). CT scan was performed using an iterative reconstruction algorithm (ASIR-V; GE Healthcare) at 50% level, 100 kV tube voltage and a body-mass index adapted protocol for the tube current. A 16-cm scan volume including the whole cardiac volume during one single-heart beat was used with prospective ECG-gating. All patients received a 50mL bolus of contrast medium (Iomeprol, 400 mg/mL, Bracco, Milan, Italy) at an infusion rate of 5 mL/s followed by 50 mL of saline solution at 5 mL/s. CT dataset were post-processed on a dedicated workstation (Advantage Workstation Volume Share 4.6, GE Healthcare), with specific dedicated software enabling reconstruction of the aortic root in different orthogonal planes.
Based on the definition of PVD given for each imaging modality, the presence of PVD was retrospectively assessed by 2 cardiologists with more than 10 years of experience for analyzing the TTE and 2D/3D TEE data (M.M. and G.T.), by 2 interventional cardiologists for F data (G.T. and P.M.) and by 2 radiologists for CT data (A.A. and G.P.). The different experts were blinded to each other. Discrepancies between the 2 observers were resolved using joint reanalysis and discussion, and consensus values were used for statistics.
Continuous variables were presented as mean ± SD and categorical variables as absolute numbers and percentages. Analysis of variance for independent measurements were used to assess differences between the study subgroups. The association between categorical variables was examined using χ 2 test. A post hoc analysis for significant results were performed using the Bonferroni correction. Receiver operating characteristic curves (ROC) were computed to determine the optimal threshold for TTE, F and TTE + F to diagnose PVD using the Youden index and the area under the curves (AUC) were compared with the DeLong Method. All results were considered significant with P value <0.05. Statistical analysis was performed with SPSS, version 25 software (SPSS Inc, Chicago, IL) and R version 3.5.1.
Results
The study population comprised of 354 patients with suspected PVD (63±11years, 127 [36%] male; 178 A-PV and 176 M-PV) and were classified into 3 groups according to PV function (154 N-PVF, 131 O-PVD, and 69 P-PVD) as detected by the combined information from 2D TEE (n = 341), 3D TEE (n = 101), CT (n = 87), surgical inspection (n = 105), or thrombolysis response (n = 34). Main clinical characteristics of M-PV and A-PV are summarized in Table 1 . Both in M-PV and A-PV, clinical parameters were similar between groups, except that NYHA III-IV was significantly more frequent in O-PVD and P-PVD and hemolytic anemia was significantly more frequent in P-PVD. INR was in range in all patients.
| Mitral prosthetic valve(n = 176) | Aortic prosthetic valve(n = 178) | |||||||
|---|---|---|---|---|---|---|---|---|
| Variable | N-PVF(n = 75) | O-PVD (n = 57) | P-PVD (n = 44) | p-value | N-PVF(n = 79) | O-PVD (n = 74) | P-PVD (n = 25) | p value |
| Age (year) | 61 ± 11 | 65 ± 11 | 65 ± 9 | 0.030 | 66 ± 9 | 65 ± 12 | 62 ± 11 | 0.292 |
| Male | 15 (20%) | 12 (21%) | 18 (41%) * † | 0.026 | 42 (53%) | 23 (31%) * | 17 (68%) † | 0.012 |
| BSA (m2) | 1.7 ± 0.18 | 1.73 ± 0.24 | 1.74 ± 0.19 | 0.427 | 1.79 ± 0.21 | 1.72 ± 0.18 | 1.84 ± 0.18 † | 0.009 |
| NYHA functional class III-IV | 9 (12%) | 34 (60%) * | 34 (77%) * | <0.001 | 6 (8%) | 30 (41%) * | 11 (44%) * | <0.001 |
| Symptoms | ||||||||
| Dyspnea | 67 (89%) | 43 (75%) | 34 (77%) | 0.081 | 73 (92%) | 69 (93%) | 19 (76%) * † | 0.029 |
| Embolic event | 2 (3%) | 5 (9%) | 1 (2%) | 0.175 | 0 | 2 (3%) | 0 | 0.241 |
| Infective-like disease | 4 (5%) | 8 (14%) | 3 (7%) | 0.186 | 5 (6%) | 3 (4%) | 5 (20%) * † | 0.027 |
| Hemolytic anemia | 2 (3%) | 1 (2%) | 6 (14%) * † | 0.012 | 1 (1%) | 0 | 1 (4%) | 0.257 |
| INR | 2.75 (2.19;3.35) | 2.16 (1.92;3.40) | 2.61 (2.10;3.33) | 0.451 | 2.59 (2.12;3.16) | 2.48 (2.07;3.03) | 2.59 (2.03;3.28) | 0.962 |
| Rhythm | 0.808 | 0.227 | ||||||
| Sinus Rhythm | 18 (24%) | 11 (19%) | 10 (23%) | 47 (60%) | 35 (49%) | 15 (65%) | ||
| Atrial Fibrillation | 57 (76%) | 46 (81%) | 34 (77%) | 31 (40%) | 37 (51%) | 8 (35%) | ||
| A-PV plus M-PV | 35 (47%) | 18 (32%) | 20 (45%) | 0.181 | 33 (42%) | 31 (42%) | 6 (24%) | 0.239 |
| Coronary artery disease | 1 (1%) | 5 (9%) | 3 (7%) | 0.132 | 17 (22%) | 10 (14%) | 7 (28%) | 0.215 |
| Prosthetic type | 0.363 | 0.873 | ||||||
| Single disk | 15 (20%) | 9 (16%) | 12 (27%) | 13 (16%) | 10 (13%) | 4 (16%) | ||
| Bileaflet | 60 (80%) | 48 (84%) | 32 (73%) | 66 (84%) | 64 (87%) | 21 (84%) | ||
| Time from implantation (years) | 14 (7;22) | 15 (12;21) | 21 (10;25) | 0.513 | 13 (8;19) | 12 (9;17) | 8 (2;20) † | 0.003 |
F was feasible in 95% of A-PV (93% of single disc PV, 95% of bileaflet PV) and in 89% of M-PV (83% of single disc PV, 91% of bileaflet PV). Opening and closing angles at F are reported in Table 2 . In particular, F examination showed abnormal leaflet motion in 104 out of 131 patients with O-PVD (79%). opening and closing angles were not significantly different in P-PVD as compared to N-PVF.
