Normal Values of the Mitral-Aortic Intervalvular Fibrosa Thickness: A Multimodality Study.

BACKGROUND: The avascular region of the fibrous body between the mitral and aortic valves, named mitral-aortic intervalvular fibrosa (MAIVF), is often involved in the periaortic diffusion of infective endocarditis (IE), resulting in abscess or pseudoaneurysm formation. The early recognition of these life-threatening complications is of crucial importance, as urgent surgical correction is necessary. In the first stages of the abscess formation, the only sign is an increased thickness of the MAIVF. To the best of our knowledge, normal transesophageal echocardiography (TEE) examination reference values for MAIVF thickness has not yet been established. The aim of the study was to define the normal ranges of MAIVF thickness in a population of healthy adults who underwent a TEE examination.
MATERIALS AND METHODS: A population of consecutive adult patients who underwent a TEE examination was enrolled in the study. Measurement was performed in short-axis (SAX) and long-axis (LAX) views. Mean-2 standard deviations (mean-2SDs) and 5%, 10%, 90%, and 95% confidence intervals were evaluated. A comparison with MAIVF thickness in patients affected by aortic IE complicated by abscess formation was performed, and receiver operating characteristic (ROC) curves were constructed to achieve the optimal cutoff value of normality.
RESULTS: A total of 477 consecutive Caucasian adult patients were enrolled (mean age: 69 years, range: 27-93 years). Mean-2SD MAIVF measurement in SAX view was 0.325 cm (95% confidence interval [CI]: 0.319-0.330 cm) and in LAX view was 0.340 cm (95% CI: 0.334-0.346 cm). Computed tomography-MAIVF mean measurement (±2SD) was 0.237 cm (95% CI: 0.110-0.340 cm). ROC curves showed that a cutoff SAX value measurement of 0.552 (area under the curve [AUC]: 95.2%) had a sensibility of 88.2% and a specificity of 92.4%; a LAX measurement value of 0.623 (AUC: 93.3%) had a sensibility of 82.7% and a specificity of 85.7%. The multivariate analysis showed no significant correlation between MAIVF thickness, age, and sex.
CONCLUSION: In healthy patients, MAIVF thickness should not exceed 0.600 cm. Above these values, the suspicion of a periaortic abscess formation should be raised. MAIVF increased thickness may be an early sign of perivalvular diffusion requiring an urgent endocarditis team evaluation. Copyright:

INTRODUCTION

The mitral-aortic intervalvular fibrosa (MAIVF) is a fibrous region connecting the anterior mitral leaflet and the aortic valve. Due to its avascular structure, it is the weakest part of the aortic root and more susceptible to infiltration by an infective process.[1] Periannular extension and abscess are frequent complications of infective endocarditis (IE), accounting for 9.8%–40% of native valve and 17%–89% of prosthetic valve endocarditis in surgical series and 86% in autopsy series.[23456] Diagnosis can be challenging even with transesophageal echocardiography (TEE) imaging with a sensitivity of 80%–87%.[7] The periaortic infective diffusion is a dynamic process in which the inflammation of the deep tissue causes, in the first stage, a MAIVF thickening, eventually progressing with the formation of an abscess and afterward with a pseudoaneurysm[1] [Figure 1]. Such complications are the expression of a locally uncontrolled infection resistant to antibiotic therapy and with a poor prognosis. This is why current guidelines recommend urgent surgical treatment (Class I, level of evidence B).[8] The proximity to the left atrium and aorta can result in fistula formation within these structures. In the first stages of the abscess formation, the only sign is an increased thickness of the MAIVF. To the best of our knowledge, normal TEE reference values for MAIVF thickness have not yet been established. TEE is much more sensitive than transthoracic echocardiography (TTE) allowing better image and anatomical definition.[7] Methodological limitations of current echocardiographic measure nomograms have been underlined in the literature.[910] Several adult and pediatric echocardiographic nomograms for cardiac chamber quantifications have been published allowing normal value estimates of ventricles, atria, valves, pulmonary arteries, and aorta.[91011] However, nomograms describing the normal values for the MAIVF thickness have never been evaluated. The aim of our work was to establish echocardiographic normal values of MAIVF thickness estimated by TEE in a population of healthy adults to detect the early sign of perivalvular infective diffusion in patients with the suspicious of IE.

Figure 1

Graphical representation of the normal and pathological mitral-aortic intervalvular fibrosa; (a and b) normal thickening; (c and d) pathological thickening; (e and f) infective endocarditis abscess (blue arrow); (g) pseudoaneurysm (red arrow)

MATERIALS AND METHODS

Inclusion criteria

Subject enrollment

Healthy Caucasian adults, both hospital and outclinic patients, referred to the TEE laboratory of our department were included in the study. A complete TEE two-dimensional, color Doppler flow examination was performed. Videos were digitally stored and subsequently analyzed. For every participant, to avoid the collection of ambiguous images/movies, ≥1 clip for each echocardiographic projection was recorded. Among these patients, those who were candidates to a cardiac computed tomography (CT) scan underwent a MAIVF measurement evaluation, which was similar to the TEE evaluation (the same planes and position). Measurements were averaged, and standard deviations were calculated. The two measurement methodologies were compared.

Finally, a group of 21 patients with a definite diagnosis of aortic endocarditis with MAIVF abscess or pseudoaneurysm was evaluated to compare their MAIVF thickness to the cutoff values derived by the healthy population. With these values, receiver operating characteristic (ROC) curves were constructed.

Transesophageal ecocardiography examination

Echocardiographic studies were performed using a Philips Epiq 5 echocardiograph (Philips Medical Systems, Bothell, WA, USA). The TEE studies were performed by two different well-trained operators who had been certified by the Italian Society of Echocardiography and Cardiovascular Imaging with a Level III competence (in accordance with the American Society of Echocardiography). Offline measurements were carried out with manual calibration. The two-dimensional measurements were calculated according to recent guidelines.[111213] MAIVF measurements were obtained with the inner–inner method at end-diastole by TEE, with standardized landmarks as displayed in Figure 2. Care was taken not to measure on the commissural points to avoid the risk of overestimation. Measurements were performed only in case of availability of good and definite images. Measurement modalities are reported in Table 1.

Figure 2

Upper panels: basal short-axis view on the left (panel a) and long-axis view on the right (b panel) on transesophageal echocardiography examination of a normal individual. Lower panels: Computed tomography scan of a normal individual, basal short-axis view (c, on the left) and long-axis view (d, on the right). The figures are displayed with the aortic valve in closed position. The white arrows show where and how to assess the mitral-aortic fibrosa thickness. Ao = Aortic valve, LA = Left atrium, LV = Left ventricle, MV = Mitral valve, RA = Right atrium

Table 1

Description of two-dimensional and m-mode echocardiographic measurements

Region of interest	View	Measurement method
Interventricular septal end-diastolic thickness	Long-axis M-mode TTE	Point-to-point measurements
Left ventricular wall end-diastolic thickness	Long-axis M-mode TTE	Point-to-point measurements
Left ventricular end-diastolic volume	Apical four-chamber view TTE	Planimetric measurements with manual tracing of the endocardial border
Left ventricular ejection fraction	Apical four-chamber view TTE	Simpson derived measurement
Sinuses of Valsalva	135° mid-esophageal 2D TEE	Point-to-point measurements
Junction	135° mid-esophageal 2D TEE	Point-to-point measurements
Ascending aorta	135° mid-esophageal 2D TEE	Point-to-point measurements
MAIVF-LAX view	120°-140° mid-esophageal 2D TEE	Point-to-point measurements during diastole
MAIVF-SAX view	25°-45° mid-esophageal 2D TEE	Point-to-point measurements during diastole

TTE=Transthoracic echocardiography, TEE=Transesophageal echocardiography, MAIVF=Mitral-aortic intervalvular fibrosa, LAX=Longitudinal axis, SAX=Short axis, 2D=Two-dimensional

Computed tomography examination

Cardio-CT examinations were performed with a 64-slice scanner (Light Speed VCT, General Electric Healthcare, Milwaukee, USA) with electrocardiography (ECG) gating. A bolus of 90 ml of contrast medium was injected intravenously at a rate of 5 ml/s, followed by 30 ml of normal saline. When ECG-triggered scanning protocol was performed, the following parameters were used: collimation width, 40 mm; slice thickness, 0.5 mm; rotation time, 0.4 s; tube voltage, 120 kV; maximum effective tube current, 890 mA; and table feed, 0.4 mm/rotation. CT data sets were transferred to a dedicated workstation for image analysis, and the perivalvular thickness was manually traced on multiplanar reconstruction [Figure 2].

Statistical methods

Continuous variables are reported as mean and 2 standard deviations (2SDs), in case of normally distributed variables, or as median and quartiles, in case of nonnormally distributed variables. Normality of distribution was assessed using the Kolmogorov–Smirnov test. Continuous variables between age groups were compared by t-test, Mann–Whitney U–test, or analysis of variance (ANOVA) as appropriate. In the univariate analysis, variables found to be significant (P <0.05) were included in multivariate linear regression analysis. MAIVF in short-axis (SAX) and long-axis (LAX) views was correlated with either Pearson's or Spearman's coefficients. Measurements were repeated five times and then averaged. Inter- and intra-observer variabilities were calculated by overall agreement and were tested using repeated-measures ANOVA test in all participants and with the Bland–Altman plot. Body mass index (BMI) was calculated using the Haycock formula. The accuracy of the upper normality value was tested by plotting the ROC curve. SPSS version 21.0 (SPSS, Inc., Chicago, IL, USA) and Stata version 10 for IOS (StataCorp LP, College Station, TX, USA) software were used for analysis.

RESULTS

A total of 576 patients were evaluated. During the enrolment phase, 170 patients were excluded (29.5%) due to IE in 21 (3.7%), severe aortic stenosis in 64 (11.1%), and the presence of normal aortic prosthesis in 85 (14.7%, biological in 49 and mechanical in 36). Therefore, 477 patients with a normal, and not infected, native aortic valve were enrolled. The mean age of the study population was 69.3 ± 15.5 years (range: 27–93 years). Male/female ratio was 1.3. Clinical, demographic, and echocardiographic characteristics of the patients are reported in Table 2. Indications for the TEE study were atrial fibrillation in 242 (51%), native mitral valvular disease in 134 (28.4%), research of embolic sources in 84 (17%), intracavitary mass in 10 (2%), or congenital heart disease in 7 (1.6%) patients. The mean body surface area was 1.86 ± 0.21 m2. MAIVF mean measurements (±2SD) were 0.325 cm (95% confidence interval [CI]: 0.319–0.330 cm) in SAX view and 0.340 cm (95% CI: 0.334–0.346 cm) in LAX view. SAX and LAX measurement percentile values for MAIVF are shown in Table 3. No significant differences were present in MAIVF thickness in males versus females in both views [Table 4]. The upper limit of normality was considered a value above the 95th percentile (>0.425 cm for SAX view and >0.475 cm for LAX view).

Table 2

Clinical, demographic, and echocardiographic characteristics of the patients

	Mean±2SD
Population	576
Population enrolled (%)	406 (70.5)
Age (years)	69.3±15.5
Sex (%)
Male	231 (56.9)
Female	175 (43.1)
BMI (Haycock formula), cm2	1.86±0.21
Interventricular septal end-diastolic thickness, mm	11.35±2.25
Left ventricular wall end-diastolic thickness, mm	10.36±3.03
Sinuses of Valsalva, mm	34.6±5.32
Aortic junction, mm	27.69±4.73
Ascending aorta, mm	34.41±5.34
Bicuspid AV (%)	11 (2.7)
Left ventricular end-diastolic volume, ml	113.11±51.90
Left ventricular ejection fraction (%)	58.15±11.90

BMI=Body mass index, SD=Standard deviation, AV=Aortic valve

Table 3

Mitral-aortic intervalvular fibrosa thickness measurement percentiles

	SAX (mean±2SD)	LAX (mean±2SD)
Sex group, cm
Male	0.328 (0.321-0.335)*	0.339 (0.313-0.329)**
Female	0.321 (0.332-0.347)*	0.342 (0.333-0.351)**
Age groups (years), cm
Group 1 (<50)	0.307 (0.299-0.316)	0.319 (0.310-0.329)
Group 2 (51-70)	0.326 (0.317-0.335)	0.342 (0.333-0.352)
Group 3 (>70)	0.344 (0.335-0.353)	0.367 (0.353-0.372)
Total	0.325 (0.319-0.330)	0.340 (0.334-0.346)
Percentiles (years), cm
Group 1 (<50)
Percentile 5	0.181	0.210
Percentile 10	0.230	0.250
Percentile 90	0.371	0.393
Percentile 95	0.385	0.448
Group 2 (51-70)
Percentile 5	0.227	0.248
Percentile 10	0.245	0.275
Percentile 90	0.386	0.422
Percentile 95	0.419	0.440
Group 3 (>70)
Percentile 5	0.266	0.256
Percentile 10	0.279	0.280
Percentile 90	0.409	0.445
Percentile 95	0.425	0.475
Men
Percentile 5	0.232	0.215
Percentile 10	0.258	0.262
Percentile 90	0.400	0.418
Percentile 95	0.419	0.447
Female
Percentile 5	0.222	0.248
Percentile 10	0.244	0.257
Percentile 90	0.381	0.431
Percentile 95	0.389	0.461
All
Percentile 5	0.222	0.248
Percentile 10	0.255	0.262
Percentile 90	0.391	0.421
Percentile 95	0.410	0.450

*P=0.011, **P=0.84. MAIVF=Mitral-aortic intervalvular fibrosa, LAX=Long axis, SAX=Short axis, SD=Standard deviation

Table 4

Mitral-aortic intervalvular fibrosa measurements in sex groups in both views

Mean values in sex age groups among normal native valve patients

Predictors of mitral-aortic intervalvular fibrosa thickness

There was a positive, although weak, correlation between MAIVF in SAX view and age (r=0.290, P < 0.001), sinuses of Valsalva (r=0.194, P < 0.001) and ascending aorta (r=0.180, P < 0.001), on the contrary no correlation was found between MAIVF and BMI (r=0,113, P = 0.065). A positive correlation was also present between MAIVF in LAX view and age.

The population was divided into three age groups (Group 1,<50 years; Group2, 51–70 years; and Group 3,>70 years. One hundred and sixty-nine patients were enrolled in the first group (35.5%), 157 in the second group (33.0%), and 151 in the third group(31.5%). An increasing thickness was present in older patients, as shown for both views by the univariate analysis (P <0.001) [Table 5]. To evaluate the influence of BMI and left ventricular end-diastolic volume on measurements, we used a multiple linear regression model including both of them as covariates. We did not find significant correlations in measurements.

Table 5

Mitral-aortic intervalvular fibrosa measurements in both views (short-axis and long-axis view)

Mean values in sex age groups among normal native valve patients

Computed tomography versus transesophageal echocardiography measurement

The patients included in the CT scan analysis were 96. The mean age of this subgroup population was 64.7 ± 13.2 years (range: 24–82 years). Male/female ratio was 1.4. CT-MAIVF mean measurement (±2SD) was 0.237 cm (95% CI: 0.110–0.340 cm) in single view. No significant differences were present in MAIVF thickness in males versus females [Table 4]. Compared with the echocardiographic measurements, CT thickness was significantly lower (P <0.001).

Comparison between normal mitral-aortic intervalvular fibrosa and abscess/pseudoaneurysm thickness

Clinical, demographic, and echocardiographic characteristics of 21 patients with MAIVF abscess are reported in Table 6 and Figure 3, Videos 1 and 2. The results showed that MAIVF thickness in the setting of IE had increased and was always above normal values [Table 7]. This was true even for patients affected by severe aortic stenosis and biological or mechanical prosthetic valves (P <0.001) and in both echographic views [Figure 4]. ROC curves showed that the best cutoff point was a SAX measurement value of 0.552 (area under the curve [AUC]: 89.9%) had a sensitivity of 82.7% and a specificity of 84.7%%; a LAX measurement value of 0.623 cm (AUC 86.4%) had a sensitivity of 88.2% and a specificity of 73.2% [Table 8].

Table 6

Clinical, demographic, and echocardiographic characteristics of the patients with mitral-aortic intervalvular fibrosa abscess

Characteristics	Values
Number of cases	21
Mean age (years)	61.65±15.59
Female sex	7/21
MAIVF in SAX view, cm	0.947±0.270
MAIVF in LAX view, cm	0.877±0.252
Native AV	9/21
Aortic mechanical prosthetic valve	2/21
Aortic biological prosthetic valve	10/21
Septic shock	8/21
Emergent/urgent surgery	10/21
Death	9/21

MAIVF=Mitral-aortic intervalvular fibrosa, SAX=Short axis, LAX=Long axis, AV=Aortic valve

Figure 3

Basal short-axis view (a) and long-axis view (b – still frame from Videos 1) on transesophageal echocardiography examination of a patient with persistent fever and positive blood cultures. An aortic valve replacement with mechanical prosthesis had been performed 5 years ago. The yellow arrows show a pathological mitral-aortic fibrosa thickness. There were no other pathological findings. (c and D) (still frame from Video 2) show the same patient 7 days after the pathological mitral-aortic fibrosa thickness degenerated in a pseudoaneurysm (d)

Table 7

Comparison of mitral-aortic intervalvular fibrosa thickness values in all groups

	n	Male/female	Age (years)	MAIVF-SAX (cm)	MAIVF-LAX (cm)
Normal valve	406	231/175	60.9±15.5	0.325 (0.319-0.330)	0.340 (0.334-0.346)
MAIVF-endocarditis	21	7/21	61.65±15.59	0.947 (0.600-1.610)	0.877 (0.567-1.512)
Moderate/severe aortic stenosis	64	16/48	68.5±2.07	0.553 (0.530-0.575)	0.681 (0.679-0.684)
Biological prosthetic valve	49	17/32	70.7±10.89	0.632 (0.594-0.670	0.692 (0.656-0.728)
Mechanical prosthetic valve	36	24/12	75.00±3.79	0.597 (0.552-0.642)	0.648 (0.584-0.711)

P<0.001. MAIVF=Mitral-aortic intervalvular fibrosa, SAX=Short axis, LAX=Long axis

Figure 4

Comparison of mitral-aortic intervalvular fibrosa thickness in different groups: Normal patients and infective endocarditis patients with mitral-aortic intervalvular fibrosa abscess or pseudoaneurysm, aortic stenosis, and normal biological and mechanical prosthetic aortic valve. The increased thickness in patients affected by a periaortic valve diffusion of the infective process is evident

Table 8

Receiver operating characteristic curves

ROC curves showed that the best cut-off point was a SAX measurement value of 0.552 cm (AUC: 89.9%, sensitivity 82.7% and specificity 84.7%), and a LAX measurement value of 0.623 cm (AUC: 86.4%, sensitivity 88.2% and specificity 73.2%)

Inter- and intra-observer variability

The intra- and inter-observer agreement was high. Intra-observer variability of MAIVF-SAX measurements showed an intraclass correlation coefficient (ICC) value of 0.929 (95% CI: 0.897–0.951). The intra-observer variability ICC value calculated for MAIVF-LAX was 0.972 (95% CI: 0.894–0.92). The calculated ICC value for inter-observer variability was 0.97 (95% CI: 0.970–0.987) for MAIVF-SAX and 0.949 (95% CI: 0.911–0.970) for MAIVF-LAX measurements.

Bland–Altman plot showed that MAIVF-LAX and MAIVF-SAX measurements had a small but significant inter- and intra-observer variability for both views (SAX: P=0.001: LAX: P=0.009) [Tables 9 and 10], but it should be noted that reporting mean difference (or bias by Bland–Altman analysis) between repeated measurements as a measure of reproducibility can be misleading because the bias will always be zero in a sufficiently large sample, since in each pair of repeated measurements, one will be larger than the other in no particular order.

Table 9

Bland-Altman t-test

	t	Df	Significant (2-codes)	Mean difference	95% CI
					Inferior	Superior
MAIVF-SAX	−3.651	405	0.000	−0.002938	−0.00452	−0.00136
MAIVF-LAX	−2.642	405	0.009	−0.001067	−0.00186	−0.00027

MAIVF=Mitral-aortic intervalvular fibrosa, SAX=Short axis, LAX=Long axis, CI=Confidence interval

Table 10

Bland-Altman plot

DISCUSSION

The MAIVF is a thin membranous structure that forms the junction between the left half of the noncoronaric cusp, the adjacent third of the left coronary cusp of the aortic valve, and the anterior mitral leaflet.[14] The MAIVF roof is composed of the pericardium, and its ventricular side forms the posterior portion of the left ventricular outflow tract.[14] MAIVF is made up of a relatively avascular tissue and is prone to the formation of an abscess, a pseudoaneurysm, or a fistula. These life-threatening complications begin with an increased thickness of the fibrosa. Abscesses are nonpulsatile cavities without visible flow on the color Doppler imaging. MAIVF pseudoaneurysms, which have been first described by Waldhausen et al.,[15] most frequently occur in patients with a prosthetic aortic valve[161718] but can also be the consequence of a chest trauma, Takayasu's arteritis,[19] or congenital abnormalities.[20] TEE is more sensitive than TTE to detect the increased MAIVF thickness, abscess, pseudoaneurysm, and fistulae.[212223] An early recognition of the thickening is of crucial importance, because, as already said, urgent surgical correction is recommended. Our study is the first to provide reference normal values to define an “increased” MAIVF thickness. An abnormal thickness may be defined as >0.425 cm in SAX view and >0.475 cm in LAX view. An upper value in SAX measurement of 0.552 and in LAX of 0.623 was the most accurate at the ROC curve analysis. On the basis of these data, we believe that an echographic cutoff value of 0.600 cm could be proposed. above this value, a perivalvular diffusion of the infective process may be suspected. In the subgroup of patients who underwent a CT scan, the mean MAIVF thickness was 0.237 cm (95% CI: 0.110–0.340 cm). The upper second value of 0.340 is similar to the cutoff value proposed by de Kerchove et al.[24] who found a maximum aortic root wall thickness at surgery of 0.360 cm at the level of the left coronary sinus in 58 human donors. In a CT study, performed in a population of patients affected by prosthetic valve endocarditis, Fagman et al.[25] proposed a cutoff value of 0.5 cm (sensitivity 67% and specificity 95%). In our study, CT thickness was significantly lower than the echographic one, as a result of the higher CT spatial resolution. Multimodality imaging and a diagnostic algorithm have been already underlined by current guidelines,[262728] and other studies comparing multislice CT and TEE have never assessed this topic.

Limitations

Our study has some limitations. The normal values were derived from TEE images, and therefore, they cannot be used for transthoracic studies. Non-Caucasian participants were excluded to avoid racial variability bias; therefore, the results of our study only apply to Caucasians. We prospectively enrolled a homogeneous cohort of healthy adults; therefore, these measurements could not be necessarily used in prosthetic valve patients, and this topic needs to be evaluated in further studies. No patient has been evaluated within the first 3 months after an aortic valve replacement to avoid overmeasurement of the MAIVF thickness due to edema; therefore, our results could not apply to this population. Another limitation is the lack of data from participants of different ethnic origins. It is important to note that even slight differences in view angle may result in substantial differences in diameter quantification. As a consequence, to obtain good reproducibility of chamber dimensions, we recommend storing several images and performing measurements only in unambiguous views. The Bland–Altman plot showed a difference in inter-observer variability. However, since this is a very small measurement difference in the order of <½ mm, this can be regarded as clinically insignificant. Finally, CT measurements of thickness in patients affected by prosthetic valve IE were not available; this is why further studies assessing this issue are necessary.

CONCLUSION

Our study shows that MAIVF normally is a thin wall that should not exceed 0.600 cm. In patients with a possible or definite diagnosis of IE, MAIVF thickening should be assumed as an early sign of perivalvular diffusion. If this is the case, an endocarditis team evaluation should be promptly performed as urgent surgery may be necessary. The nomograms provided in the present report may be considered a helpful objective tool for clinicians to perform quantitative MAIVF measurements in adults with various acquired and congenital heart diseases. In particular, this work covers the gap of knowledge on MAIVF thickness in normal adults. Further studies, including a multimodality imaging evaluation in patients with a prosthetic aortic valve, are required.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.