Methods

From December 21, 2004 to May 27, 2010, 687 operatively excised stenotic aortic valves (± AR) in patients having isolated AVR (± coronary bypass) without associated mitral stenosis or mitral valve replacement were received in the surgical pathology department at Baylor University Medical Center (BUMC). The valve was weighed and its structure was determined by W.C.R. in all cases during the week that the valve was excised. Any patient having replacement of a cardiac valve other than the aortic valve was excluded from the study.

Patients’ isolated aortic valve disease was confirmed by examination of their preoperative cardiac catheterization and/or echocardiographic reports, which were available on the Baylor Physician Portal and/or cardiovascular computer database (Apollo Advance C/S 4.2.13, LUMEDX Corp., Oakland, California). Patients were considered to have AS if the peak systolic transvalvular gradient was >10 mm Hg. Aortic valve gradient and valve area were obtained from the cardiac catheterization report.

The December 2004 date was chosen because that was the time the echocardiograms were available on the Cardiology Consultants of Texas computerized database and/or on the Baylor Physician Portal (Baylor electronic medical record).

Of the 687 patients, 100 were >50 years of age at the time of AVR and had preoperative echocardiograms available on the Cardiology Consultants of Texas computerized database and/or on the Baylor Physician Portal. Echocardiograms were recorded from 2 to 1,460 days (mean 89, median 36) before AVR. Of 100 echocardiograms, 80 were recorded <120 days before AVR. At least 1 parasternal long-axis and 1 parasternal short-axis view of the aortic valve were available in each patient.

Echocardiograms were interpreted by a staff cardiologist (P.A.G.) who was blinded to the surgical, pathological, and echocardiographic reports. His interpretations were the primary focus of our analysis. Echocardiograms were also independently reviewed and interpreted by a second cardiologist (R.F.A.) who is a fellow in training (third year) and these interpretations were used to measure interobserver agreement.

Aortic valve structure was interpreted as unicuspid, bicuspid, tricuspid, or indeterminate. The latter was because of heavy calcium and/or because of poor image quality. A valve was considered bicuspid if 2 of 3 echocardiographic criteria were present: (1) elliptical or slitlike orifice in ventricular systole in the parasternal short-axis view, (2) cuspal doming in ventricular systole, or (3) an asymmetrical closure line in the parasternal long-axis view. There were no prespecified echocardiographic criteria for unicuspid or tricuspid valves.

We also reviewed echocardiographic reports made at the time the echocardiograms had originally been recorded. In addition, we reviewed the surgeons’ interpretations of the aortic valve structure from the operative reports. We used the echocardiographic report closest to the date of AVR.

Associations between various factors (demographic, clinical, and anatomic) and echocardiographic interpretations of aortic valve structure (uninterpretable, interpretable, concordant with morphologic interpretation, and discordant with morphologic interpretation) were tested with a Wilcoxon (for continuous factors) or a chi-square (for categorical factors) test. A Bonferroni correction was employed to account for multiplicity. Kappa tests, specificity, and sensitivity were used to assess agreement between echocardiographic and morphologic interpretations of aortic valve structure.

This study was approved by the institutional review board at BUMC .

Results

Table 1 presents demographics and pertinent cardiac findings in all patients, those with interpretable versus uninterpretable echocardiograms for aortic valve structure, and those with concordant versus discordant echocardiographic (P.A.G.) and morphologic (W.C.R.) interpretations. Patients’ age ranged from 51 to 91 years (mean 70 ± 10); 61 were men and 39 were women. Body mass index ranged from 19.7 to 47.1 kg/m ² (mean 29.0 ± 5.8, median 27.5). There were 3 unicuspid, 51 bicuspid, and 46 tricuspid valves as determined morphologically (W.C.R.).

Aortic valve weights ranged from 0.81 to 18.38 g (mean 3.03 ± 2.11). Peak systolic transvalvular gradients (by cardiac catheterization) ranged from 11 to 125 mm Hg (mean 49 ± 22). Aortic valve area (by cardiac catheterization) ranged from 0.25 to 1.69 cm ² (mean 0.81 ± 0.29). Sixteen valves had areas >1 cm ² but only 2 of these 16 valves had valve areas >1 cm ² when measured by echocardiogram. Of these 16 patients, peak systolic gradients ranged from 19 to 96 mm Hg (mean 40 ± 18), mean transvalvular gradients ranged from 14 to 77 mm Hg (mean 31 ± 15), and aortic valve weights ranged from 1.47 to 6.33 g (mean 3.22 ± 1.43, normal aortic valve weight ≤0.50 g).

Of the 100 patients, 86 had echocardiograms that were interpretable for aortic valve structure (P.A.G.). In 57 of these 86 patients (66%), echocardiographic interpretation was concordant with morphologic interpretation (kappa test = 0.33). Aortic valve structure in patients with interpretable compared to uninterpretable echocardiograms is shown in Figure 1 . Ten of the 14 patients (71%) who had echocardiograms that were uninterpretable (P.A.G.) had congenitally malformed valves ( Figure 1 ).

Each pie chart displays echocardiograms in which each echocardiographic reader (P.A.G. or R.F.A.) could (A) or could not (B) interpret aortic valve structure. In each chart, the morphologic interpretations (W.C.R.), not echocardiographic interpretations, of valve structure are shown. Most valves that were uninterpretable (P.A.G. or R.F.A.) were congenitally malformed valves (10/14 [71%] and 14/17 [82%], respectively). *Echocardiographic reader.

Graphical comparisons of echocardiographic interpretability and concordance of echocardiographic (P.A.G.) and morphologic (W.C.R.) interpretations of aortic valve structure in different groups according to valve weight, gender, age at time of AVR, body mass index, peak systolic gradient, and valve area are shown in Figures 2 to 5 . Unadjusted associations between these factors and echocardiographic interpretability and accuracy are presented in Table 1 .

In each weight group of men (A) and women (B) , the total number of patients (white bar) and the number of patients in whom the echocardiographic reader was able to interpret valve structure (concordant) [diagonal hatched bar] or discordant [horizontal hatched bar] with that of the morphologic reader [W.C.R.]) or unable to interpret valve structure (black bar) are shown. Numbers of tricuspid (T), bicuspid (B), and unicuspid (U) aortic valves (determined by W.C.R.) in each group are displayed. Valve weight was determined primarily by quantity of calcium in valve cusps. Frequency of congenitally malformed valves was higher (90% vs 44%) in patients with heavier valves (>4 g in men and >3 g in women) compared to lighter valves (<4 g in men and <3 g in women). Echocardiographic interpretability was lower (71% vs 90%) but echocardiographic accuracy was identical (66%) in patients with heavier valves compared to lighter valves. Men had heavier valves than women (3.76 ± 2.36 vs 1.86 ± 0.69 g). Echocardiographic interpretability was similar (88% vs 82%) but echocardiographic accuracy was higher (76% vs 48%) in men than women. *Echocardiographic reader (P.A.G.).

(A) In each age group, the total number of patients (white bar) and the number of patients in whom the echocardiographic reader was able to interpret valve structure (concordant [diagonal hatched bar] or discordant [horizontal hatched bar] with that of the morphologic reader [W.C.R.]) or unable to interpret valve structure (black bar) are shown. Numbers of tricuspid, bicuspid, and unicuspid aortic valves (determined by W.C.R.) in each group are displayed. Echocardiographic interpretability was higher (100% vs 84%) in patients >80 than in patients ≤80 years of age at time of aortic valve replacement. Frequency of tricuspid valves was higher in patients >80 years old (80%) than in patients ≤80 years (33%). Echocardiographic accuracy was similar in the 2 age groups (73% vs 65%). (B) In each age group, the number of heavily calcified valves (men with valves weighing >4 g [diagonal hatched bar] and women with valves weighing >3 g [black bar] ) are shown. Seven percent of patients >80 years old had heavily calcified valves versus 37% of patients ≤80 years. *Echocardiographic reader (P.A.G.). Abbreviations as in Figure 2 .

In each body mass index group, the total number of patients (white bar) and the number of patients in whom the echocardiographic reader was able to interpret valve structure (concordant [diagonal hatched bar] or discordant [horizontal hatched bar] with that of the morphologic reader [W.C.R.]) or unable to interpret valve structure (black bar) are shown. Numbers of tricuspid, bicuspid, and unicuspid aortic valves (determined by W.C.R.) in each group are displayed. Of the 12 patients with body mass index >35 kg/m ² , in only 5 (42%) was the echocardiogram interpretable and accurate; in contrast, of the 33 patients with body mass index ≤25 kg/m ² , in 17 (52%) the echocardiogram was interpretable and accurate. *Echocardiographic reader (P.A.G.). Abbreviations as in Figure 2 .

In each peak systolic gradient group (A) and valve area group (B) , the total number of patients (white bar) and the number of patients in whom the echocardiographic reader was able to interpret valve structure (concordant [diagonal hatched bar] or discordant [horizontal hatched bar] with that of the morphologic reader [W.C.R.]) or unable to interpret valve structure (black bar) are shown. Numbers of tricuspid, bicuspid, and unicuspid aortic valves (determined by W.C.R.) in each group are displayed. Echocardiographic interpretability and accuracy decreased slightly (86% vs 94% and 59% vs 68%, respectively) with higher peak systolic transvalvular gradient (>50 vs ≤50 mm Hg). In patients with valve area ≤0.7 compared to >0.7 cm ² , echocardiographic accuracy was lower (50% vs 80%), but interpretability was similar (86% vs 91%). *Echocardiographic reader (P.A.G.). Abbreviations as in Figure 2 .

At heavier valve weights (>4 g in men and >3 g in women), stenotic valve structure was less interpretable (71% vs 90%) but equally accurate (66%) by echocardiogram compared to lighter weights (<4 g in men and <3 g in women; Figure 2 ). Frequency of congenitally malformed valves in these heavier valves was also higher than in the lighter valves (90% vs 44%; Figure 2 ).

Men had higher mean valve weights (quantity of calcium) than women (3.76 ± 2.36 vs 1.86 ± 0.69 g; Figure 2 ). Aortic valve structure was similarly interpretable by echocardiography (82% vs 88%) but less accurately interpreted (48% vs 76%) in women than in men ( Figure 2 ).

Echocardiographic interpretability appeared higher in patients >80 years old (100% vs 84%) than in patients ≤80 years old at time of AVR ( Figure 3 ). Echocardiographic accuracy was similar in the 2 age groups (73% vs 65%; Figure 3 ). Older patients (>80 years old) had a lower frequency of heavily calcified valves (weight >4 g in men and >3 g in women) than younger patients (<80 years old, 7% vs 37%; Figure 3 ). Older patients also had a higher frequency of tricuspid valves than younger patients (80% vs 33%; Figure 3 ).

Overall, echocardiographic interpretability and accuracy did not change with body mass index ( Figure 4 ). Nevertheless, comparison of the largest and smallest body mass index groups showed differences. Of the 12 patients with body mass index >35 kg/m ² , in only 5 (42%) was the echocardiogram interpretable and accurate; in contrast, of the 33 patients with body mass index ≤25 kg/m ² , in 17 (52%) the echocardiogram was interpretable and accurate.

Echocardiographic interpretability and accuracy decreased slightly (86% vs 94% and 59% vs 68%, respectively) with higher peak systolic transvalvular gradient (>50 vs ≤50 mm Hg; Figure 5 ). In patients with valve area ≤0.7 compared to >0.7 cm ² , echocardiographic accuracy was lower (50% vs 80%), but interpretability was similar (86% vs 91%; Figure 5 ).

Comparisons of aortic valve structure determined by echocardiogram by the 2 readers to each other and to that determined by examination of the operatively excised stenotic valve are presented in Table 2 . Sensitivity and specificity of echocardiography (P.A.G.) were 69% and 67% for detection of a bicuspid and 67% and 69% for detection of a tricuspid aortic valve, respectively. Neither echocardiographic reader accurately identified any of the 3 unicuspid valves. Both readers accurately identified 28 (61%) of the 46 tricuspid valves. P.A.G. accurately identified 29, and the fellow in training (R.F.A.), 24 of the 51 bicuspid valves (57% vs 47%). P.A.G. and R.F.A. were concordant with morphologic findings in 69% and 63% of bicuspid valves, respectively, that were interpretable by echocardiography. Of tricuspid valves interpretable by echocardiography, their rates of concordance were 67% and 65%, respectively. In 62% of interpretable echocardiograms, the 2 readers agreed in their interpretation of aortic valve structure (kappa test = 0.20).

Table 2

Comparison of aortic valve structure determined by echocardiogram by the two readers to each other and to that determined by examination of operatively excised stenotic aortic valves (n = 100)

Echocardiographic Interpretation	Morphologic Interpretation ⁎
	Unicuspid (n = 3)			Bicuspid (n = 51)			Tricuspid (n = 46)
	P.A.G. †	R.F.A. †	P.A.G. = R.F.A. ‡	P.A.G. †	R.F.A. †	P.A.G. = R.F.A. ‡	P.A.G. †	R.F.A. †	P.A.G. = R.F.A. ‡
Bicuspid	1			29	24	15	14	15	6
Tricuspid	1	1	1	13	15	5	28	28	18
Indeterminate–heavily calcified	1	2	1	7	6	2	4	2
Indeterminate–poor image quality				2	6	1		1

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Pulmonary Function in Patients With Acute Coronary Syndrome Treated With Ticagrelor or Clopidogrel (from the Platelet Inhibition and Patient Outcomes [PLATO] Pulmonary Function Substudy)
2. Predictors of Response to Cardiac Resynchronization Therapy in Patients With a Non-Left Bundle Branch Block Morphology
3. Comparison of Coronary Microcirculation in Female Nurses After Day-Time Versus Night-Time Shifts
4. Threshold for the Upper Normal Limit of Indexed Epicardial Fat Volume: Derivation in a Healthy Population and Validation in an Outcome-Based Study

Stay updated, free articles. Join our Telegram channel

Join