Background
Recent studies have emphasized the importance of quantitative assessment of the degree of aortic regurgitation (AR). However, semiquantitative methods have remained mainly used despite their unclear diagnostic value. The aim of this study was to define the sensitivity and specificity of semiquantitative methods compared with the proximal isovelocity surface area method as a reference for the diagnostic of severe AR.
Methods
The degree of AR was evaluated using the proximal isovelocity surface area method and four semiquantitative measurements (left ventricular cardiac output, pressure half-time, diastolic flow reversal, and vena contracta) in 224 patients with a wide range of AR severity.
Results
The mean effective regurgitant orifice area was 25 ± 14 mm 2 (range, 3–69 mm 2 ), the mean regurgitant volume was 57 ± 31 mL (range, 9–183 mL), and 100 patients (44%) had severe AR (effective regurgitant orifice area ≥ 30 mm 2 or regurgitant volume ≥ 60 mL). Overall, semiquantitative methods had good specificity but poor sensitivity, except the vena contracta, which had good sensitivity and specificity. Sensitivity, specificity, and positive and negative predictive values of the recommended thresholds for severe AR of the four semiquantitative methods were 53%, 89%, 77%, and 73% for left ventricular cardiac output ≥ 10 L/min; 12%, 100%, 100%, and 52% for pressure half-time < 200 msec; 45%, 87%, 79%, and 60% for diastolic flow reversal ≥ 18 cm/sec; and 81%, 83%, 78%, and 85% for vena contracta ≥ 6 mm, respectively.
Conclusions
For the assessment of AR severity, current thresholds appear specific but poorly sensitive, except for vena contracta, which provides good discriminative value. Semiquantitative methods should be integrated into the comprehensive evaluation of AR severity, but severe AR should not be excluded only on the basis of semiquantitative criteria. These results emphasize the need for the quantitative assessment of AR severity.
Assessment of the degree of regurgitation is paramount to clinical decision making in patients with aortic regurgitation (AR). Patients with severe AR and either symptoms or left ventricular (LV) dysfunction should be promptly referred to surgery. Doppler echocardiography has become the mainstay of the assessment of AR severity. Pioneering studies have shown good correlations between semiquantitative grading using color or Doppler echocardiography and angiography. More recently, quantitative echocardiography with measurements of effective regurgitant orifice (ERO) area, a measure of lesion severity, and of regurgitant volume (RVOL), a measure of LV volume overload, has been developed and validated. The proximal isovelocity surface area (PISA) or flow convergence method is the simplest, most widely used, and most reproducible quantitative method. Of importance, the prognostic value of quantitative measurements has been shown to supersede that of semiquantitative methods.
However, in clinical practice, the evaluation of AR severity has remained often based only on semiquantitative methods, despite their unclear diagnostic value and the recommendations of the European Association of Echocardiography and American Society of Echocardiography to also consider quantitative assessment. Thus, the aim of the present study was to evaluate the sensitivity and specificity of four semiquantitative methods—LV cardiac output (LVCO), the pressure half-time (PHT), diastolic flow reversal (DFR), and the vena contracta (VC)—for the diagnostic of severe AR compared with the PISA method as a reference.
Methods
Study Population
We retrospectively included patients with at least mild AR on the basis of color flow Doppler who underwent assessments of AR degree using both semiquantitative methods (LVCO, PHT, VC, and DFR) and the PISA method at two centers, the Mayo Clinic (Rochester, MN) and Bichat Hospital (Paris, France). Exclusion criteria were not feasible or not attempted PISA. Patients with associated valvular diseases were not excluded. All echocardiographic studies were clinically indicated. All data were collected from the echocardiographic reports. The study was approved by institutional review boards.
Doppler Echocardiography
Comprehensive Doppler echocardiography was performed. Semiquantitative assessment of AR severity was based on four methods. LVCO was calculated as follows: LVCO = Π/4 × LVOT 2 × LVOT TVI × heart rate, where LVOT is the LV outflow tract diameter measured at the insertion of the leaflets in midsystole from the parasternal long-axis view, and LVOT TVI is the LVOT time-velocity integral recorded using pulsed Doppler just proximal to the valve orifice ( Figures 1 A and 1 B). The PHT was measured using continuous-wave Doppler ( Figure 1 C). The Doppler ultrasound beam should be carefully aligned with the AR jet as visualized in color Doppler in the most appropriate view (the apical view in central jets and the parasternal long-axis view in eccentric jets). DFR was assessed in the upper descending aorta at the aortic isthmus level using a suprasternal view by pulsed Doppler. Measurements were performed in end-diastole ( Figure 1 D). The VC, defined as narrowest portion of the regurgitant jet downstream from the regurgitant orifice, was measured perpendicular to the jet, from the parasternal long-axis view in central jets ( Figure 1 E) and from the apical view in eccentric jets. Recommended semiquantitative thresholds for severe AR are LVCO ≥ 10 L/min, PHT < 200 msec, DFR ≥ 18 cm/sec, and VC > 6 mm.
The quantitative assessment of AR severity was based on the PISA method. The PISA method is based on the principle of conservation of mass and is presented elsewhere. Briefly, Doppler color flow images of the flow convergence were obtained from the apical view in central jets and from the parasternal long-axis view in eccentric jets to align the flow convergence and the ultrasound beam. A large zoom of the flow convergence was obtained and the baseline velocity shifted in the direction of the regurgitant flow until the flow convergence region was clearly visualized, with an optimal hemispheric shape. The radius of the flow convergence was measured between the first aliasing contour and the aortic regurgitant orifice in early diastole at the same time as the peak regurgitant velocity ( Figure 2 ). In a subset of patients, AR was also quantified on the basis of LV volume (biplane method of disks, quantitative two-dimensional echocardiography) or aortic stroke volume and mitral stroke volume measurements (quantitative Doppler). Severe AR was defined as an ERO area ≥ 30 mm 2 or an RVOL ≥ 60 mL. The number of aortic cusps (bicuspid or tricuspid aortic valve) as well as associated valvular diseases were recorded. Ejection fraction was assessed either visually or using the biplane Simpson’s method. Severe LV enlargement was defined as an LV end-diastolic diameter ≥ 62 mm. Systolic pulmonary artery pressure was measured from the tricuspid regurgitation using continuous-wave Doppler.
Statistical Analysis
Data are presented as mean ± SD or as numbers and percentages. Comparisons between semiquantitative measurements and ERO area or RVOL were performed using linear regressions. The diagnostic value of each semiquantitative criterion for the diagnosis of severe AR as assessed by the PISA method (ERO area ≥ 30 mm 2 or RVOL ≥ 60 mL) was analyzed. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were determined for various thresholds, and the area under the curve (AUC) on receiver operating characteristic analysis was calculated. Analyses were performed in the entire population and in subgroups according to valve anatomy (bicuspid or tricuspid valve), associated diseases, and quantitative measurement performed on the basis of methods other than the PISA method. P values <.05 were considered statistically significant.
Results
Baseline Characteristics
Two hundred twenty-four patients with at least mild AR were enrolled in the present study. The mean age was 60 ± 17 years, 64% were men, and 93% were in sinus rhythm. The aortic valve was tricuspid in 179 patients (80%) and bicuspid in 46 (20%). AR was pure in 185 patients (82%) and with associated disease in the remaining 40 patients (aortic stenosis or mitral valve disease). The mean ejection fraction was 60 ± 9%, and depressed ejection fractions (<55%) were observed in 26% of patients. On the basis of the PISA method, the mean ERO area was 25 ± 14 mm 2 (median, 23 mm 2 ; range, 3–69 mm 2 ), the mean RVOL was 57 ± 31 mL (median, 53 mL; range, 9–183 mL), and 100 patients (44%) presented with severe AR (ERO area ≥ 30 mm 2 or RVOL ≥ 60 mL). LV end-diastolic diameters were larger in patients with severe AR than in patients with nonsevere AR (62 ± 8 vs 54 ± 6 mm, P < .0001), but severe LV enlargement was observed in only 51% of patients with severe AR.
Diagnostic Value of Semiquantitative Measurements
LVCO
Mean LVCO, measured in 199 patients, was 8.4 ± 2.6 L/min (median, 8.0 L/min; range, 4.0–19.7 L/min) and was ≥10 L/min in 29%. Correlations between LVCO and ERO area ( Figure 3 A) and between LVCO and RVOL were good ( r = 0.57, P < .0001, and r = 0.49, P < .0001, respectively). The analysis of the diagnostic value of LVCO for severe AR, presented in Table 1 , was good overall (AUC, 0.80), and a value of 9.2 L/min provided the best sum of sensitivity and specificity ( Figure 4 ). However, although LVCO ≥ 10 L/min provided 89% specificity and 77% PPV, although sensitivity was low (53%). Use of LV cardiac index instead of LVCO did not improve accuracy (AUC, 0.78). The sensitivity, specificity, PPV, and NPV of a threshold value of LVCI of 4.2 L/min/m 2 of body surface area that provided the best sum of sensitivity and specificity were 79%, 68%, 64%, and 82%, respectively.
| LVCO (L/min) | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) |
|---|---|---|---|---|
| 6 | 95 | 26 | 48 | 88 |
| 7 | 90 | 53 | 58 | 88 |
| 8 | 76 | 69 | 64 | 80 |
| 9 | 67 | 83 | 74 | 78 |
| 9.2 ∗ | 65 | 87 | 78 | 78 |
| 10 | 53 | 89 | 77 | 73 |
| 11 | 34 | 95 | 82 | 67 |
| 12 | 16 | 96 | 72 | 61 |
∗ The value providing the best sum of sensitivity and specificity.
PHT
Mean PHT, measured in 96 patients, was 484 ± 193 msec (median, 451 msec; range, 121–1,075 msec). The correlation between PHT and ERO area ( Figure 3 B) was good ( r = 0.53, P < .0001), but the correlation with RVOL was poor ( r = 0.22, P = .03). A value of 443 msec provided the best sum of sensitivity and specificity ( Table 2 ). A PHT < 200 msec provided excellent 100% specificity but poor sensitivity (12%). The AUC on receiver operating characteristic analysis for the diagnosis of severe AR was 0.73 ( Figure 4 ).
| PHT (msec) | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) |
|---|---|---|---|---|
| 200 | 12 | 100 | 100 | 52 |
| 300 | 24 | 96 | 86 | 55 |
| 400 | 53 | 83 | 76 | 63 |
| 443 ∗ | 69 | 74 | 74 | 70 |
| 500 | 78 | 60 | 67 | 72 |
| 600 | 86 | 36 | 58 | 71 |
∗ The value providing the best sum of sensitivity and specificity.
DFR
DFR velocity was measured in 142 patients (mean, 16 ± 7 cm/sec; median, 15 cm/sec; range, 5–54 cm/sec). Correlations between DFR and ERO ( Figure 3 C) and between DFR and RVOL were good ( r = 0.55, P < .0001, and r = 0.43, P < .0001). The AUC on receiver operating characteristic analysis for the diagnosis of severe AR was 0.77 ( Figure 4 ), and a value of 15.6 cm/sec provided the best sum of sensitivity and specificity ( Table 3 ). A DFR ≥ 18 cm/sec provided 87% specificity and 79% PPV but low sensitivity (45%). The use of a threshold of 20 cm/sec improved specificity (96%) and PPV (90%) but lowered sensitivity (36%) and NPV (58%).
| DFR (cm/sec) | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) |
|---|---|---|---|---|
| 10 | 99 | 25 | 58 | 94 |
| 12 | 93 | 41 | 62 | 85 |
| 14 | 75 | 52 | 63 | 67 |
| 15.6 ∗ | 68 | 71 | 71 | 68 |
| 16 | 67 | 71 | 71 | 67 |
| 18 | 45 | 87 | 79 | 60 |
| 20 | 36 | 96 | 90 | 58 |
| 22 | 26 | 97 | 90 | 55 |
∗ The value providing the best sum of sensitivity and specificity.
VC
Mean VC, measured in 173 patients, was 5.4 ± 1.8 mm (median, 5.5 mm; range, 2.0–10.0 mm). Close correlations were found between VC and ERO ( r = 0.74, P < .0001; Figure 3 D) and between VC and RVOL ( r = 0.68, P < .0001). The analysis of the diagnostic value of VC for severe AR is presented in Table 4 (AUC, 0.89; Figure 4 ). A value of 6 mm provided the best sum of sensitivity (81%) and specificity (83%). The 5-mm threshold provided high sensitivity and NPV, whereas the 7-mm threshold provided high specificity and PPV.
| VC (mm) | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) |
|---|---|---|---|---|
| 3 | 100 | 13 | 47 | 100 |
| 4 | 97 | 32 | 52 | 94 |
| 5 | 93 | 61 | 65 | 92 |
| 6 ∗ | 81 | 83 | 78 | 85 |
| 7 | 51 | 96 | 90 | 72 |
| 8 | 23 | 99 | 94 | 63 |
| 9 | 8 | 100 | 100 | 59 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
