3 Results
Patients’ backgrounds (n=81) are shown in Table 1 . Mean age was 79+ 10 year-old. Female patients (57%) were more than male patients (43%). Echocardiographic data were also shown in Table 1 . LV was not dilated, LVMI was increased due to AS, and LVEF and SVI was preserved because of inclusion criteria. AVAi by continuous equation was limited (0.63+ 0.25 cm2/m2) and peak V was increased (4.0+ 1.1 m/sec). Distribution of AS severity was as follows: 17 cases of mild AS, 19 cases of moderate AS, and 45 cases of severe AS.
Peak velocity of 3 approaches: Mean values of peak V were 3.65+ 1.06 m/sec by 3C, 3.65+ 1.06 m/sec by 5C, and 3.83+1.17 m/sec by R, namely, peak V of R approach was slightly greater than other approaches (Figure 2 , ns).
cET, cAT, and AT/ET of 3 approaches: Figure 3 shows cET, cAT, and AT/ET obtained from 3C, 5C, and R approaches. cET was no significant difference between approaches. cAT was significantly longer in R (115+ 23 msec) than 3C (105+ 21 msec) and 5C approaches (105+ 20 msec) (p<0.05). Therefore, AT/ET was significantly greater in R (0.34+ 0.058: p<0.05) than 3C (0.317+ 0.053) and 5C (0.316+ 0.055).
AT/ET and severity of AS 1) AT/ET-peak V relation: Next, we analyzed the relation between AT/ET and severity of AS. First, we plotted AT/ET-peak V relation obtained by 3C, 5C and R approach (Figure 4 ). For each approach, linear regression analysis was performed. Regression equations were y = 0.0341x +0.1862 for 3C, y = 0.0364x +0.1750 for 5C, and y = 0.0374x +0.1887 for R, respectively. Regression line for R was positioned significantly higher than that of 3C or 5C by ANCOVA (p<0.05).
AT/ET and severity of AS 2) AT/ET-AVAi relation: Then, we plotted AT/ET-AVAi relation (Figure 5 ). For each approach, linear regression analysis was performed. Regression equation was y = -0.1763x +0.4314 for 3C, y = -0.1824x +0.4337 for 5C, and y = -0.1753x +0.4480 for R, respectively. Regression line for R was also positioned significantly higher than that of 3C or 5C by ANCOVA (p<0.05).
ROC analysis of 3 approaches: Therefore, we obtained 3 ROC curves of each approach to differentiate severe AS from moderate AS (Figure 6 ). According to the ROC analysis, cutoff AT/ET values for moderate and severe AS were 0.305 (AUC = 0.94) by 3C approach, 0.308 (AUC = 0.94) for 5C approach, and 0.331 (AUC = 0.91) for R approach, respectively.
Intra-observer variability of time measurements:Regression coefficients of AT were 0.994, 0.990, and 0.997 for 3C, 5C, and R, respectively. Regression coefficients of ET were 0.996, 0.997, and 0.990 for 3C, 5C, and R, respectively. Regression coefficients of AT/ET were 0.993, 0.991, and 0.998 for 3C, 5C, and R, respectively.
Concerning to the mean of difference (+ 2SD) of AT were 0.4+ 1.8 msec, -0.3+ 2.7 msec, and 0.5+ 1.5 msec for 3C, 5C, and R, respectively. Mean of difference (+ 2SD) of ET were 0.4+ 2.4 msec, -0.5+ 2.6 msec, and 1.0+ 3.5 msec for 3C, 5C, and R, respectively. Mean of difference (+ 2SD) of AT/ET were -0.002+ 0.005, -0.001+ 0.007, and 0.001+ 0.003 for 3C, 5C, and R, respectively.
Inter-observer variability of time measurements:Although inter-observer variability was slightly larger than that of intra-observer variability, the measurement errors were considered within the clinically accepted ranges. Regression coefficients of AT were 0.910, 0.880, and 0.914 for 3C, 5C, and R, respectively. Regression coefficients of ET were 0.951, 0.969, and 0.892 for 3C, 5C, and R, respectively. Regression coefficients of AT/ET were 0.902, 0.837, and 0.880 for 3C, 5C, and R, respectively.
Concerning to the mean of difference (+ 2SD) of AT were 0.8+ 7.8 msec, -2.7+ 9.2 msec, and -1.6+ 9.9 msec for 3C, 5C, and R, respectively. Mean of difference (+ 2SD) of ET were -1.1+ 10.4 msec, -0.1+ 8.0 msec, and 0.9+ 15.7 msec for 3C, 5C, and R, respectively. Mean of difference (+ 2SD) of AT/ET were 0.003+ 0.020, 0.008+ 0.027, and 0.007+ 0.027 for 3C, 5C, and R, respectively.
Additionally, comparison between apical (3C and 5C) and R approach was done (Figure 7 ). By 102 beats (3C and 5C) analysis, r=0.893(p<0.001), CV=6.76% for AT; r=0.957(p<0.001), CV=2.02% for ET; r=0.867(p<0.001), CV=6.11% for AT/ET. Then, by 99 beats (R) analysis, r=0.849(p<0.001), CV=8.33% for AT; r=0.887(p<0.001), CV=3.05% for ET; r=0.826(p<0.001), CV=7.92% for AT/ET.
Inter-observer variability of peak V (Vmax): We also checked the inter-observer variability in both of apical approach (3C and 5C) and R approach. Concerning to the regression coefficients of peak V in apical approach was 0.997 (n=32, p<0.01), and that in R approach was 0.980 (n=29, p<0.01). The CV was 3.51% for apical approach, and 3.41% for R approach, respectively, thus inter-observer variability of Vmax was no significant difference between apical and right parasternal approach. They were also considered to be within the clinically accepted ranges.