RESULTS:
The mean age of patients in our study was 31.66 ± 8.69 years and 56%
were females. Forty-six percent patients were in NYHA class II, 52% in
NYHA class III and 2% in class IV. (Table I)
Based on LAAEV of <25 cm/sec or more, the patients were
divided into 2 groups: Group A with LAA inactivity (LAAEV <25
cm/sec) and group B without LAA inactivity (LAAEV ≥ 25 cm/sec). There
were 73 patients in our study who had evidence of LAAI (Group A) while
the remaining 27 patients had normal LAA contractile function (Group B,
Table II). There was no significant difference in height, weight, age or
gender between the two groups. Hemoglobin and erythrocyte sedimentation
rate (ESR) were also comparable between the 2 groups. However, serum
fibrinogen was significantly higher in group A compared to group B (340
± 86.56 vs. 266.12 ± 62.84 mg/dL, p<0.001). [Table II]
There was significant difference in MVA (p<0.001), mean
gradient across MV (p<0.001) and mitral valve lateral annulus
systolic velocity (p<0.001) between the two groups. A
significantly higher number of patients in group A had evidence of smoke
in LA/LAA than group B (67 [90.5%] vs. 13 [48.1%], p
<0.001). Two patients in group A also had associated LAA
thrombus compared to none in group B. [Table II]
Majority of patients in both groups had grade 1 or 2 SEC (47 out of 67
patients i.e. 70.1% in group A and 10 out of 13 patients i.e. 76.9% in
group B) [Table II]. Twenty patients (29.9%) in group A had grade 3
or 4 SEC compared to 3 (23.1%) patients in group B having grade 3 SEC
only and none having grade 4 SEC.
Inter-observer difference in grading of SEC was observed in 8 patients
primarily while determining between grade 1 and grade 2 and was 10% in
our study. The mean co-efficient of variation for measurement of left
atrial peak emptying velocity between observers was 1.8% (range 0 to
18.6%)
Only four variables, having the p<0.2 in univariable analysis,
were included in the multivariable analysis: MVA, mean trans-mitral
gradient (MVMG), Sa wave amplitude and serum fibrinogen level.
On multivariable regression analysis, mean gradient across mitral valve
(p=0.001), Sa (p=0.02) and serum fibrinogen (p= 0.005) were found to be
independent predictors of LAAI. However, the MVA assessed by planimetry
failed to achieve statistical significance to predict LAAI in our study
(P=0.06). (Table III)
We found significant positive correlation between LAAEV and Sa-wave by
using Pearson correlation analysis (P <0.001, r=0.475) (Figure
2A). The optimal cut-off value of Sa-wave obtained was 6.82 cm/sec, from
the ROC curve analysis, for predicting presence of inactive LAA with
sensitivity of 66.6% and specificity of 61.8%. The area under ROC
curve of 0.67 revealed moderate discrimination. The positive predictive
value of Sa- wave to predict LAAI was 44.7% and negative predictive
value was 79.2% (Figure 2B).
A significant negative correlation was found between LAAEV and MVMG by
using Pearson correlation analysis (P <0.018, r= -0.235)
(Figure 3A). The optimal cut-off value of MVMG, obtained from the ROC
curve analysis, for prediction of LAAI was 11.5 mm Hg with sensitivity
of 82.2% and specificity of 70%. The area under ROC curve of 0.836
revealed good discrimination. The positive predictive value of MVMG to
predict LAAI was 73.5% and negative predictive value was 79.8% (Figure
3B).
We also found a significant negative correlation between LAAEV and
plasma fibrinogen level by using Pearson correlation analysis (P
<0.005, r= -0.391) (Figure 4A). The optimal cut-off value of
serum fibrinogen obtained from ROC curve for predicting LAAI was 300
mg/dl with sensitivity of 71.0% and specificity of 74.1%. The area
under ROC curve of 0.780 revealed good discrimination. The Positive
predictive value of serum fibrinogen to predict LAA inactivity was
73.27% and negative predictive value was 71.87% (Figure 4B).
We also assessed the combined predictive value of aforementioned
variables (serum fibrinogen, MVMG and Sa-wave). The AUC for predicting
LAAI by combining these three variables improved to 0.89 (95% CI- 0.81
to 0.96) with improvement in sensitivity to 87.7% and positive
predictive value to 79.8% (Figure 5, Table IV).
We also compared 80 patients in our study with LA/LAA smoke with
associated thrombus (Group C) with 20 patients(Group D) without LA/LAA
smoke or thrombus. On univariate analysis, MVA by planimetry (p=0.007),
MVMG (p = 0.002), serum fibrinogen (p=0.001) and percentage of patients
with inactive LAA (p<0.001) had p value < 0.2 (Table
V). Multivariable binary logistic regression analysis was done on these
four variables to identify the independent predictors of LA/LAA smoke.
Only LAAI was found to be an independent predictor of LA/LAA smoke with
associated thrombus (p=0.04) (Table VI)