In patients with long-standing persistent AF, the median heart-to-mediastinum ratio in the early and late phase of the study before the intervention did not differ significantly from the group of patients with SR (1.75 (1.59; 1.91) and 2.06 (1.7; 2.1) respectively, p = 0.1). The washout rate also did not significantly differ between the groups (24.5 (10.8; 41.1) % and 24.3 (12.3; 29.9) %, respectively, p = 0.15). The 123I-MIBG uptake defect before surgery in both groups was comparable and did not differ significantly (13.0 (7.0; 24.0) % and 19.0 (9.5; 23.5%) %, respectively, p = 0.59).
Based on the preoperative data, there were no statistically significant differences in sympathetic nervous system tonus in both groups.
Patients underwent a postoperative examination 4 week after surgery, which included several studies mentioned above (Table 2).
Table 2
Main postoperative clinical and instrumental parameters of patients with AF (group I) and SR (control group II), who underwent sympathetic nervous system examination (n = 68)
Parameter | Goup I. AF: AHD treatment + RFA (n = 53) Me (Q1; Q3) | Goup II. Control group with SR (n = 15) Me (Q1; Q3) | p-value |
LA diameter before the surgery (mm) | 45.0 (44; 49) | 39.0 (25; 41) | 0.005 |
RV diameter before the surgery (mm) | 23.5 (20; 27) | 20.0 (20; 22) | 0.13 |
IVS before the surgery (mm) | 10.0 (9; 10) | 9.0 (9; 10) | 0.5 |
LVEDD before the surgery (mm) | 49.0 (47; 54) | 46.0 (44; 48) | 0.1 |
LVESD before the surgery (mm) | 33 (30; 37) | 30.0 (26; 35) | 0.2 |
LVEDV before the surgery (mL) | 119.5 (90; 127) | 87.0 (72; 98) | 0.2 |
LVESV before the surgery (Ml) | 40.5 (34; 53) | 31.0 (27; 37) | 0.03 |
LVEF (Teicholz) before the surgery (%) | 61.5 (57; 65) | 64.0 (61; 75) | 0.04 |
LVEF (Simpson biplane) before the surgery (%) | 60.5 (58; 66) | 67.0 (64; 70) | 0.08 |
RVSP before the surgery (mmHg) | 35.0 (30; 40) | 35.0 (28; 37) | 0.5 |
CPB time (min) | 120.0 (110; 161) | 120.0 (113; 140) | 0.6 |
ACC time (min) | 70.0 (57; 98) | 92.0 (79; 110) | 0.3 |
6-minute walking test after the surgery (m) | 406.5 (380; 435.5) | 412.0 (290; 450) | 0.7 |
Comments. LA - left atrium; RV – right ventricle; IVS – interventricular septum; LVEDD – left ventricular end-diastolic diameter; LVESD – left ventricular end-systolic diameter; LVEF – left ventricular ejection fraction; LVEDV – left ventricular end-diastolic volume; LVESV – left ventricular end-systolic volume; RVSP – right ventricular systolic pressure; CPB – cardiopulmonary bypass; ACC – aortic cross-clamp. |
The efficacy of AF RF ablation was evaluated using postoperative 123I-MIBG scintigraphy and the neurotransmitters (metanephrine, normetanephrine, NE) levels assessment in the blood from Ao and CS were assessed.
In patients with atrial RF fragmentation (group I), the heart-to-mediastinum ratio after intervention in the late phase of the study was significantly lower compared to the group of patients with SR: (1.5 (1.4; 1.6) and 1.8 (1.56; 1.83), respectively, p = 0.02). In group I, the uptake defect was significantly higher than in control group II (25.0 (24.0; 35.0) % and 15.0 (12.0; 20.0) %, respectively, p = 0.01).
A statistically significant (p = 0.0098) decrease in the NE level in patients with RF ablation was found in the CS blood collected after the main stage of surgery compared to preoperative level. In the control group, there was no significant difference (p = 0.2) in the NE level in the CS blood.
There was a statistically significant intergroup difference (p = 0.0039) in the postoperative NE level (Fig. 2). This correlates with scintigraphy data: 1) the decrease in the heart-to-mediastinum ratio and reduction of the CS NE level in patients with RF ablation (group I) compared to group II after the intervention – a direct correlation, 2) the increased RP uptake defect and reduced CS NE level – an inverse relationship. Meanwhile, the differences between the subgroups with the ‘Maze-IV’ procedure in the heart-to-mediastinum ratio (Fig. 3), the RP washout rate (Fig. 4) and the RF uptake defect (Fig. 5) were statistically insignificant.
Discriminant analysis was used to build a model for the assessment of the RF ablation efficacy based on neurohumoral factors. The Wilks` Lambda criterion was p = 0.0037, indicating the adequacy of the created model (Table 3).
Table 3
Assessment of the adequacy of the RF denervation efficacy model based on the Wilks` Lambda criterion
Parameter | Wilks' Lambda: 0.56315 approx. F (3.23) = 5.9471. p < 0.0037 |
Wilks` Lambda | Partial Lambda | F-remove (1.23) | p-value | Toler. | 1-Toler. (R-Sqr.) |
NE level gradient (Ao/CS) (pg/mL) | 0.89 | 0.64 | 13.2 | 0.001 | 0.99 | 0.01 |
Normetanephrine level gradient (Ao/CS) (pg/mL) | 0.59 | 0.94 | 1.41 | 0.24 | 0.99 | 0.006 |
Normetanephrine level gradient (Ao/CS) (pg/mL) | 0.59 | 0.94 | 1.34 | 0.25 | 0.99 | 0.005 |
The difference in the NE levels after the main stage of the surgery is a significant factor. This parameter was calculated separately for each patient as follows:
\(\Delta\) NA = NAaa-NAcs,
where NA – noradrenaline; aa – ascending aorta; cs – coronary sinus.
Distribution in groups was carried out based on linear classification. The total number of correct relationships to a group was 89%, that indicates a high sensitivity and specificity of the created model (Table 4).
Table 4
Group | The percentage of correct assignments | G_1:1 p = 0.7037 | G_2:2 p = 0.2963 |
G_1:1 | 94.7 | 50 | 3 |
G_2:2 | 75.0 | 3 | 12 |
Total | 88.9 | 53 | 15 |
Based on the coefficients for the canonical linear discriminant function (CLDF), the following equation was calculated: |
CLDF = 0.77 + 0.003\(\Delta\)NAafter
The values of the centroid function were obtained (Fig. 6) by substituting the average values of predictors in the group of patients with postganglionic nerve plexi ablation (control group) into the discriminant equation.
Thus, the centroid was 0.55 in patients with ablated postganglionic plexi, and minus 1.36 in the group without ablation. The cut off value ΔНА = − 0.405 pg/mL. The efficacy of cardiac denervation can be calculated using the CLDF equation and the ratio of the obtained data to the centroids for each group and the group separation boundary (Fig. 6).