4 DISCUSSION
The main findings were as follows: (1) HPSD approach was associated with higher first-pass PVI, and recurrence of atrial arrhythmias; (2) HPSD approach could significantly reduce procedural time, ablation time and fluoroscopy time compared with the LPLD approach; and (3) Major complications and ETI were similar between two groups.
Thermal injury by RF ablation comprises two consecutive phases: resistive and conductive. The balance between power and duration parameters in resistive and conductive heating has a significant influence on lesion creation. The resistive phase has a resistive component adjacent to the catheter tip, which results in local heating and dissipation of energy as a heat source. Resistive heating probably occurs in the first few seconds during the RF application. With immediate heating, the electrical current is delivered at the catheter–tissue interface. The tissue necrosis is confined only to the first 1–1.5 mm from the catheter tip; however, the temperature always rises above 50°C with a conventional power setting17. Greater resistive heating can be achieved using higher-power delivery. In the conductive phase, the resistive heat source then extends energy passively to deeper tissues. Conductive heating is time dependent. The heating of deeper tissues increases with longer-duration RF applications2, 7, 21-23. The LPLD approach is associated with longer conduction heating.
Low power longer duration time of 10–30 s per site is based on earlierin vivo studies on the ventricular tissue 7. Simmers et al. showed that the lesion dimensions increased in 30 s using 25 W power; however, the mean lesion depth already reached 7.25 mm in 30 s 24. In the left atrium, where the mean posterior wall thickness is 1.5–2.5 mm and the esophageal distance from the posterior wall is as small as 2.5 mm, the LPLD approach might cause serious injury to the esophagus7, 18, 24, 25. The incidence of atrio-esophageal fistula (AEF) is reported as 0.1%–0.25%, and the incidence of esophageal thermal injury (ETI) is 2%–50% for the LPLD approach18. In addition, longer procedural time, ablation time, and fluoroscopy time are necessary for the LPLD approach due to its inherent characteristics. Prolonged procedural time would inevitably lead to a longer anesthesia duration, which also increases the procedural risk, especially for elderly patients 26.
One potential approach to optimize the longer ablation time of LPLD is to modify the relationship between the resistive and the conductive heating phases by increasing the resistive heating phase and reducing the conductive hearting phase to deliver immediate heating to the full thickness of tissues and limit deep tissue injury2, 17, 27. To achieve this, high energy must be delivered in a short duration. Thus, the strategy of HPSD was proposed and applied for AF ablation.
The HPSD approach is largely based on immediate heating during the resistive phase, however, whether high power with shorter duration might create trans-mural lesions in the left atrium? The average thickness of the left atrium is 2.8 ± 1.1 mm in humans28. For patients with persistent AF, the mean atrial wall thickness is only 1.89 ± 0.5 mm and never exceeds 3.5 mm25.The HPSD approach affects a tissue depth of 3.5–4 mm2, thus, the left atrial wall thickness is well within the depth for the HPSD approach. Hence, HPSD is well suited to AF ablation. Several previous experiments evaluated the utility of HPSD approach. Bourier et al evaluated the lesion metrics created by HPSD compared with LPLD application. They found that the HPSD approach created lesions similar in volume but wider and shallower compared with the LPLD approach 29. In fresh killed porcine ventricles, Goyal et al. showed that for a given CF, 20 s were needed to create a 4 mm deep lesion using 20 W ablation, while only 6–7 s were enough for 50 W ablation22. Bhaskaran et al. examined several combinations of high-power ablation in vitro and in a sheep model. The study showed that the use of 50 W for 5 s resulted in a similar lesion depth compared with the use of 40 W for 30 s. Steam pops occurred in 8% of the 40 W (30 s) ablation and in none of the 50 W (5 s) ablation7. So, HPSD approach could achieve rapid, more controlled, resistive tissue heating, and avoiding deeper collateral injury.
For clinical studies, Winkle et al showed that HPSD approach using CF sensing catheter was safe and result in excellent long-term freedom from AF with short procedure times and delivery of small amounts of total RF energy. No complications were reported in this trial27. Other studies also showed the safety and efficacy of HPSD approach in an abstract form. Nonetheless, these studies are too small to evaluate the safety and efficacy of HPSD approach, and the comparison data between HPSD and LPLD were still limited. So, the present meta-analysis was performed.
The present study showed a higher first-pass PVI in the HPSD group when compared to the LPLD group. The main reasons may due to greater size, more uniform and better consistency of the lesions created by HPSD approach. Catheter-tissue contact stability is an important factor contributing to lesion creation, and catheter instability in a constantly moving heart may account for the difficulty to transmit heat to the tissue17. This shortening of HPSD approach may mitigate the negative effects of catheter instability and probably optimizes lesion creation by increasing the likelihood to keep the catheter stable throughout the entire application, before stability becomes a consideration27. In the LPLD group, the catheter stability was an issue when longer single-lesion ablation was needed, leading to unevenness of lesions, tissue edema, and lower rate of first-pass PVI. Moreover, HPSD approach could reduce the rate of long-term recurrence of atrial arrhythmias. Complete PVI with trans-mural injury is most important for the freedom from AF during the long-term follow-up30, 31. Hence, HPSD approach trend to form more trans-mural, continuous and permanent lesions.
The pooled analysis showed significant advantages of the HPSD approach. The approach could shorten procedural time and ablation time compared with the LPLD approach, thus limiting patient exposure to intravenous fluids and anesthesia. Additionally, fluoroscopy time was also shorter in the HPSD group, which had a direct favorable impact on the patient, operator, and supporting staff. These results were consistent with previously published findings4, 16, 20, 22, 23, 32. A significant reduction in the procedural time was observed in the HPSD group because of shorter ablation time. The shorter ablation time for HPSD were due to the shorter time required for lesion creation, higher first-pass PVI, and fewer acute PV reconnections2. In the LPLD group, additional ablations were required for gap ablation in non-transmural lesions and achieving biphasic block of PVs so as to achieve completed PVI17. During the procedure, catheter movement mostly relies on the combination of x-ray and 3D electro-anatomic mapping system. Hence, longer ablation time in the LPLD group inevitably led to greater fluoroscopy time to locate and move the catheter in the left atrium. Additionally, less irrigation fluid was needed during HPSD due to shorter ablation time, making HPSD more suitable for patients with impaired left ventricular function.
Previous studies involved an increased power of 40–50 W for AF ablation. One study using HPSD showed improved outcomes but an increase in complications, such as ETI, cardiac tamponade, and so forth. The safety of using high power for AF ablation, especially on the posterior wall, was a concern33. However, another study using HPSD ablation reported no increase in complications 34. According to the principle of HPSD with increasing resistive and reducing conductive phases, minimizing damage to collateral tissues has been a crucial consideration in HPSD. Several animal studies suggested that HPSD was superior to LPLD with lower complication rate2, 21. Some human studies using HPSD showed excellent clinical outcomes with fewer complications15, 35, 36. In the present study, the pooled analysis of included studies also showed that the rate of complications similar between two groups. The AEF or ETI was rarely reported because most of the studies were single-center studies, and none of them were large enough to evaluate infrequent serious complications. A large observational study focused on the complications of the HPSD approach. In this study, 13,974 ablations were performed on 10,284 patients, revealing an extremely low complication rate in the HPSD group. Only 1 AEF occurred in 11,436 ablations using HPSD; however, 3 AEFs occurred in 2538 ablations using LPLD32. There were two studies included in our analysis discussed ETI in patients received ablation treatment. The results showed low and similar rates of ETI in the two groups. The subgroup analysis showed that HPSD could reduce mild ETI. No AEFs were reported in the included studies. Thus, the HPSD approach was safe enough for AF ablation18.
Additionally, another study reported by Reddy et al compared very high power short-duration (vHPSD, a delivery of 90 W for 4 s) approach using a QDOT microcatheter (Biosense Webster, Inc., CA, USA) with the LPLD approach. They demonstrated that vHPSD was an efficient, feasible, and safe strategy for AF ablation. However, this study was not included in the present meta-analysis due to a completely different setting of delivery of the vHPSD and HPSD approaches. Nevertheless, vHPSD may be another promising strategy in the future23.
Our study has some limitations. First, publication bias could not be completely excluded, the inclusion of only published data contributed to bias. Second, the number of included studies was limited to only seven, and most of the studies were designed as nonrandomized, thus, more well-designed and large-scale RCTs are required to confirm the findings. Third, limited collateral tissue damage is one of the important advantages of HPSD. However, in the present meta-analysis, this damage was not completely reflected due to limited endpoints reported from the included studies. Fourth, in some studies, the catheters applied in the HPSD and LPLD groups were different, thus may affect the outcomes of the pooled analysis. Fifth, the ablation power and duration settings in the included studies were not completely consistent.