3.4.2 Analysis of outcomes with repeated procedures included
The mean follow-up for both groups was 424 ± 338 days. In the ZF group, there were fewer patients on AAD therapy (12.9% vs 36.1 %; p< 0.001). The overall success rate was higher in the ZF group (98.3% vs 93.5%, p < 0.004). A Kaplan-Meier curve of arrhythmia-free survival is shown in Figure 5. This difference was present in the ZF adult subgroup only. The overall success rate was also higher in patients without AAD therapy (98.4% vs 90.8%, p< 0.001), a difference present in all ZF subgroups. More specifically, there was a higher AVRT-free survival rate (99.0% vs 90.9%, p = 0.011) in the ZF group. This difference was present only in the ZF adult subgroup. There were also more procedures per patient in the ZF group (1.13 ± 0.356 vs 1.05 ± 0.241, p = 0.002). Detailed information is presented in the supplementary data in Table S4.
4.0 Discussion
The results of our analysis show comparable outcomes between the ZF and CF approaches for the treatment of SVTs with regards to efficacy and safety. The rate of complications was low among both groups, with no major procedural complications occurring.
The differences in the mean age and mean BMI of patients between groups in our study can be explained by the inclusion of paediatric patients in the ZF group only, an explanation confirmed by the ZF subgroup analysis. The rationale behind such a decision is that the stochastic and deterministic effects of ionising radiation are especially harmful to paediatric patients. Due to their developing bodies and long life expectancy post-exposure, they are more susceptible to developing cancer [9]. As such, we referred them to ZF CA as a first choice. Interestingly, some differences in the use of AAD therapy before and after the CA between both groups were discovered. These differences cannot be explained by the inclusion of paediatric patients as evidenced by the analysis of the ZF paediatric and adult subgroups. It can be hypothesised that, since the same physicians both referred patients for CA procedures and performed post-procedural follow-up, there was some degree of individual physician preference for the use of AADs.
Regarding the procedural data, we found a statistically significant difference in procedural duration, with the mean duration being lower in the ZF group (94.2 ± 50.4 min vs 104.0 ± 54.0 min, p = 0.002). Our findings are in line with previous studies investigating the ZF approach in various SVTs, which reported the mean procedural duration to range from 50 to 129 minutes [6,10–13]. Perhaps in contrast with the available literature, we found the procedural duration to be lower in almost all types of arrhythmias and locations of the substrate except in right-sided APs. We hypothesise that the features of the 3D EAM system not present in the CF approach are the cause of the shorter procedure times. Firstly, the ability to annotate the location of previously effective or ineffective ablation lesions facilitates the decision-making process during the CA procedure in almost all cases. Secondly, the distance between the His-bundle annotation and ablation catheter can be constantly monitored from two different views, which can be especially helpful in AVNRT, as well as for mid-septal and para-Hisian location of APs. Finally, local activation time mapping can identify the location of APs or the origin of AT both more rapidly and accurately. Additionally, our analysis of the learning curve showed that procedure time drops significantly after the first year of experience. Features of the 3D EAM system can perhaps explain the higher median number of RF lesions in the ZF group (8 (4 – 17) vs 7 (4 – 13),p = 0.015). We hypothesise that the ability to localise the successful ablation site with the 3D EAM system can prompt the operator to add additional adjacent bonus lesions. Without the use of the 3D EAM system, the exact location characterisation can be challenging. However, the difference was only present in the ZF adult subgroup (11 (6 – 20) vs 7 (4 – 13), p < 0.001). We suggest that the smaller hearts of paediatric patients as well as the potential for important heart structures such as the AV node and His bundle to be closer to the successful ablation site may prevent the operator from performing additional ablations despite the advantages of using the 3D EAM system. There were no statistically significant differences between the two groups regarding the overall procedural success rates. In the subgroup analysis, the procedural success rate in patients who underwent RF ablation for AVNRT was lower in the ZF adult subgroup (95.8% vs 99.4 %, p = 0.040). However, recurrence rates during follow-up after index procedure in those groups were similar.
A relatively lower procedural success rate for AT in the ZF group (70.3% vs 83.8%, p = 0.167) and a relatively higher procedural success rate for right-sided APs in the ZF adult subgroup (100% vs 57.1%, p = 0.091) did approach but not reach statistically significant difference. The number of patients with AT and right-sided APs was also too low to be representative and useful for any further statistical analysis. In addition, we found that both groups had no major complications and no statistically significant difference in minor complication rate (CF vs ZF; 0.4% vs 0.0%, p = 0.304). Importantly, the reduction of radiation dose was substantial (CF; DIA: 13.9 ± 11.0 minutes; DAP: 606 ± 1003 mGym2; p< 0.001). These results are in line with the findings from previously published studies, adding to the body of evidence on the feasibility and safety of ZF SVT ablation [6, 10–15].
Regarding long-term outcomes, there were no statistically significant differences between groups when only the data from the index procedure was analysed. The overall success rate did not differ between groups (ZF vs CF; 87.1% vs 89.2%, p = 0.422). This is also in line with previously published studies that reported no differences between the ZF and CF approach [10, 11, 13]. When the data on repeated procedures was included, the overall long-term outcomes differed between the groups, being significantly higher in the ZF group (98.3% vs 93.5%;p = 0.004). This was also true when only patients with AVRT were analysed (99.0% vs 90.8%; p = 0.010). In the subgroup analysis of procedures for AVRT, this difference was notable only in the ZF adult subgroup (100% vs 90.8%, p =0.031). These differences can be attributed to the higher number of procedures per patient in the ZF group (1.13 ± 0.356 vs 1.05 ± 0.234; p = 0.001) improving the overall success rate with included repeat procedures. As with the explanation for differences in the use of AADs, these differences in outcomes cannot be explained by the inclusion of paediatric patients, as evidenced by the analysis of the ZF paediatric and adult groups. It can again be hypothesised that in cases where tachycardia recurred, there was some individual physician preference for referring patients for repeat procedures versus the use of AADs.
An important finding that we would like to address is that CA procedures of AVNRT performed with cryoenergy had successful long-term outcomes in both analyses (100% after index procedure; 100% after repeated procedures were included). This is perhaps in contrast with the available published data. In a meta-analysis of cryoablation versus RF ablation for AVNRT, the long-term recurrence rate was significantly higher (9.7% vs 3.8%; p = 0.003) in patients ablated with cryoenergy [16]. The high success rate of ZF cryoablation in our study can, at least partially, be explained with the use of the 3D EAM system, which enables the operator to mark the location of a possible mechanical termination of the targeted slow pathway by the cryocatheter [17]. With this approach, the site of the mechanical block can be accurately ablated, in contrast to the CF approach where the exact location of the mechanical block can be more difficult to localise. An additional factor adding to the high success rate may be the utilisation of cryomapping during the ongoing AVNRT with tachycardia termination during ongoing cryoablation serving as an endpoint. This technique was previously published by Eryazici et al. and demonstrated a high long-term success rate [18].
Lastly, there are some limitations of the ZF approach when only a 3D EAM system is used for guidance. One such instance is that of transseptal punctures. To overcome the limitations of the 3D EAM system, the use of ICE for transseptal puncture is mandatory in our institution [19].
With ICE guidance, the operator can successfully identify true endovascular borders and anatomical structures as well as their relation to the transseptal needle. The usefulness of ICE, however, goes beyond guiding the transseptal puncture; with visual inspection, the presence of possible masses in the left atrium, on the long sheath, and on the catheter inserted into the left atrial cavity can be identified and further appropriate actions can be taken, possibly preventing certain potential procedural complications. It also offers precise guidance in manipulating catheters in the heart, including supervising the stability of the ablation catheter during lesion formation. Importantly, the pericardial space can be readily visually inspected for early detection of pericardial effusion.
5.0 Limitations
Our study has several limitations. Due to the nature of retrospective studies, our results may have been influenced by selection bias. Two factors need to be addressed here: first, the inclusion of paediatric patients exclusively in the ZF group, which might have influenced the study results. However, several studies showed similar procedural outcomes compared to adult patients [20, 21]. Furthermore, additional analysis of our ZF subgroups comparing baseline characteristics and procedural outcomes did not show significant differences between adult and pediatric patients. Second, cryoenergy was used only in the ZF group. This may also have an impact on the procedural success rate. Also, the learning curve related to the ZF procedures might have impacted procedural parameters and outcomes in comparison to the CF procedures. Finally, all data was collected at a single centre; as such, the results may not be directly applicable to experiences in other populations.
6.0 Conclusions
Our study demonstrated that the safety and efficacy of ZF CA procedures for right and left-sided SVTs is comparable to that of the CF CA approach. To further assess the non-inferiority of the ZF approach compared to the CF approach, a randomised trial comparing both methods is needed.