The results of the CRAFT extension study show that in patients presenting with isolated typical AFL with no previously documented AF, cryoballoon PVI resulted in a similar reduction in symptomatic atrial arrhythmia as compared to RF CTI ablation. Importantly, the incidence of AF was significantly reduced with an upfront PVI approach, without creating any significant safety signals. The extended follow-up of 3 years showed that whilst the Kaplan–Meier curves for AF occurrence diverged early, this difference was preserved over time. This is an important observation and refutes a potential concern with the 12-month results [11] that a PVI-only approach may have merely delayed the development of arrhythmia.

As described earlier, it is theorised that all patients with AFL have underlying PV triggers, and those susceptible to developing an intercaval functional line of block may organise into a typical flutter circuit [2,3,4]. The results of the CRAFT study demonstrate that addressing the arrhythmia trigger, without targeting the mechanism, can result in similar arrhythmia freedom to the elimination of the mechanism itself. Indeed, 78.8% (Kaplan–Meier estimate) of patients in the PVI group stayed free of symptomatic arrhythmia recurrence in spite of no CTI ablation having been performed, giving support to this hypothesis. It should be acknowledged, however, that symptomatic recurrence can be underestimated when using ILR follow-up, as it relies upon patient reporting.

Such an approach may be beneficial in selected cases, with shared decision-making. For example, a patient with isolated AFL, but with comorbidities and left atrial enlargement, may be more likely to re-present with AF and may benefit from early elimination of PV triggers. Whilst it can be argued that PVI is a higher-risk procedure—due to the necessity for transseptal puncture and left atrial access—we did not find a difference in safety outcomes (accepting that our study was not powered to detect such rare events). Furthermore, a recently published comprehensive review has shown that AF ablation has become an increasingly safe procedure over time [13], which may support a lower threshold for considering a trigger-based PVI strategy.

Although not explicitly studied here, in those patients where a PVI approach to AFL is considered, it may be prudent to perform both PVI and CTI ablation in the same procedure. This is because AFL recurrences do happen, particularly after a single PVI procedure; thus, the patient may gain maximal benefit from targeting both trigger and mechanism.

Although recurrences and redo procedures were relatively infrequent, it is notable that PVI patients were numerically more likely to experience AFL and require CTI ablation, and CTI patients were more likely to experience AF and require PVI. Aside from highlighting the benefits of targeting both, as mentioned above, it is important to consider the mechanisms at play when arrhythmia recurs. We did not perform routine re-mapping of the left atrium to check for PV reconnections in this study. It is therefore possible that PVI patients presenting with recurrent AFL simply had a PV reconnection, thus facilitating re-initiation of the unablated CTI mechanism. Hence, robust PVI may, in theory, eliminate typical AFL. However, given the well-known difficulty in achieving durable PVI at present, as well as the potential for non-PV triggers, it again stands to reason that CTI ablation should be recommended for the majority. This is in line with prior studies assessing the combined procedure versus CTI ablation alone [14,15,16,17,18,19,20]. Routinely ablating the CTI in patients with isolated AF who have not demonstrated clinical AFL is not recommended [21,22,23,24,25].

It is notable that a large percentage of patients have short bursts or asymptomatic atrial tachyarrhythmia, as shown in Fig. 4. Atrial high-rate episodes have recently been discussed as pre-cursors to AF [26]—in our study, ablation reduced the amount of sustained arrhythmia, but these episodes may still be present as shown by ILR interrogation. These sub-clinical episodes may reflect the ongoing presence of arrhythmogenic substrate despite symptomatic resolution. This may be important in terms of decision-making around long-term anticoagulation. Alternatively, these may be unimportant features which are only detected due to continuous ILR monitoring. The recent NOAH-AFNET 6 study found that patients with incidentally detected atrial high-rate episodes did not benefit from oral anticoagulation [26]; however, these patients had not been diagnosed with AF or AFL, nor had they undergone prior catheter ablation. Further research would be beneficial.

The CRAFT trial provides supportive evidence for considering PV trigger elimination in patients presenting with isolated typical AFL, as well as providing insight into the relationship between mechanisms and triggers of atrial arrhythmia. This may support decision-making in patients at high risk of presenting with subsequent AF. Equally, our results may give reassurance in the setting where bidirectional CTI block cannot be achieved but PVI is feasible.

The field of cardiac electrophysiology has evolved considerably over time [13], especially since the CRAFT study began, particularly with the recent advent of pulsed-field ablation (PFA). Future work may include assessing the effect of PFA-based PVI on AFL. Ongoing work to understand the optimal approach to atrial arrhythmia ablation—be it trigger or mechanism ablation—will be crucial in the years to come.