Diagnosis of atrial fibrillation
81-year-old man participating in the LBBB-TAVI study.
Episode diagnosed as AF;
- the tachogram and Lorenz plot are highly suggestive of rapidly conducted AF with significant cycle-to-cycle variability in sensed ventricular intervals;
- the tracing confirms the diagnosis of AF;
- diagnosis of AF by the device.
Long-term electrocardiogram monitoring in order to diagnose asymptomatic atrial fibrillation in patients with stroke constitutes a new preferential indication for the implantation of an implantable loop recorder. Indeed, approximately 25% of ischemic strokes remain unexplained after performing a complete work-up, including telemetric monitoring and 24- to 48-hour loop recording, and are thus labeled cryptogenic. The probability of recording an episode of atrial arrhythmia increases with the prolonging of the electrocardiogram monitoring time. Integrating an automatic AF detection algorithm by the device in order to allow diagnosis even when episodes are asymptomatic is a necessary prerequisite for the implantable loop recorder to be positioned in the decision tree of cryptogenic stroke. Evidence of AF cannot be based on a direct analysis of atrial activity (rapid and uncontrolled), the atrial signals not being sufficiently voluminous and the various settings rather favoring the sensing of QRS complexes. For the Biomonitor, the diagnosis of AF is based on the analysis of the stability of QRS complexes, the difficulty being to differentiate AF from other conditions associated with an irregular rhythm (atrial extrasystoles, ventricular extrasystoles, etc.).
The RR intervals are continuously analyzed and intervals classified as noise (Vn) are excluded from the analysis. Various parameters are programmable:
- RR interval variability: this parameter corresponds to the maximum value in terms of interval-to-interval variation between 2 Vs-Vs intervals for the rhythm to be considered as stable by the device; the lower the programmed value, the higher the sensitivity of the AF diagnosis; all intervals with greater variability than the programmed value are considered AF intervals and thus increment the episode onset counter.
- the episode onset window and the number of intervals required at the onset of an episode: it is possible to determine the number of consecutive intervals on which rhythm stability is studied (onset window) as well as the minimum number of unstable intervals required in this window for the rhythm to be considered unstable; programming at 8 and 5 implies that the device continuously analyzes the stability over 8 consecutive intervals and that the rhythm is deemed to be unstable in this window if at least 5 intervals are considered unstable; when two consecutive windows are considered unstable, the device suspects an AF episode.
- an AF episode is diagnosed and recorded, when instability persists for a programmable period of time; the nominal confirmation time is relatively long (6 minutes) although the latter can be changed (between 1 and 30 minutes); an episode is only recorded when this confirmation criterion is completed.
- the end-of-episode window and the number of intervals necessary to determine the end of an episode: it is possible to determine the number of consecutive intervals on which the study of stability is performed for establishing the end of the episode (termination window) as well as the minimum number of unstable intervals required in this window for the rhythm to be considered as systematically unstable; a programming of 16 and 1 implies that the device continuously analyzes the stability over 16 consecutive intervals and that the rhythm is considered to be systematically unstable in this window if at least 1 interval is considered unstable (all intervals must be stable for the episode to be considered terminated); 2 consecutive windows with rhythm deemed stable are necessary for the end-of-episode diagnosis.
It is possible to program 3 sensitivity levels for the diagnosis of AF (low, intermediate and high) with predetermined values for the various parameters of the algorithm. It is also possible to program these different parameters independently. Low is the least sensitive level; therefore, more variability between RR intervals is required over a longer period of time to arrive at a diagnosis of AF. Conversely, less variability over fewer intervals is required when a high level is programmed.