BHRS editorial – When loss of capture leads to VF detection


Sarah Whittaker-Axon

Chief Cardiac Physiologist, Barts Heart Centre



A 68 year old male received multiple shocks from his device, while an inpatient following a ventricular tachycardia (VT) ablation 1 week previously. The cardiac monitor displayed a paced rhythm whilst receiving shocks.

He had a biventricular cardiac resynchronisation device (CRTD) in situ, implanted in 2007 with an indication of impaired left ventricular (LV) function due to dilated cardiomyopathy (DCM). Box change in 2012 was complication-free, although right ventricular (RV) lead replacement was required in 2013 due to failure of a Riata lead.  He had also undergone an atrioventricular node (AVN) ablation previously due to rapidly conducting atrial fibrillation (AF) leading a reduction in biventricular pacing.

All device and lead measurements were stable and within acceptable limits post VT ablation (Table 1), with documented underlying rhythm of permanent AF with complete heart block, and ventricular escape rhythm of 30bpm.

RV threshold

0.75V at 0.8ms

LV threshold

1.75V at 0.8ms

RV impedance


LV impedance


R wave sensing


Table 1: Device measurements post VT ablation

Interrogation revealed the following presenting rhythm (Fig 1), with 20 episodes labelled as ventricular fibrillation (VF) in the arrhythmia logbook, some of which received a shock, some of which anti-tachycardia pacing (ATP) during charge was delivered but the shock diverted (Fig 2).

Fig 1: presenting electrogram (EGM) on interrogation. BP=biventricular pace. F=sensed interval in the VF zone.

Fig 2: EGM from one of the 20 stored VF episodes; a VF episode is declared, ATP during charge is delivered (STIM) but the shock diverted due to paced beats causing ‘Return to Sinus’.

Closer inspection of the presenting EGM (Fig 3) reveals that every other beat demonstrates true biventricular pacing with appropriate capture (red arrows), whilst the first, third and fifth beats demonstrate loss of capture on the RV lead. Capture occurs at the LV lead, with depolarisation eventually reaching the RV lead approx 170ms later (green arrows). This conduction is delayed enough for the signal to fall outside the ventricular blanking period and to be sensed as a VF event. This loss of biventricular capture is reflected in a different morphology farfield EGM, and would be seen as a broader QRS complex on a surface ECG.  The ‘F’ interval resets the timing clock, therefore the following VV interval is slightly longer than the interval following true biventricular capture.


Fig 3: red arrows indicate true BP, the green arrows demonstrate late depolarisation of the RV due to RV lead failure to capture, which is subsequently sensed.

If this loss of capture occurred on the LV lead instead this would not be sensed as a VF event, as ICDs do not use the LV lead for detection.  The way SJM detects arrhythmias leads to these FS intervals being classified as a VF event, and the device charges (with ATP during charge).

When analysing device episodes and understanding why a therapy is delivered it is important to be aware how each of the manufacturers detect a tachycardia. These are summarised below.


Intervals are classified by comparing the current VV interval with an average interval (calculated as a mean of the current interval and previous 3 VV intervals).

·         If both intervals fall below the tachycardia detection rate, the beat is classified as ‘VS’ (venticular sense).

·         If both intervals fall within a tachycardia zone the beat is classified as ‘F’ (VF), or ‘T’ (VT).

·         If the intervals do not agree (e.g. one interval is VF, the other is VS), the beat is not classified and is labelled ‘-‘ (see Table 2).

Table 2: Classification of beats in a SJM ICD (Mansour & Khairy, 2008).

Beats classified as ‘T’ or ‘F’ are ‘binned’, i.e. added to a VF or VT bin accordingly. The relevant bin increments with each binned interval, and can only be reset if a ‘Return to Sinus’ is detected.  The binned intervals do not need to be consecutive, and there is no sliding window.  ‘Return to Sinus’ is nominally 5 consecutive sinus (or paced) intervals, but is programmable between 3 and 7. If no ‘Return to Sinus’ occurs, whichever bin first reaches the programmed number of intervals to detect determines which rhythm is declared.

In the case described above, the VF bin increments with each loss of capture and subsequent FS, leading to VF detection. Although there were some ‘BP’ intervals during each epiosde, there were not 5 consecutive BP/VS intervals for ‘Return to Sinus’ to be declared, and thus the VF bin did not reset and allowed the number of intervals to detect to be reached.


The VF zone uses a probabilistic counter, e.g. a programmed detection of 9 out of 12 intervals uses a 12 beat “sliding window” (demonstrated in Fig 4). If 9 or more of 12 intervals in the sliding window is within the VF zone, VF detection is met.

The VT zone uses a consecutive counter, e.g. a programmed detection of 12 intervals requires 12 consecutive intervals falling in the VT zone for a VT episode to be detected. A slower interval will reset the counter, while a VF interval does not reset or increment the counter.

For a 3 therapy zone set up, choosing ‘FVT via VF zone’ gives a fast VT zone using the probabilistic counter from the VF zone, while choosing ‘FVT via VT zone’ gives a fast VT zone using the consecutive counter from the VT zone.

Fig 4: Example of the sliding window used in Medtronic detection; here a VF detection of 9 out of 12 intervals is programmed (Mansour & Khairy, 2008).

Boston Scientific

3 consecutive intervals falling within any tachycardia zone start a detection window (a sliding window).

The detection window of 10 intervals is applied with 8 out of 10 fast intervals resulting in an episode being declared. The beats are classed according to the rate (e.g. VF, VT or VT-1 zone). Intervals in a faster zone will also count as an interval in the slower zones, however slower zone intervals do not count towards faster zones.  Once 8 out of 10 intervals is reached in a zone, the respective zone’s duration window begins.

The duration window is programmable (e.g. 2.5s in the VF zone). The ‘sliding window’ continues, and must detect 6 out of 10 intervals in that zone until the duration window expires, after which therapy for that zone is delivered (see Fig 5).

If the number of tachycardia intervals drops below the 6 out of 10 intervals, the duration window resets to zero, and will only start again after a further 8 out of 10 interval detection window is re-met. If this has not occurred within a set time interval the episode is declared over.

Fig 5: detection in a Boston Scientific device with 2 zones programmed. Although the VT detection window starts before the VF detection window, VF duration expires first and therefore rhythm is treated as VF (Zanker et al, 2016).


Mansour, F. & Khairy, P (2008) ICD Monitoring Zones: Intricacies, Pitfalls and Programming Tips. Journal of Cardiovascular Electrophysiology. 19 (5) 568-574

Zanker, N. Schuster, D. Gilkerson, J. Stein, K. (2016) Tachycardia Detection in ICDs by Boston Scientific. Herzschrittmachertherapie + Elektrophysiologie. 27 (3) 186-192

Click here to download the editorial as a PDF file