Why does the AV delay shorten with ventricular pacing?

Someone presented a conundrum last week, when he queried a pacing ECG which violated a fundamental principle of dual chamber pacing. In the DDD pacing mode, the A-V delay is always longer with ventricular pacing.

AV delay 1.jpg

The A-V delay is a programmable physiological period, where after atrial sensing/pacing, A-V conduction is allowed to depolarize the ventricle (Vs) similar to the PR interval. If conduction fails, then after a fixed time, the ventricle is paced (Vp).

Obviously Vs will occur before Vp resulting in the shorter A-V delay.

The case in concern showed the opposite and thus violated a fundamental principle of dual chamber pacing.

AV delay 2.jpg

The ECG shows simultaneous rhythm strips from 4 leads. The rhythm unexpectedly changes from ApVs to ApVp and the A-V delay shortens from about 360 ms to 280 ms. 

How do we explain this?

This is an auto-threshold measuring algorithm and is rarely captured on a 12-lead ECG, but is very common with ambulatory Holter monitoring (if you look for it).

All companies have such a testing algorithm for ventricular pacing and some for atrial as well. All use the principle of an “evoked response” where depolarization is detected after pacing. Failure to recognise depolarization results in a back-up pulse of higher voltage delivered immediately after, so that no pauses are seen on the ECG and thus no symptoms. Every company’s algorithms are different and minor changes are made more frequently than I change my underwear, so what occurs today may be different tomorrow.

 Once the threshold for pacing is determined by the pacemaker, the voltage is adjusted to a level just above the threshold and thus will conserve energy and extend the life of the pulse generator.

A summary of the two most important auto-threshold features with all the companies is shown. Again, they change regularly.

AV delay 3.jpg

With two companies, the A-V delay shortens modestly (up to 65 ms) as in our case and for the other three companies, the total A-V delay is very short; 15 to 60 ms. 

Here are some examples:

The finish of a test. Cannot see the loss of capture as the back-up pulse is very close to the primary stimulus artefact and both are not seen.


Another example of a modest reduction of 60 ms in the A-V delay at 1 am and is thus Medtronic.


Here is an example of a very short A-V delay.

AV delay 6.jpg

During the test, the A-V delay is only 30 ms and the upstroke of the P wave can be just seen before Vp. Note at the end of the test, capture fails and there is a delay of 100 ms with the P wave appearing before the back up pulse, which in this illustration is unipolar.

With a little detective work, this case can only be from two companies although the A-V delay measurements are a little off. Seeing it was an ECG and done during the day, it must be from MicroPort (headquarters in Shanghai and previously Sorin or Liva Nova). Small number sold in Australia.

Thank you

Harry Mond

Chelsea Cunningham