Another way of saying the same thing:  I tell my students that in AVD the P waves simply never get the chance to conduct.  They are always immediately in front, buried inside, or behind the QRS or somewhere in the ST segment or T wave (refractory period).  I tell them that in order to make the call of complete block, you must find P waves in places where they have a legitimate opportunity to conduct - but they still fail to.

This still raises the possibility of Second degree AV block combined with junctional escapes which are dissociated from some of the P's.  To rule this out as well, I tell them that the ventricular rhythm in complete block ALSO has to remain regular (with the exception of any pvc's) throughout the period of AV Block.  This regularity is due to the ventricular pacing being done by either a junctional or ventricular escape mechanism.  If there is any irregularity at any point, then that indicates capture of the ventricles by a P wave, and that eliminates complete block as a possibility.

Important topic brought up by Christopher! - with GREAT input from Walter Mueller. The clinical issue that presents is whether a rhythm with P waves that do not conduct is the result of "AV dissociation" vs complete AV block.

AV dissociation (ie, P waves that don't conduct) may be due to: i) Usurpation; ii) Default; or iii) AV block. Christopher provides the classic example of AV dissociation with VT - which is by USURPATION from the malignant ventricular rhythm that takes over.

A common example of AV dissociation by DEFAULT occurs when there is sinus bradycardia (say down to a rate of 50/min) - and because of this slowing of the sinus rate, an appropriate junctional response (say at a rate of ~55/min) may take over. You might see 10 or more beats on such a rhythm strip - and NONE of the P waves are conducting. However, this is NOT complete AV block - because as Walter emphasizes, NONE of the P waves has a "chance to conduct". Clue to when conduction IS occurring will be a CHANGE in the rhythm that should occur when a P wave falls at a point in the cardiac cycle when as Christopher emphasizes it is not refractory. Several points to further expand on Christopher/Walter's comments:

• The easiest way to tell at a glance if complete AV block is present is to check out the ventricular rhythm - which most of the time should be regular with complete block (because most of the time the escape focus be it AV nodal or ventricular will have a fairly regular rate). However - the escape rate might not be precisely regular all of the time. This is especially true with code tracings - in which the crashing heart does not always "read the textbook".  That said - almost always you WILL be able to tell at a glance if there is a fairly regular escape site at work vs a clearly unexpected earlier-than-should-be beat. That earlier-than-expected beat is a HUGE clue that the block is probably NOT complete (but rather high-grade 2nd degree).
• Most of the time - the ATRIAL rate will be regular (or fairly regular) when there is AV block. Realizing that slight irregularities (ie, from underlying sinus arrhythmia) may occur - this is helpful to know because it facilitates recognizing blocked PACs and sinus pauses in which the atrial rhythm is not regular.
• Keep in mind that in addition to Walter's excellent point about looking for ventricular regularity and P waves falling at all points in the cycle - that usually the ventricular rate must be SLOW ENOUGH for this to happen before one can confidently call "complete" AV block. Ideally - the ventricular rate should be LESS than 50/minute. This simply means that IF you have no conducted P waves and a faster regular ventricular rate (say 80/min or more) - that it is often VERY difficult to be certain one is dealing with "complete" AV block vs AV dissociation by usurpation because of how difficult it is to be sure P waves truly have "opportunity to conduct yet still fail to do so ... ".

Great discussion by Christopher & Walter!

This is a copy of some comments that I recently posted on Facebook and have reposted here in response to this topic.

This is a subject that I’m probably more passionate about than any other in electrocardiography. It’s important to remember that A-V dissociation is NEVER a primary disturbance of rhythm nor is it a diagnosis or an individual rhythm although all too often it is stated as such. A-V dissociation is always secondary to some other primary disturbance of rhythm. I've always thought that the late Dr. Henry J. L. Marriott summed up the definition of A-V dissociation the best. He maintained that A-V dissociation, . . . and I quote,: "is the independent beating of the atria and the ventricles, period! – No more, no less!". Marriott said A-V dissociation is analogous to a "symptom" while Mike Taigman (of the book Advanced Cardiology in Plain English) likened it to a "description". Some authors erroneously defined “complete A-V block” as the total failure of A-V conduction in the presence of an atrial rate that is faster than the ventricular rate. These same authors also incorrectly define “A-V dissociation” as the absence of A-V conduction in the presence of a ventricular rate that is faster than the atrial rate. The respective atrial and ventricular rates are NOT rate-specific in defining A-V dissociation. The atrial rate can be faster than the ventricular rate OR the ventricular rate can be faster than the atrial rate. Rate is not important just so long as the single criteria of independent beating is satisfied. In my personal collection of 12-lead ECGs, I have several examples of monomorphic ventricular tachycardia dissociated from atrial flutter. In each case, the atrial rate (e.g., ~300/min) far exceeds the ventricular rate (e.g., ~160/min). By some authors’ narrow definition of A-V dissociation, these would qualify as “complete A-V block” because there is no A-V conduction AND the ventricular rate is slower than the atrial rate. To refer to these examples of ventricular tachycardia as being “complete A-V block” would be utterly absurd! This page on Medscape is just one example of a misleading definition of A-V dissociation:
http://emedicine.medscape.com/article/151715-overview
A-V dissociation is not synonymous or interchangeable with Complete A-V block. Complete (i.e., “third-degree”) A-V block is just one of several causes of A-V dissociation and the list of causes is quite short and easy to memorize. Quite often, you’ll hear healthcare providers, who are well-versed in the subtle nuances of A-V dissociation, referring to it as being due to either “default” or “usurpation”. For those who have access to it, this is an outstanding paper by Marriott, et al. titled "A-V dissociation revisitied": http://www.ncbi.nlm.nih.gov/pubmed/5932450