Does the classsification of atrial fibrillation need to be revised?

I think so , here is why,

 Paroxsymal atrial fibrillation  (APF) is currently defined a a spontaneous remitting episode of atrial fibrillation (AFIB) lasting less than seven days and resolving without intervention. Current U.S. guidelines recommend oral anticoagulation therapy (OAC) for patients with all types of afib based on the stroke risk estimate using  CHA2DS2-Vasc score.

Modern pacemakers (PM) can now record tracings continuously and store them for long periods of time. Several studies have analyzed these tracings and we have learned that commonly PM patients have short runs of Afib,( or more accurately  atrial high rate episodesAHRE-some of which are in fact  not AF)  varying greatly in duration and frequency .

It does not make sense to assume we should have the same prognostic concerns  and therapeutic recommendations   to someone with a episode of afib lasting one minute as to  someone with afib lasting  five days without good data validating that position.

Several studies have attempted to correlate these runs of afib with stroke  rate. With one exception data  (see footnote 1) have not been analyzed if these episodes last less than 5 or 6 minutes.These studies have shown that there is a  coarse grain dose response relationship  between duration or  "burden" of afib and stroke risk. Attempts have been made to determine an afib burden level above which there is a significant increase in risk worth of initiating anticoagulation  and no consensus has been reached.

Various investigations have yielded different burden levels .

The ASSERT trial is widely quoted as showing a two fold increase in stroke risk for episodes of afib last over  6 minutes. See footnote 2 for a discussion of the surprisingly and diametrically  different ways these results were described in medical publications.

Yet another article discussing ASSERT  claims that a fib greater than 24 hours does have increased risk of stroke but for shorter episodes no statistically significant relationship was shown.





footnotes:

 1. The Rate Registry study reported no increase risk of stroke for afib duration episodes less than 10-20 seconds.But what about those episodes between 20 seconds and six minutes?

2. Chin-Scarabellit et al including senior author K. Ellenbogen  describe the ASSERT trial simply as showing  an increased hazard ratio (HR) even though it was not statistically significant while Conn  et al  say "surprisingly,no correlation was found in the ASSERT between AHRE (atrial high rate episodes) and a history of stroke" and speculated that this might have resulted from the fact that the sampling period was only the first 3 months after PM implantation. A third spin on the ARREST was offered by Dr B P Knight (EPLab Digest March 2012) who said that only the  highest quartile of duration ( more than 18 hours) was statistically significant.He did not discuss the problems involving in relying on subgroup analysis to reach a conclusion regarding outcome validity less alone  making therapeutic decisions on that basis. Data is often hard to obtain but sometimes  the real difficulty is in finding agreement as to what the data signify.