Steven Mickelsen M.D., Pragnesh Parikh M.D., Rebecca McNeil PhD., and Fred Kusumoto M.D.

1
Diagnosis and Clinical Management of Atrial
Fibrillation After ICD Implantation.
Steven Mickelsen M.D., Pragnesh Parikh M.D.,
Rebecca McNeil PhD., and Fred Kusumoto M.D.
J a c k s o n v i l l e
2
INTRODUCTION

• Atrial fibrillation is common
• Atrial fibrillation and heart failure often
  coexist
• ICD-event-data can detect atrial fibrillation

Benjamin et al. Circ 1998, Go et al JAMA 2001,
Miyasaka Eu Heart J 2005 Patton et al. J Card
Electro 2006 Wang et al. Circ 2007
3
BACKGROUND
Prevalence of atrial fibrillation and NYHA class
NYHA-IV
50
NYHA-III

NYHA-II
NYHA-I
5
SOLVED, Vheft, CHF-STAT,
DIAMOND CHF, and GESICA
Benjamin et al. Circ 1998, Go et al JAMA 2001,
Miyasaka Eu Heart J 2005 Patton et al. J Card
Electro 2006 Wang et al. Circ 2007
Benjamin et al. Circ 1998, Go et al JAMA 2001,
Miyasaka Eu Heart J 2005 Patton et al. J Card
Electro 2006 Wang et al. Circ 2007
4
BACKGROUND I
Wojciech et al 2006, Saxon et al 2006,
Freudenberger et al. 2007, Hoppe et al 2008
5
HYPOTHESIS

• We think ICD-event-data is being used more
  frequently to diagnose atrial fibrillation in the
  clinical setting.
• We expected to find more atrial fibrillation in
  our patients than suggested by the published
  device trials.

• We wanted to characterize the clinical impact AF
  had on clinical outcomes and medical decision
  making in this patient group.

6
METHODS
Retrospective chart review of all ICD implants
(01/01/2002-2008)

• Demographics baseline characteristics at
  implant and follow-up
• Arrhythmia burden
• All cause mortality
• Hospitalization
• Defibrillator discharge
• Temporal relationship of clinical AF diagnosis to
  ICD-ED

7
RESULTS I
SUBJECTS
165 implants (January 1, 2002-January 1, 2008)
88 patients in the Mayo device clinic
Median follow-up 2.77 years (range 0-19.2 years)
8
RESULTS II
SUBJECTS
9
RESULTS V
PREVELENCE
ATRIAL FIBRILLATION
HX AF
40/165
Paroxismal
82/165
43/165
NSR
New AF
10
RESULTS VII
INCIDENCE
One year

Mean time to detection of new AF after implant
1.3 years (0.07-9.6 years)
11
RESULTS VI
SUBJECTS
12
RESULTS VIII
CLINICAL END POINT
13
RESULTS X
ALL CAUSE MORTALITY
All cause mortality HR 1.72
(p0.007)
Overall median survival 6.36 years, 95 CI
4.3-9.2 years With AF
4.30 years, 95 CI 3.5-8.1 years NSR
6.90 years,
95 CI 5.1-12.1 years
14
RESULTS XI
CLINICAL END POINTS

15
RESULTS IX
COMBINED END POINT
ATPShockDeath
HR 2.0 (p0.02)
16
RESULTS XII
CLINICAL END POINTS
65 ICD-ED led to AF diagnosis in most cases.
82 ICD-ED diagnosis of AF changed medical
management most of the time. 53 started on
warfarin after diagnosis.
17
LIMITATIONS
Retrospective chart review in heterogeneous
population No formal criteria for diagnosis of
AF by event recorder Incomplete follow-up in a
large proportion of the patients
18
DISCUSSION
Presence of AF is a bad prognostic marker in ICD
recipients. Devices will pick up asymptomatic AF
and potentially clinically insignificant
events. Does the information lead to improved
management?
19
CONCLUSIONS
Event data from ICDs is being used to help
establish the diagnosis of new onset atrial
fibrillation and influence clinical
decision-making.  Prevalence and incidence of AF
appears much higher in our patient population
than that found in published device trials.  We
found patients with atrial fibrillation were more
likely to experience episodes of ventricular
tachycardia, device discharges, and/or death.
Future studies should address whether the use
of ICD event data leads to improved clinical
outcomes.
20
THANK YOU.
J a c k s o n v i l l e