Clinical Predictors of Cardiovascular Implantable
Electronic Device-Related Infective Endocarditis
KATHERINE Y. LE, M.D.,* MUHAMMAD R. SOHAIL, M.D.,†
PAUL A. FRIEDMAN, M.D.,‡ DANIEL Z. USLAN, M.D.,§ STEPHEN S. CHA,¶
DAVID L. HAYES, M.D.,‡ WALTER R. WILSON, M.D.,† JAMES M. STECKELBERG, M.D.,†
LARRY M. BADDOUR, M.D.,† and for the Mayo Cardiovascular Infections Study Group
From the *Mayo School of Graduate Medical Education, Mayo Clinic, Rochester, Minnesota; †Divisions of
Infectious Diseases and ‡Cardiovascular Diseases, Department of Medicine, Mayo Clinic, Rochester, Minnesota;
§Division of Infectious Diseases, Department of Medicine, David Geffen School of Medicine, University of California
Los Angeles, California; and ¶Division of Biostatistics and Informatics, Mayo Clinic, Rochester, Minnesota
Background: Cardiovascular implantable electronic device (CIED)-related infective endocarditis (CIEDIE) is a serious complication of cardiac device infection and is associated with increased mortality. At
present, there exist no criteria to predict CIED-IE in patients who present with CIED infection.
Methods: We retrospectively reviewed all cases of CIED infection seen at Mayo Clinic Rochester between
1991 and 2008. CIED-IE was classified using pathologic and clinical criteria. Clinical predictors of CIED-IE
were identified using logistic regression, and quantified using a summary score and plotted against the
distribution of CIED-IE.
Results: Ninety-three (22.4%) of the 416 patients with CIED infection had CIED-IE. Host factors including
chronic immunomodulator therapy exclusive of corticosteroid (odds ratio [OR], 3.79 [confidence interval
(CI) 1.10, 13.04]), chronic corticosteroid therapy (OR, 2.15 [CI 0.93, 5.00]), hemodialysis (OR, 3.24 [CI
1.39, 7.55]), or remote infection (OR, 1.77 [CI 0.99, 3.14]) were associated with increased odds of CIED-IE.
Patients with CIED-IE were at increased odds of presenting with fever (OR, 3.78 [CI 1.93, 7.40]), or malaise
(OR, 1.87 [CI 1.02, 3.41]), and have findings of leukocytosis (OR, 3.61 [CI 1.51, 8.62]). In marked contrast,
they were at decreased odds of exhibiting signs/symptoms of infection at the generator pocket site (OR,
0.19 [CI 0.10, 0.36]). Summary scores of 6 and 11 predicted CIED-IE in approximately 50% and 90% of
cases, respectively.
Conclusions: Certain comorbid conditions and systemic manifestations of infection were associated
with CIED-IE. In contrast, pocket site infection was negatively associated with CIED-IE. These findings
should assist clinicians in identifying patients who would more likely benefit from further investigation
of CIED-IE with transesophageal echocardiography. (PACE 2011; 34:450–459)
endocarditis, predictors, infection, pacemaker, defibrillator, cardiovascular implantable electronic
device
Introduction
Implantation
of cardiovascular implantable
Disclosures: All < $10,000.
electronic
devices
(CIEDs), that include permaPAF: Honoraria/Consultant: Medtronic, Guidant, Astra Zeneca.
nent pacemakers and implantable cardioverterSponsored research: Medtronic, Astra Zeneca via Beth Israel,
Guidant, St. Jude, Bard.
defibrillators, has rapidly increased in the United
Intellectual property rights: Bard EP, Hewlett Packard,
States over the past two decades.1,2 According
Medical Positioning, Inc.
to one estimate, the implantation rate of CIEDs
DZU: Research: American Heart Association.
increased by 42% among Medicare beneficiaries
Honoraria/Consultant: Biotronik, Cubist, TyRx Pharma, Inc.
from 1990 to 1999.3 Unfortunately, increasing
DLH: Honoraria: Medtronic, Boston Scientific, St. Jude
rates
of CIED infection have also been observed,
Medical, ELA Medical, Biotronik.
and are much higher than expected based on the
Royalty payments: UpToDate; Wiley-Blackwell.
Medical advisory board: Boston Scientific, St. Jude Medical,
implantation rates alone.3,4
Pixel Velocity.
CIED infection can present as a pulseSteering committee member: Medtronic, St. Jude Medical.
generator pocket infection or as bloodstream
LMB: Royalty payments: UpToDate.
Editorship: Massachusetts Medical Society (Journal Watch
Infectious Diseases); ACP/PIER editorial consultant.
MRS: Honoraria/Consultant: TyRx Pharma, Inc.
All other authors: No disclosures.
Address for reprints: Katherine Y. Le, M.D., M.P.H., Mayo
School of Graduate Medical Education, 200 First Street
SW, Rochester, MN 55905. Fax: 507-255-7767; e-mail:
le.katherine1@mayo.edu
Received July 19, 2010; revised October 12, 2010; accepted
October 20, 2010.
doi: 10.1111/j.1540-8159.2010.02991.x
C 2010 Wiley Periodicals, Inc.
C 2010, The Authors. Journal compilation
450
April 2011
PACE, Vol. 34
PREDICTORS OF CARDIAC DEVICE-RELATED ENDOCARDITIS
infection, or both, with or without CIED-related
infective endocarditis (CIED-IE).5,6 Reported rate
of CIED-IE in earlier publications ranges from 10%
to 23% among patients with CIED infection.7,8
The prevailing hypothesis regarding pathogenesis
of CIED-IE includes the initial development
of infection at the device generator site with
subsequent extension of infection to involve the
leads. Hematogenous seeding of the device from
a distant infectious focus is thought to be a less
common occurrence.9
In most cases, a diagnosis of CIED-IE requires
transesophageal echocardiogram (TEE) to detect
infected vegetations on electrode leads and/or
cardiac valves. Transthoracic echocardiogram
(TTE) has limited sensitivity (30% to 40%) when
compared to TEE (90% to 95%) in detecting
vegetations in adults.5,8,10–12 However, TEE is an
invasive procedure with potential complications
from conscious sedation, insertion, and manipulation of the TEE probe, and is not readily
available at some medical centers.13 For these
reasons, it is neither practical nor desirable for
every patient who presents with CIED infection
to undergo investigation with TEE. Consequently,
it is imperative that patients who are more likely
to have CIED-IE be identified so that timely
investigation with TEE is done.
Prompt identification of CIED-IE has important therapeutic and prognostic implications.
Experts have recommended that patients with
CIED-IE that involves cardiac valves receive
early and complete explantation of the infected
device with a prolonged duration of antimicrobial
therapy.6,14,15 In addition, prompt and appropriate
management of CIED-IE is crucial because patients
with this complication can have worse outcomes.8
A systematic analysis to identify clinical
predictors associated with CIED-IE has not been
published. Moreover, earlier investigations that
addressed the subject of CIED-IE have used
modified Duke criteria16 that were originally
intended to classify cases of valvular endocarditis
and has limitations when applied to CIED-IE cases.
The aim of our investigation was to identify
patients who were more likely to have CIED-IE,
so that optimal management, which includes the
selective use of TEE, is achieved.
Methods
We retrospectively reviewed all cases of
CIED infection at Mayo Clinic Rochester (MCR)
between January 1, 1991, and December 31, 2008.
Cases were identified in the Mayo Clinic Heart
Rhythm Device Database, the surgical index, and
the computerized central diagnostic index. All
patients consented to use of their medical records
for research purposes. The Mayo Foundation
PACE, Vol. 34
Institutional Review Board approved the study
protocol.
Definitions
CIED infection was previously defined by our
group6,8 as clinical evidence of device infection
at the generator pocket, or microbiologically
confirmed device infection based on positive
cultures from the generator pocket, lead(s), or
blood (in the presence of local inflammatory
signs at generator pocket or absence of another
source of bloodstream infection and resolution of
bloodstream infection after device explantation).
CIED-IE cases were classified using pathologic
evidence of CIED lead or valve infection for
patients who underwent surgical removal, or
using a clinical criterion for a patient who
underwent percutaneous device removal. Clinically, patients were classified as having CIEDIE if they had positive echocardiographic findings and two or more positive blood cultures
for typical skin organisms (coagulase-negative
staphylococci [CoNS], Corynebacterium species,
Propionibacterium species), or one positive blood
culture for all other microorganisms. Positive
echocardiographic findings for CIED-IE were
defined as presence of an oscillating intracardiac
mass on cardiac valve or supporting structures (in
the path of regurgitant jets), or CIED leads in the
absence of an alternative anatomic explanation,
or visualization of a cardiac abscess, or new
dehiscence of prosthetic valve. Patients with
CIED-IE as defined by the clinical criteria were
further classified as having left-sided CIED-IE if
they had any positive left-sided echocardiographic
findings, and those having right-sided CIEDIE if they had positive right-sided valvular or
lead-associated echocardiographic findings. The
remaining patients were rejected as having CIEDIE.
Demographic and Device-Related Variables
Demographic variables included the patient’s
age in years, gender, and race/ethnicity. Body
mass index (BMI) was calculated as kilograms
per meter squares. Device characteristics included
device type, device age, number of procedures, the
indication for CIED implantation, procedure type,
and anatomical site of implantation.
Comorbid Conditions and Clinical Features
at Presentation
Patients’ medical records were reviewed
for comorbid conditions, presenting systemic
signs/symptoms, inflammatory signs at the generator pocket site, and laboratory parameters for
potential predictors of CIED-IE.
April 2011
451
LE, ET AL.
Statistical Analyses
Statistical analyses were conducted using
the SAS version 9.1.3 software (SAS Institute
Inc., Cary, NC, USA). Bivariate comparisons were
performed using Student’s t-test for continuous
and Pearson χ 2 or Fisher’s Exact test for categorical
variables. Clinical predictors of CIED-IE were
first identified using univariate logistic regression
and later summarized by multivariate logistic
regression using stepwise elimination.
Using the findings from our multivariate
logistic regression models, we then devised a
scoring system containing both statistically and
clinically meaningful predictors to better quantify
their individual and cumulative effects on the
probability of CIED-IE. The resulting distribution
of the summary score was plotted against the
distribution of CIED-IE cases.
To evaluate our CIED-IE classification
scheme’s ability to discriminate outcomes, survival curves were generated. For the survival
analysis, we considered two primary end points:
infection-related death during index hospitalization and all-cause mortality. For infection-related
death, survival was considered from time of
presentation to MCR to time of hospital discharge,
and data were censored at the time of infectionrelated death. For all-cause mortality, survival
was considered from time of presentation to
MCR to time of last follow-up, and data were
censored at time of death. Kaplan-Meier survival
curves were plotted and log-rank test was used
to determine the univariate significance of CIEDIE on outcome of interest. A two-sided P value of
0.05 or less was considered to indicate statistical
significance.
Results
We identified 416 patients who met our case
definition for CIED infection (Table I) and elderly
white males were the predominant group within
this cohort. Sixty percent of device-related infections involved permanent pacemakers. The subset
of patients with CIED-IE had devices implanted
for a longer period of time than did the overall
cohort. Device placement indications included
heart block, sinus node disease, or ventricular
arrhythmia in the bulk of patients. Most were
initial implants or system revisions/upgrades and
were located at the original site of implantation on
the left chest wall.
Diagnostic echocardiography was performed
in 82% of patients. Of those who underwent
echocardiography, 22.8% underwent TTE alone,
while 77.1% had TEE (47.2% TEE only, 29.9%
both). An echocardiographic study was obtained
452
in 96.9% of patients who had bloodstream
infection.
Of the 416 patients with CIED infection,
47 underwent surgical (27 immediate and 20
after failed percutaneous extraction), 368 underwent percutaneous device removal, and 21 were
managed conservatively without device removal.
Among patients who underwent immediate surgical device removal, only two were due to concerns
regarding vegetation size. Many of these patients
had additional indications for surgical removal of
devices via median sternotomy. Lead vegetation
specimens were available for 23 of 47 patients
who underwent surgical removal. Of these, 18
demonstrated histopathologic evidence of CIEDIE.
Twenty-four patients who underwent cardiothoracic surgery for lead removal did not have
lead specimens submitted for histopathologic
evaluation. Clinical criteria were used in these 24
cases, the 348 cases that underwent percutaneous
device removal, and the 21 cases that did
not undergo device removal to define CIEDIE. Of these 393 patients, 26 (6.6%) had leftsided valvular IE and 49 (12.4%) had right-sided
valvular IE or had lead-associated IE (Fig. 1). When
combined with pathologic criteria, 93 (22.4%)
were accepted and 323 (77.6%) were rejected cases
of CIED-IE (Fig. 1, Table II).
CoNS (38.2%) and Staphylococcus aureus (31.0%) caused the large majority of
CIED infections; gram-negative bacilli and nonstaphylococcal gram-positive organisms were less
commonly identified. Patients with CIED-IE were
at increased odds of infection by S. aureus (43.0%
vs 27.5%), or gram-negative bacilli (11.8% vs
5.8%). There was no difference in the distribution
of CoNS between the two groups (Fig. 2).
In the univariate setting, host factors including chronic immunomodulator therapy exclusive of corticosteroid (>1 month prior to
onset of infection), chronic corticosteroid therapy
(>1 month prior to onset of infection), hemodialysis, implanted central venous catheter, and the
presence of remote focus of primary infection
were positively associated with CIED-IE (Table II).
Similarly the presenting systemic signs/symptoms
of fever, tachycardia, chills, diaphoresis, malaise,
nausea, anorexia, hypotension, congestive heart
failure symptoms, and metastatic infectious foci
were also associated with CIED-IE. In contrast,
inflammatory signs/symptoms at the generator
pocket site were less commonly associated with
CIED-IE. Of the laboratory parameters, leukocytosis, anemia, elevated serum creatinine, and
high erythrocyte sedimentation rate (ESR) were
positively associated with diagnosis of CIED-IE
(Table II).
April 2011
PACE, Vol. 34
PREDICTORS OF CARDIAC DEVICE-RELATED ENDOCARDITIS
Table I.
Comparison of Patient Demographics and CIED Features in Cases with and Without Complicating Infective
Endocarditis*
Demographics
Age (years), mean ± standard deviation (SD)
Male
Race/ethnicity (white)
BMI, mean ± SD
Device characteristics
Permanent pacemakers
Single chamber
Implanted at MCR
Age of device (years), mean ± SD
Number of procedures, mean ± SD
Indication for CIED implantation
Heart block
Sinus node disease
Ventricular arrhythmia
Hypertrophic cardiomyopathy
Supraventricular tachycardia
Syncope
Other
Procedure type
Initial implant
System revision/upgrade
Lead revision/insertion
Generator replacement
Other
Site of generator
At original site of implantation†
Left chest
CIED-IE
(n = 93)
CIED-I
(n = 323)
P-Value
67.6 ± 13.5
65 (69.8)
85 (91.4)
29.3 ± 5.9
69.2 ± 15.4
249 (77.0)
292 (90.4)
28.2 ± 6.2
0.334
0.155
0.771
0.128
61 (65.5)
25 (26.8)
26 (27.9)
3.0 ± 2.8
1.9 ± 1.1
196 (60.6)
65 (20.1)
106 (32.8)
2.1 ± 2.8
2.1 ± 1.3
0.390
0.163
0.374
0.023
0.163
0.682
30 (32.2)
29 (31.1)
20 (21.5)
0 (0)
0 (0)
2 (2.1)
12 (12.9)
97 (30.0)
89 (27.5)
83 (25.7)
4 (1.2)
2 (0.6)
14 (4.3)
34 (10.5)
44 (49.4)
19 (21.3)
7 (7.8)
19 (21.3)
4 (4.3)
131 (42.8)
68 (22.2)
24 (7.8)
80 (26.1)
20 (6.1)
62 (69.6)
27 (51.9)
223 (70.5)
113 (58.8)
0.696
0.868
0.395
*All values are expressed as number (percentage), unless indicated otherwise.
† Most recent device-related procedure before presenting with CIED infection.
After adjusting for age, gender, race/ethnicity,
and BMI, we incorporated significant associations demonstrated by univariate analyses of
comorbid conditions, systemic signs/symptoms,
and laboratory parameters in multivariable logistic regression models. The odds of CIEDIE remained elevated among patients receiving
chronic immunomodulator therapy exclusive of
corticosteroid (odds ratio [OR] 3.79 [1.10, 13.04]),
chronic corticosteroid therapy (OR 2.15 [0.93,
5.00]), hemodialysis (OR 3.24 [1.39, 7.55]), or had
remote infection (OR 1.77 [0.99, 3.14]). These
patients were at increased odds of presenting with
systemic signs/symptoms of fever (OR 3.78 [1.93,
7.40]), or malaise (OR 1.87 [1.02, 3.41]), but in
the absence of infectious signs/symptoms at the
generator pocket site (OR 0.19 [0.10, 0.36]). The
most meaningful laboratory parameter remained
leukocytosis (OR 3.61 [1.51, 8.62]) (Table III).
PACE, Vol. 34
The predictive summary score reflecting the
negative and positive effects of significant host
factors, presenting sign/symptoms and laboratory
predictors on CIED-IE diagnosis, ranged from −5
to +21 (Table III). When plotted against the
distribution of CIED-IE, at a summary score of
6, approximately 50% of patients had CIED-IE. A
sigmoidal relationship was seen with 75% of cases
classified as CIED-IE at a summary score of 9, and
90% at a summary score of 11 (Fig. 3).
Patients were hospitalized for an average
of 17 days, during which time 5.7% of them
experienced CIED infection-related death. After
hospital discharge, patients were followed for
an average of 2 years. At the time of last
follow-up, the all-cause mortality rate was 33.5%.
Survival curves for those classified as having leftsided valvular IE and right-sided valvular IE or
lead-associated CIED-IE were similar. Likewise,
April 2011
453
LE, ET AL.
CIED infection
N=416
Immediate surgical extraction
n=27
Percutaneous extraction
n=368
No device removal
n=21
(Clinical criteria)
Surgical extraction after failed
percutaneous extraction
n=20
Successful
percutaneous extraction
n=348
Lead specimen
n=23
No lead specimen
n=24
(Pathologic criteria)
(Clinical criteria)
(Clinical criteria)
Left-sided valvular IE
n=26
Lead-associated
CIED-IE
n=18
Right-sided valvular IE,
or lead-associated CIED-IE
n=49
No lead-associated
CIED-IE
n=5
Rejected cases of
CIED-IE
n=318
Figure 1. Schema outlining device removal technique, availability of a histopathologic specimen,
and criteria applied in cases of CIED infection.
survival was similar among patients with implantable cardioverter defibrillators when compared to those with permanent pacemakers. Both
in-hospital (P = 0.005) and overall survival (P =
0.025) were higher among patients who did not
have CIED-IE (Figs. 4 and 5).
We also examined survival in the 75 patients
who did not undergo echocardiography as part
of the diagnostic evaluation for CIED infection
during hospital admission and observed no
significant differences in index hospitalization
(96.0% vs 92.0%, P = 0.235) or overall survival
(73.3% vs 64.5%, P = 0.144) among them as
compared to that of patients who underwent
echocardiography.
Discussion
The current investigation includes one of
the largest cohorts of CIED-IE cases analyzed to
date, and offers several key contributions to a
contemporary clinical profile of CIED infection.
First, we propose clinical criteria to better
define CIED infection cases that are complicated
by CIED-IE, and examine short- and long-term
patient outcomes under this classification scheme.
Second, we identify clinical predictors of CIED-IE,
and develop a summary score to quantify these
predictors in our study population. Third, our
predictive model may provide valuable insights
into CIED-IE disease pathogenesis.
454
While our CIED-IE rate of 22.4% appears similar to reported rates ranging from 10% to 23%,7,8,14
meaningful comparisons across study populations
have been difficult to perform due to lack of a
uniform CIED-IE classification scheme. Previous
investigators have utilized either the original17
or modified Duke criteria16 to classify CIED-IE.
However, Duke criteria were not developed to define CIED-IE17,18 and alternative criteria designed
specifically for CIED infection are needed.
The majority of infected CIED leads in contemporary practice are removed percutaneously.
During percutaneous device removal, vegetations
usually become dislodged from the lead19–21 and
are not available for histopathologic confirmation
of CIED-IE. This is in contrast to valvular
endocarditis where surgically resected tissue
from valves, or less often emboli, are available
for histopathologic examination6,8 and can be
used to validate proposed clinical criteria. Some
investigators have attempted to overcome this
limitation by using positive lead tip cultures (as
equivalent of valve tissue cultures in valvular
endocarditis) to define CIED-IE.8,10,22,23 However,
lead tips can become contaminated when extracted through infected generator pockets8 and
subsequent cultures may only be reliable when
leads are removed via other techniques that are
less frequently utilized for lead removal, such as
via femoral vein incision24 or median sternotomy.
April 2011
PACE, Vol. 34
PREDICTORS OF CARDIAC DEVICE-RELATED ENDOCARDITIS
Table II.
Univariate Comparisons of Comorbid Conditions and CIED Features in Cases with and Without Complicating
Endocarditis*
CIED-IE
(n = 93)
CIED-I
(n = 323)
Comorbid conditions
Active malignancy
12 (12.9)
39 (12.0)
Autoimmune disease
10 (10.7)
17 (5.2)
Immunomodulator therapy
8 (8.6)
5 (1.5)
Corticosteroid therapy
13 (13.9)
17 (5.2)
Splenectomy
3 (3.2)
4 (1.2)
Diabetes mellitus
29 (31.1)
86 (26.6)
Hemodialysis
13 (13.9)
13 (4.0)
CAD
52 (55.9)
185 (57.2)
Status-post CABG
26 (27.9)
91 (28.1)
CHF
59 (63.4)
175 (54.1)
EF
44.7 ± 16.9
42.5 ± 16.6
Atrial fibrillation
31 (33.3)
120 (37.1)
COPD
19 (20.4)
48 (14.8)
Cirrhosis
1 (1.0)
3 (0.9)
Chronic skin condition
8 (8.6)
22 (6.8)
Implanted CVC
11 (11.8)
16 (4.9)
Prosthetic heart valve
15 (16.1)
46 (14.2)
Vascular graft
3 (3.2)
6 (1.8)
Previous device infection
11 (11.8)
37 (11.4)
Remote infection
26 (27.9)
51 (15.7)
Presenting signs/symptoms – systemic
Fever
74 (79.5)
94 (29.1)
Tachycardia
15 (16.1)
18 (5.5)
Chills
54 (58.0)
73 (22.6)
Diaphoresis
25 (26.8)
32 (9.9)
Malaise
57 (61.2)
79 (24.4)
Nausea
15 (16.1)
15 (4.6)
Anorexia
19 (20.4)
30 (9.2)
Hypotension
21 (22.5)
25 (7.7)
CHF symptoms
33 (35.4)
78 (24.1)
Metastatic infectious foci
11 (11.8)
9 (2.7)
Presenting signs/symptoms – inflammatory signs at generator pocket site
Erythema
20 (21.5)
190 (58.8)
Pain
16 (17.2)
103 (31.8)
Swelling
19 (20.4)
169 (52.3)
Warmth
14 (15.0)
73 (22.6)
Tenderness
15 (16.1)
119 (36.8)
Drainage
2 (6.0)
16 (11.5)
Purulent drainage
11 (11.8)
88 (27.2)
Skin ulceration
7 (7.5)
61 (18.8)
Erosion of lead/generator
3 (3.2)
91 (28.1)
Intraoperative signs of infection
30 (32.6)
225 (70.0)
Laboratory parameters
13.7 ± 7.9
9.2 ± 4.6
WBC (×109 cells/L), mean ± SD
Leukocytosis
57 (61.2)
108 (33.4)
Hematocrit (%), mean ± SD
34.8 ± 5.5
36.9 ± 5.8
Anemia
70 (75.2)
195 (60.3)
OR [95% CI]
P-Value
1.07 [0.54, 2.15]
2.16 [0.95, 4.91]
5.98 [1.90, 18.75]
2.92 [1.36, 6.27]
2.65 [0.58, 12.09]
1.24 [0.75, 2.06]
3.87 [1.72, 8.68]
0.94 [0.59, 1.50]
0.98 [0.59, 1.65]
1.46 [0.91, 2.36]
1.01 [0.99, 1.02]
0.84 [0.52, 1.37]
1.47 [0.81, 2.65]
1.15 [0.11, 11.27]
1.28 [0.55, 2.99]
2.57 [1.15, 5.75]
1.15 [0.61, 2.18]
1.76 [0.43, 7.18]
1.03 [0.50, 2.12]
2.07 [1.20, 3.56]
0.830
0.063
0.002
0.005
0.206
0.387
0.001
0.815
0.967
0.113
0.286
0.500
0.199
0.898
0.557
0.021
0.650
0.429
0.920
0.008
9.48 [5.42, 16.58]
3.25 [1.57, 6.75]
4.74 [2.91, 7.72]
3.34 [1.86, 6.00]
4.89 [3.00, 7.96]
3.94 [1.85, 8.42]
2.50 [1.33, 4.70]
3.47 [1.84, 6.55]
1.72 [1.05, 2.83]
4.68 [1.87, 11.67]
<0.001
0.001
<0.001
<0.001
<0.001
<0.001
0.004
<0.001
0.030
<0.001
0.19 [0.11, 0.33]
0.44 [0.24, 0.79]
0.23 [0.13, 0.40]
0.60 [0.32, 1.13]
0.33 [0.18, 0.59]
0.49 [0.10, 2.25]
0.35 [0.18, 0.70]
0.35 [0.15, 0.79]
0.08 [0.02, 0.27]
0.20 [0.12, 0.33]
<0.001
0.006
<0.001
0.117
<0.001
0.360
0.002
0.011
<0.001
<0.001
1.13 [1.08, 1.18]
3.15 [1.95, 5.07]
0.93 [0.89, 0.97]
1.99 [1.18, 3.36]
<0.001
<0.001
0.002
0.009
Continued.
PACE, Vol. 34
April 2011
455
LE, ET AL.
Table II.
Continued.
Creatinine (mg/dL), mean ± SD
ESR (mm/h), mean ± SD†
CIED-IE
(n = 93)
CIED-I
(n = 323)
OR [95% CI]
1.8 ± 1.5
49.5 ± 31.6
1.4 ± 0.7
32.6 ± 28.2
1.34 [1.09, 1.65]
1.01 [1.00, 1.03]
P-Value
0.004
0.004
*All values are expressed as number (percentage) unless indicated otherwise.
CAD = coronary artery disease; CABG = coronary artery bypass grafting; CHF = congestive heart failure; EF = ejection fraction;
COPD = chronic obstructive pulmonary disease; CVC = central venous catheter; WBC = white blood cell count; ESR = erythrocyte
sedimentation rate; SD = standard deviation.
Immunomodulator therapy (for >1 month preceding CIED infection, exclusive of corticosteroid); corticosteroid therapy (prednisone use
for >1 month preceding CIED infection); chronic skin condition (dermatitis, psoriasis, pressure ulcers); remote infection (concomitant or
preceding infection at a distant site such as skin and soft tissue infection, deep abscess, dental infection); hypotension (systolic blood
pressure <90 or diastolic blood pressure <60 mmHg); CHF symptoms (diagnosed by a physician); leukocytosis (peripheral WBC count
>10×109 cells/L); anemia (hematocrit <40% for men and 36% for women).
† ESR n = 36 for CIED-IE; n = 116 for CIED-I.
Therefore, we developed clinical criteria
consisting of bloodstream infection and echocardiographic findings to define cases of CIEDIE. Indirect support for the accuracy of our
classification scheme is evident in short- and longterm survival data. Patients who were classified
as having CIED-IE had decreased survival, both
during the index hospitalization period and
overall, when compared to patients who were
rejected as cases of CIED-IE (Figs. 4 and 5).
Having established the clinical relevance of
our diagnostic criteria, we next sought to identify
clinical predictors for CIED-IE. Host factors
including receipt of immunomodulator therapy,
corticosteroid therapy, hemodialysis, or presence
of remote primary sources of infection increased
the odds of having CIED-IE. Patients with these
risk factors, particularly when presenting with
50
p=0.004
Distribution percentage (%)
CIED-IE
CIED-I
Overall p<0.001
p=0.271
40
30
p=0.002
20
p=0.050
10
0
CoNS
Staph aureus
GNB
Others
Organism
Figure 2. Distribution of pathogens in cases of CIED
infection and CIED-IE.
456
fever, malaise, or with leukocytosis, should undergo early evaluation for endocarditis with TEE.
Our data are supported by a recent investigation
in which the absence of fever, leukocytosis, and
elevated ESR had a negative predictive value
of 100% in ruling out non-staphylococcal CIEDIE.25 Although complete removal of an infected
CIED system is a requisite for cure, regardless
of the clinical manifestations of infection (pocket
infection vs CIED-IE), patients with device-related
endocarditis have higher in-hospital mortality and
the diagnosis should prompt urgent removal of
the infected device and longer duration of antimicrobial therapy to achieve cure.6,15,26 Expedited
diagnosis and appropriate management is crucial
due to the decreased in-hospital and overall
survival in this group of patients (Figs. 4 and 5).
When these clinical predictors were quantified alone and through a summary predictive
score and plotted against the distribution of CIEDIE cases, the contribution of each variable was
remarkable. At relatively low predictive summary
scores of 6, 9, and 11, substantial portions (50%,
75%, and 90%) of the study population had
CIED-IE.
Our work also prompts a reevaluation of the
current understanding of CIED-IE pathogenesis.
At present, direct extension of infection from the
generator pocket to CIED leads and/or cardiac
valves27–30 is thought to be operative in most CIED
infection cases. However, the low likelihood of
CIED-IE among patients with clinical findings of
pocket site infection was striking and suggests that
novel pathogenic mechanisms, possibly related
to hematogenous seeding, may be operative.
Moreover, the finding of a negative association
may be extremely valuable in defining patient
April 2011
PACE, Vol. 34
PREDICTORS OF CARDIAC DEVICE-RELATED ENDOCARDITIS
Table III.
Multivariable Logistic Regression Models Predicting CIED-IE (Adjusted for Age, Gender, Race/Ethnicity, and BMI)
OR [95% CI]
Model 1—Comorbid conditions
Immunomodulator therapy
Corticosteroid therapy
Hemodialysis
Remote infection
3.79 [1.10, 13.04]
2.15 [0.93, 5.00]
3.24 [1.39, 7.55]
1.77 [0.99, 3.14]
Model 2—Presenting signs/symptoms
Fever
Malaise
Metastatic infectious foci
Signs/symptoms at generator pocket
Model 3—Laboratory parameters
Leukocytosis
Anemia
Serum creatinine (mg/dL)
ESR (mm/h)
+4
+2
+3
+2
3.78 [1.93, 7.40]
1.87 [1.02, 3.41]
1.92 [0.65, 5.67]
0.19 [0.10, 0.36]
<0.001
0.041
0.234
<0.001
+4
+2
3.61 [1.51, 8.62]
2.29 [0.78, 6.67]
0.92 [0.61, 1.40]
1.01 [0.99, 1.02]
0.003
0.127
0.717
0.323
100
CIED-IE (%)
75
50
25
0
-1
1
3
5
7
9
11
13
15
17
19
Predictor summary score
Figure 3. Association between predictor summary score
and probability of CIED-IE diagnosis.
PACE, Vol. 34
−5
+4
−5–21
management schemes that include the optimal use
of TEE.
Our observations support limited preliminary
data that described discordant findings between
signs of generator pocket infection and echocardiographic evidence of lead infection among
patients with S. aureus bacteremia,31,32 and observations that implicate soft tissue infections and
intravascular catheters as sources of bloodstream
infection that resulted in hematogenous seeding of
device lead(s) and cardiac valves(s).32
Our investigation has several limitations
inherent in retrospective study designs, including
biases. We attempted to minimize bias by using
objective criteria and standardized definitions to
-3
Predictor Score
0.034
0.073
0.006
0.050
Predictor summary score
-5
P-Value
categorize cases. However, lack of a uniform and
consistent case definition and diagnostic criteria
for CIED-IE in earlier publications limits our
ability to make direct comparisons. No patient was
excluded based on age, gender, ethnicity, or severity of illness. Laboratory and echocardiographic
investigations were not uniformly applied in all
cases, but were conducted at the discretion of
the primary physician. However, our subgroup
analysis suggests that this had a nondifferential
impact on short- and long-term patient outcomes.
As a tertiary care center, referral bias could
have impacted our clinical data. Patients in our
cohort, for example, could have more comorbid
conditions as compared to that in the general
population. Thus, selection biases might limit the
generalization of our findings to other populations.
We believe, however, that many of the factors
related to the development of CIED infection
are operative in all patients, and this notion
is supported by findings from other medical
centers.9,31,32 Nevertheless, validation of both
our modified CIED-IE case definition and the
predictive scoring system in cohorts at other
institutions is needed.
In conclusion, we defined CIED-IE cases
by using clinical criteria that accommodated
the uniqueness of lead-related endocarditis; in
addition, a scoring system was developed that
predicted patients with CIED-IE. Appropriate and
judicious use of invasive diagnostic procedures,
such as TEE, is more relevant than ever with the
current focus on cost-effectiveness and optimal
April 2011
457
LE, ET AL.
100
CIED-I
Hospital Survival (%)
80
CIED-IE
60
40
20
Log Rank p=0.005
0
0
20
40
60
80
100
Time (days)
Figure 4. Unadjusted CIED infection-related in-hospital survival curves.
utilization of health-care resources. Our investigation, which included outcomes data, suggests
that TEE may not be needed in patients with
negative blood cultures and clinical evidence of
CIED infection limited to the generator pocket
site. Moreover, the negative correlation between
pocket site changes and likelihood of CIED-IE
challenges our current dogma that the generator
pocket site serves as the predominant source
of lead infection and prompts us to reevaluate
alternative mechanisms of infection pathogenesis.
The results of our investigation, however, should
prompt additional study to define the optimal use
of TEE in the management of CIED infection. At
present, it would be premature to provide specific
recommendations in the management of CIED
infection based on the findings demonstrated in
our current investigation.
100
Overall Survival (%)
80
60
CIED-I
CIED-IE
40
20
Log Rank p=0.025
0
0
2
4
6
8
10
Time (years)
Figure 5. Unadjusted overall survival curves at the time of last follow-up.
458
April 2011
PACE, Vol. 34
PREDICTORS OF CARDIAC DEVICE-RELATED ENDOCARDITIS
Acknowledgments:
This investigation utilized resources of the Mayo Cardiovascular Infections Study Group
(Le KY, Sohail MR, Baddour LM, Steckelberg JM, Wilson WR,
Enzler MJ, Chung HH, Friedman PA, Hayes DL, Anavekar
NS, Nkomo VT, Dib C, Madhavan M, Thomas JM, Raza SS,
Correa de Sa DD, Bachuwar A, Sultan OW, Abou Ezzeddine
OF, Habib A, Patel R, Williamson EE, Kalra M, Edwards
WD, Maleszewski JJ [Mayo Clinic Rochester]; Hellinger WC,
Lynady CJ, Kusumoto F [Mayo Clinic Florida]; Virkram HR,
and Keckich DW [Mayo Clinic Arizona]).
This study was supported, in part, by “Small Grants”
award from the Division of Infectious Diseases, Department
of Medicine, Mayo Clinic College of Medicine.
We wish to thank Joanne E. Spencer for her important
contribution in data collection.
References
1. Goldberger Z, Lampert R. Implantable cardioverter-defibrillators:
Expanding indications and technologies. JAMA 2006; 295:809–818.
2. Birnie D, Williams K, Guo A. Reasons for escalating pacemaker
implants. Am J Cardiol 2006; 98:93–97.
3. Cabell CH, Heidenreich PA, Chu VH, Moore CM, Stryjewski
ME, Corey GR, Fowler VG Jr. Increasing rates of cardiac device
infections among Medicare beneficiaries: 1990–1999. Am Heart J
2004; 147:582–586.
4. Uslan DZ, Tleyjeh IM, Baddour LM, Friedman PA, Jenkins SM,
St Sauver JL, Hayes DL. Temporal trends in permanent pacemaker
implantation: A population-based study. Am Heart J 2008; 155:896–
903.
5. Klug D, Balde M, Pavin D, Hidden-Lucet F, Clementy J, Sadoul
N, Rey JL, et al.; PEOPLE Study Group. Risk factors related to
infections of implanted pacemakers and cardioverter-defibrillators:
Results of a large prospective study. Circulation 2007; 116:1349–
1355.
6. Sohail MR, Uslan DZ, Khan AH, Friedman PA, Hayes DL, Wilson
WR, Steckelberg JM, et al. Management and outcome of permanent
pacemaker and implantable cardioverter-defibrillator infections.
J Am Coll Cardiol 2007; 49:1851–1859.
7. Arber N, Pras E, Copperman Y, Schapiro JM, Meiner V, Lossos IS,
Militianu A, et al. Pacemaker endocarditis. Report of 44 cases and
review of the literature. Medicine 1994; 73:299–305.
8. Sohail MR, Uslan DZ, Khan AH, Friedman PA, Hayes DL, Wilson
WR, Steckelberg JM, et al. Infective endocarditis complicating
permanent pacemaker and implantable cardioverter-defibrillator
infection. Mayo Clin Proc 2008; 83:46–53.
9. Camus C, Leport C, Raffi F, Michelet C, Cartier F, Vilde JL. Sustained
bacteremia in 26 patients with a permanent endocardial pacemaker:
Assessment of wire removal. Clin Infect Dis 1993; 17:46–55.
10. Cacoub P, Leprince P, Nataf P, Hausfater P, Dorent R, Wechsler B,
Bors V, et al. Pacemaker infective endocarditis. Am J Cardiol 1998;
82:480–484.
11. Victor F, De Place C, Camus C, Le Breton H, Leclercq C, Pavin D,
Mabo P, et al. Pacemaker lead infection: Echocardiographic features,
management, and outcome. Heart 1999; 81:82–87.
12. Vilacosta I, Sarriá C, San Román JA, Jiménez J, Castillo JA, Iturralde
E, Rollán MJ, et al. Usefulness of transesophageal echocardiography
for diagnosis of infected transvenous permanent pacemakers.
Circulation 1994; 89:2684–2687.
13. Peterson GE, Brickner ME, Reimold SC. Transesophageal echocardiography: Clinical indications and applications. Circulation 2003;
107:2398–2402.
14. Chua JD, Wilkoff BL, Lee I, Juratli N, Longworth DL, Gordon
SM. Diagnosis and mangement of infections involving implantable
electrophysiologic cardiac devices. Ann Intern Med 2000; 133:604–
608.
15. Baddour LM, Epstein AE, Erickson CC, Knight BP, Levison
ME, Lockhart PB, Masoudi FA, et al. Update on cardiovascular
implantable electronic device infections and their management.
A scientific statement from the American Heart Association.
Circulation 2010; 1:1–20.
16. Li JS, Sexton DJ, Mick N, Nettles R, Fowler VG Jr, Ryan T, Bashore T,
et al. Proposed modifications to the Duke criteria for the diagnosis
of infective endocarditis. Clin Infect Dis 2000; 30:633–638.
17. Durack DT, Lukes AS, Bright DK. New criteria for diagnosis of
PACE, Vol. 34
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
April 2011
infective endocarditis: Utilization of specific echocardiographic
findings. Duke Endocarditis Service. Am J Med 1994; 96:200–
209.
Klug D, Wallet F, Kacet S, Courcol RJ. Detailed bacteriologic tests to
identify the origin of transvenous pacing system infections indicate
a high prevalence of multiple organisms. Am Heart J 2005; 149:322–
328.
Novaro GM, Saliba W, Jaber WA. Images in cardiovascular
medicine. Fate of intracardiac lead vegetations after percutaneous
lead extraction. Circulation 2002; 106:e46.
Rizzello V, Dello Russo A, Casella M, Biddau R. Residual fibrous
tissue floating in the right atrium after percutaneous pacemaker lead
extraction: An unusual complication early detected by intracardiac
echocardiography. Int J Cardiol 2008; 127:e67–e68.
Thuny F, Le Dolley Y, Mancini J, Casalta J, Gouriet F, Riberi A,
Avienrinos J, et al. Ghost of infected leads: A new critrion of
cardiac device-rated infective endocarditis. Poster presented at the
10th International Symposium on Modern Concepts in Endocarditis
and Cardiovascular Infections, April 26–28, 2009, Naples, Italy,
2008.
Massoure PL, Reuter S, Lafitte S, Laborderie J, Bordachard
P, Clementy J, Roudaut R. Pacemaker endocarditis: Clinical
features and management of 60 consecutive cases. Pacing Clin
Electrophysiol 2007; 30:12–19.
Sohail MR. Concerning diagnosis and management of pacemaker
endocarditis. Pacing Clin Electrophysiol 2007; 30:829.
Jarwe M, Klug D, Beregi JP, Le Franc P, Lacroix D, Kouakam C,
Guédon-Moreau L, et al. Single center experience with femoral
extraction of permanent endocardial pacing leads. Pacing Clin
Electrophysiol 1999; 22:1202–1209.
Viola GM, Awan LL, Darouiche RO. Nonstaphylococcal infections
of cardiac implantable electronic devices. Circulation 2010;
121:2085–2091.
Darouiche RO. Treatment of infections associated with surgical
implants. N Engl J Med 2004; 350:1422–1429.
Bluhm, G. Pacemaker infections. A clinical study with special
reference to prophylactic use of some isoxazolyl penicillins. Acta
Med Scand Suppl 1985; 699:1–62.
Da Costa A, Lelièvre H, Kirkorian G, Célard M, Chevalier P,
Vandenesch F, Etienne J, et al. Role of the preaxillary flora
in pacemaker infections: A prospective study. Circulation 1998;
97:1791–1795.
Klug D, Wallet F, Lacroix D, Marquié C, Kouakam C, Kacet S,
Courcol R. Local symptoms at the site of pacemaker implantation
indicate latent systemic infection. Heart 2004; 90:882–886.
Golzio PG, Vinci M, Anselmino M, Comoglio C, Rinaldi M, Trevi
GP, Bongiorni MG. Accuracy of swabs, tissue specimens, and
lead samples in diagnosis of cardiac rhythm management device
infections. Pacing Clin Electrophysiol 2009; 32(Suppl 1):S76–S80.
Greenspon AJ, Rhim ES, Mark G, Desimone J, Ho RT. Leadassociated endocarditis: The important role of methicillin-resistant
Staphylococcus aureus. Pacing Clin Electrophysiol 2008; 31:548–
553.
Chamis AL, Peterson GE, Cabell CH, Corey GR, Sorrentino RA,
Greenfield RA, Ryan T, et al. Staphylococcus aureus bacteremia in
patients with permanent pacemakers or implantable cardioverterdefibrillators. Circulation 2001; 104:1029–1033.
459