Heart Rhythm Disorders
Extracellular Matrix Alterations in Patients
With Paroxysmal and Persistent Atrial Fibrillation
Biochemical Assessment of Collagen Type-I Turnover
Eleftherios M. Kallergis, MD,* Emmanuel G. Manios, MD,* Emmanuel M. Kanoupakis, MD,*
Hercules E. Mavrakis, MD,* Dimitris A. Arfanakis, MD,* Niki E. Maliaraki, MD,†
Chrisovalantis E. Lathourakis, MD,* Gregory I. Chlouverakis, PHD,*
Panos E. Vardas, MD, PHD, FESC, FACC*
We investigated whether the serum markers of collagen turnover differed in various forms of atrial fibrillation
(AF) and in sinus rhythm (SR) in humans.
Structural alterations and fibrosis have been implicated in the generation and perpetuation of AF.
Serum C-terminal propeptide of collagen type-I (CICP), C-terminal telopeptide of collagen type-I (CITP), matrix
metalloproteinase-1, and tissue inhibitor of matrix metalloproteinases-1 were measured as markers of collagen
synthesis and degradation in 70 patients with AF and 20 healthy control subjects in SR.
C-terminal propeptide of collagen type-I and CITP were significantly higher in AF patients than in control subjects
(91 ? 27 ng/ml vs. 67 ? 11 ng/ml, p ? 0.001 and 0.38 ? 0.20 ng/ml vs. 0.25 ? 0.08 ng/ml, p ? 0.001, re-
spectively). Persistent AF patients had higher levels of CICP (105 ? 28 ng/ml vs. 80 ? 21 ng/ml, p ? 0.001),
but not CITP, compared with those with paroxysmal AF. Patients with persistent AF had lower levels of matrix
metalloproteinase-1 but increased levels of tissue inhibitor of matrix metalloproteinases-1 compared with pa-
tients with paroxysmal AF (11.90 ? 4.79 ng/ml vs. 14.98 ? 6.28 ng/ml, p ? 0.03 and 155 ? 45 ng/ml vs.
130 ? 38 ng/ml, p ? 0.001, respectively). Tissue inhibitor of matrix metalloproteinases-1 levels were signifi-
cantly lower in control subjects compared with those in both paroxysmal and persistent AF patients (102 ? 15
ng/ml vs. 130 ? 38 ng/ml vs. 155 ? 45 ng/ml, respectively, p ? 0.001).
Serum markers of collagen type-I turnover differed significantly between patients with AF and SR. Furthermore,
these markers also differed significantly between paroxysmal and persistent AF patients, suggesting that the
intensity of the extracellular synthesis and degradation of collagen type-I may be related to the burden or type
of AF. (J Am Coll Cardiol 2008;52:211–5) © 2008 by the American College of Cardiology Foundation
Recent data from cardiac biopsies of patients with atrial
fibrillation (AF) have demonstrated a substantial increase in
collagen deposition in the atria of humans with AF as opposed
to those in sinus rhythm (SR) (1–3). Preliminary experimental
and clinical data suggest that biochemical markers of collagen
turnover correlate significantly with fibrosis in endomyocardial
biopsies in hypertensive patients (4). However, although such
markers have been shown to differ between hypertensive and
normotensive individuals, studies of their association with
hypertrophy or other echocardiographic parameters have pro-
duced inconsistent results (5,6).
Collagen type-I is the major collagenous product of
cardiac fibroblasts (7). We assessed the amount of fibrosis in
patients with paroxysmal or persistent lone AF using serum
analysis of peptides derived from the tissue synthesis (C-
terminal propeptide of collagen type-I [CICP]) and degra-
dation (C-terminal telopeptide of collagen type-I [CITP])
of collagen type-I (4). Enzymes that control collagen type-I
turnover, specifically matrix metalloproteinase (MMP)-1
and tissue inhibitor of matrix metalloproteinases (TIMP)-1,
were also measured (8).
The ethics committee of our institution approved the study,
which conformed to the principles outlined in the Declara-
tion of Helsinki. Informed written consent was obtained
from all subjects.
From the *Department of Cardiology and †Laboratory of Biochemistry, University
Hospital of Heraklion, Crete, Greece.
Manuscript received October 29, 2007; revised manuscript received February 13,
2008, accepted March 18, 2008.
Journal of the American College of Cardiology
© 2008 by the American College of Cardiology Foundation
Published by Elsevier Inc.
Vol. 52, No. 3, 2008
The study population con-
sisted of 70 ambulatory patients
(age 24 to 78 years) who suffered
from lone AF, defined as AF
without clinical or echocardio-
graphic evidence of cardiopul-
monary disease, including hyper-
tension. The arrhythmia was
considered paroxysmal when it
was self-terminating in ?24 h
and persistent when it was doc-
umented on sequential 12-lead
electrocardiograms, without any
intervening periods of SR, for at
least 3 months preceding enroll-
ment. Patients with persistent
AF had been referred to our
department for electrical cardio-
version; none had undergone
such a procedure before. No patient had permanent AF,
defined as sustained arrhythmia despite cardioversion.
Twenty healthy subjects in SR with no history of atrial
arrhythmias served as a control group.
Exclusion criteria were conditions associated with ele-
vated serum concentrations of myocardial or tissue fibrosis
markers, such as liver disease, renal impairment, pulmonary
fibrosis, extensive wounds, metabolic bone disease, malig-
nancy, connective tissue disorders, chronic inflammatory
disease, and recent infection or surgery. Patients who were
over 80 years old or had a pacemaker/implantable
cardioverter-defibrillator were also excluded.
Using a case-control study design, serological markers of
collagen type-I turnover, echocardiographic maximal left
atrial (LA) diameter, and left ventricular ejection fraction
(LVEF) were compared in patients and control subjects.
Diltiazem and beta-blockers were allowed for ventricular
rate control, and all AF patients received the indicated
All patients were in AF at the time of blood sampling.
Blood samples were obtained at the time of the clinical
studies and were immediately placed on ice and centrifuged
within 1 h. Specimens were stored at ?80°C until analysis.
Serum TIMP-1 and MMP-1 levels were assayed by
enzyme-linked immunoadsorbent assay with commercially
available kits (Human Biotrack enzyme-linked immunoad-
sorbent assay system, Amersham Biosciences, Piscataway,
New Jersey). C-terminal propeptide of collagen type-I levels
were determined by a sandwich enzyme immunoassay with
a commercially available kit (enzyme immunoassay, Metra
CICP, Quidel, San Diego, California), while CITP was
measured using the Elecsys ?-CrossLaps/serum assay
(Roche Diagnostics, Mannheim, Germany). Measurements
were performed by personnel blinded to the patients’ clinical
details. The intra- and interassay coefficients of variation of
all assays were ?8% and ?10%, respectively, in our laboratory.
Statistical analysis. Summary descriptive statistics for con-
tinuous parameters are presented as mean ? SD. Categor-
ical variables were compared among the persistent, parox-
ysmal AF, and control groups using the chi-square test.
One-way analysis of variance was used to test whether these
3 groups differed with respect to various continuous param-
eters of interest. Where findings were significant, post-hoc
Tukey tests were performed to pinpoint differences. If the
homogeneity of variance assumption was violated, the non-
parametric Kruskal-Wallis test was used instead. Stepwise
logistic regression analysis was used to assess whether age,
gender, LA, and fibrosis markers were independently asso-
ciated with AF. Age, LA, and fibrosis markers were treated
as continuous variables in the model. The thresholds for
entry into and removal from the model were 5% and 10%,
respectively. Associations between continuous variables of
interest were assessed with correlation and linear regression
techniques. Values of p ? 0.05 were considered statistically
Patients. The baseline clinical and demographic features of
the study population are shown in Table 1. Thirty-two
patients had persistent and 38 paroxysmal AF. There were
no significant differences in gender (p ? 0.40) or age (p ?
0.058) between the AF groups. Control subjects were
younger than AF patients (p ? 0.001), but were comparable
in terms of gender distribution. Patients with persistent AF
had lower LVEF (p ? 0.038) and a larger LA (p ? 0.001)
compared with paroxysmal AF patients.
Serum markers of collagen turnover. The findings are
summarized in Table 1. Both CICP and CITP levels were
significantly higher in persistent AF patients than in control
subjects (p ? 0.016). Persistent AF patients also had higher
CICP levels compared with those with paroxysmal AF (p ?
0.001), whereas there was no significant difference in CITP
levels (p ? 0.57) (Table 1, Figs. 1 and 2).
Patients with persistent AF had lower levels of MMP-1
(p ? 0.026) but higher levels of TIMP-1 (p ? 0.013)
compared with patients with paroxysmal AF (Table 1,
Figs. 3 and 4). Tissue inhibitor of matrix metalloproteinases-1
levels were significantly lower (p ? 0.001) in control subjects
compared with both paroxysmal and persistent AF patients
(Fig. 4). Plasma MMP-1 levels did not differ significantly (p ?
0.20) between persistent AF patients and control subjects
Finally, in all AF patients taken together, there was a
positive correlation between both CICP and TIMP levels
and LA dimension (r ? 0.635, p ? 0.001 and r ? 0.563,
p ? 0.001, respectively) (Fig. 5), while there was a weak,
inverse relation between CICP and TIMP levels and LVEF
(r ? ?0.234, p ? 0.05 and r ? ?0.278, p ? 0.020,
respectively). A positive correlation was also observed be-
tween MMP-1 levels and LVEF (r ? 0.30, p ? 0.012)
while there was an inverse relation between MMP-1 levels
AF ? atrial fibrillation
CICP ? C-terminal
propeptide of collagen
CITP ? C-terminal
telopeptide of collagen
LA ? left atrial
LVEF ? left ventricular
MMP ? matrix
SR ? sinus rhythm
TIMP ? tissue inhibitor of
212Kallergis et al.
Fibrosis and AF
JACC Vol. 52, No. 3, 2008
July 15, 2008:211–5
and LA dimension (r ? ?0.615, p ? 0.001). Logistic
regression analysis showed that, among age, gender, LA and
collagen markers, TIMP-1, MMP-1, age, and CICP were
independently associated with AF (Table 2).
Patients initially presenting with paroxysmal AF often
progress to persistent and eventually develop permanent
AF. Although the exact pathophysiological mechanisms are
poorly understood, persistence of AF is thought to result
from atrial remodeling (9,10). However, pure electrical
remodeling cannot explain the development of sustained AF
(10,11). Atrial fibrosis, a factor with a slower time course,
could be involved (3,12,13).
In this study we demonstrated an elevation in CICP and
CITP in the AF cohort taken as a whole compared with
those in SR. Interestingly, persistent AF patients had the
highest serum concentrations of CICP, whereas there was
no difference in CITP levels between patients with persis-
tent and paroxysmal AF. Thus, CICP demonstrated a
gradual increase from control subjects to paroxysmal and
then to persistent AF, but CITP did not, suggesting that
the intensity of the extracellular degradation of collagen
type-I was insufficient to compensate for its increased
synthesis, resulting in augmented fibrosis in patients with
Furthermore, in patients with persistent AF, MMP-1
levels were decreased, whereas TIMP-1 levels were in-
creased compared with paroxysmal AF patients. Levels of
TIMP-1 were also higher in paroxysmal AF patients than in
control subjects. In addition, control subjects had lower
levels of MMP-1 than paroxysmal AF patients but higher
Baseline Characteristics and Serum Markers of Collagen Type-I Turnoverin Patients With Persistent or Paroxysmal AF and Healthy Control Subjects in SR
Baseline Characteristics and Serum Markers of Collagen Type-I Turnover
in Patients With Persistent or Paroxysmal AF and Healthy Control Subjects in SR
Persistent AF (n ? 32) Paroxysmal AF (n ? 38)Control Subjects (n ? 20) p Value
Age (yrs)65.34 ? 13.3161.79 ? 12.3756.75 ? 5.210.025
Gender (M/F)23/927/1112/8 0.621
LVEF (%)56.25 ? 9.41 60.13 ? 4.42 60.00 ? 4.58 0.227
44.37 ? 4.6237.23 ? 3.6737.15 ? 3.54
27.12 ? 1.7127.02 ? 1.4426.95 ? 1.390.919
SBP (mm Hg) 138.59 ? 12.52 138.55 ? 11.90 131.75 ? 10.030.079
DBP (mm Hg)83.91 ? 4.3584.74 ? 4.3483.75 ? 4.55 0.633
CICP (ng/ml)105.06 ? 27.5779.84 ? 21.2367.3 ? 11.05
CITP (ng/ml) 0.40 ? 0.210.37 ? 0.190.25 ? 0.08 0.016
MMP-1 (ng/ml)11.90 ? 4.789 14.98 ? 6.28 12.63 ? 1.210.033
TIMP-1 (ng/ml) 154.90 ? 44.91 129.75 ? 37.92 102.10 ? 15.13
Values are expressed as mean ? SD. Post-hoc analysis showed that: 1) age and left atrial diameter (LA) differed significantly between control
subjects and persistent atrial fibrillation (AF) patients; 2) C-terminal propeptide of collagen type-I (CICP) levels in persistent AF patients differed
significantly compared with both paroxysmal and control subjects; 3) C-terminal telopeptide of collagen type-I (CITP) levels in control subjects
differed significantly compared with both paroxysmal and persistent AF patients; 4) matrix metalloproteinase (MMP)-1 levels differed significantly
between persistent and paroxysmal AF patients; and 5) tissue inhibitor of matrix metalloproteinase (TIMP)-1 levels differed significantly in all
BMI ? body mass index; DBP ? diastolic blood pressure; LVEF ? left ventricular ejection fraction; SBP ? systolic blood pressure; SR ? sinus
ControlsPersistent AF Paroxysmal AF
Figure 1Comparative Levels of Serum Concentration of CICP
C-terminal propeptide of collagen type-I (CICP) showed a gradual increase from
control subjects to paroxysmal and then to persistent atrial fibrillation (AF)
patients as shown in this box plot.
Controls Persistent AF
Figure 2Box Plot of Serum Concentrations of CITP
C-terminal telopeptide of collagen type-I (CITP) in control subjects differed sig-
nificantly compared with that in both paroxysmal and persistent atrial fibrillation
JACC Vol. 52, No. 3, 2008
July 15, 2008:211–5
Kallergis et al.
Fibrosis and AF
levels than persistent AF patients (although the differences
did not reach statistical significance). That seems paradox-
ical, but it could be the result of the activation manner of
MMP-1, which depends on the nature of the stimulus and
differs in the setting of acute or chronic stimulation (14).
Thus, short-lasting paroxysmal AF could result acutely in
pressure or volume overload, activating the MMP-1 system,
which then becomes compromised with the prolongation
and stabilization of the stimulus.
Another interesting finding was that both CICP and
TIMP levels positively correlated with LA diameter and
inversely related to LVEF, whereas AF patients with a
larger LA and a smaller LVEF, probably as a result of a
longer arrhythmia duration, had lower MMP-1 levels.
Since our study included patients with lone AF only, we
can speculate that the aforementioned alterations were
attributable to the arrhythmia itself, and not to the presence
or absence of any confounding factor, with a progressive
increase in fibrosis from paroxysmal to persistent AF. In
addition, the augmented fibrosis, especially in patients with
the persistent form of AF, could further suggest that if
collagen modification plays a causal role, this relates to both
the initiation and the maintenance of AF.
Study limitations and clinical implications. Serum mark-
ers of collagen turnover are not heart specific. In addition,
we did not support our findings with atrial tissue biopsy data
or coronary sinus sampling. However, we made strenuous
efforts to exclude subjects with conditions associated with
Despite our efforts to match control subjects with AF
patients, the latter were older. However, collagen markers
differed significantly between patients with persistent and
paroxysmal AF, despite the lack of any significant difference
in age, suggesting that age differences did not influence the
interpretation of our results.
Serial measurements of collagen indexes after SR resto-
ration to evaluate the potential temporal alterations of
collagen turnover are not available, although these data
would undoubtedly have been a valuable addition to our
study and could have reinforced our findings.
Paroxysmal AFPersistent AF Controls
Figure 3 Intergroup Serum Concentrations of MMP-1
This box plot depicts matrix metalloproteinase (MMP)-1 levels that differed sig-
nificantly between persistent and paroxysmal atrial fibrillation (AF) patients.
Controls Persistent AFParoxysmal AF
Box Plot Depiction of the
Differences in Levels of Serum TIMP-1
Tissue inhibitor of matrix metalloproteinase (TIMP)-1
levels differed significantly in all intergroup comparisons. AF ? atrial fibrillation.
Direct Correlation Between Serum Concentrations
of CICP and LA Diameter Dimensions
In all AF patients taken together, there was a positive correlation between CICP
and left atrial diameter (LA) dimension (r ? 0.635, p ? 0.001). Abbreviations
as in Figure 1.
Independent Predictors of AF Accordingto Stepwise Logistic Regression Analysis
Independent Predictors of AF According
to Stepwise Logistic Regression Analysis
Variable Odds Ratio95% CI p Value
TIMP-1 1.17 1.06–1.280.002
MMP-1 1.771.28–2.46 0.001
Age 1.171.05–1.32 0.005
CI ? confidence interval; other abbreviations as in Table 1.
214Kallergis et al.
Fibrosis and AF
JACC Vol. 52, No. 3, 2008
July 15, 2008:211–5
Finally, the small patient sample size does not allow
major conclusions regarding the relationship between sys-
temic fibrosis and AF.
Given these limitations, this is the first study to estimate
the relationship between AF burden and serum markers of
collagen type-I synthesis and degradation, demonstrating a
potential role for these markers in AF evaluation.
Serum markers of collagen type-I turnover may provide a
noninvasive method of documenting and monitoring both
the extent and the mechanisms of myocardial fibrosis in AF
patients and of evaluating pharmacological measures de-
signed to treat this arrhythmia. However, further investiga-
tion and randomized trials are needed to elucidate the exact
role of fibrosis in AF and to evaluate the clinical importance
and value of biochemical monitoring of collagen turnover in
this clinical setting. Although cardiac biopsy is the gold
standard for documenting and monitoring myocardial fibro-
sis, noninvasive methods offer a particularly attractive alter-
native that could have a broader application.
Reprint requests and correspondence: Dr. Panos E. Vardas,
Department of Cardiology, Heraklion University Hospital, 71100,
Heraklion, Crete, Greece. E-mail: firstname.lastname@example.org.
1. Frustaci A, Chimenti C, Bellocci F, Morgante E, Russo MA, Maseri
A. Histological substrate of atrial biopsies in patients with lone atrial
fibrillation. Circulation 1997;96:1180–4.
2. Boldt A, Wetzel U, Lauschke J, et al. Fibrosis in left atrial tissue of
patients with atrial fibrillation with and without underlying mitral
valve disease. Heart 2004;90:400–5.
3. Bishop JE. Regulation of cardiovascular collagen deposition by me-
chanical forces. Mol Med Today 1998;4:69–75.
4. Lo ´pez B, Gonza ´lez A, Varo N, Laviades C, Querejeta R, Dı ´ez J.
Biochemical assessment of myocardial fibrosis in hypertensive heart
disease. Hypertension 2001;38:1222–6.
5. Laviades C, Varo N, Fernandez J, et al. Abnormalities of the
extracellular degradation of collagen type-I in essential hypertension.
6. Lindsay MM, Maxwell P, Dunn FG. TIMP-1: a marker of left
ventricular diastolic dysfunction and fibrosis in hypertension. Hyper-
7. Lijnen P, Petrov V, Fagard R. Induction of cardiac fibrosis by
transforming growth factor-b1. Mol Genet Metab 2000;71:418–35.
8. Li YY, McTiernan CF, Feldman AM. Interplay of matrix metallo-
proteinases, tissue inhibitors of metalloproteinases and their regulators
in cardiac matrix remodeling. Cardiovasc Res 2000;46:214–24.
9. Wijffels MC, Kirchhof CJ, Dorland R, Allessie MA. Atrial fibrillation
begets atrial fibrillation. A study in awake chronically instrumented
goats. Circulation 1995;92:1954–68.
10. Allessie M, Ausma J, Schotten U. Electrical, contractile and
structural remodeling during atrial fibrillation. Cardiovasc Res
11. Yu WC, Lee SH, Tai CT, et al. Reversal of atrial electrical remodeling
following cardioversion of long-standing atrial fibrillation in man.
Cardiovasc Res 1999;42:470–6.
12. Kostin S, Klein G, Szalay Z, Hein S, Bauer E, Schaper J. Structural
correlate of atrial fibrillation in human patients. Cardiovasc Res
13. Janse MJ. Why does atrial fibrillation occur? Eur Heart J 1997;
14. Nagatomo Y, Carabello BA, Coker ML, et al. Differential effects of
pressure or volume overload on myocardial MMP levels and inhibitory
control. Am J Physiol 2000;278:H151–61.
Key Words: atrial fibrillation y collagen y fibrosis.
JACC Vol. 52, No. 3, 2008
July 15, 2008:211–5
Kallergis et al.
Fibrosis and AF