Credits:Venkata M Alla MD a, Senthil Thambidorai MD a, Kishlay Anand MD, MS a, Aryan N Mooss MD a, Richard Baltaro MD b, Syed M Mohiuddin MD a.
Division of Cardiology, Department of Medicine, Creighton University, Omaha, Nebraska.
Department of Pathology, Creighton University, Omaha, Nebraska.
Corresponding Address : Venkata M Alla, 3006, Webster Street, Omaha, NE- 68131. USA.
There is increasing evidence linking C-reactive
protein (CRP) and atrial fibrillation (AF). Despite the abundance of
literature, confusion exists regarding this association because of inconsistent
results. MEDLINE and Cochrane Controlled Trials Register databases were carefully
searched through July, 2009 combining the following terms “C-reactive protein”
and “atrial fibrillation”. Reference lists of selected articles and reviews were
also screened to identify additional relevant studies. Of the 129 studies
initially identified, 8 studies with 7507 subjects (719 with AF) were included
in the meta-analysis. Analysis yielded a relative risk of 1.63 (1.43, 1.86) for
occurrence of AF when CRP level was above a cut off of 3-3.5 mg/l. When 3
studies with data on a higher cut off of 4.5-5.0 mg/l were analyzed separately,
the relative risk was 4.03 (3.1, 5.25). Our study suggests that elevated CRP is
associated with increased risk for AF. The risk appears incremental with
higher CRP levels conferring proportionately increased risk. There is an urgent
need for further large scale, well designed prospective studies to assess the
relationship between CRP and AF.
Keywords: Arrhythmia, Atrial Fibrillation, CRP,
Atrial fibrillation (AF) is the most common
arrhythmia encountered in everyday clinical practice and affects approximately
0.9% of the general population .It is associated with
significant cardiovascular morbidity and mortality and also has an adverse
impact on the quality of life . The prevalence of AF
increases with increasing age. With the demographic curve leaning towards the
elderly, the burden imposed by this disease on healthcare systems across the
western world is expected to increase significantly .It is therefore imperative to devise new ways to
prevent, detect and treat this condition. There is growing evidence linking
inflammation to a variety of cardiovascular diseases. C-reactive protein (CRP)
is an excellent marker of inflammation and has been linked to the pathogenesis
and prognosis in patients with coronary artery disease, congestive heart failure,
AF, myocarditis and aortic valve disease . The increasing
body of evidence linking CRP and AF has opened a new door of opportunity in our
understanding of AF and will potentially lead to new ways of managing this
common problem. The association between CRP and AF has been demonstrated in
various settings and has been previously reviewed . The
aim of this study is to systematically review published data on the association
between CRP and AF and study the strength of this association. We used a
meta-analytic approach to estimate the relative risk of AF associated with
MEDLINE and Cochrane Controlled Trials
Register databases were carefully searched through July, 2009 combining the
following terms “C-reactive protein” and “Atrial Fibrillation”. One of the
authors (V.A) screened the studies for potential inclusion. Reference lists of
the identified reports, reviews and letters were also screened to include potentially
relevant studies. Studies were selected for further review after letters,
reviews and irrelevant articles were excluded from the search results [flow
sheet in Figure 1]. The manuscripts of the short listed
studies were then separately reviewed by two of the authors (V.A, S.T) for
inclusion in the systematic review. Studies which fulfilled the following
criteria were included for systematic review. 1. Availability of baseline CRP
levels. 2. Availability of duration of follow up 3. Exclusion of atrial
arrhythmias other than AF. 4. AF diagnosed by a physician based on EKG or
telemetry strip, or by coded diagnosis of AF (ICD-9) on discharge records. 5.
Availability of the absolute number of subjects with AF in the high and low CRP
groups, and CRP cut offs. 6. CRP measurement using high sensitive assay. 7.
Quality score ≥ 7. The quality of the selected studies was assessed using
the Newcastle-Ottawa quality assessment scale .
Discrepancies were resolved by consensus after review by the third author
(K.A). If more than one study from the same authors fulfilled the inclusion
criteria, the larger of the two was included in the review so as to avoid
duplication of data sets.
Figure 1: Flow diagram of study identification and selection.
Using a standardized data extraction form the
two authors extracted the following data from each of the eligible studies:
first author, citation, year of publication, study population, study design,
ascertainment of AF, method of CRP assay, baseline CRP distribution, CRP cut
off, number of subjects in the high and low CRP groups, incidence or prevalence
of AF in the high and low CRP groups, relative risk for AF based on CRP level,
adjusted covariates and brief results. A CRP value of 3mg/l was empirically
chosen for stratification into high and low CRP groups based on the AHA
statement on CRP and increased risk of cardiovascular disease .
From among the studies selected for systematic review, those in which high and
low CRP groups could be stratified into a combinable format (CRP cut off close
to 3 mg/l) were included in the final meta-analysis. Letters were mailed to the
authors regarding the above details when the same were not available in the
The absolute number of subjects who developed AF
in the high and low CRP groups, and the total number of subjects in each group
were obtained for each individual study and the pooled data was used to obtain
the cumulative relative risk. Meta-analysis was performed adhering to
Meta-analysis Of Observational Studies in Epidemiology (MOOSE) guidelines . Analyses were conducted based on intention to treat principle.
Each study was considered as a single stratum. We obtained the pooled relative
risks with 95% confidence interval (CI) for development of AF using
the random effects model of Der Simonian and Laird .
Random effects model was preferred over fixed effects model because of the
significant heterogeneity between the included trials. We used inverse-variance
weighting to calculate random effects summary estimates, and used the Q test to
test for heterogeneity of study results. We performed an influence
analysis and assessed the influence of individual studies on the summary effect
estimate. A funnel plot was also done to look for any publication bias in the
studies. All statistical analyses were performed using Stata 9.0
(Stata Corporation, College Station, TX).
The literature search yielded 129 potentially
relevant studies. 84 of these were excluded after scrutiny of the abstract
because they were letters, replies, reviews or irrelevant. Manuscripts of the
remaining 45 were reviewed separately by two of the authors (V.A, S.T). 34 more
were subsequently excluded as they did not meet the inclusion criteria and the
remaining 11 were selected for systematic review [10-20]. The concise detail of these studies is shown in [Table 1]. Of these, only 8 studies in which absolute numbers
(of subjects at risk for AF and subjects with AF) for a CRP cut off of around 3
mg/l could be obtained were included in the final meta-analysis [10-13,
15, 17-19 ].
The 8 studies included 7507 subjects of which 719 had
AF. Of these, one study (n=5806) pertained to prevalent AF in patients enrolled
in a Cardiovascular study registry , 2 studies (n=282)
dealt with post cardiac surgery AF [11, 13], one (n=1011) pertained to incident AF in normal
and the remaining 4 studies
(n=408) pertained to AF recurrence after successful cardioversion [
15, 17-19]. The mean age of the
study populations varied between 45-70 and males constituted 50-75%.
Table 1: Characteristics of the studies selected for Systematic Review.
¶ Not included in the final meta-analysis.
§ Studies reporting no association between CRP and AF.
RR=relative risk; CI= confidence interval; CRP=C-reactive protein; EKG= electrocardiogram; CABG=coronary artery bypass grafting; ACS=acute coronary syndrome; CAD= coronary artery disease.
Meta-analysis of the 8 studies yielded a relative
risk of 1.63 (1.43, 1.86) for occurrence of AF when CRP was elevated above
3-3.5 mg/l [figure 2]. In an influence analysis, none of
the individual studies had an overwhelming effect on the summary effect
estimate and it remained relatively stable and significant on excluding one
study at a time. There was no publication bias observed on the funnel plot [figure 3] and Egger’s weighted regression method p-value was
Figure 2: Forest plot of AF risk associated with CRP elevation (>3-3.5mg/l).
Figure 3: Funnel plot of the selected studies to assess for publication bias.
Three of the studies selected for review were not
included in the final meta-analyses as the cut off CRP used was quite
dissimilar [14, 16, 20]. The studies by >Korantzopouloset
al  and Kotsakaou et al used CRP cutoffs of 4.3 and 4.9 mg/l respectively.
Secondary analysis using the data from these studies and that from Dernellis et
al ,where numbers
for a similar CRP cut off (4.85 mg/l) were available, yielded a relative risk
of 4.03 (3.1, 5.25) for AF [Figure 4].
Figure 4: Forest plot of AF risk associated with CRP elevation (>4.5-5mg/l).
Two of the 11 studies selected for systematic
review accounting for 1141 subjects reported lack of independent association
between CRP and AF [12, 13].In the study
by Dernellis et al ,elevated CRP was predictive of AF
only in the presence of concomitant elevation of complement. The poor ability
of CRP in predicting incident AF in this study was probably due to the low risk
population studied (relatively young, exclusion of those with CAD, CHF or other
heart diseases). Additionally, measuring downstream products of inflammation
(complement components) could have led to underestimation of the association
between CRP and AF in this study. It is known that women have more elevations
in inflammatory markers at baseline .This might have
been a potential reason for the poor predictive value of CRP in the study by
Hogue et al ,which was done in a small and select group
of women (post menopausal women > 55 years of age undergoing cardiac
surgery). Further studies evaluating the sex specific limitations in the
utility and applications of CRP are therefore warranted.
The 5 studies that addressed recurrent AF
following electrical cardioversion [15, 17-20]
were limited by their observational
design, lack of uniformity in the study population, follow up, and their small
size. In the study by Aviles et al 
,elevated CRP was predictive of both prevalent and incident AF. Watanabe et al  used a CRP cutoff of 0.6 mg/l (falls into low risk category
by AHA definition) and demonstrated a relative risk of 5.3 for post cardio
version AF recurrence. This was included neither in the primary nor the
secondary analysis because of the extremely low cut off used (which was very
dissimilar to most other studies). It is well known that the median CRP level
in healthy Japanese and Chinese subjects is much lower compared to their
western counterparts [22-23]. The high
risk of AF despite relatively low CRP levels in Asian populations could be
related to differences in body mass index and genetic constitution which alter
inflammatory response and CRP levels [23-25].
The pathophysiology of AF is complex and to date
is not completely understood. It is now known that pulmonary veins serve a
crucial role in the initiation of this arrhythmia .Once
initiated, AF sets in motion a process of self propagation through electrical,
biophysical and structural remodeling of the atria [27-30]. There is a strong association between AF and inflammation.
Numerous serum markers of inflammation like TNF α, IL-6, leukocyte count
and CRP have been shown to be elevated in AF [5, 30].It is known that AF and inflammation alter myocardial
energy kinetics and increase oxidative stress which can further perpetuate the
arrhythmia [29, 31].In addition, CRP
leads to complement activation and tissue damage locally in the atrial
myocardium further increasing the substrate for AF .Moreover,
CRP levels progressively increase with increasing AF burden 
but it is unclear whether inflammation is a cause or consequence of AF. Some
investigators have shown that anti inflammatory therapy reduces recurrences of
AF with parallel reductions in CRP suggesting a possible cause effect
relationship [34, 35]. However, this
The present systematic review and meta-analysis
supports the strong association between elevated CRP and occurrence of AF
across a variety of clinical settings. Moreover, our study suggests possible incremental
relationship with higher CRP levels conferring a relatively higher risk of AF.
It is important to bear in mind a number of limitations while interpreting and
applying results of meta-analyses. The potential for publication bias against
negative studies and small studies is the foremost concern. This might have
caused our study to overestimate the risk attributable to elevated CRP. Thirteen
out of the 45 studies initially selected for review, showed no independent
association between AF and CRP. Only 2 of these 13 studies met the criteria for
inclusion [12-13].On the other hand, 32
out of the 45 studies reported a significant independent association between
CRP and AF and 11 of these met the inclusion criteria [10, 11,
Arial'>. It is apparent that publication bias against negative studies would
not have a major impact on our study as more than a third of the reviewed
studies reported no association between CRP and AF. In addition, both Begg’s
Funnel plot and Eggers test did not reveal any effect of publication bias on
our results. However, a potential shortcoming of our study could be the
inclusion of more studies reporting a positive result in the final analysis. Of
note, a majority of the negative studies that were excluded were small and
poorly designed. On the other hand, the studies that reported a positive
association had to be excluded because of lack of absolute numbers in the high
and low CRP groups or absolute values for CRP cut off despite better design and
higher numbers. Thus, with the relatively small number of subjects in the
excluded negative studies, it is unlikely that we would have erroneously
detected an association between CRP and AF in the true absence of one.
Another major concern would be the issue of bias.
As with any study on AF, all the studies in our analysis had a propensity for
ascertainment bias because the follow up for AF was periodic and not
continuous. Additionally, despite adjustment for confounding factors like
coronary artery disease, hypertension, diabetes, heart failure, age and smoking
status by the individual investigators, residual confounding cannot be
excluded. The lack of original data from the included trials precluded our
ability to perform a logistic regression analysis to independently assess the
effects of these confounders. Moreover, it is both impractical and impossible
to adjust for the confounding effects of the innumerable inflammatory markers
and cytokines; leaving enormous scope for ‘residual confounding’. Other potential
limitations of our analysis are exclusion of studies not published in English,
wide variation in the study populations and the heterogeneity of CRP assay
among the individual studies. All of these inherent problems adversely
influence the applicability of the results and make it difficult to make
general conclusions regarding the relation between CRP and AF. Finally, all
the included studies utilized single measurements of CRP for stratification;
whereas, it is generally recommended that for improved specificity, CRP be
measured at least 2 different times (2 weeks apart) when being used for risk
assessment of cardiovascular diseases .
Only a small number of patients with elevated CRP
have AF and not all patients with AF have elevated CRP. Thus, the lack of
specificity limits the general applicability of CRP in predicting the presence
or the occurrence of AF. As in the case of coronary artery disease where CRP
measurement is best used for further risk stratification of patients at
intermediate risk, it is imperative to identify appropriate populations for CRP
testing in the context of AF. Review of literature supports a potential role
for CRP testing in the following scenarios. In patients with a prior history of
AF, CRP can help discriminate between those who will and will not have a
recurrence and identify patients at a higher risk for complications like
embolism .CRP testing can potentially identify patients
at risk for developing postoperative AF and AF following acute myocardial
infarction .In addition, recent data suggests that CRP
can predict recurrence of AF following the first episode of paroxysmal AF, success
of cardioversion for persistent AF and recurrent AF following successful
cardioversion [38, 39].In our study, secondary
analysis using studies that addressed post cardioversion AF recurrence [15, 17-19] yielded a
relative risk of 1.49 (1.23, 1.79) for AF recurrence when CRP was >3-3.5
mg/l. This is consistent with the findings of previously published studies
assessing the association between CRP and AF recurrence following cardioversion
. Thus, CRP has the potential to be an invaluable tool in
identifying patients who would require increased surveillance and serve as a
guide to determine the intensity of treatment and follow up.
Overall the evidence linking elevated CRP and AF
is robust and the strength of association is strong. There is fair consistency
in the data supporting this association and there seems to be an incremental
relationship with higher CRP levels conferring proportionately increased risk.
Elevated CRP (usually above 3.0-3.5 mg/l) is associated with about 1.6 times
increased risk of AF compared to CRP < 3 mg/l. The major limitation in the
clinical utility of CRP is its poor specificity (as it could be elevated in
multiple other disease states). The potential benefit of CRP lies in our
ability to use it in selected groups of patients at risk for AF (heart failure,
acute myocardial infarction) and in specific settings like post operative state,
post cardioversion etc. The importance of using the mean value of CRP
measurements made over time as opposed to single measurements should be
emphasized and encouraged. There is therefore, a pressing need for further
large and well designed prospective studies to test the association of CRP with
AF and its utility in clinical practice.
Feinberg WM, Blackshear JL, Laupacis A, Kronmal R, Hart RG. Prevalence, age distribution, and gender of patients with atrial fibrillation. Analysis and implications. Arch Intern Med. 1995;155:469-73.
Kannel WB, Wolf PA, Benjamin EJ, Levy D. Prevalence, incidence, prognosis, and predisposing conditions for atrial fibrillation: population-based estimates. Am J Cardiol. 1998;82(8A):2N-9N.
Go AS, Hylek EM, Phillips KA, et al. Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study. JAMA. 2001;285:2370-5.CrossRef
Osman R, L'Allier PL, Elgharib N, Tardif JC. Critical appraisal of C-reactive protein throughout the spectrum of cardiovascular disease. Vasc Health Risk Manag. 2006;2:221-37.
Boos CJ, Anderson RA, Lip GY. Is atrial fibrillation an inflammatory disorder? Eur Heart J. 2006;27:136-49.
Wells GA, Shea B, Peterson J, Welch V, Losos M, Tugwell P. The Newcastle-Ottawa Scale for assessing the quality of nonrandomised studies in meta-analyses. http://www.ohri.ca/programs/clinicalepidemiology/oxford.htm
Pearson TA, Mensah GA, Alexander RW, et al Centers for Disease Control and Prevention; American Heart Association. Markers of inflammation and cardiovascular disease: application to clinical and public health practice: A statement for healthcare professionals from the Centers for Disease Control and Prevention and the American Heart Association. Circulation. 2003;107:499-511CrossRef
Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA. 2000; 283:2008-12.
Der Simonian R, Laird N. meta-analysis in clinical trials. Controlled Clin Trials, 1986; 7: 177-188.
Aviles RJ, Martin DO, Apperson-Hansen C, et al. Inflammation as a risk factor for atrial fibrillation. Circulation. 2003;108:3006-10.
Lo B, Fijnheer R, Nierich AP, Bruins P, Kalkman CJ. C-reactive protein is a risk indicator for atrial fibrillation after myocardial revascularization. Ann Thorac Surg. 2005;79:1530-5.
Dernellis J, Panaretou M. Effects of C-reactive protein and the third and fourth components of complement (C3 and C4) on incidence of atrial fibrillation. Am J Cardiol. 2006;97:245-8.CrossRef
Hogue CW Jr, Palin CA, Kailasam R, et al. C-reactive protein levels and atrial fibrillation after cardiac surgery in women. Ann Thorac Surg. 2006; 82:97-102.CrossRef
Hatzinikolaou-Kotsakou E, Tziakas D, Hotidis A, et al. Relation of C-reactive protein to the first onset and the recurrence rate in lone atrial fibrillation. Am J Cardiol. 2006;97:659-61.CrossRef
Loricchio ML, Cianfrocca C, Pasceri V, et al. Relation of C-reactive protein to long-term risk of recurrence of atrial fibrillation after electrical cardioversion. Am J Cardiol. 2007;99:1421-4.CrossRef
Watanabe E, Arakawa T, Uchiyama T, Kodama I, Hishida H. High-sensitivity C-reactive protein is predictive of successful cardioversion for atrial fibrillation and maintenance of sinus rhythm after conversion. Int J Cardiol. 2006;108:346-53.CrossRef
Wazni O, Martin DO, Marrouche NF, et al. C reactive protein concentration and recurrence of atrial fibrillation after electrical cardioversion. Heart. 2005;91:1303-5.CrossRef
Zarauza J, Rodriguez Lera MJ, Farinas Alvarez C, et al. Relationship between C-reactive protein level and early recurrence of atrial fibrillation after electrical cardioversion. Rev Esp Cardiol. 2006;59:125-9.
Mazza A, Bendini MG, Cristofori M, et al. Baseline apnoea/hypopnoea index and high-sensitivity C-reactive protein for the risk of recurrence of atrial fibrillation after successful electrical cardioversion: a predictive model based upon the multiple effects of significant variables. Europace. 2009;11:902-9.CrossRef
Korantzopoulos P, Kalantzi K, Siogas K, Goudevenos JA. Long-term prognostic value of baseline C-reactive protein in predicting recurrence of atrial fibrillation after electrical cardioversion. Pacing Clin Electrophysiol. 2008; 31:1272-6.
Khera A, McGuire DK, Murphy SA, et al. Race and gender differences in C-reactive protein levels. J Am Coll Cardiol. 2005;46:464-9.CrossRef
Yamada S, Gotoh T, Nakashima Y, et al. Distribution of serum C-reactive protein and its association with atherosclerotic risk factors in a Japanese population: Jichi Medical School Cohort Study. Am J Epidemiol. 2001; 153:1183-90.
Lear SA, Chen MM, Birmingham CL, Frohlich JJ. The relationship between simple anthropometric indices and C-reactive protein: ethnic and gender differences. Metabolism. 2003;52:1542-6.
Zee RY, Ridker PM. Polymorphism in the human C-reactive protein (CRP) gene, plasma concentrations of CRP, and the risk of future arterial thrombosis. Atherosclerosis. 2002;162:217-9.CrossRef
MacGregor AJ, Gallimore JR, Spector TD, Pepys MB. Genetic effects on baseline values of C-reactive protein and serum amyloid a protein: a comparison of monozygotic and dizygotic twins. Clin Chem. 2004;50:130-4.CrossRef
Haissaguerre M, Marcus FI, Fischer B, Clementy J. Radiofrequency catheter ablation in unusual mechanisms of atrial fibrillation: report of three cases. J Cardiovasc Electrophysiol 1994;5:743–751.CrossRef
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;96:1180–1184.
Lau CP, Tse HF. Electrical remodelling of chronic atrial fibrillation. Clin Exp Pharmacol Physiol 1997;24:982–983.CrossRef
Korantzopoulos P, Kolettis T, Siogas K, Goudevenos J. Atrial fibrillation and electrical remodeling: the potential role of inflammation and oxidative stress. Med Sci Monit 2003;9: 225–9.
Engelmann MDM, Svedsen JH. Inflammation in the genesis and perpetuation of atrial fibrillation. Eur Heart J 2005;26:2083–92.CrossRef
Mihm MJ, Yu F, Carnes CA, et al. Impaired myofibrillar energetics and oxidative injury during human atrial fibrillation. Circulation 2001;104:174–180.
Kaplan MH, Volanakis JE. Interaction of C-reactive protein complexes with the complement system. I. Consumption of human complement associated with the reaction of C-reactive protein with pneumococcal C-polysaccharide and with the choline phosphatides, lecithin and sphingomyelin. J Immunol 1974;112:2135–2147.
Chung MK, Martin DO, Sprecher D, et al. C-reactive protein elevation in patients with atrial arrhythmias: inflammatory mechanisms and persistence of atrial fibrillation. Circulation. 2001;104:2886-91.
Korantzopoulos P, Kolettis TM, Kountouris E, et al. Oral vitamin C administration reduces early recurrence rates after electrical cardioversion of persistent atrial fibrillation and attenuates associated inflammation. Int J Cardiol. 2005;102:321-6.
35Dernellis J, Panaretou M. Relationship between C-reactive protein concentrations during glucocorticoid therapy and recurrent atrial fibrillation. Eur Heart J. 2004;25:1100-7.
Lip GY, Patel JV, Hughes E, Hart RG. High-Sensitivity C-Reactive Protein and Soluble CD40 Ligand as Indices of Inflammation and Platelet Activation in 880 Patients With Nonvalvular Atrial Fibrillation. Relationship to Stroke Risk Factors, Stroke Risk Stratification schema, and Prognosis. Stroke. 2007;38:1229-37.
Gedikli O, Orem C, Baykan M, et al. Association between serum C-reactive protein elevation and atrial fibrillation after first anterior myocardial infarction. Clin Cardiol. 2008 Oct;31(10):482-7.
Liu T, Li G, Li L, Korantzopoulos P. Association between C-reactive protein and recurrence of atrial fibrillation after successful electrical cardioversion: a meta-analysis. J Am Coll Cardiol. 2007; 49:1642-8.
Liu T, Li L, Korantzopoulos P, Goudevenos JA, Li G. Meta-analysis of association between C-reactive protein and immediate success of electrical cardioversion in persistent atrial fibrillation. Am J Cardiol. 2008 15;101:1749-52