OUP user menu

Role of inflammation in early atrial fibrillation recurrence

Marcelle D. Smit , Alexander H. Maass , Anne Margreet De Jong , Anneke C. Muller Kobold , Dirk J. Van Veldhuisen , Isabelle C. Van Gelder
DOI: http://dx.doi.org/10.1093/europace/eur402 810-817 First published online: 11 January 2012


Aims Outcome of rhythm control in atrial fibrillation (AF) is still poor due to various mechanisms involved in the initiation and perpetuation of AF. Differences in timing of AF recurrence may depend on different types of mechanisms. The aim of this study was to assess the mechanisms involved in early AF recurrence in patients with short-lasting AF.

Methods and results Patients with short-lasting persistent AF undergoing rhythm control (n= 100) were included. Markers of mechanisms involved in the initiation and perpetuation of AF were assessed, including clinical factors, echocardiographic parameters, and biomarkers. Primary endpoint was early AF recurrence (recurrence <1 month). Secondary endpoint was progression to permanent AF. Median total AF history was short: 4.2 months. Early AF recurrences occurred in 30 patients (30%) after a median of 6 (inter-quartile range 2–14) days. Baseline log2 interleukin (IL)-6 [adjusted hazard ratio (HR) 1.3, 95% confidence interval (CI) 1.0–1.7, P= 0.02] and present or previous smoking (adjusted HR 3.6, 95% CI 1.2–10.9, P= 0.03) were independently associated with early AF recurrence, suggesting that inflammation played an important role in early recurrences. Atrial fibrillation became permanent in 29 patients (29%). Baseline transforming growth factor-β1, left ventricular ejection fraction, and early AF recurrence were independently associated with progression to permanent AF.

Conclusion In patients with short-lasting AF, early AF recurrence seemed to be associated with inflammation as represented by IL-6. Treatment aimed against inflammation may therefore prevent early AF recurrences, which can improve rhythm control outcome.

  • Short-lasting atrial fibrillation
  • Inflammation
  • IL-6


Outcome of rhythm control in patients with persistent atrial fibrillation (AF) is still poor, despite the attempts that have been made to improve rhythm control therapy.14 We previously failed to show that immediate cardioversion of early recurrences increased maintenance of sinus rhythm.5 In addition, short-term peri-cardioversion amiodarone therapy could not reduce AF recurrences.6 Atrial fibrillation, however, is a complex condition with multiple interacting mechanisms involved in the initiation and perpetuation of the arrhythmia, including acute triggers, changes in electrical properties, and structural remodelling.7,8 The mechanisms leading to the initiation and perpetuation of AF may contribute differently to the onset and persistence of AF in different patients, which could imply that there are distinct types of AF requiring specific types of treatment.9 Furthermore, these mechanisms may vary according to the timing of AF recurrence and progression. Most AF recurrences occur within 1 month after cardioversion (early AF recurrences).1012 To our knowledge, mechanisms involved in early AF recurrences have not been studied before in persistent AF. Furthermore, most studies investigating rhythm control in persistent AF have included patients in whom the extent of remodelling was severe due to a long history of AF or underlying disease. In patients with a short history of AF, i.e. short-lasting AF, electrical and structural remodelling processes are assumed to be less advanced,13 providing opportunities for rhythm control to be more effective.14 This patient category has not been studied before. The aim of this study was therefore to investigate the mechanisms involved in early AF recurrences in a patient population with short-lasting AF.


Study population and study protocol

This was a prospectively designed observational study performed in the University Medical Center Groningen. Recruitment started in January 2008 and ended in July 2010. Patients were included if they had short-lasting persistent AF, defined as a total AF history of <2 years, a total persistent AF history of <6 months, and ≤1 previous electrical cardioversion.

At enrolment, a detailed medical history was obtained in all patients, transthoracic echocardiography was performed, and blood samples for biomarker analyses were obtained. Patients were treated according to our standardized rhythm control strategy which involved causal treatment of the underlying disease, anticoagulation as indicated, adequate rate control with negative dromotropic drugs, and rhythm control in accordance with the guidelines.1,2,4,15,16 Initial rhythm control consisted of electrical cardioversion, chemical cardioversion with amiodarone, or pulmonary vein ablation. Pre-treatment with ion-channel antiarrhythmic drugs was started only in patients in whom rhythm control was expected to be less successful, i.e. due to underlying disease.15,16 Time to AF recurrence was carefully monitored by frequent outpatient visits and 24 h Holter monitoring (1, 3, 6, 9, and 12 months after start of rhythm control therapy). In case of recurrence of AF after initial rhythm control therapy, ion-channel antiarrhythmic drugs were instituted as soon as possible after documentation of the recurrence, in combination with electrical cardioversion if required.1,2,4,15,16 Amiodarone loading was started 4 weeks before the next planned electrical cardioversion, sotalol was started on the day of the next planned cardioversion. Atrial fibrillation was accepted in case of failure of at least one ion-channel antiarrhythmic drug and/or if patient declined to pursue normal sinus rhythm in the absence of severe symptoms of AF.1,2,4,15,16

Definitions of atrial fibrillation recurrence

The primary endpoint consisted of early AF recurrence, defined as any (a)symptomatic recurrence of AF within the first month after cardioversion lasting ≥30 s.10,11,17 Secondary endpoint was progression to permanent AF within 1 year. Patients were defined to have permanent AF when rhythm control interventions were no longer pursued.15,16 Shock failure was defined as no single sinus beat seen after cardioversion, immediate reinitiation of AF (IRAF) was defined as AF recurrence within 2 min after electrical cardioversion.11,18


Echocardiographic evaluation was conducted at baseline and included atrial dimensions and volumes, atrial ejection fractions, septal and posterior wall thicknesses, ventricular dimensions, and left ventricular ejection fraction. Left atrial volume was measured using the biplane Simpson's method and the right atrial volume was measured using the single plane area–length method.19 Left and right atrial volumes were additionally indexed to body surface area. Left and right atrial ejection fractions were calculated using end-systolic and end-diastolic left and right atrial volumes, respectively. Echocardiograms were performed in accordance with standard recommendations.

Biomarker analyses

Pre-specified biomarkers analysed for this study were biomarkers of haemodynamic stress, i.e. atrial natriuretic peptide (ANP), N-terminal pro-B-type natriuretic peptide (NT-proBNP), and apelin; and biomarkers of fibrosis and inflammation, i.e. growth differentiation factor (GDF)-15, matrix metalloproteinase (MMP)-1, MMP-2, MMP-9, tissue inhibitor of metalloproteinase (TIMP)-1, TIMP-2, transforming growth factor (TGF)-β1, and interleukin (IL)-6. Atrial natriuretic peptide and NT-proBNP are natriuretic hormones released from ventricular and atrial cells in response to volume expansion and increased wall stress.8,20 Apelin is a component of the apelin-angiotensin receptor-like one pathway that plays an important counter-regulatory role in the effects of angiotensin.21 Growth differentiation factor-15 is a member of the TGF-β cytokine family and is secreted during periods of ischaemia and reperfusion and is also an antihypertrophic regulating factor in the heart.22 Matrix metalloproteinases are associated with degradation of collagen, while TIMPs inhibit the activity of MMPs.8,23 Transforming growth factor-β1 is an inflammation-associated cytokine that is central to signalling cascades stimulating cardiac fibrosis, and may be a key mediator of fibrosis.8 Interleukin-6 is an inflammatory cytokine that is also a potent regulator of extracellular protein metabolism through MMPs and collagen.8,23,24 Venous blood samples for biomarker analyses were obtained at enrolment. Ethylenediaminetetraacetic plasma, lithium-heparin plasma, and serum samples were stored at −80°C until further analysis. Biomarker analyses were conducted using commercially available kits. All biomarkers except apelin were analysed using enzyme-linked immunosorbent assays according to the manufacturer's instructions (GE Healthcare UK Ltd, Buckinghamshire, UK, for MMP-1, MMP-2, MMP-9, TIMP-1, TIMP-2; R&D systems, Minneapolis, MN, USA, for ANP, NT-proBNP, GDF-15, TGF-β1, and IL-6). Apelin-12 was analysed using enzyme immunoassay according to the manufacturer's instructions (Phoenix Pharmaceuticals, Inc., Burlingame, CA, USA).

Statistical analysis

Baseline descriptive statistics are presented as mean ± standard deviation or median (inter-quartile range) for continuous variables and numbers with percentages for categorical variables, as required. We evaluated differences between groups using χ2 test and Fisher's exact test for categorical data, and Student's t-test and Mann–Whitney U test for continuous data, dependent on whether data were normally distributed. Cumulative event proportions were calculated using Kaplan–Meier analyses. Cox proportional hazards regression analyses were conducted to evaluate predictors of time to early AF recurrence and time to progression of permanent AF. Univariate Cox proportional hazards regression analysis was performed on all baseline variables shown in Tables 13. For permanent AF, early AF recurrence was added as time-dependent covariate. Stepwise multivariable hazards regression analysis was conducted using all baseline variables with P≤ 0.1 in univariate analysis. The final multivariate model included all variables with P< 0.05. Analyses were performed with STATA 11.0 for Windows. In all statistical analyses P< 0.05 was considered statistically significant.

View this table:
Table 1

Baseline characteristics

Study population (n= 100)Early AF recurrence (n= 30)No early AF recurrence (n= 70)
Age—mean ± SD (years)65 ± 963 ± 865 ± 10
Male sex—no. (%)74 (74.0)22 (73.3)52 (74.3)
AF characteristics
 Total AF history—median (IQR) (months)4.2 (1.9–9.0)5.6 (3.0–11.1)*3.9 (1.5–7.5)
 Current AF duration—median (IQR) (months)3.0 (1.3–4.7)3.2 (1.8–5.4)2.9 (1.1–4.5)
 Previous electrical cardioversion—no. (%)26 (26.0)12 (40.0)*14 (20.0)
 Previous chemical cardioversion—no. (%)4 (4.0)4 (5.7)
CHADS2 scorea—mean ± SD1.5 ± 1.01.4 ± 0.91.6 ± 1.0
Hypertension—no. (%)67 (67.0)19 (63.3)48 (68.6)
Previous admission for heart failure—no. (%)20 (20.0)8 (27.7)12 (17.1)
Coronary artery disease—no. (%)18 (18.0)6 (20.0)12 (17.1)
 Previous myocardial infarction8 (8.0)3 (10.0)5 (7.1)
History of valve dysfunction—no. (%)22 (22.0)6 (20.0)16 (22.9)
History of valve surgery—no. (%)4 (4.0)1 (3.3)3 (4.3)
Other medical history
 Diabetes mellitus—no. (%)
  Type I1 (1.0)1 (1.4)
  Type II13 (13.0)2 (6.7)11 (15.7)
 Hypercholesterolaemia—no. (%)30 (30.0)8 (26.7)22 (31.4)
 Smoking—no. (%)
  Previous smoking49 (49.0)19 (63.3)*30 (42.9)
  Present smoking13 (13.0)6 (20.0)*7 (10.0)
 Chronic obstructive pulmonary disease—no. (%)10 (10.0)1 (3.3)9 (12.9)
 History of transient ischaemic attack/stroke—no. (%)7 (7.0)1 (3.3)6 (8.6)
 History of thyroid disease—no. (%)
  Hypothyroidism1 (1.0)1 (1.4)
  Hyperthyroidism5 (5.0)5 (7.1)
 Sleep apnoea—no. (%)1 (1.0)1 (1.4)
AF EHRA class—no. (%)
 I18 (18.0)7 (23.3)11 (15.7)
 II49 (49.0)15 (50.0)34 (48.6)
 III30 (30.0)7 (23.3)23 (32.9)
 IV3 (3.0)1 (3.3)2 (2.9)
Physical examination
 Body mass index—mean ± SD (kg/m2)28 ± 428 ± 427 ± 5
 Systolic blood pressure—mean ± SD (mmHg)130 ± 16130 ± 17130 ± 16
 Diastolic blood pressure—mean ± SD (mmHg)81 ± 1281 ± 1080 ± 13
 Heart rate—mean ± SD (b.p.m.)101 ± 2893 ± 20*104 ± 30
 QRS duration—mean ± SD (ms)97 ± 1596 ± 1098 ± 17
 QTc duration—mean ± SD (ms)440 ± 44442 ± 37439 ± 47
Medication at cardioversion—no. (%)
 Beta-blocker89 (89.0)27 (90.0)62 (88.6)
 ACE-inhibitor/angiotensin receptor blocker74 (74.0)17 (56.7)*57 (81.4)
 Aldosterone receptor antagonist15 (15.0)3 (10.0)12 (17.1)
 Diuretic43 (43.0)11 (36.7)32 (45.7)
 Verapamil/diltiazem12 (12.0)4 (13.3)8 (11.4)
 Dihydropyridine calcium channel blocker9 (9.0)1 (3.3)8 (11.4)
 Digitalis13 (13.0)4 (13.3)9 (12.9)
 Amiodarone12 (12.0)2 (6.7)10 (14.3)
 Vitamin K antagonist99 (99.0)30 (100.0)69 (98.6)
 Platelet aggregation inhibitor10 (10.0)3 (10.0)7 (10.0)
 Statin38 (38.0)10 (33.3)28 (40.0)
 Nitrate1 (1.0)1 (3.3)
  • ACE, angiotensin-converting enzyme; AF, atrial fibrillation; EHRA, European Heart Rhythm Association; IQR, inter-quartile range; LA, left atrial; LV, left ventricular; NYHA, New York Heart Association; RA, right atrial; SD, standard deviation.

  • aThe CHADS2 score is a measure of the risk of stroke in patients with AF, with scores ranging from 0 to 6 and higher scores indicating a greater risk.16 Congestive heart failure, hypertension, an age of 75 years or older, and diabetes are each assigned one point, and previous stroke or transient ischaemic attack is assigned two points; the score is calculated by summing all the points for a given patient.

  • *P < 0.05.

View this table:
Table 2

Baseline laboratory values

Study population (n= 100)Early AF recurrence (n= 30)No early AF recurrence (n= 70)
White blood cell count—mean ± SD (109/L)7.7 ± 1.97.7 ± 1.57.6 ± 2.0
eGFR—mean ± SD (mL/min/1.73 m2)77 ± 1980 ± 1876 ± 19
ANP—median (IQR) (pg/100 µL)144 (76–232)125 (71–213)151 (81–232)
NT-proBNP—median (IQR) (pg/mL)1022 (552–1933)735 (535–1775)1110 (650–2139)
GDF-15—median (IQR) (pg/mL)1049 (796–1604)1010 (767–1479)1050 (840–1608)
Apelin—median (IQR) (pg/mL)258 (145–308)260 (102–355)242 (160–308)
MMP-1—median (IQR) (ng/mL)1.7 (1.7–8.3)1.7 (1.7–8.3)1.7 (1.7–6.1)
MMP-2—mean ± SD (ng/mL)2198 ± 6502194 ± 6202201 ± 672
MMP-9—mean ± SD (ng/mL)28.6 ± 15.328.4 ± 17.228.7 ± 14.5
TIMP-1—median (IQR) (ng/mL)154 (130–182)155 (138–188)149 (130–172)
TIMP-2—median (IQR) (ng/mL)97 (90–108)95 (88–100)102 (93–114)
TGF-β1—mean ± SD (ng/mL)24.1 ± 10.425.1 ± 13.223.6 ± 8.8
IL-6—median (IQR) (pg/mL)2.5 (1.6–4.8)2.9 (1.8–5.8)2.3 (1.6–3.9)
  • ANP, atrial natriuretic peptide; eGFR, estimated glomerular filtration rate; GDF, growth differentiation factor; IL, interleukin; IQR, inter-quartile range; MMP, matrix metalloproteinase; NT-proBNP, N-terminal pro-B-type natriuretic peptide; SD, standard deviation; TGF, transforming growth factor; TIMP, tissue inhibitor of metalloproteinase.

  • *P < 0.05.

View this table:
Table 3

Baseline echocardiography

Study population (n= 100)Early AF recurrence (n= 30)No early AF recurrence (n= 70)
LA size, parasternal axis—mean ± SD (mm)45 ± 645 ± 645 ± 7
LA volume—mean ± SD (mL)91 ± 3396 ± 4089 ± 30
LA volume index—mean ± SD (mL/m2)45 ± 1646 ± 1844 ± 15
LA ejection fraction—mean ± SD (%)19 ± 1317 ± 1220 ± 14
RA size, length—mean ± SD (mm)62 ± 662 ± 763 ± 6
RA volume—mean ± SD (mL)73 ± 2874 ± 2773 ± 28
RA volume index—mean ± SD (mL/m2)36 ± 1436 ± 1336 ± 14
RA ejection fraction—mean ± SD (%)18 ± 1915 ± 2019 ± 19
Septum—mean ± SD (mm)10 ± 210 ± 210 ± 2
Posterior wall—mean ± SD (mm)9 ± 19 ± 19 ± 2
LV end-diastolic diameter—mean ± SD (mm)51 ± 750 ± 751 ± 7
LV end-systolic diameter—mean ± SD (mm)37 ± 937 ± 838 ± 10
LV ejection fraction—mean ± SD (%)48 ± 1348 ± 1247 ± 13
Valve dysfunction—no. (%)16 (16.0)4 (13.3)12 (17.1)
Mitral regurgitation14 (14.0)4 (13.3)10 (14.3)
Tricuspid regurgitation9 (9.0)3 (10.0)6 (8.6)
  • LA, left atrial; LV, left ventricular; RA, right atrial; SD, standard deviation.

  • *P < 0.05.


Baseline characteristics and rhythm control therapy

One hundred patients were included. Median total AF history was short: 4.2 months (Table 1). Sixty-seven patients (67%) had their first episode of persistent AF. Baseline laboratory values are shown in Table 2, baseline echocardiography in Table 3. Patients with an early AF recurrence had a significantly longer total AF history, were more often present or previous smokers, and less frequently used angiotensin-converting enzyme-inhibitors/angiotensin receptor blockers at time of cardioversion (Table 1).

Eighty-six patients (86%) underwent electrical cardioversion, which was successful in 76 patients (88%) and unsuccessful in 10 patients (12%) due to shock failure (n= 4, 5%) and IRAF (n= 6, 7%). Four patients (4%) underwent chemical conversion, three patients (3%) pulmonary vein ablation, and in seven patients (7%) spontaneous conversion to sinus rhythm under antiarrhythmic drugs occurred.

Atrial fibrillation recurrence during follow-up

Within 1-year follow-up, AF recurred in 59 patients (59%), being an early recurrence of AF in 30 patients (30%) (Figure 1). Twelve of these 30 patients (40%) had a history of a previous electrical cardioversion, significantly more than patients with no early AF recurrence (Table 1). Most early AF recurrences occurred within 1 week after rhythm control (Figure 2); median time to early AF recurrence was 6 days (inter-quartile range 2–14 days).

Figure 1

Time to atrial fibrillation recurrence (continuous line) and time to permanent atrial fibrillation (dashed line).

Figure 2

Daily incidence of atrial fibrillation recurrence during the first month after rhythm control (early atrial fibrillation recurrence). Early atrial fibrillation recurrence occurred in 30 of the 59 patients with atrial fibrillation recurrence within 1 year. Most of the early recurrences occurred during the first week after rhythm control. Day 0 = day of conversion to sinus rhythm (i.e. day of electrical cardioversion, chemical conversion, pulmonary vein ablation, or spontaneous conversion to sinus rhythm under antiarrhythmic drugs).

Patients underwent a median of one (inter-quartile range 1–2) electrical cardioversions during follow-up. Seventeen of 59 patients with a first recurrence (29%) had a second AF recurrence during a median follow-up of 7 days (inter-quartile range 1–14 days). In 55 patients electrical cardioversion was the only rhythm control therapy. Antiarrhythmic drugs were instituted in a total of 37 patients (37%); 12 were already using amiodarone at baseline (Table 1), while an additional 25 patients received sotalol (n= 16) or amiodarone (n= 9) because of recurrent AF. Twenty-three patients (23%) were using ion-channel antiarrhythmic drugs at 1-year-follow-up. Atrial fibrillation became permanent in 29 patients (29%) (Figure 1), of whom 16 patients (55%) had had an early AF recurrence.

Independent predictors of time to early AF recurrences were baseline IL-6 and present or previous smoking (Table 4). Independent predictors of time to progression of permanent AF within 1 year were baseline TGF-β1, left ventricular ejection fraction, and early AF recurrence (Table 4). None of the other biomarkers were predictive of early AF recurrence or progression to permanent AF.

View this table:
Table 4

Predictors of early atrial fibrillation recurrence and of permanent atrial fibrillation within 1 year

Univariate analysisMultivariable analysis
Hazard ratio (95% CI)P valueHazard ratio (95% CI)P value
Early AF recurrence
 Log2 IL-61.3 (1.0–1.6)0.071.3 (1.0–1.7)0.02
 Present or previous smoker3.8 (1.4–9.9)0.0073.6 (1.2–10.9)0.03
Permanent AF within 1 year
 TGF-β1 per 5 ng/mL1.2 (1.0–1.5)0.091.2 (1.0–1.5)0.04
 LV ejection fraction1.0 (1.0–1.1)0.091.1 (1.0–1.1)0.05
 Early AF recurrence42.6 (5.8–314.4)<0.000145.5 (3.4–611.1)0.004
  • CI, confidence interval; IL, interleukin; LV, left ventricular; TGF, transforming growth factor.


We investigated mechanisms involved in early AF recurrences in patients with short-lasting persistent AF. We found that early AF recurrence was associated with elevated IL-6 levels, possibly representing inflammation.

Mechanisms involved in atrial fibrillation recurrence and progression

We found that IL-6 independently predicted early AF recurrences in short-lasting AF patients, suggesting that inflammation may play an important role in the mechanisms triggering early AF recurrence. Inflammation has been associated with a variety of cardiovascular conditions including AF.8,25 The exact mechanism linking inflammation with (non-operative-related) AF is unknown, and it is unclear whether inflammation is an initiator or consequence of AF though it is not unlikely that both mechanisms are interrelated. Inflammation may cause AF, as one study demonstrated that short-term hypertension led to increased inflammatory cell infiltrates in the atria and increased inducibility of AF,26 and inflammation has been shown to be associated with incident AF.2730 On the other hand, a recent study demonstrated that intra- and extra-cardiac markers of inflammation were increased during AF itself, instead of being higher in patients with vs. without previous AF, suggesting that AF causes inflammation.24 Markers of inflammation including IL-6, IL-8, C-reactive protein, and tumour necrosis factor-α have been shown to be increased in patients with AF and with AF recurrence.24,25,3133 Interleukin-6 is a pro-inflammatory cytokine that stimulates the synthesis of acute phase proteins and is a potent regulator of extracellular protein metabolism through MMPs and collagen which could induce fibrosis.8,23 Interleukin-6 has been correlated with the presence and duration of AF and increased left atrial diameter.34 The association between IL-6 and AF recurrence after cardioversion has not been robust in other studies,35 but a study population consisting of short-lasting persistent AF patients has not been studied before.

Present or previous smoking was also independently associated with an early AF recurrence. Smoking is not an established risk factor for incident AF,36 though recently a study did observe an increased risk of AF in current and former smokers.37 Another study found that women had a greater risk on arrhythmia recurrence after cardioversion for atrial flutter if they were current smokers.38 Smoking could have deleterious effects regarding AF through several mechanisms including direct toxicity of nicotine and carbon monoxide, sympathetic neural stimulation, regional myocardial hypoperfusion, and perhaps induction of an inflammatory state and/or fibrosis.

Early AF recurrence predicted progression to permanent AF, which is logical because AF recurrence is required in order for AF to be progressive, and most AF recurrences occur within 1 month after rhythm control.10,11 Independent of early AF recurrence, we found that TGF-β1 was a predictor of development of permanent AF within 1 year. Transforming growth factor-β1 is a central determinant in the signalling cascade of cardiac (atrial) fibrosis leading to structural remodelling, though it is also involved in inflammatory processes.8 Markers of fibrosis including TGF-β1 have been associated with AF recurrence and failure of electrical cardioversion.3941 Fibrosis may therefore be an important contributor to progression to permanent AF. Higher left ventricular ejection fraction was also independently associated with progression to permanent AF which seems to be a counterintuitive observation, as heart failure is a well-known predictor of AF progression.42 Progression of AF, however, is not just a passive diagnosis as it also requires decision making by patient and physician. It would seem plausible that, in case of no or mild symptoms of AF, a physician is less reluctant to accept AF when ventricular function is preserved rather than decreased. Furthermore, it is unknown whether known risk factors for AF progression also pertain to patients with short-lasting AF who have not been studied before.

Atrial size has been shown to be a predictor of AF recurrences and progression,43 but this was not observed in our study. However, the degree of atrial structural remodelling may not be extensive in patients with a short history of AF, so the predictive value of atrial size regarding outcome of rhythm control may perhaps be absent in these patients.

Rhythm control and atrial fibrillation recurrence in short-lasting atrial fibrillation

To our knowledge, this is the first study investigating rhythm control in patients with short-lasting persistent AF, in whom we may expect that remodelling processes are less widespread and in whom restoration of sinus rhythm may halt disease progression.14 The median duration of AF history was only 4 months in the present population. Early AF recurrence, chosen as one of the primary endpoints for this study because most AF recurrences occur within 1 month after cardioversion,1012 occurred in 30% of these patients with short-lasting AF. At first glance 30% seems to be a high recurrence rate, but in other studies in which patients had a much longer AF history, early recurrences (without amiodarone treatment) were observed in 56–68% of patients.5,12 Early AF recurrence rates are lower with amiodarone, however, ∼20%.44 Despite a low early AF recurrence rate, persistent AF progressed to permanent AF in 29% of these short-lasting AF patients, which is comparable with other studies in which 25–30% of persistent AF patients had permanent AF after 1 year.5,45 However, the number of electrical cardioversions during follow-up in our study was quite low with a median of one cardioversion as compared with a median of two to three cardioversions in other studies.35 This implies that relatively less of an effort was made in order to keep short-lasting AF patients in sinus rhythm during 1-year follow-up due to the data of the rate vs. rhythm control trials, especially if symptoms were mild in case of AF recurrence.3,4 Perhaps AF should be accepted less easily and more attempts should be done to restore and maintain sinus rhythm in patients with short-lasting AF who have never been studied before. Success rates may be higher, although it still remains to be established if permanent sinus rhythm will improve prognosis in these patients.

Clinical implications

The most important finding of our study is that inflammation, as represented by IL-6, may be involved in early AF recurrences. Antiinflammatory therapies such as statins25,46,47 and corticosteroids,25,48,49 provided, e.g., shortly before and during the first month after electrical cardioversion, may therefore prevent early AF recurrences through direct antiarrhythmic effects. Prevention of early AF recurrences may subsequently prevent progression to permanent AF.

Strengths and limitations

The novel study population taken from clinical practice, the thorough assessment of parameters reflecting various mechanisms involved in the initiation and perpetuation of AF, and the clinically valuable new results constitute the most important strengths of this study. Because some patients had a prior history of AF while others presented with their first episode, it is uncertain whether the variables we measured were associated with early AF recurrence or with the progression to persistent AF in new-onset AF patients who might have had paroxysmal AF. However, early AF recurrences occurred in a substantial proportion of patients with a previous electrical cardioversion, i.e. in patients with established persistent AF, implying that the variables studied were for an important part associated with AF recurrences in persistent AF. In addition, we may have missed (asymptomatic) early AF recurrences because we only performed a 24 h Holter at 1 month. Detection of early recurrences would have improved if we had used an external device giving continuous monitoring during the first month after rhythm control. Furthermore, not all possible biomarkers that have been associated with AF were analysed in this study, which means that results could have differed if a different subset of biomarkers had been used. Regarding the potential antiarrhythmic effect of statins, we did not find that statin use prevented early AF recurrences or progression to permanent AF in our study population. However, these patients had an indication for statin use due to their underlying disease, which may have overruled the antiarrhythmic and antiinflammatory effects. Although set-up prospectively, a power analysis was not conducted because there were no prior data concerning patients with short-lasting AF. Our study is therefore limited by the small study population. This could have led to false positive results and potential overfitting of the multivariable models. On the other hand, the primary endpoint still occurred in a substantial amount of patients. Because of the small patient numbers, the results and especially the prediction models should be interpreted with caution. Larger studies are required to validate our results.


In patients with short-lasting AF, early AF recurrence seemed to be associated with inflammation as represented by IL-6. Treatment aimed against inflammation may therefore prevent early AF recurrences, which can improve rhythm control outcome.


This work was supported by the Netherlands Heart Foundation [Grant number 2009T014] and unrestricted educational grants from Medtronic and Biotronik.


The authors thank Silke U. Oberdorf-Maass for her excellent technical assistance.

Conflict of interest: none declared.


View Abstract