Prognosis, disease progression, and treatment of atrial fibrillation patients during 1 year: follow-up of the Euro Heart Survey on Atrial Fibrillation

Abstract

Aims

To gain insight in the prognosis and treatment of atrial fibrillation (AF) patients during 1-year follow-up in the Euro Heart Survey (EHS) on AF.

Methods and results

The EHS enrolled 5333 AF patients in 2003–2004. One-year follow-up data were available for 80%. Of first detected AF patients, 46% did not have a recurrence during 1 year, paroxysmal AF largely remained paroxysmal AF (80%), and 30% of persistent AF progressed to permanent AF. Many treatment changes occurred since baseline. Oral anticoagulation was started in 19% and discontinued in 16% of all patients. Of patients initially on rhythm control 27% did not receive rhythm control during follow-up, whereas 15% of patients initially on rate control received rhythm control. Mortality was highest in permanent AF (8.2%), but also substantial in first detected AF (5.7%). In multivariable analysis, sinus rhythm at baseline was associated with lower mortality, but no significant effect was observed regarding the application of either rhythm or rate control.

Conclusion

The EHS on AF provides unique prospective observational data on AF progression, long-term treatment, prognosis, and determinants of adverse outcome of the total clinical spectrum of AF in a European cardiology-based patient cohort.

Atrial fibrillation, Mortality, Prognosis, Progression, Management, Anticoagulation, Rhythm control, Rate control

Introduction

The Euro Heart Survey (EHS) provided a unique snapshot of the characteristics and management of all clinical atrial fibrillation (AF) types in cardiology practices across Europe.¹ Some discordance with the 2001 ACC/AHA/ESC management guideline recommendations² was observed and management differed highly among patients, mainly due to the variety of clinical AF types and differences in their patient profiles as encountered in cardiology practice. The question remains to what extent prognosis varies among the AF types and management strategies.

Current observational data on mortality and morbidity in AF patients were mainly derived from retrospective analyses in general populations and general practices and few were hospital-based.^3–11 Real-life AF progression has been described before,^10,12 but not in cardiology practice. Also, little information is available on AF treatment changes in daily practice over time.¹³

In order to gain insight in the prognosis, disease progression and treatment changes in the EHS cohort, a follow-up survey was performed after 1 year.

Methods

In the EHS on AF, 5333 patients were enrolled in 2003 and 2004. Survey methods, centre participation, patient characteristics, management, and definitions have previously been described.¹ In the preparation phase of the survey a follow-up was planned at 1 year following the baseline survey, containing questions on survival status, morbidity, re-admissions, diagnostics, interventions, drug therapy, AF characteristics, and QoL. Data were collected through searching medical records and patient interview. By using a validation plan, integrated in the data entry software, data were checked for missing or contradictory entries and values out of the normal range. Additional edit checks were performed by the EHS staff at the European Heart House and by the EHS-AF Data Analysis Centre at the University Hospital Maastricht.

Definitions

Cardiovascular mortality

Mortality due to any cardiovascular reason such as myocardial infarction, heart failure, sudden cardiac death (all sudden deaths without any other known reason), stroke, or rupture of an aortic aneurysm.

Ischaemic stroke

Focal neurological deficit of sudden onset as diagnosed by a neurologist, lasting >24 h and caused by ischaemia.

Transient ischaemic attack

Focal neurological deficit of sudden onset as diagnosed by a neurologist, lasting <24 h.

Coronary artery disease

Myocardial infarction (new or presumed new ST-segment elevation in two or more contiguous leads of ≥0.2 mV in leads V1, V2, or V3, and ≥0.1 mV in other leads, and/or presumably new left bundle branch block, and/or cardiac enzyme rise more than two times the upper values), or unstable angina.

Peripheral embolism

Embolism outside the heart, brain, eyes, and lungs.

Haemorrhagic stroke

Focal neurological deficit of sudden onset as diagnosed by a neurologist, lasting >24 h and caused by bleeding.

Other major bleeding

Major bleeding, other than haemorrhagic stroke, requiring hospitalization and/or causing a drop of haemoglobin level of >2 g/L and/or requiring blood transfusion.

Heart failure

Onset of new heart failure, or worsening of known heart failure.

Atrial fibrillation progression

First detected becoming paroxysmal, persistent, or permanent AF; paroxysmal becoming persistent or permanent AF; persistent becoming permanent AF.

Rhythm control strategy

Applying a pharmacological conversion, electrical cardioversion or catheter ablation during 1 year, or prescribing a class IA, IC, or III antiarrhythmic drug (Vaughan Williams classification¹⁴) at 1 year. Of note, rhythm control patients may be on typical rate control drugs, because the latter drugs are frequently unavoidable as background therapy in rhythm control patients.

Rate control strategy

Prescribing digitalis, class II, or class IV antiarrhythmic drugs at 1 year, and not applying any of the ‘rhythm control’ procedures during 1 year and not prescribing a class IA, IC, or III antiarrhythmic drug at 1 year.

Sinus rhythm at discharge

We considered patients to be in sinus rhythm at discharge of the baseline visit when the first known baseline ECG showed sinus rhythm and no episode of AF was reported, or when patients who were in AF underwent a successful procedure to restore sinus rhythm.

Statistical analysis

Data analysis was performed with SPSS statistical software (SPSS Inc., release 12.01). In Tables 1and2 continuous variables are reported as mean (SD) or median (25–75th percentile) and categorical variables as observed number (percentage within AF type). Total counts of the four AF types might be slightly lower than the total of the overall population due to a small proportion of patients who had an unknown AF type at baseline. Whether there was any difference among the four AF types was tested with one-way ANOVA or Kruskal–Wallis H for continuous variables and with χ² statistic for categorical variables. If statistically significant (P < 0.05), we investigated post hoc which AF types were different when clinically relevant, and report these specifications in the text. Survival from all-cause mortality per AF type during 1 year is depicted with Kaplan–Meier curves. Other events could not be analyzed in a time-dependent manner.

Table 1

Patient characteristics and management

	Type of AF at baseline				P-value
	First detected (n = 708)	Paroxysmal (n = 1170)	Persistent (n = 886)	Permanent (n = 1126)
Age, years	65 ± 14	64 ± 13	66 ± 11	70 ± 11	<0.001
Female gender	300 (40)	520 (43)	348 (38)	523 (43)	0.088
Follow-up duration, days	376 (364–430)	376 (364–411)	375 (363–410)	374 (359–406)	<0.001
AF characteristics at 1 year
Symptoms at 1 year	93 (15)	349 (34)	283 (35)	403 (38)	<0.001
AF at last recorded ECG	154 (24)	185 (18)	378 (45)	998 (92)	<0.001
Drug therapy at 1 year
Antithrombotic drugs
Oral anticoagulation	302 (45)	557 (50)	614 (71)	895 (79)	<0.001
Antiplatelet drug	264 (38)	424 (37)	235 (27)	265 (23)	<0.001
Heparin	9 (1)	12 (1)	9 (1)	12 (1)	0.944
Nothing	132 (20)	176 (16)	66 (8)	57 (5)	<0.001
Rhythm/rate control drugsa
Class Iab	2 (0)	9 (1)	5 (1)	1 (0)	0.068
Class Icc	36 (5)	157 (14)	86 (10)	14 (1)	<0.001
Class IId	230 (34)	479 (43)	362 (42)	442 (39)	<0.001
Class IIIe	181 (27)	346 (31)	323 (37)	190 (17)	<0.001
Class IVf	40 (6)	76 (7)	70 (8)	129 (11)	<0.001
Digitalisg	97 (15)	147 (13)	174 (20)	512 (45)	<0.001
Interventions during 1 year
Pharmacological cardioversion	89 (13)	237 (21)	143 (16)	39 (3)	<0.001
Electrical cardioversion	97 (14)	152 (14)	211 (24)	34 (3)	<0.001
Catheter ablation	4 (1)	62 (6)	32 (4)	12 (1)	<0.001
Implantable defibrillator	2 (0)	6 (1)	2 (0)	6 (1)	0.643
Pacemaker	12 (2)	53 (5)	34 (4)	57 (5)	0.005
AF surgery	0 (0)	3 (0)	4 (0)	5 (0)	0.342

Results are reported as mean ± SD, median (25–75th percentile) or observed number (%). ^aAntiarrhythmic drugs according to the Vaughan Williams classification,¹⁹ if possible; ^bbepridil, cibenzoline, disopyramide, procainamide, quinidine; ^cflecainide, propafenone; ^dbeta blocker for antiarrhythmic indication (excluding sotalol); ^eamiodarone, sotalol; ^fdiltiazem, verapamil; ^gdigoxin, digitoxin.

Table 2

Major adverse events and hospital admissions

	Type of AF at baseline				P-value
	First detected (n= 708)	Paroxysmal (n= 1170)	Persistent (n= 886)	Permanent (n= 1126)
Major adverse events during 1 year
All-cause death	43 (5.7)	43 (3.5)	27 (3.0)	100 (8.2)	<0.001
Cardiovascular death	14 (1.9)	15 (1.3)	19 (2.1)	43 (3.6)	0.001
Ischaemic stroke	9 (1.3)	22 (1.9)	11 (1.2)	19 (1.6)	0.582
Transient ischaemic attack	5 (0.7)	9 (0.8)	12 (1.4)	30 (2.5)	0.001
Coronary artery disease	46 (6.6)	63 (5.6)	38 (4.3)	71 (6.1)	0.005
New CADa	12 (2.6)	19 (2.5)	10 (1.7)	20 (2.8)	0.567
CAD worseningb	34 (14.9)	44 (11.8)	28 (10.4)	51 (11.8)	0.482
Other thromboembolismc	1 (0.1)	1 (0.1)	1 (0.1)	12 (1.0)	0.001
Intracranial haemorrhage	1 (0.1)	3 (0.3)	1 (0.1)	4 (0.3)	0.702
Other major bleeding	11 (1.6)	13 (1.1)	7 (0.8)	28 (2.4)	0.018
Heart failure	66 (9.5)	109 (9.6)	75 (8.5)	195 (16.6)	<0.001
New onset heart failured	35 (6.8)	44 (5.0)	18 (3.0)	37 (6.0)	0.025
Worsening of existing heart failuree	31 (17.7)	64 (26.9)	57 (20.3)	158 (28.6)	0.006
Hospital admissions during 1 year
For atrial fibrillation	120 (18)	372 (33)	250 (29)	197 (17)	<0.001
For any cardiovascular reason	204 (30)	488 (43)	339 (39)	403 (35)	<0.001
For non-cardiovascular reason	74 (11)	127 (11)	87 (10)	173 (15)	0.002

Results are reported as observed number (proportion within column). ^aProportion calculated among patients without coronary artery disease at baseline; ^bProportion calculated among patients with coronary artery disease at baseline; ^cpulmonary or peripheral embolism; ^dProportion calculated among patients without heart failure at baseline; ^eProportion calculated among patients with heart failure at baseline. CAD, coronary artery disease.

Multivariable logistic regression was performed to assess which factors were associated with all-cause mortality and the combined endpoint of all-cause mortality, any thromboembolism, major bleeding, and heart failure during 1 year follow-up. Patient characteristics and management factors were taken in account in these models. We did not aim to build prognostic models, but rather explorative models to generate hypotheses. We stress that caution is warranted with drawing conclusions on causal relations. Further, the effect of management might be biased due to their observational character, the short follow-up duration and the absence of randomization. For all variables incorporated in the two multivariable logistic regression models, we report the net odds ratio and its 95% confidence interval, and P-value.

Results

Population

Of the 5333 patients who were initially enrolled, 5272 were alive and 43 deceased at the end of the baseline admission/visit, whereas survival status was unknown for the remaining 18 patients. At 1-year follow-up, survival status was known for 4192 (80%) patients and these patients are described in this manuscript. Median follow-up duration (25–75th percentile) was 379 (366–418) days for patients alive at follow-up and 164 (78–312) days until decease. Patient characteristics per AF type were previously reported for all patients enrolled at baseline.¹ Compared with patients with follow-up data available, patients lost to follow-up were at baseline more often enrolled in central European countries (36 vs. 27%; P < 0.001), less often at the outpatient clinic (21 vs. 37%; P < 0.001), more often had heart failure (42 vs. 32%; P < 0.001), more often were symptomatic (76 vs. 67%; P < 0.001), and less often received oral anticoagulation (OAC) (54 vs. 68%; P < 0.001). No association was observed between AF type and proportion of patients lost-to-follow-up.

Atrial fibrillation progression

Figure 1 shows that 46% of patients with initially a first detected AF episode did not have an AF recurrence during 1 year hereafter. Further, paroxysmal mainly remained paroxysmal AF (80%), 30% of persistent AF progressed to permanent AF and permanent largely remained permanent AF (96%). Of all patients, 6% was considered cured. Patients with initially first detected AF were least often symptomatic at the time of follow-up (Table 1). Of permanent AF patients, 8% was not in AF according to the last recorded ECG.

Figure 1

AF type categorization after 1 year per baseline AF type

Compared with the patients remaining in the category of first detected, paroxysmal or persistent AF, patients who progressed to a more persistent form of AF were older (68 ± 12 vs. 64 ± 13 years; P < 0.001), more often had heart failure (39 vs. 21%; P < 0.001), and valvular heart disease (28 vs. 19%; P < 0.001), and had a larger left atrium at baseline (47 ± 8 vs. 44 ± 8 mm; P < 0.001). During 1-year follow-up, patients who progressed to persistent forms of AF more often had new onset (10 vs. 4%; P < 0.001) or worsening of existing heart failure (30 vs. 19%; P = 0.001) and a thromboembolism (4.5 vs. 2.0%; P = 0.001), and they more often had AF at the last known ECG (85 vs. 12%; P < 0.001) and symptoms at 1 year (41 vs. 26%; P < 0.001). These patients equally often underwent a cardioversion or catheter ablation (33 vs. 37% in the remaining patients; P = 0.051) during follow-up.

Management

Of all patients, 53% underwent chest X-ray, 62% echocardiography, and 38% thyroid hormone levels measurement during follow-up. When additionally taking in account these proportions in the baseline survey, respectively, 88, 94, and 68% had undergone these procedures at least once at any time point.

Any antithrombotic drug in general, and specifically OAC, was applied less often in first detected and paroxysmal AF (Table 1). After 1 year, OAC was discontinued in 16% and started in 19% of all patients (Figure 2). OAC was discontinued in 45% of patients with first detected AF and no detected recurrence since, and overall in 63% of patients who were considered cured. The most frequently reported reasons not to prescribe OAC at 1 year in patients at high risk for stroke were ‘no indication’ and ‘chronic sinus rhythm’ (both 26%). The main reason for starting OAC in patients at high risk for stroke was ‘persistent/permanent AF’ (37%), followed by ‘high risk factor for stroke’ (28%).

Figure 2

Antithrombotic therapy at 1 year per baseline antithrombotic regimen. OAC, oral anticoagulation; AP, antiplatelet drug

Class I and III antiarrhythmic drugs were most often prescribed in paroxysmal and persistent AF, but also 17% of permanent AF patients received a class III drug. Of surviving patients who underwent rhythm control at baseline, 27% did not receive rhythm control during 1-year follow-up and of patients under rate control at baseline 15% underwent rhythm control (Figure 3). Either rhythm or rate control was applied in 31% of patients who initially did not receive any heart rhythm therapy.

Figure 3

Heart rhythm strategies at 1 year per baseline heart rhythm strategy

Morbidity and mortality

In the overall population, 63 patients (1.6%) suffered an ischaemic stroke, 173 patients (4.3%) any thromboembolic complication, 9 patients (0.2%) an intracranial haemorrhage (ICH), and 60 patients (1.5%) another type of major bleeding during 1 year. Furthermore, 135 patients (5.0%) had new onset heart failure, and 314 (24.7%) worsening of existing heart failure. Symptomatic coronary artery disease was observed in 160 patients (12.1%), and newly diagnosed coronary artery disease in 62 patients (2.4%). Table 2 shows that in general permanent AF patients most often suffered from these morbidities.

During 1-year follow-up 221 patients (5.3%) died. For 156 of these patients the cause of death was reported and was cardiovascular in 97 patients (67%). Mortality was clearly highest in permanent AF and also higher in first detected AF compared with paroxysmal and persistent AF (Table 2 and Figure 4). The increased mortality in first detected AF persisted beyond the acute phase, and levelled after approximately 6 months. Mortality among patients with an ischaemic stroke was 21%, with ICH 67%, and with another type of major bleeding 10%.

Figure 4

Survival from all-cause mortality during 1 year per baseline AF type

Old age, heart failure, a prior major bleeding, ventricular tachycardia, renal failure, and malignancy were strongly associated with mortality and the occurrence of any event (Table 3). Patients having sinus rhythm after the baseline visit had a lower risk of mortality and any event than patients who were in AF, but no significant effect of rhythm and rate control was observed. Also the prescription of an angiotensin converting enzyme (ACE) inhibitor was associated with decreased all-cause mortality, whereas prescription of a statin was associated with a reduced risk for any event. Prescription of a diuretic was associated with an increased risk for both endpoints.

Table 3

Multivariable logistic regression results regarding factors associated with all-cause mortality and the combined endpoint of all-cause mortality, any thromboembolism, major bleeding and heart failure during 1 year

	All-cause mortality			Combined endpointa
	OR	95% CI	P-value	OR	95% CI	P-value
Patient characteristics
Age, per year increase	1.04	1.02–1.06	<0.001	1.02	1.01–1.03	<0.001
Female gender	0.78	0.53–1.16	0.221	0.91	0.72–1.14	0.398
Coronary artery disease	1.01	0.67–1.51	0.971	1.19	0.94–1.52	0.154
Valvular heart disease	1.01	0.68–1.48	0.979	1.17	0.92–1.47	0.204
Hypertension	0.69	0.46–1.03	0.070	1.04	0.81–1.34	0.751
Diabetes	1.24	0.82–1.88	0.320	1.04	0.80–1.36	0.755
Heart failure	1.92	1.25–2.93	0.002	2.80	2.18–3.58	<0.001
Stroke/TIA	1.42	0.87–2.30	0.167	1.06	0.77–1.47	0.732
Other thromboembolism	1.15	0.47–2.80	0.758	1.26	0.73–2.19	0.409
Major bleeding	3.98	1.89–8.39	0.001	4.68	2.52–8.69	<0.001
Minor bleeding	1.80	0.76–4.23	0.202	1.70	0.93–3.09	0.089
Sick sinus syndrome	1.07	0.54–2.10	0.854	1.12	0.73–1.72	0.606
Thyroid disease	0.54	0.28–1.05	0.053	1.04	0.73–1.47	0.835
Ventricular fibrillation	0.85	0.25–2.83	0.784	0.92	0.39–2.21	0.856
Ventricular tachycardia	2.74	1.31–5.73	0.012	2.74	1.62–4.64	<0.001
Renal failure	1.82	1.06–3.12	0.034	1.84	1.26–2.68	0.002
Malignancy	2.46	1.46–4.17	0.002	1.68	1.13–2.51	0.012
Peripheral vascular disease	1.13	0.64–2.00	0.670	1.30	0.91–1.87	0.157
COPD	1.55	0.99–2.41	0.059	1.67	1.26–2.21	<0.001
Admission characteristics
Reason for admission/visit			0.050			0.202
AF only	1	Reference		1	Reference
AF and other reason	1.29	0.76–2.17		1.20	0.91–1.60
Other reason only	1.90	1.09–3.31		1.35	0.97–1.89
Type of AF			0.083			0.046
First detected AF	1	Reference		1	Reference
Paroxysmal AF	0.69	0.39–1.23		0.96	0.68–1.34
Persistent AF	0.50	0.27–0.94		0.68	0.48–0.98
Permanent AF	0.52	0.30–0.89		1.06	0.74–1.51
AF related symptoms			0.163			0.802
Current symptoms	1	Reference		1	Reference
Previous symptoms	1.26	0.78–2.01		0.95	0.71–1.27
Never symptoms	1.59	0.98–2.56		0.90	0.65–1.26
Sinus rhythm at discharge	0.40	0.24–0.65	<0.001	0.73	0.55–0.96	0.026
Management
Minimal recommended diagnosticsb	1.00	0.68–1.46	0.992	0.87	0.70–1.09	0.230
Heart rhythm strategy			0.796			0.305
Rhythm control	1	Reference		1	Reference
Rate control	1.12	0.72–1.75		0.83	0.63–1.09
No therapy	0.92	0.45–1.89		0.76	0.48–1.22
Oral anticoagulation	0.99	0.60–1.63	0.966	0.89	0.66–1.21	0.460
Antiplatelet drug	1.42	0.88–2.29	0.153	1.09	0.81–1.46	0.578
Beta blocker	1.08	0.73–1.61	0.690	0.97	0.77–1.23	0.803
ACE inhibitor	0.61	0.40–0.91	0.016	1.06	0.82–1.37	0.651
AT II receptor blocker	0.63	0.36–1.10	0.093	1.01	0.72–1.42	0.953
Diuretic	1.61	1.02–2.54	0.039	1.38	1.07–1.79	0.014
Dihydropyridin Ca2+ antagonist	0.98	0.56–1.72	0.938	1.03	0.75–1.43	0.845
Statin	0.77	0.50–1.20	0.247	0.76	0.58–0.98	0.031

^aCombined endpoint of all-cause mortality, any thromboembolism, major bleeding and heart failure; ^bMinimal diagnostic work-up according to the 2001 AF guidelines: chest X-ray, echocardiography, and thyroid hormone levels measurement bold printed P-values indicate significant differences. ACE, angiotensin converting enzyme; AT, angiotensin; OR, odds ratio; CI, confidence interval; COPD, chronic obstructive pulmonary disease.

Discussion

The EHS on AF provides unique prospective data on management and prognosis of AF during 1 year in a European cardiology-based patient cohort. These results stress the difficulty and dynamics of treating AF with its many faces and accomplices and provide an insight in the prognosis of AF patients in current cardiology practice.

Atrial fibrillation progression

Progression of the arrhythmia to more intractable forms related to underlying disease and the extent of structural atrial remodelling. In addition, progression seems associated with persistence of AF symptoms as well as occurrence of heart failure and strokes. Our findings on AF progression are largely in agreement with previous findings of the ALFA and CARAF studies.^10,12

Half of the patients with initially first detected AF did not have a recurrence during 1-year follow-up, only one out of four had AF on their last recorded ECG and only 15% was symptomatic at 1 year. This favourable arrhythmia outcome might partly be due to simultaneous detection and successful management of AF or the associated disease. Nevertheless, survival was not as favourable as in paroxysmal and persistent forms of AF and undetected—asymptomatic—AF episodes may have had deleterious effects.

Only a minority of paroxysmal AF patients progressed to persistent or permanent AF. This might indicate that from a clinical perspective paroxysmal AF just slowly progresses to more persistent forms of AF, which may relate to slow structural deterioration of the atria. In contrast, once having persistent AF, many patients progress to permanent AF. In only one-third of patients progressing to persistent forms of AF rhythm control intervention was attempted during follow-up, indicating that other factors play a role when the physician indicates that the AF has become more relentless.

Remarkably, permanent AF patients were most often symptomatic at 1 year, which is in contrast to the lowest proportion observed at baseline. This possibly relates to the fact that permanent AF patients more often have chronic symptoms because of the higher prevalence of associated diseases, rather than intermittent symptoms due to AF itself. Also, 8% of permanent AF patients was reported not to be in AF at the last recorded ECG. This could mean that either recurrence of AF was not recorded on ECG, but probably noted by symptoms or on simplistic rhythm recordings, which obviously would not be recommended practice, or that there still is some indistinctness about the definitions of persistent and permanent AF.

Management

Although almost all patients underwent chest X-rays and echocardiographic tests, 30% did not undergo thyroid hormone levels measurement at any point in time, which is highly recommended by the guidelines.² Nevertheless, performance of a minimal recommended diagnostic work-up did not affect prognosis. This finding may relate to the use of alternative diagnostic procedures or a low yield of these investigations in general, thus precluding finding a beneficial effect.

In agreement with the baseline survey, first detected and paroxysmal AF less often received OAC than persistent and permanent AF.¹⁵ Also, at 1 year OAC seemed mainly discontinued when there was no AF recurrence observed or when AF was considered cured. Apparently, physicians weigh the perceived AF burden when deciding on antithrombotic treatment, which is however unreliable due to the high incidence of asymptomatic attacks and lack of adequate non-invasive monitoring tools. The current guidelines dissuade to consider the type of AF for selecting the appropriate antithrombotic drug.¹⁶ Although we could not show this, it is tempting to assume that the low OAC application rate and frequently stopping of OAC in patients with first detected AF contributed to their high mortality rate.

During 1 year, one-third of patients initially on rhythm control did not receive rhythm control therapy anymore and of patients initially on rate control 20% did receive rhythm control. These results stress the dynamics of assessing on-treatment rhythm and rate control therapy, rather than querying the physician on the intention-to-treat. That 17% of patients with permanent AF at follow-up did receive a class III antiarrhythmic drug might relate to some lack of clarity about the definitions of persistent and permanent AF, but possibly also to the fact that these drugs have rate control properties and are given for this purpose.

Mortality and morbidity

One year mortality in patients suffering from AF was high and the risk is continuously present. Mortality (5.3%) was comparable with results of previous studies, which is also the case for the observed higher mortality in permanent AF than in other AF types.^4,8,10 Patients lost to follow-up were sicker or less well anticoagulated or both, suggesting that the event rate might have been even higher than we recorded. Patients with first detected AF had a worse prognosis than patients with paroxysmal or persistent AF. This might relate to the new detection of both AF itself and possibly also associated diseases, which implies more often unstable and less well managed patients when compared with paroxysmal and persistent AF. Survival of first detected AF patients during 1 year (94%) was in general much higher than in previous studies on patients hospitalized with newly detected AF, which was around 80%.^5,6,9,12,17 This difference might be due to EHS patients partly being enrolled at the outpatient clinic and few at the first (heart) aid department.

Old age, malignancy, renal failure, and heart failure, as well as major bleeding all significantly contributed to the high mortality rate, whereas sinus rhythm at baseline improved survival. Many of these parameters are not readily modifiable and should be seen as a marker of bad prognosis rather than causative. Similarly, presence of sinus rhythm at inclusion may represent a marker of good prognosis rather than being instrumental in enhancing survival, especially since rhythm control did not contribute to survival. Nevertheless, we feel that aggressive management of modifiable cardiovascular factors should be undertaken to improve prognosis in AF patients. One of these options is prescription of a blocker of the renin–angiotensin–aldosteron system. The lower mortality in patients receiving an ACE inhibitor is probably due to their proven protective effect in heart failure and possibly hypertension, but might also be mediated by a lower recurrence rate of AF.^18,19 The same effect might be true for AT II receptor blockers, although their effect on mortality was near significant in our analysis. We suspect that the negative associations of diuretics with both endpoints represent the fact that especially patients in an advanced state of heart failure received this drug, which is supported by the fact that the negative association of heart failure with both endpoints strongly increased when we left diuretics out of the analyses.

Incidence of ischaemic stroke was as low as that observed in randomized controlled trials (RCT) among patients on OAC, which is probably partially due to the high OAC application rate in our survey.²⁰ Also in agreement with RCT results was that ICH rarely occurred, but when it occurred it was more lethal than ischaemic stroke and other major bleedings. Heart failure worsened in one out of five heart failure patients and in agreement with previous findings patients with newly detected AF had approximately a 7% risk of developing new HF within the first year after diagnosis.²¹

Strengths and limitations

The EHS is a unique large observational study that prospectively assesses mortality, morbidity, AF progression, and treatment changes across the broader spectrum of AF patients in daily cardiology practice. Follow-up was available for 80% of initially enrolled patients, which is adequate for this kind of study.

We might have underestimated mortality and incidence of adverse events during 1 year in this cohort, since patients lost to follow-up were more often admitted in the hospital at the time of the baseline survey, more often had heart failure and less often received OAC. In addition, it is plausible that lost to follow-up was caused by death, transfer to a nursing home (possibly due to disability as a result of ischaemic stroke), or refusal to participate due to a bad health state.

Clinical implications

Progression of AF to more intractable forms relates to underlying disease and extent of structural atrial remodelling and heralds complaints, heart failure, and strokes. We feel that focusing on associated cardiovascular diseases may not only retard AF progression, but also reduce the disease burden in patients with first detected and paroxysmal AF.

One year mortality in patients suffering from AF is high. Most important it is also high in first detected AF. Most death factors are not well modifiable. Although sinus rhythm heralded a good prognosis, we do not advocate rhythm control to improve prognosis since presence of sinus rhythm most probably is a marker for easy-to-treat AF as well as less severe associated cardiovascular disease. Nevertheless, we feel that aggressive management of modifiable cardiovascular factors—including anticoagulation as needed—should be undertaken to improve prognosis, which probably is especially promising in first detected AF.

Performing tests as recommended by the Guidelines does not contribute to improved prognosis, probably relating to low yield of abnormalities of some tests (like thyroid function tests), and unfavourable prognosis of abnormalities seen on other tests like echocardiography. Nevertheless, these tests should be done to find correctable causes of AF, to assess feasibility and necessity of specific therapies as well as to follow effects of therapy.

The many treatment changes during 1 year indicate the difficulty of treating a progressive disease like AF with its variety of clinical presentations and many associated diseases. In addition, these changes might relate to the fact that current treatment measures are not ideal or that research evidence is inconclusive for subgroups of AF patients that are underrepresented in clinical trials. Research to develop anticoagulation not requiring constant INR monitoring, more efficacious and safer antiarrhythmic drugs and to tailor management of specific patient subgroups might lead to an enhanced physicians’ confidence to apply medical therapy and could make treatment more stable.

Funding

AstraZeneca, Sanofi-Aventis, Eucomed.

Acknowledgements

We thank the Euro Heart Survey team, national co-ordinators, investigators and data collection officers for performing the baseline and follow-up surveys. In addition, we thank the main sponsor AstraZeneca, major sponsor Sanofi-Aventis, and sponsor Eucomed for their support to the Euro Heart Survey on Atrial Fibrillation. Finally, we thank the following institutions for their national support: Austrian Heart Foundation, Austrian Society of Cardiology, French Federation of Cardiology, Hellenic Cardiological Society, Netherlands Heart Foundation, Portuguese Society of Cardiology, Spanish Cardiac Society, and the Swedish Heart and Lung Foundation.

Conflict of interest: none declared.