OUP user menu

Nurse-led care vs. usual care for patients with atrial fibrillation: results of a randomized trial of integrated chronic care vs. routine clinical care in ambulatory patients with atrial fibrillation

Jeroen M.L. Hendriks, Rianne de Wit, Harry J.G.M. Crijns, Hubertus J.M. Vrijhoef, Martin H. Prins, Ron Pisters, Laurent A.F.G. Pison, Yuri Blaauw, Robert G. Tieleman
DOI: http://dx.doi.org/10.1093/eurheartj/ehs071 2692-2699 First published online: 27 March 2012


Aims The management of patients with atrial fibrillation (AF) is often inadequate due to deficient adherence to the guidelines. A nurse-led AF clinic providing integrated chronic care to improve guideline adherence and activate patients in their role, may effectively reduce morbidity and mortality but such care has not been tested in a large randomized trial. Therefore, we performed a randomized clinical trial to compare the AF clinic with routine clinical care in patients with AF.

Methods and results We randomly assigned 712 patients with AF to nurse-led care and usual care. Nurse-led care consisted of guidelines based, software supported integrated chronic care supervised by a cardiologist. The primary endpoint was a composite of cardiovascular hospitalization and cardiovascular death. Duration of follow-up was at least 12 months. Adherence to guideline recommendations was significantly better in the nurse-led care group. After a mean of 22 months, the primary endpoint occurred in 14.3% of 356 patients of the nurse-led care group compared with 20.8% of 356 patients receiving usual care [hazard ratio: 0.65; 95% confidence interval (CI) 0.45–0.93; P= 0.017]. Cardiovascular death occurred in 1.1% in the nurse-led care vs. 3.9% in the usual care group (hazard ratio: 0.28; 95% CI: 0.09–0.85; P= 0.025). Cardiovascular hospitalization amounted (13.5 vs. 19.1%, respectively, hazard ratio: 0.66; 95% CI: 0.46–0.96, P= 0.029).

Conclusion Nurse-led care of patients with AF is superior to usual care provided by a cardiologist in terms of cardiovascular hospitalizations and cardiovascular mortality.

Trial registration information: Clinicaltrials.gov identifier number: NCT00391872.

  • Atrial fibrillation
  • Mortality
  • Hospital admissions
  • Adherence to guidelines
  • Nursing


Atrial fibrillation (AF) is the most common cardiac arrhythmia and increasing in prevalence with age.1,2 It is becoming an extremely costly public health problem,3,4 with hospitalization as the primary cost driver.5 Atrial fibrillation is often inadequately managed due to deficient adherence to guidelines.6,7 Non-adherence relates to disease complexity with large differences between patients concerning risks of stroke, heart failure, and coronary syndromes. The growing number of patients, the complexity of the disease and the frequent non-adherence ask for a change in management of AF. Nurse-led multidisciplinary heart failure clinics were shown to improve clinical outcomes.8,9 In contrast, it has never been investigated whether a multidisciplinary AF clinic may likewise improve cardiovascular outcomes in patients with AF. Also, to our knowledge, the effects of decision support software in the management of these patients have not been described. Previously, we demonstrated that a nurse-led multidisciplinary AF clinic can be implemented safely,10 but a prospective comparison with standard care is lacking. Therefore, we performed a randomized trial to compare nurse-led care with usual care to prevent cardiovascular hospitalization or death in patients with AF in the outpatient setting. We hypothesized that nurse-led care is superior to usual care.


Study design

This prospective randomized clinical trial took place in the Maastricht University Medical Centre and recruited patients between January 2007 and December 2008. Follow-up was at least 1 year. The design of the study has been described previously.11 The study complies with the Declaration of Helsinki and ethical approval was obtained from the Institutional Review Board. The study was registered at Clinicaltrials.gov (identifier: NCT00753259). The steering committee was responsible for the design and conduct of the study, data analysis and manuscript preparation. All authors reviewed previous versions of the manuscript and vouch for accuracy and completeness of data and analyses.

Study participants and randomization

All patients ≥18 years referred for AF (documented on an electrocardiogram) by general practitioners or non-cardiology specialists to our outpatient department were eligible.11 Patients were asked permission for use of their personal clinical data to be collected during their future visits. At the same time, we informed them about the AF clinic and nurse-led care, as well as the possibility of participation in a clinical trial. Patients were then randomly assigned to nurse-led care or usual care, following a computerized one to one randomization. Patients were excluded in case of any comorbidity which is unsatisfactorily treated, e.g. unstable and uncontrolled hypertension, unstable heart failure defined as NYHA IV or necessitating hospital admission <3 months before inclusion, untreated hyperthyroidism, current or foreseen pacemaker, internal cardioverter defibrillator or cardio resynchronization therapy, or cardiac surgery <3 months before inclusion.11


The care provided in the AF clinic was based on the chronic care model.12,13 It consisted of nurse-led outpatient care steered by decision support software based on the guidelines2 and supervised by a cardiologist.11 Before the first visit, patients underwent laboratory testing, electrocardiogram, Holter monitoring and echocardiography. At the first visit, a nurse specialist took the patient's history and informed them about the pathophysiology of AF, its symptoms and possible complications, the results of the diagnostic tests and treatment options. The dedicated software CardioConsult AF® (Curit BV, Groningen, The Netherlands) was used to guide comprehensive management of AF and associated cardiovascular conditions. The software determined the individual patient profile based on symptoms, type of AF, and stroke risk. Thereafter, it proposed the most appropriate management. To further empower patients they were instructed about rate and rhythm control as well as prophylactic vascular therapy (including strict anticoagulation monitoring), and about when to report to the hospital. At the end of the consultation, the nurse specialist was supervised by a cardiologist, endorsing the proposed diagnoses and treatments. Visits to the nurse were scheduled to last 30 min. Follow-up visits were planned at 3, 6, and 12 months, and every 6 months thereafter. Patients could contact the nurse in person or by telephone between planned visits as needed. During follow-up visits, psychosocial support and educational interventions were repeated.

Patients in the control group received usual care by a cardiologist in the outpatient clinic during visits scheduled to last 20 min for the first visit and 10 min for follow-up visits.

During follow-up visits, patients were questioned for major adverse cardiovascular events and hospitalizations. Also all medical records were reviewed for such events after 1 and 2 years, and at the end of follow-up.


The primary outcome was a composite of death from cardiovascular causes, and cardiovascular hospitalization (i.e. hospitalization with an overnight stay) for heart failure, ischaemic stroke, acute myocardial infarction, systemic embolism, major bleeding, severe arrhythmic events, and life-threatening adverse effects of drugs. All primary outcome events were adjudicated on the basis of pre-specified criteria by an independent clinical endpoint committee that was not aware of the randomized treatment assignments. Cardiovascular deaths were classified according to a modified Hinkle and Thaler classification.14 Heart failure was defined as heart failure independent of the left ventricular ejection fraction, preferably confirmed by N-terminal pro-brain natriuretic peptide assessment, and necessitating hospitalization. Thrombo-embolic complications included stroke (sudden onset of focal deficit caused by occlusion of major cerebral artery, confirmed by a neurologist on the basis of computerized tomography or magnetic resonance imaging), systemic embolization (acute occlusion of an extremity or organ, confirmed by imaging or surgery) and myocardial infarction (two of three characteristics: chest pain, ischaemic electrocardiographic changes, elevated blood levels of key biochemical markers), all necessitating hospitalization. Major bleeding: fatal bleeding, or intracranial bleeding, or a reduction in the haemoglobin level >1.25 mmol/L and requiring blood transfusion with hospitalization. Arrhythmic events included AF, cardiac syncope, sustained ventricular tachycardia, and cardiac arrest, all associated with hospitalization. Life-threatening adverse effects of rate or rhythm controlling drugs included pro-arrhythmia of Vaughan-Williams classes I and III anti-arrhythmic drugs, digitalis intoxication, drug-induced heart failure, or conduction disturbances, all necessitating hospitalization.

Statistical analysis

The primary analysis was performed to test whether nurse-led care was superior to usual care. A sample size of 349 patients in both the intervention and control groups was calculated on a reduction from 35 to 25% for the primary endpoint, using an alpha of 0.05 and a power of 0.80. All analyses were performed on an intention-to-treat basis. We calculated Kaplan–Meier estimates and their 95% confidence intervals (CIs). To assess the risk for cardiovascular hospitalization or cardiovascular death in both groups, Cox proportional hazards modelling was used to calculate hazard ratios and their 95% CIs, both crude and adjusted for sex, age, and the presence of cardiovascular disease at the time of the patient's referral. The latter included heart failure, hypertension, diabetes and prior stroke, or transient ischaemic attack. The assumptions of the Cox proportional hazards model was checked by visual inspection of the log–log survivor function-by-time curve. To assess implementation of care as recommended by the AF guidelines in both groups, we tested for differences using Fisher's exact test after calculating the cumulative percentage of patients in whom one through six recommendations were followed. We did all calculations with SPSS, version 16.0 (SPSS, Chicago, IL, USA).


Patients and implementation of care

A total of 712 patients were enrolled in the study: 356 in the nurse-led care group and 356 in the usual care group (Figure 1). The groups were well matched without significant differences in baseline characteristics (Table 1). None of the patients was lost to follow-up. Figure 2 shows the adherence to guidelines recommendations. At the end of the first visit appropriate prescription of anti-thrombotic treatment was seen in 99 and 83% of patients in the nurse-led care and usual care groups, respectively. Thyroid function testing was done in 91 vs. 54%, and echocardiography in 91 vs. 82% of patients, in both treatment groups, respectively. Appropriate application of rhythm control was more frequent in the nurse-led care group, i.e. it was not applied in asymptomatic patients (95 vs. 85%), no rhythm control drugs were prescribed if contraindicated (87 vs. 82%), and no such drugs in patients with permanent AF (97 vs. 93%, respectively). In total 2732 outpatient department visits, 551 first heart aid visits and 362 telephone contacts were produced from inclusion to end of follow-up. In the nurse-led care arm, an estimated 407h were spent in total by the nurse specialist and 265 h by the supervising cardiologist. In the control group, an estimated 436 h were spent in total by the cardiologist. The level of knowledge the patients had of their disease and its management was scored using a heart failure questionnaire15 modified for AF. At 1 year after inclusion, patients in the nurse-led group scored significantly better than usual care patients (data not shown).

View this table:
Table 1

Characteristics of the patients according to treatment group

CharacteristicNurse-led care (n = 356)Usual care (n = 356)
Age, year66 ± 1367 ± 12
Male sex, n (%)197 (55.3)221 (62.1)
Type of AF, n (%)
 Paroxysmal190 (53.4)203 (57.0)
 Persistent68 (19.1)44 (12.4)
 Permanent75 (21.1)84 (23.6)
 Symptomatic294 (82.6)296 (83.1)
History of underlying disease, n (%)
 Hypertension187 (52.5)193 (54.2)
 Diabetes mellitus50 (14.0)46 (12.9)
 Previous stroke/TIA44 (12.4)45 (12.6)
 Coronary artery disease33 (9.3)38 (10.7)
 Myocardial infarction19 (5.3)22 (6.2)
 Chronic obstructive lung disease29 (8.1)31 (8.7)
 Congestive heart failure25 (7.0)25 (7.0)
 Peripheral vascular disease13 (3.7)20 (5.6)
 Hyperthyroidism12 (3.4)12 (3.4)
 Mitral valve heart disease7 (2.0)10 (2.8)
 Aortic valve heart disease5 (1.4)11 (3.1)
 No underlying heart disease6 (1.7)7 (2.0)
CHADS2 score, n (%)a
 0107 (30.0)95 (26.7)
 1122 (34.3)135 (37.9)
 >1127 (35.7)126 (35.4)
Treatment, n (%)
 Beta-blocker164 (46.1)187 (52.5)
 Digitalis59 (16.6)43 (12.1)
 Verapamil44 (12.4)18 (5.1)
Vaughan-Williams class I and III antiarrhythmic
 Drugs105 (29.1)88 (24.7)
Other medication, n (%)
 Angiotensin receptor blocker88 (24.7)80 (22.5)
 Angiotensin-converting enzyme inhibitor72 (20.2)67 (18.8)
 Diuretic56 (15.7)67 (18.8)
 Statin119 (33.4)99 (27.8)
 Vitamin K antagonist218 (61.2)188 (52.8)
 Aspirin118 (33.1)108 (30.3)
Heart rate
 Mean, b.p.m.80 ± 2285 ± 26
 >100 b.p.m., n (%)54 (15.2)78 (21.9)
 Body-mass index, kg/m227.1 ± 4.927.3 ± 5.2
Blood pressure, mmHg
 Systolic141 ± 20.6143 ± 24.7
 Diastolic79 ± 10.883 ± 13.7
Echocardiographic findings
 Size of left atrium, long axis, mm42 ± 643 ± 8
 Left ventricular end-diastolic diameter, mm49 ± 649 ± 6
 Left ventricular end-systolic diameter, mm34 ± 634 ± 6
 Septal thickness, mm9 ± 19 ± 1
 Posterior wall thickness, mm9 ± 19 ± 1
 Left ventricular ejection fraction, %57 ± 1056 ± 12
  • aThe CHADS2 score is a stroke risk classification scheme, using a point system ranging from 0 to 6, to determine the yearly risk index.31 Congestive heart failure, hypertension, age 75 years or above, and diabetes are assigned one point each. Previous stroke or transient ischaemic attack is assigned two points. The score is calculated by summing all points for a patient.

Figure 1

CONSORT flow diagram showing flow of patients and reasons for exclusion and protocol deviations.

Figure 2

(A) The cumulative adherence to six guidelines recommendations in the nurse-led care group vs. the usual care group, see text for details concerning recommendations tested. P-values represent statistical differences concerning guidelines implementation between the two groups. (B) The distribution of the proportion of patients adhering to only one through all six guidelines recommendations in the two arms of the study.

Primary outcome

During a mean follow-up of 22 months, the primary outcome was reached in 125 patients: 51 patients (14.3%) in the nurse-led care and 74 (20.8%) in the control group (hazard ratio: 0.65; 95% CI 0.45–0.93) (Table 2). After adjustment the hazard ratio was 0.63; 95% CI 0.44–0.90. Therefore, nurse-led care was superior to usual care with regard to prevention of a composite of cardiovascular hospitalization and mortality. Figure 3 shows Kaplan–Meier curves for the estimates of the first occurrence of the primary outcome over time in both groups.

View this table:
Table 2

Incidence of the primary outcome and its components according to treatment groupa

End pointNurse-led care (n = 356)Usual care (n = 356)Hazard ratiob (95% CI)
Composite end point (%)51 (14.3)74 (20.8)0.65 (0.45–0.93)
Cardiovascular death (%)4 (1.1)14 (3.9)0.28 (0.09–0.85)
 Cardiac arrhythmic (%)1 (0.3)2 (0.6)
 Cardiac non-arrhythmic (%)1 (0.3)4 (1.1)
 Vascular non-cardiac (%)2 (0.6)8 (2.3)
Cardiovascular hospitalization (%)48 (13.5)68 (19.1)0.66 (0.46–0.96)
 Arrhythmic events (%)18 (5.1)33 (9.3)
  Atrial fibrillation (%)15 (4.2)23 (6.5)
  Syncope (%)3 (0.8)7 (2.0)
  Sustained ventricular tachycardia (%)1 (0.3)
  Cardiac arrest (%)2 (0.6)
 Heart failure (%)14 (3.9)19 (5.3)
 Acute myocardial infarction (%)4 (1.1)2 (0.6)
 Stroke (%)3 (0.8)5 (1.4)
 Systemic embolism
 Major bleeding (%)6 (1.7)6 (1.7)
 Life-threatening effects of drugs (%)3 (0.8)3 (0.8)
  • aThe tabulation of the composite primary outcome includes the first event for each patient, whereas the tabulations of component events include all such events.

  • bHazard ratios from the univariate analysis.

Figure 3

Kaplan–Meier estimates of the cumulative incidence of the primary outcome in both groups. The primary outcome is a composite of the first occurrence of cardiovascular hospitalization or cardiovascular death. NLC, nurse-led care; UC, usual care.

Other outcomes

The rate of cardiovascular death was significantly lower in the nurse-led care group compared with the usual care group (1.1 and 3.9%, respectively, hazard ratio: 0.28; 95% CI 0.09–0.85) (Table 2). Heart failure death was seen in one patient in the nurse-led care group compared with four patients in usual care. One and two patients died suddenly, one and three patients had fatal pulmonary embolism and zero and three patients suffered from fatal stroke, in the nurse-led care and usual care groups, respectively. One nurse-led care patient succumbed from subdural haematoma. Fatal gastro-intestinal bleeding and unspecified cardiovascular mortality in a vascular compromised patient occurred in one patient each and only in the usual care group. Also the number of cardiovascular hospitalizations was significantly lower in the nurse-led care group (13.5 vs. 19.1%, respectively, hazard ratio: 0.66; 95% CI: 0.46–0.96). The adjusted hazard ratio is 0.64; 95% CI 0.44–0.93. Some patients were hospitalized twice or more for cardiovascular reasons leading to a total number of 55 and 87 hospitalizations in the nurse-led care and usual care groups, respectively. In total 453 hospitalizations and 42 deaths were reported during follow-up, including 269 non-cardiovascular hospitalizations and 24 non-cardiovascular deaths.

The hazard ratios and their 95% confidence intervals for the treatment effect on the primary outcome according to subgroups are presented in Figure 4. It shows that the relative effect between nurse-led care and usual care was consistent among subgroups, with the exception of female patients.

Figure 4

Hazard ratios and their 95% confidence intervals of the treatment effect on the primary outcome according to subgroups. See text for details.


We compared nurse-led care with usual care for stable patients with AF in the outpatient setting. Nurse-led care was associated with a higher relative efficacy with respect to prevention of the composite of cardiovascular death and cardiovascular hospitalization. The primary outcome occurred in 14.3% of patients with nurse-led care compared with 20.8% when usual care was delivered. The relative risk reduction in 35% by nurse-led care was substantial and represents the combined beneficial effects of our disease management system. Nurse-led care patients were better informed about their disease and its management. Also, guidelines recommendations were more comprehensively implemented in the intervention group representing the impact of the decision support software we used.

To the best of our knowledge, this is the first large outcomes study in nurse-led care for patients with AF. In a pilot study, Inglis et al.16 suggested that a nurse-led, multidisciplinary home-based intervention—using a heart failure algorithm—reduces hospital readmission and mortality in patients with AF but results were only convincing in the subset with heart failure. Obviously, this may relate to low number of patients, but also to lack of an integrated approach specifically directed to AF. Even though we included on average less sick patients who were not recently discharged from the hospital like in the study by Inglis et al., we still were able to demonstrate favourable effects of our intervention.

The rate of the primary endpoint was lower than expected. Nevertheless, we were able to show superiority of nurse-led care because of a slightly larger than expected relative risk reduction in outcome events in the intervention group. The rate of the primary endpoint was also lower than in previous reports with a similar composite endpoint,17,18 presumably because we included stable patients and did not exclusively select high-risk patients. Nonetheless, nurse-led care retained its beneficial effects in complex AF patients since our post hoc analysis shows that it is effective in elderly with heart failure and previous stroke (Figure 4). However, the trend in favour of nurse-led care was not seen in female patients (interaction P-value 0.007). Female patients had similar characteristics at baseline, a similar level of knowledge of their disease and similar implementation of guideline recommendations compared with male patients (data not shown). All the same, compared with men, women receiving usual care had a relatively low event rate for which we lack an explanation but which may have been a matter of chance. Similarly, the trend in favour of nurse-led care in patients with diabetes did not seem as strong as compared with the other medical conditions in the analysis but the number of events was low and the interaction appeared non-significant. Of note, all of these post hoc subanalyses should be interpreted with caution for low numbers of patients per subgroup.

Atrial fibrillation is a complex condition and frequently associated with admissions for stroke, heart failure, and AF as well as cardiovascular death.17,18 Also in the present study the composite primary endpoint was driven by these clinical events. Rhythm control is not helpful in this respect,19,20 although dronedarone may be an exception.18 In stead, comprehensive management of underlying cardiovascular disease should come first.21,22 Nevertheless, rhythm control is often applied when not indicated.23 Even attaining the optimal heart rate in permanent AF may be difficult, and—expectedly—recent new data will not find their way easily into clinical practice.24,25 Also, many patients in whom AF is associated with heart failure do not receive recommended drug treatment.6 We reasoned that—like in other chronic conditions26—the use of an electronic patient record with incorporated dedicated decision support software based on the guidelines may downsize complexity and improve adherence to recommendations. In addition, a multi-disciplinary approach in which cardiologists and nurses work closely together and have to justify reciprocally deviations from the protocol may largely preclude treatment decisions which do not comply with the guidelines. Moreover, using an electronic patient record in the outpatient setting challenges information transfer between multiple care providers and medication safety.27,28 The outcomes of the present study justify the notion that like in the aviation industry, modern medicine benefits from protocolized procedures and presence of a co-pilot, thereby preventing medical accidents. From the present study, we cannot tell whether straightforward application of checklists by stand-alone cardiologists would have yielded the same results as nurse-led co-piloted care, but for now it seems appropriate to strongly advocate the use of checklists whatever the care programme type.

The American Heart Association developed an eight domain classification system for disease management programmes for chronic diseases.29 AF is such a chronic condition. We studied the effects of disease management in consecutive stable outpatients with electrocardiographically documented AF. Key to the success of our approach, we believe, was the comprehensive intervention focusing on patient education, reassurance, prophylactic measures guided by electronic decision support based on the guidelines, time spent with the patients, and the teamwork between the nurse specialist and the cardiologist. These elements are essential in the chronic care model on which we based our intervention.11,13 The relative contribution of the different elements to lowering the composite primary endpoint is difficult to establish, but most tangible is the improved adherence to guidelines recommendations (Figure 2). Presumably, the intervention content may improve from involving general practitioners’ practices and intensifying education and peer support for patients. Novel communication systems may further ameliorate morbidity in AF but results may be disappointing—if used too much in isolation—as was recently the case with telemonitoring for heart failure.30 From the above it is clear that disease management programmes trigger costs, which relates among others to increased time spent with the patients. The latter was also the case in the present study. However, in the end these systems also save costs by preventing complications and hospitalizations. Obviously, it is this balance which will drive decisions to implement disease management systems like that presented here for AF.

Study limitations

The mechanisms through which the beneficial effect of nurse-led AF care was obtained are not immediately clear. Traditionally, stroke prevention using antithrombotic drugs has received much attention, including the major bleedings associated with it. Implementation of appropriate antithrombotic treatment was indeed more comprehensive in the nurse-led care group, but that did not lead to significantly fewer strokes and expectedly also not to fewer major bleeds. On the other hand, the enhanced implementation of appropriate antithrombotic treatment may have contributed to fewer vascular deaths. Nevertheless, our findings must be interpreted with caution, also because they derive from cardiology practice and may not hold for the general AF population.


Nurse-led care for stable AF is superior to usual care in terms of major clinical events. These findings should trigger disease management for AF similar to other chronic cardiovascular conditions like heart failure and diabetes.


The trial was supported by the University Hospital Maastricht as well as by unrestricted educational grants from Boehringer Ingelheim and Medtronic Bakken Research Centre. The sponsors were not in any way involved in the study.

Conflict of interest: none declared.


Members of the study group are as follows: Writing Committee: H.J.G.M.C. (Chair), J.M.L.H., R.G.T., M.H.P., H.J.M.V.; Steering Committee: H.J.G.M.C. (Chair), R.G.T., R.d.W., H.J.M.V., M.H.P.; Adjudication Committee: C. Franke, neurologist (Chair), H. ten Cate, internist, haematologist, V. van Ommen, cardiologist, R. Rennenberg, internist, vascular medicine.


View Abstract