Care and referral patterns in a large, dedicated nurse-led atrial fibrillation outpatient clinic

This single-centre, prospective cohort study was conducted in the Amsterdam University Medical Centre, a tertiary academic centre, in Amsterdam, the Netherlands from 2014 to 2018. Patients were eligible for inclusion if they were diagnosed with AF (paroxysmal or persistent) as confirmed with electrocardiography (ECG). Postoperative AF was diagnosed in patients with a new diagnosis of AF within 30 days of operation. The study was conducted according to the World Medical Association Declaration of Helsinki.

Consecutive patients underwent detailed cardiac work-up comprising registration of demographic characteristics, medical history, medication history and current prescription, 24-hour Holter rhythm monitoring and echocardiography (Fig. 1). Symptoms related to AF were classified according to the European Heart Rhythm Association (EHRA) symptom classification score, in which class I indicates no symptoms and while class IV indicates disabling symptoms [10]. All patients also completed a validated AF knowledge questionnaire prior to consultation with the nurse practitioner (Atrial Fibrillation Knowledge Scale) [11]. This questionnaire consists of multiple-choice questions in four domains: AF in general, AF therapy, AF symptom recognition and AF general attitudes. The initial questionnaire consisted of 11 questions; however, due to the current policy to prescribe a non-vitamin K antagonist oral coagulant (NOAC) to all newly diagnosed AF patients, one question concerning the function of the Thrombosis Centre with regard to regular international normalized ratio (INR) control in the Netherlands was no longer relevant.

Patients came in for an initial evaluation visit, after which they were either referred to a cardiologist or back to their GP or follow-up was continued at the nurse-led outpatient clinic. In principle, patients receiving class I or III antiarrhythmic drugs were not referred back to their GP. However, a few patients did not attend consecutive visits in the outpatient clinic and therefore stayed under control of their GP. In these cases, the GP was informed and advised to switch to another class of antiarrhythmic drugs. All newly referred patients were discussed by the nurse practitioner with a supervising cardiologist.

Anticoagulants, including NOACs, were prescribed to all patients with CHA₂DS_2-VASc score ≥ 1 (unless solely based on female sex), irrespective of the (presumed) absence of AF after ablation or the exclusion or closure of the left atrial appendage, according to current guidelines [12]. Rate control comprised antiarrhythmic drugs class II and IV and digoxin, while rhythm control comprised antiarrhythmic drug class I (i.e. flecainide) and III (i.e. sotalol, amiodarone), according to current guidelines [12]. A rate or rhythm control strategy was chosen based on the patient’s symptoms and preference and in deliberation with a supervising cardiologist. In selected cases where there was doubt about the burden of symptoms, electrical cardioversion was attempted to assess relieve of symptoms during sinus rhythm.

Patients who remained under nurse-led care were followed up with yearly outpatient clinic visits or more frequently if deemed necessary. Patients could also directly contact the nurse practitioner by telephone or email. A clinical consultation consisted of symptom assessment, rhythm monitoring (ECG) and blood screening. When necessary, exercise ECG or cardiac imaging was performed. Furthermore, the management of comorbidities associated with AF (mainly hypertension and obesity) was addressed in the nurse-led AF outpatient clinic. If patients reported symptoms related to obstructive sleep apnoea syndrome, polysomnography was conducted.

For the comparison of normally distributed continuous variables, an unpaired sample t-test was used; the results are expressed as mean ± standard deviation (SD). In case of not normally distributed continuous variables, the Mann-Whitney U test was used; the results are expressed as median with interquartile range (IQR). Categorical variables were compared with the Pearson χ² test and are expressed as frequency with percentage.

Patient characteristics are displayed in Tab. 1. After initial cardiac assessment, 139 patients (29.1%) remained under nurse-led care and 121 (25.3%) were referred to a cardiologist and 218 (45.6%) to a GP. Notably, patients who remained under nurse-led care were significantly younger and had less comorbidities than the other two groups. They more often had paroxysmal AF as their presenting arrhythmia. On the other hand, patients who remained under control of the nurse practitioner suffered more often from AF symptoms (higher EHRA score) than those who were referred back to their GP.

At the initial evaluation visit, 6 patients (4.3%) who remained under nurse-led care were prescribed class 1c antiarrhythmic drugs and 11 patients (7.9%) received a class III antiarrhythmic drug prescription (Tab. 2). Oral anticoagulants were prescribed to 328 patients (68.6%), of whom 249 (75.9%) received a NOAC prescription. NOACs were most frequently prescribed to patients under nurse-led care (65.5%) as opposed to 55.4% who were referred to a cardiologist and 41.7% who were referred to a GP (p = 0.004) (Tab. 3).

In a univariable analysis, age under 75 years and CHA₂DS₂-VASc score < 2 were associated with remaining under nurse-led care. Indeed, of the patients under nurse-led care, 115 (82.7%) were < 75 years, whereas 241 patients (71.1%) referred to a cardiologist of GP were younger than 75 years. Furthermore, 69 patients (49.6%) under nurse-led care had a CHA₂DS₂-VASc score ≥ 2, while this was true for 230 patients (67.8%) referred to a cardiologist or GP. After multivariable testing, only CHA₂DS₂-VASc score ≥ 2 was associated with not remaining under nurse-led care (OR 0.57, 95% CI 0.34–0.95 for remaining under nurse-led care) (see Fig. S1a in Electronic Supplementary Material).

Persistent AF, congestive heart failure and presence of valvular disease were associated with referral to a cardiologist in a univariable analysis. Persistent AF was seen in 53 patients (43.8%) referred to a cardiologist and in 108 patients (30.3%) who remained under nurse-led care or who were referred to a GP. Furthermore, 5 (4.1%) of the patients referred to a cardiologist had heart failure compared with 2 patients (0.6%) who remained under nurse-led care or were referred to a GP. Significant valvular disease was present in 6 patients (5.0%) referred to a cardiologist and in 4 patients (1.1%) who remained under nurse-led care or were referred to a GP. After multivariable testing, heart failure (OR 6.43, 95% CI 1.34–45.85) and valvular disease (OR 4.48, 95% CI 1.23–18.11) were strongly associated with referral to a cardiologist (see Fig. S1b in Electronic Supplementary Material).

Similarly, CHA₂DS₂-VASc score ≥ 2, postoperative AF and presence of diabetes mellitus were associated with referral back to a GP in a univariable analysis. Of the patients who were referred to a GP, 41 (18.8%) had postoperative AF compared with 22 patients (8.5%) who remained under nurse-led care or who were referred to a cardiologist. In addition, diabetes mellitus was present in 34 patients (15.6%) who were referred to a GP and in 21 patients (8.1%) who remained under nurse-led care or who were referred to a cardiologist. After multivariable testing, postoperative AF (OR 2.02, 95% CI 1.21–4.07) and CHA₂DS₂-VASc score ≥ 2 (OR 1.92, 95% CI 1.26–2.93) were associated with referral back to a GP (see Fig. S1c in Electronic Supplementary Material).

Since patients with postoperative AF may possess different characteristics than those without postoperative AF, a subanalysis was performed. This resulted in similar outcomes, although persistent AF was a stronger risk factor for referral to a cardiologist in patients without postoperative AF (data not shown).

In total, 417 patients (87.2%) filled out the questionnaire before the initial evaluation visit, of whom 327 (78.4%) completed all 10 questions. Patients scored differently in the four domains: (1) AF general attitudes domain (89.1% correct answers), (2) AF symptom recognition (79.8%), (3) AF therapy (75.4%) and (4) AF in general (49.5%) (data not shown).

Obstructive sleep apnoea syndrome was previously diagnosed in 24 patients (5.0%). However, of the remaining 454 patients, 94 (20.7%) reported symptoms related to sleep apnoea, of whom 72 (76.6%) underwent polysomnography. This resulted in newly diagnosed sleep apnoea in 45 patients (9.9%).

One of our initial questions was whether the nurse-led clinic would result in a reduction of the number of referrals to the general cardiologist’s outpatient clinic or if it would mainly result in redundancy and postponement of the visit to the cardiologist. We showed that with a standardised approach, most patients were referred back to their GP after evaluation by a dedicated nurse practitioner or remained under nurse-led care. One explanation for this finding is that most AF patients, even the (very) elderly, have a limited number of cardiovascular comorbidities. The most frequently encountered comorbidities were hypertension and obesity, which are chronic conditions that, in general, can be treated in accordance with established protocols [13]. We speculate that most patients with first-time AF do not need to consult a cardiologist once the arrhythmia is managed according to current guidelines.

The seminal paper by Hendriks et al. demonstrated that in a randomised setting, patients who were treated in a nurse-led setting suffered from fewer complications and had a lower mortality rate compared with usual care [5]. In particular, adherence to six guideline recommendations (i.e. oral anticoagulation when indicated, thyroid function assessment, echocardiography, and no prescription of rhythm control in asymptomatic patients, patients with a contraindication and patients with permanent AF) was more than 80% in the nurse-led outpatient clinic versus less than 40% in the usual care arm [5]. This observation has fuelled the development of a nurse-led programme that strictly follows the European Society of Cardiology guidelines and assesses all patient in a protocolled manner. Interestingly, in a multicentre study, the same authors attempted to demonstrate superiority of the nurse-led clinic [14]. The results of that study were, surprisingly, neutral, and it was shown that the lack of benefit was driven by those clinics in which there was no or only limited experience with a nurse-led clinic.

Due to the design and the time per patient available at the nurse-led clinic, more attention can be paid to the management of comorbidities. Indeed, AF is highly prevalent among those with obesity, and weight loss is associated with a considerable rate of freedom of AF [15, 16]. Although it is not easy to adjust lifestyle, many patients achieve a more favourable weight when subjected to strict programmes [17]. Also noteworthy was our finding that when asked, a significant proportion of the patients mentioned symptoms of sleep apnoea. In 63% of the patients who were subjected to polysomnography, obstructive sleep apnoea was indeed demonstrated.

It has been shown previously that a large proportion of patients with AF are unable to name more than two AF symptoms and that AF knowledge is better among patients under nurse-led care compared with those under usual care [18]. Improved understanding of the arrhythmia translates to fewer negative emotions and a higher health-related quality of life [7]. There is a trend towards fewer hospitalisations and presentations to the emergency room in patients visiting the nurse-led clinic compared with those receiving usual care [19]. Moreover, knowledge combined with a plan for self-management improves lifestyle in AF patients [20]. This emphasises the notion that more time and attention are needed for the care of this chronic condition and for the coordinating role of a dedicated nurse practitioner.

The improvements associated with a nurse-led AF outpatient clinic may lead to a reduction in the burden of AF on the healthcare system and consequently a reduction in healthcare costs. The current high costs are expected to rise even further in an aging society with more subjects burdened with chronic disease, which calls for redesigning our healthcare system [6]. The costs of outpatient care are driven by diagnostic procedures, consultations and medications. Nurse-led care is associated with slightly more life-years and also more quality-adjusted life-years at a lower cost than usual care [21]. In addition, the healthcare budget is consumed by hospitalisation costs, which are reduced in patients under nurse-led care [9]. Simultaneously, an improved lifestyle is expected to have a favourable effect on a patient’s cardiovascular status as a whole.

This study lacked follow-up data on cardiovascular morbidity and mortality, which could have been used to assess the consequences of a decision for referral. Importantly, the wide confidence intervals in the univariable and multivariable analyses indicated that conclusions should be drawn with caution.