Screening for Actionable Atrial Fibrillation in a Community Hospital Eva Roseboom 1965204 Faculty supervisor: Dr. R. G. Tieleman Location: Martini Hospital Groningen, Department of Cardiology September 2016 – March 2017
Screening for Actionable Atrial
Fibrillation in a Community Hospital
Eva Roseboom 1965204
Faculty supervisor: Dr. R. G. Tieleman
Location: Martini Hospital Groningen, Department of Cardiology
September 2016 – March 2017
2
Table of Contents
Summary 3
Samenvatting 4
1. Introduction 1.1 Epidemiology 5
1.2 Risk factors 5
1.3 Pathophysiology 5
1.4 Diagnosis 5
1.5 Thromboembolic complications 6
1.6 Treatment 6
1.7 Actionable AF 7
1.8 Screening 8
1.9 Relevance 10
2. Objectives 2.1 Aim 11
2.2 Research questions 11
3. Material & Methods 3.1 Study design 12
3.2 Study population 12
3.3 Primary study parameters 12
3.4 Secondary study parameters 12
3.5 Statistical analysis 13
3.6 Trial registration 13
4. Results 4.1 Study population characteristics 14
4.2 Objectified parameters 14
4.3 Primary research question 15
4.4 Secondary research questions 17
5. Discussion 5.1 Utility of screening 21
5.2 Detection methods 21
5.3 Study population 22
5.4 Limitations 23
5.5 Acknowledgements 23
6. Conclusions 24
7. References 25
8. Appendix 30
3
Summary
Aim
To investigate the yield of screening for silent and undertreated atrial fibrillation (AF) in a
community hospital using a single-lead hand-held ECG recording device with automatic
detection (MyDiagnostick). Results of MyDiagnostick’s algorithm to detect AF are
compared with pulse palpation and subsequent 12-lead ECG recording.
Methods
During preoperative consultation in patients >65 years, nurses carried out AF screening by
pulse palpation and, in case of pulse irregularity, subsequent 12-lead ECG recording.
Simultaneously, heart rhythm was checked by the MyDiagnostick, a hand-held single-lead
ECG recorder with automated AF detection. Nurses were blinded from the results of the
MyDiagnostick. The rhythm strips from the MyDiagnostick were reviewed by a cardiologist
and compared to the verdict of manual pulse palpation (regular or irregular) and
MyDiagnostick outcome (AF or no AF).
Results
In total, 505 consecutive patients (48.7% male, 51.3% female, median age 72 years) were
screened. Forty-seven patients (9.3%) had a history of AF, all anticoagulated according to
most recent guidelines. Screening detected two new patients with AF, both requiring
anticoagulant therapy, and 19 patients with known AF. 28 patients had paroxysmal AF and
did not display the arrhythmia during screening. The sensitivity of MyDiagnostick’s
algorithm compared to manual pulse palpation was 100% versus 42.9%. Specificity was
95.2% and 94.1%, respectively. Independent predictors for the presence of AF were heart
failure (OR 7.3, CI 1.35 – 39.77), male gender (OR 2.91, 95% CI 1.52 – 5.54), BMI >35 (OR
2.64, 95% CI 1.01 – 6.88) and increasing age (OR of 1.07 per year, 95% CI 1.01 – 1.13).
Conclusion
The prevalence of AF in history during in-hospital preoperative screening is high (9.3%), due
to a combination of a population with more risk factors for the development of AF and more
incidental findings due to frequent examination prior screening. Screening revealed new AF
in only 0.4% of the patients. Manual pulse palpation for the detection of AF appears to be an
inadequate screening method, compared with the MyDiagnostick.
4
Samenvatting
Doel
Het onderzoeken van de opbrengst van screenen naar subklinisch en onderbehandeld
atriumfibrilleren (AF) in een groot perifeer ziekenhuis, met behulp van een single-lead
draagbaar ECG opnameapparaat met automatische detectie (MyDiagnostick). De resultaten
van de MyDiagnostick als screeningsmethode worden vergeleken met het voelen van de pols
en alleen in het geval van irregulariteit vervaardigen van een 12-lead ECG.
Methodes
Doktersassistentes screenden gedurende een preoperatief spreekuur bij patiënten >65 jaar naar
AF middels pols voelen, waarbij er bij een onregelmatige pols een 12-lead ECG werd
vervaardigd. Tegelijkertijd werd het hartritme beoordeeld door de MyDiagnostick. De uitslag
van de MyDiagnostick werd afgeschermd voor de doktersassistentes. De ritmestroken
geproduceerd door de MyDiagnostick werden beoordeeld door een cardioloog en vergeleken
met de uitkomst van zowel het pols voelen (regulair of irregulair) en de MyDiagnostick (AF
of geen AF).
Resultaten
In totaal werden er 505 patiënten gescreend (48.7% vrouw, 51.3% man, mediane leeftijd 72
jaar). Zevenenveertig patiënten (9.3%) waren al bekend met AF, zij gebruikten allemaal
antistolling volgens de recentste richtlijn. De screening detecteerde twee (0.4%) patiënten met
nieuw AF, die beiden antistolling nodig hadden en 19 patiënten met bekend AF. 28 patiënten
waren bekend met paroxysmaal AF en toonden de aritmie niet ten tijde van de screening. De
sensitiviteit van het algoritme van de MyDiagnostick vergeleken met pols voelen was 100%
versus 42.9%. De specificiteit was 95.2% en 94.1% respectievelijk. Onafhankelijke
voorspellers voor de aanwezigheid van AF waren hartfalen (OR 7.3, CI 1.35 – 39.77),
mannelijk geslacht (OR 2.91, 95% CI 1.52 – 5.54), BMI >35 (OR 2.64, 95% CI 1.01 – 6.88)
en oplopende leeftijd (OR of 1.07 per jaar, 95% CI 1.01 – 1.13).
Conclusie
De prevalentie van AF is hoog tijdens screenen in een ziekenhuis (9.3%), door een combinatie
van een populatie met meer risicofactoren voor het ontwikkelen van AF en voorafgaand meer
incidentele vangsten als gevolg van frequent onderzoek. Gericht screenen onthulde in 0.4%
van alle patiënten onbekend AF. Pols voelen blijkt door zijn lage sensitiviteit een
ontoereikend screeningsinstrument vergeleken met de MyDiagnostick.
5
1. Introduction
Atrial fibrillation (AF) is a condition which causes the atria of the heart to contract irregularly.
It is the most common arrhythmia in the world.1 It carries great public health significance
because of the number of patients, additional morbidity and mortality.
1.1 Epidemiology
In Europe, the prevalence of AF ranges from 1.9% - 2.3%, resulting in 10 million European
patients.2,3
Men are 1.5 times more at risk for developing AF compared to women.4 It is
known that the disease occurs more often as age progresses: 70% of affected cases are aged
between 65 and 85.2 Below the age of 65 the prevalence is 0.34%, while if the age is over 80,
the prevalence is 10-18%.4,5
The total number of patients with AF is increasing, because of
population ageing and better treatments for cardiac diseases. The prevalence of AF is
estimated to grow each year with 0.04%, meaning that in 2050 approximately 3.6% of the
total European population will be affected.2 American studies revealed that the summed
annual costs per individual with AF amounted up to 7841$, in comparison to 2622$ for non-
AF subjects.6
1.2 Risk factors
Evidently, gender and age are key determinants for developing AF. Beyond that, quite a few
other risk factors have been exposed. A vast majority consists of diseases that bring changes
in the myocardium, like hypertension (HR 1.42),7 myocardial infarction (HR 1.46),
8 heart
failure (HR 1.43),9 non-rheumatic and rheumatic valvular heart disease (HR 2.42) especially
mitral stenosis10
, and hypertrophic cardiomyopathy (RR 10-28%).11
Numerous other non-
cardiac associations have been discovered increasing the risk for development of AF: obesity
(HR 1.22),12
diabetes mellitus (HR 1.25),13
chronic obstructive pulmonary disease (HR 1.28 –
2.53)14
and chronic kidney disease (1.68 – 3.52).15
However not all risk factors are of chronic
nature. A few reversible conditions are known for initiating AF, like smoking (HR 2.05),16
inflammation (a CRP in the most upper quintile provides a HR of 1.77),17
and heavy alcohol
abuse (HR 1.01 – 1.39).18
Even though blacks more often have risk factors for developing AF,
they have a much lower lifetime risk than whites (11% versus 24% respectively), a concept
called the racial paradox.19
1.3 Pathophysiology
The exact mechanism how these risk factors intensify the tendency for developing AF are not
completely clear. It seems that there is an interplay of ‘triggers’, which initiate AF and
‘substrate’, which is responsible for the maintenance of AF. Triggers are electrical impulses
that interrupt the normal cardiac conduction. They mostly originate from excitable foci around
the site were the pulmonary veins are embedded.20,21
For this reason, AF will usually start out
as being paroxysmal. The atria are relatively undamaged, and episodes of AF will terminate
on their own, generally within seven days. Atrial remodelling is a slow process caused by
periods of paroxysmal AF, hypertension or structural cardiac disease and it leads eventually to
a short refractory period and atrial fibrosis, both perpetuating AF.22
AF then may become
persistent, where the condition is not self-limiting and if sinus rhythm is pursued, it has to be
done by either electrical or chemical cardioversion. If after repeated attempts sinus rhythm
turns out not to be feasible, AF becomes permanent and rate controlling medication is used
for management of the symptoms. AF has a variety of symptoms, of which palpitations are
the most common. The atrial frequency can reach up to 300-600 impulses per minute and is
irregularly transmitted by the atrioventricular node. In turn, this causes the ventricles to
6
contract irregular and often faster than normal, leading to a tachyarrhythmia. Accompanying
symptoms can be shortness of breath, dyspnoea on exertion or fatigue.23
1.4 Diagnosis
The gold stand for diagnosing AF remains rhythm determination by a 12-lead
electrocardiogram (ECG). Characteristic for AF is the absence of distinct P-waves and a RR
interval that is irregularly irregular: it follows no repetitive pattern. The ventricular rate
generally ranges from 90-170 beats/min. An episode of more than 30 seconds of AF on an
ECG recording is diagnostic.24
1.5 Thromboembolic complications
Even brief periods of AF can lead to atrial remodelling. This, in collaboration with loss of
organized atrial contraction may provide for a prothrombotic environment. The damaged atria
express prothrombotic factors, activate thrombocytes and inflammatory cells and thus a
thrombus can be formed.25
The formation of thromboemboli is the most feared complication
of AF, for this can result in an ischaemic stroke or transient ischaemic attack (TIA). In
comparison with unaffected persons, patients with AF are four to five times more likely to
endure an ischaemic stroke.26
Furthermore, strokes in AF patients tend to be more severe.27
Of
all patients hospitalized with an ischaemic stroke, 11% is diagnosed with AF de novo.28
Though AF has been proven to be a risk factor for the development of a stroke, recently
questions have been raised about the time ratio of AF and subsequent stroke. This was
demonstrated by a study in which cardiac implantable electronic devices (CIED) were read
out to detect the presence of AF. It showed that in only 8% of the cases, AF was present in the
month prior to the stroke. It also established that in 16% of the cases, AF was only seen after
stroke, but never one year preceding. This makes AF a risk factor and a risk marker for
ischaemic stroke.29
1.6 Treatment
Atrial fibrillation treatment consists of a dual policy with one component handling the
prognostic factors while the other focusses on treating the symptoms. The latter can be done
by either rate control or rhythm control.
Prognostic factors
Management of the prognostic factors partly consist of monitoring and controlling
comorbidities, such as maintenance of adequate blood pressure, treating diabetes mellitus or
optimizing therapy for heart failure. Another major component on prognosis is the
prescription of anticoagulants for reducing the risk of a stroke. Oral anticoagulant therapy
(OAC), with either Vitamin-K antagonists (VKAs) or Novel Oral Anticoagulants (NOACs)
results in a 70% risk reduction of having a thromboembolic event.30
The CHA2DS2-VASc
Score (figure 1) displays the risk score of getting an ischaemic stroke in patients with non-
rheumatic AF. The ESC Guidelines suggest that with non-rheumatic AF and a CHA2DS2-
VASc Score > 2 in men and > 3 in women, antithrombotic therapy is beneficial, thus
recommended. Patients with no risk factors do not need to receive anticoagulation. An
individual assessment of advantages and disadvantages should be made for males with a
CHA2DS2-VASc Score of 1 and women with a CHA2DS2-VASc Score of 2.24
When
prescribing antithrombotic medication, there is always an accessory bleeding risk. Scoring
systems have been developed to estimate this risk. One of these scoring methods is called the
HAS-BLED, and it estimates the one-year risk of major bleeding (i.e. intracranial bleeding,
hospitalisation, haemoglobin decrease > 2g/dl or the need of transfusion). The HAS-BLED
covers hypertension, liver and kidney dysfunction, stroke, labile INR, elderly and the use of
7
Figure 1: Table to calculate CHA2DS2-VASc
Score.
drugs and/or alcohol. The scoring system is
intended more as a warning to correct those
risk factors, not so much as an advice to
withhold OAC.31
OAC
VKAs have the preference over aspirin, as they further reduce the risk of stroke with 30% and
all-over mortality with 25%. The limitation with VKA is that is has a small therapeutic range
which requires regular check-ups with a Thrombosis Service, aiming at an International
Normalized Ratio (INR) between 2,0 and 3,0. The quality of INR control is expressed as
Time in Therapeutic Range.32-34
With the arrival of a new group of anticoagulants, the
NOACs (Factor Xa-inhibitors Apixaban, Edoxaban and Rivaroxaban, direct thrombin
inhibitor Dabigatran), a shift in treatment of AF is seen.35
All NOACs have been tested in
comparison to warfarin, showing an similar or better efficacy/safety profile for the different
NOACs. In all studies, the incidence of haemorrhagic stroke was reduced by 50% in the
NOAC treated patients.36,37
When a patient is initially diagnosed with AF and is eligible for
NOACs, a NOAC is preferred to VKA. The ESC states that if a patient with AF is already on
a VKA but has a poor TTR, switching to a NOAC should be considered.26
Rate and rhythm-control
Symptom reduction of AF can be done by either rate control or rhythm control. During rate-
control, the ventricular rate is reduced to 60 – 110 beats per minute, by slowing of AV
conduction. During rhythm control, restoration of sinus rhythm is pursued by pharmacologic
or electrical cardioversion. In the long term there is no differences in stroke risk, mortality or
quality of life when comparing rate and rhythm control.38
Rate control is obtained by beta-
blockers, digoxin or calcium channel blockers or a combination of those, if rate control is not
achieved by monotherapy. Conversion to sinus rhythm can be achieved by prescription of
anti-arrhythmic drugs like flecainide, propafenone, vernakalant or amiodarone. Another
possibility to obtain sinus rhythm is electrical cardioversion, or if all else fails, catheter
ablation of the sites around the pulmonary veins.39
1.7 Actionable AF
Actionable AF is a new term devised by an international collaboration dedicated to the
prevention of AF-related strokes. This term assembles the two concepts ‘silent AF’ and
‘undertreated AF’.
Silent AF
As mentioned above, the prevalence of AF is 1.9% - 2.3%. However, the real prevalence of
AF may be much higher. This is because not all existing AF is known to its carrier. In
approximately 67% AF present itself with symptoms.40
Yet for the remaining one-third of the
Condition Points
C Congestive heart failure (or Left
ventricular systolic dysfunction)
1
H Hypertension (blood pressure
consistently above 140/90 mmHg (or
treated hypertension on medication)
1
A2 Age ≥75 years 2
D Diabetes Mellitus 1
S2 Prior Stroke or TIA or
thromboembolism
2
V Vascular disease (e.g. peripheral
artery disease, myocardial infarction,
aortic plaque)
1
A Age 65–74 years 1
Sc Sex category (female sex) 1
8
cases it proceeds asymptomatic, so called ‘Silent AF’, or ‘Subclinical AF’. These patients
have no symptoms of AF or prior history of AF and are mostly discovered by general
screening or when complications occur. AF detected on single time point screening is called
‘Screen-detected AF’. Screen-detected AF carries presumably similar risks as symptomatic
AF, and when prescribed oral anticoagulants as likely to have the same decline in stroke risk
and death.41
Undertreated AF
In spite of the antithrombotic guidelines, there still exists a group of patients who are known
to have AF and require anticoagulants according to the ESC Guidelines, but are undertreated,
and therefore vulnerable to having an ischaemic stroke. The most common reason for
withholding OAC is the fear of bleeding events . Nonetheless, the need to receive OAC for
prevention of strokes is often bigger than the risk of major bleeds. Research has shown that in
a group of patients admitted with an ischaemic stroke and known AF, 51% had not received
sufficient antithrombotic treatment.42
Swedish research revealed that out of 94.000 ischaemic
strokes, in 20% of the cases AF was known but undertreated, and in 9% of the cases AF was
silent.43
1.8 Screening
In view of the latest epidemiologic figures, knowledge of the ageing population and the
concept of actionable AF, it seems there is a need for screening for prevention of actionable
AF-associated strokes. The ESC discloses in their most recent guideline that opportunistic
screening appears feasible and cost-effective in patients older than 65 years, to be performed
by either pulse taking or ECG rhythm strip.24
Screening is defined as a medical test performed
on members of an asymptomatic population or subgroup to assess the likelihood of their
members having a particular disease.44
Screening is not to be used as a diagnostic tool, but an
indicator for further examination or test. It aims for reducing morbidity or mortality by
catching the disease, so treatment can be early and more successful. An ultimate screening
device carries a high sensitivity and specificity, is easy to use and cost-effective. Also, for
achieving a high positive predictive value, screening must only be performed in groups with
an expected high prevalence.45
In terms of diagnosing, screening tests are always measured
ATRIAL FIBRILLATION
KNOWN
UNDERTREATED* ADEQUATELY
ANTICOAGULATED*
UNKNOWN
Actionable Atrial Fibrillation
Figure 2: Distribution of AF.
*According to the most recent guidelines from the European Society of Cardiology.
9
against the gold standard, which by definition should be 100% sensitive and 100% specific.
In atrial fibrillation, the gold standard is an 12-lead ECG. Taking an ECG of every individual
is time-consuming and expensive, calling the need for other screening devices. A systematic
review analysed several methods for detecting an irregular pulse and suspected AF. An
assessment of 15,129 pulses compared to a 12-lead ECG resulted in a sensitivity of 92% and a
specificity of 82%, leaving room for potentially improved detection.46
Screening programmes
Screening programmes with the use of single-lead ECGs already have been carried out. For
example the STROKESTOP study, which screened 75-and 76-year-olds for underdiagnosed
and undertreated AF, by repeated hand-held ECG-recordings during 2 weeks in Sweden.
From a total population of 7173 participants, 12.3% (583 patients) had AF, from which 3.0%
(218) newly diagnosed. In the patients with a known history of AF (666 patients), 2.1% were
not treated according to the guidelines, resulting in actionable AF in 5.1% of the patients.47
Another example is the SEARCH-AF study, which was conducted in 10 pharmacies in
Australia using a smartphone ECG device (Alivecor) in all patients >65 years of age. In this
study a single measurement revealed silent AF in 1.5% of all patients.48
The MyDiagnostick
Another screening device designed for mass screening is the MyDiagnostick. This is a
compact, hand-held ECG recorder with automatic AF detection. It has to be held with two
hands, functioning as a conductor, producing a lead I ECG strip of 60 seconds. Subsequently,
it analyses the rhythm for the presence of AF. A special algorithm will generate a score
measuring the irregular irregularity of the RR-intervals, the so called ‘AF-score’. The
likelihood of the diagnoses AF increases as this score goes up. With a normal sinus rhythm,
the AF-score would be 0. After one minute of holding, the MyDiagnostick displays a light. It
is set to turn green when AF-score is below 10, and red when it reaches 10 or higher. The
ability of displaying this light can be switched off. A red light usually indicates the presence
of AF, but can also be false positive. Small disturbances, caused by dry hands or a too firm
grip, might produce artefacts on the rhythm strip, which will also increase the AF-score.
Furthermore, the presence of premature complexes or a sinus arrhythmia will raise the AF-
score, due to more RR-irregularity. Low voltages, caused by a vertical heart axis, can also
provoke an increase in AF-score due to maldetection of R-waves. By plugging the
MyDiagnostick in a computer, all rhythm strips stored on the device (max. 140), can be
individually analysed.49
In a study of 192 patients the MyDiagnostick showed to be 100%
sensitive for detecting AF and had a specificity of 95.9%, thereby proving to be a suitable
device for both individuals and large screening programmes.50
It was used in a large study in
which the prevalence of silent AF was examined during influenza vaccination in 10 Dutch
general practices. In a total of 3269 screened patients, 121 (3.7%) cases of AF were detected,
from which 37 (1.1%) had not yet been diagnosed before.5 From that same study, cost-
effectiveness was calculated. It was found that screening decreases the overall costs by 764€
and increases QALYs by 0.27 per AF patient diagnosed.51
Image 1: The MyDiagnostick.
Image 2: Sinus rhythm on rhythm strip.
Image 3: Atrial fibrillation on rhythm strip.
10
1.9 Relevance
For community-dwelling persons, awareness already has been raised to prevent AF-associated
strokes by mass screening, yielding a significant (1.5 – 5.1%) proportion of actionable AF.
However, data concerning the utility of screening for actionable AF in secondary care remains
unknown. Furthermore, patient specific characteristics of those with AF presented in
secondary care can provide information on potential subgroups to screen. Lastly, the benefit
of screening with a single-lead ECG recorder versus routine practice (manual pulse palpation)
in a hospital is unexplored. This knowledge can contribute to the possible need to screen for
actionable AF in a hospital setting. Therefore, this research will be engaged in exploring the
yield of screening for in-hospital actionable AF, comparing performances of two AF detection
methods and distinguishing patients vulnerable for the presence of AF.
11
2. Objectives
2.1 Aim
The intent of the present study is to investigate the utility of screening for actionable (silent
and undertreated) atrial fibrillation in a community hospital. Because utility is difficult to
objectify, the following research questions were formed to fulfil this aim.
2.2 Research questions
Primary research question
What is the yield of screening for known, silent and undertreated atrial fibrillation among
patients older than 65 in a community hospital?
Secondary research questions
1. Is the performance of the MyDiagnostick superior to manual pulse palpation when
screening for atrial fibrillation?
2. What are the differences between patients with AF and patients without AF?
3. Are there predictive variables for AF?
12
3. Material & Methods
3.1 Study design
This single-center exploratory cross-sectional study was carried out in the Martini Hospital
located in Groningen. Screening was performed with the MyDiagnostick in all patients >65
years of age visiting pre-operative consultation during a period of 10 weeks Every patient
undergoing surgery in the Martini Hospital was invited to visit the preoperative consultation.
This consultation hour consisted of a carousel of medical staff: an anaesthesiologist, a
pharmaceutical assistant, an intake nurse and a physician assistant. The screening was
conducted by the physician assistants, who prior were trained to handle the MyDiagnostick.
The assessment of the rhythm strip was in consultation with a cardiologist. Note that the
display of the light (red or green) was switched off during screening in trial context, so when
comparing performances it did not bias the outcome of manual pulse palpation.
3.2 Study population
The study population consisted of patients aged 65 or older, who visited the preoperative
consultation hours. Annually, 19,000 patients attend those hours and it was estimated that
approximately 25% of all cases were >65 years old. This came down to circa 100 eligible
patients weekly, who prior to their appointment, received a letter providing information about
the study (Appendix 1).
Inclusion criteria
Visiting preoperative consultation
Aged 65 and over
Verbally informed consent
Exclusion criteria
Aged < 65
Pacemaker or implantable cardioverter-defibrillator (ICD)
Not willing to participate in the study
Not being able to hold the MyDiagnostick with both hands (e.g. amputees)
Sample size
The estimated screen-detected incidence of AF (P) was 3.7%. With a 95% confidence interval
(1.9% - 6.9%) the sample size (N) came down to 237 patients.
3.3 Primary study parameter
The primary study parameter was the presence of atrial fibrillation, seen on the rhythm strip
provided by the MyDiagnostick. The primary investigator examined every rhythm strip,
independent of the judgement of the MyDiagnostick itself (AF-score). Any abnormal strip
was verified by a cardiologist, so the verdict atrial fibrillation was only given to those who
truly displayed it on the rhythm strip. Resulting from this, the distribution between known,
silent and undertreated atrial fibrillation came forth.
3.3 Secondary study parameters
1. Comparison of performance
During preoperative care, a physician assistant manually assessed the patient’s pulse to
determine its regularity. This was solely done to detect AF. In case of an irregular pulse, an
13
anaesthesiologist made the decision whether to take an 12-lead ECG or not. By comparing
several pillars of performance, it was assessed if the MyDiagnostick is superior to manual
pulse palpation for the detection of AF. These pillars are displayed in a ‘table of confusion’,
as shown below in figure 2. From such a table, various details could be derived. Sensitivity
and specificity gave information about the specific measuring device, indicating the chance of
incorrectly rejecting or accepting a diagnosis. Positive and negative predictive values
signified the chance that one’s test result actually told the truth, and were highly dependent on
the prevalence of the condition.
2. Differences between AF an no AF
By comparing characteristics of patients with AF and without AF, significant differences
between the two groups were demonstrated. Patient characteristics included age, gender, body
mass index (BMI), smoking, and several comorbidities: hypertension, diabetes mellitus,
ischaemic stroke or TIA in history, myocardial infarction, history of PCI/CABG, heart failure,
aortic valve surgery, and COPD. These variables were derived from patient’s medical file.
3. Predictive factors for the presence of AF
By analysing the same set of variables as mentioned above, we attempted to examine if any of
those variables predicted the presence of AF.
3.6 Statistical analysis
Statistical analysis were conducted by the usage of IBM SPSS Statistics 20. For the
comparison of the two groups (AF vs. no AF) the Chi-square test was used for the
dichotomous variables gender, hypertension, diabetes mellitus, ischaemic stroke or TIA in
history, myocardial ischaemia, history of PCI/CABG, history of aortic valve surgery, COPD
and smoking. For continue variables age, systolic and diastolic blood pressure during
preoperative consultation and BMI the Student’s t-test and Mann-Whitney U were used in
case of normal distribution and non-normal distribution, respectively. For the analysis of
possible predictive factors for the development of AF, a logistic regression was used, with the
presence of AF being the dependent variable. Outcomes of statistical test with a p-value ≥
0,05 were considered to be significant.
3.7 Trial registration
On 09-09-2016, The Medical Ethical Committee of the Martini Hospital Groningen approved
this research. They stated that it does not fall within the extent of the Act Medical Scientific
Research with Human Beings (Wet Medisch-wetenschappelijk Onderzoek met mensen,
WMO, Appendix II) On 08-11-2016, ClinicalTrials.gov Protocol Registration and Results
System reviewed this research and made it public. ClinicalTrials.gov Identifier:
NCT02960334.
Total
Population
Condition Present
Condition Absent
Prevalence =
Condition Present /
Total Population
Test
Positive
True Positive
(A)
False Positive
(B)
Sensitivity
A / (A + C)
Test
Negative
False Negative
(C)
True Negative
(D)
Specificity
D / (D + B)
Positive Predictive Value
A / (A + B)
Negative Predictive Value
D / (C + D)
Figure 2: Table of
confusion.
14
4. Results
4.1 Study population characteristics
During a ten week pilot screening procedure, 524 patients held the MyDiagnostick. 19
patients were excluded for having a pacemaker or ICD. In total, 505 patients were included in
the database. 246 (48.7%) were male and 259 (51.3%) were female. The age ranged from 65
up to 96 with a median of 72. The Kolmogorov-Smirnov and Shapiro-Wilk test for normality
were both 0, indicating a non-normal distribution. The distribution of age is shown in figure 3.
4.2 Objectified parameters
Blood pressure
From the total of 505 patients, 503 patients had their blood pressure recorded during the
preoperative consultation. In two cases, the blood pressure were not listed. The systolic blood
pressure ranged from 111 to 219, median 157 mmHg. The mean diastolic blood pressure was
83 mmHg and ranged from 44 up to 122. The systolic blood pressure was not normally
distributed according to the Kolmogorov-Smirnov and the Shapiro-Wilk criteria (sign. 0.014
and 0.013 respectively), however the diastolic blood pressure was normally distributed with a
significance of 0.190 and 0.515 respectively.
BMI
504 patients had their BMI calculated during the preoperative consultation, one case was
missing in the file. The BMI ranged from 16 – 47 kg/m2, median 27.3. The BMI was not
normally distributed, with the Kolmogorov-Smirnov and Shapiro-Wilk significance both
being 0.
Comorbidities
From the total study population (N = 505), 60.4% had hypertension, 20% had COPD, 19.2%
diabetes mellitus, 15.8% myocardial ischaemia, from which 67.5% had PCI/CABG. 9.3%
had had a TIA or ischaemic stroke. 87.7% claimed to be non-smokers. 1.4% had received
aortic valve surgery (AVS).
Figure 3: Age distribution of the total
population with normal curve.
N = 505
Median = 72 (65 – 96)
15
History of atrial fibrillation
Prior to the screening, 47 patients were known to have AF. 17 had permanent AF, in 30 cases
AF was paroxysmal.
4.3 Primary research question
Actionable atrial fibrillation
In total, 505 rhythm strips were evaluated, independent of the light shown by de
MyDiagnostick after holding. From these 505 patients, 21 (4.2%) showed AF on the rhythm
strip. AF was known in 19 of the cases (90.5%), yet for two patients, the diagnoses was new.
From all patients without AF during screening (N = 481), 28 patients were previously
diagnosed with paroxysmal AF, making a total AF prevalence of 49 (9.7%). All patients with
known AF were treated according to the most recent guidelines, creating a total number of
Actionable AF N = 2 (0.4%)
Actionable Atrial
Fibrillation N = 2
Figure 4: Quantification of
comorbidities.
N = 505
Total Screened N = 505
No AF during Screening N = 481
Paroxysmal AF N = 28
Correctly Anticoagulated
N = 28
Undertreated N = 0
AF during Screening
N = 21
Known AF N = 19
Correctly Anticoagulated
N = 19
Undertreated N = 0
Unknown AF N = 2
Figure 5: Distribution of
Atrial Fibrillation.
N = 505
0 50 100 150 200 250 300 350
TIA/Ischaemic Stroke
Smoking
Myocard Ischeamia
Hypertension
Heart Failure
DM-II
COPD
AVS
Atrial Fibrillation
Number of patients
PCI/CABG
16
The two patients with actionable AF were a 76-year old man, with a BMI of 30.5 and
hypertension (CHA2DS2VASc of 3), who after diagnosis was treated with Apixaban, and a
69-year old female, with a BMI of 31.2, with hypertension and diabetes mellitus
(CHA2DS2VASc of 4). After diagnoses she was admitted to the hospital for initiation of rate
control therapy and anticoagulated with Apixaban as well.
CHA2DS2VASc scores
47 Patients had known AF. Five of them (10.6%) had a CHA2DS2VASc score of 1, thus not
needed to be treated with anticoagulants. The remaining 42 patients (89.4%) had a
CHA2DS2VASc score > 2, an indication for anticoagulant therapy. The mean CHA2DS2VASc
score was 3.2.
Usage of oral anticoagulants
From all 47 patients with AF, 42 patients had a CHA2DS2VASc score > 2, hence were treated
with oral anticoagulants. Even though no strong indication, four patients with a
CHA2DS2VASc score of 1 also received oral anticoagulant treatment. 26 patients (56.5%)
were treated by means of a VKA, such as acenocoumarol or fenprocoumon. The remaining 20
patients (43.5%) received a NOAC, with Dabigatran being the most prescribed.
0
2
4
6
8
10
12
14
16
1 2 3 4 5 6 7 8
CH
A2D
S2V
AS
c sc
ore
s
Number of patients
Figure 6: CHA2DS2VASc scores.
N = 47
Mean = 3.2
Median = 3 (1 – 8)
Figure 7: Usage of
Oral Anticoagulants. N = 46
25
1
12
4
4
Acenocoumarol
Fenprocoumon
Dabigatran
Apixaban
Rivaroxaban
17
4.4 Secondary research questions
1. Comparison of performance
MyDiagnostick
From the total of N = 505, 461 cases (91.3%) had an AF-score < 10, so were labelled ‘Green’
by the MyDiagnostick. In retrospect, none of those cases had AF, so are true negative. 44
cases (8.7%) would have received a red light, AF-scores > 10. Of these 44, 21 (47.7%) had
AF. 23 cases had other reasons to be labelled ‘Red’, as shown in figure 8.
Screening with the MyDiagnostick generates a sensitivity of 100%, not a single case of AF
was missed, leading to a negative predictive value of 100%. This means that if the light turns
green, there is zero chance of having AF. With the prevalence of atrial fibrillation being
4.16% in the screened population, receiving a red light creates a 47.7% chance of actually
having AF. The remaining 52.3% receives a red light on other grounds than the presence of
AF.
Total Population
N = 505
AF
No AF
Prevalence 4.16%
Red Light
21
23
Sensitivity
100%
Green Light
0
461
Specificity
95.2%
PPV
47.7%
NPV
100%
Total Screened N = 505
Red Signal N = 44
AF during Screening N = 21
No AF during Screening
N = 23
Sinus Arrhytmia N = 4
Artefacts N = 9
Low Voltages N = 3
Sinus Rhythm with Premature Complexes
N = 7
Green Signal N = 461
AF during Screening N = 0
No AF during Screening N = 461
Figure 8: Evaluation of red (N = 44) and green signals (N = 461) labelled by the MyDiagnostick and assessment
of all (N = 505) rhythm strips.
Figure 9: Table of confusion for
the MyDiagnostick.
Figure X: Table of Confusion for the MyDiagnostick
18
Manual pulse palpation
From the 505 patients who visited the preoperative consultation, 499 patients had their pulse
measured by palpation, in 6 cases there was no notation in the file. From those 499 cases, 462
(92.6%) were issued as regular. From those 462 regular pulses, 12 (2.6%) had AF on the
MyDiagnostick rhythm strip, which makes 450 true negative and 12 false negative. 37 cases
were issued as irregular, from which 9 had AF and 28 had other rhythms, as shown in a figure
10 below.
Manual pulse palpation is 42.9% sensitive to finding atrial fibrillation when it is present. In
more than half of all cases, it will be missed. Due to the prevalence of 4.2%, it generates a
negative predictive value of 97.4%. The specificity of manual pulse palpation is 94.1%.
Finding of an irregular pulse indicates a chance of 24.3% of actually having AF (positive
predictive value), in 75.7% an irregular pulse is caused by other reasons. From all 37 irregular
pulses measured, in 8 cases (21.6%) the anaesthesiologist decided to record a 12-lead ECG. 4
of these ECGs showed AF.
Total Population
N = 499
AF
No AF
Prevalence 4.21%
Pulse Irregular
9
28
Sensitivity
42.9%
Pulse Regular
12
450
Specificity
94.1 %
PPV
24.3%
NPV
97.4%
Total Pulse Palpation N = 499
Pulse Irregular
N = 37
AF during Screening
N = 9
No AF during Screening
N = 28
Sinus Rhythm N = 14
Artefacts N = 1
Sinus Arrhythmia N = 1
Sinus Rhythm with Premature Complexes
N = 11
Sinus Rhythm with pauses N = 1
Pulse Regular N = 462
AF during Screening
N = 12
No AF during Screening N = 450
Figure 10: Evaluation of irregular (N = 37) and regular pulses (N = 462) measured by physician assistants and
assessment of all (N = 499) rhythm strips.
Figure 11: Table of
confusion of manual pulse
palpation.
19
Concerning actionable AF
Both cases of Actionable AF (N = 2) were picked up by the MyDiagnostick. Even though
both cases also had an irregular pulse, only in one case the anaesthesiologist decided to make
a 12-lead ECG. Therefore, without the MyDiagnostick, in one patient AF would have been
missed.
2. Difference in groups AF vs. No AF
In the total group of N = 505, patients with AF were significantly older than patients without
AF. Both medians differed four years, 71 and 75 years, respectively. Moreover, gender is
another major difference: patient with AF were in 71.4% of the male sex, where in the control
group gender was divided almost equally. Concerning significant differences in
comorbidities, hypertension is more common among people with AF (79.6 versus 58.3%
respectively), as are heart failure (8.2 vs. 0.9%), a history of PCI/CABG (20.4 vs. 9.6%) and a
history of aortic valve surgery (6.1 vs. 0.9%).
3. Predictive factors for the presence of AF
There are several significant predictors for the presence of AF, heart failure being the largest
with an OR of 7.3 (CI 1.35 – 39.77). Second largest predictor is male gender, OR of 2.91
(95% CI 1.52 – 5.54), followed by having a BMI > 35 OR 2.64 (95% CI 1.01 – 6.88) and
increasing age, with an OR of 1.07 per year (95% CI 1.01 – 1.13). Surprisingly, being diabetic
decreases the chance of developing AF, OR 0.32 (CI 0.12 – 0.88).
No AF
(N = 456)
AF
(N = 49)
P-Value
Age (Median, in years) 71 75 0.013
BMI (Median, in kg/m2) 27.3 28.6 0.092
Systolic blood pressure (Mean, in mmHg) 161 159 0.444
Diastolic blood pressure (Mean, in mmHg) 84 81 0.125
Men (%) 46.3 71.4 0.001
Hypertension (%) 58.3 79.6 0.006
Diabetes Mellitus (%) 19.7 14.3 0.466
Heart Failure (%) 0.9 8.2 0.001
TIA/Ischaemic Stroke (%) 8.6 16.3 0.128
Myocardial Infarction (%) 14.7 26.5 0.051
PCI/CABG (%) 9.6 20.4 0.038
Aortic Valve Surgery (%) 0.9 6.1 0.019
COPD (%) 20.0 20.4 1.0
Smoking (%) 12.3 12.2 1.0
BMI > 25 (%) 73.7 81.6 0.309
BMI > 35 (%) 9.9 16.3 0.250
Figure 12: Comparison of patients with AF and without AF. A p-value < 0.05 indicates significance.
20
P-Value
Odds Ratio (OR)
95% C.I. for OR
Lower Upper
Age 0.020 1.066 1.010 1.125
Aortic Valve Surgery 0.059 5.992 0.935 38.384
BMI > 25 0.363 1.469 0.641 3.368
BMI > 35 0.048 2.635 1.010 6.875
COPD 0.911 1.046 0.470 2.331
Diabetes Mellitus 0.026 0.319 0.116 0.875
Heart Failure 0.021 7.314 1.345 39.768
Hypertension 0.055 2.169 0.984 4.778
Male Gender 0.001 2.903 1.521 5.541
Myocardial Infarction 0.971 0.975 0.251 3.793
PCI/CABG 0.989 0.989 0.214 4.573
Smoking 0.934 1.045 0.374 2.916
TIA/Ischaemic Stroke 0.976 1.015 0.380 2.712
Figure 13: Logistic regression of variables on the chance of the presence of AF. A p-value < 0.05 indicates significance.
21
5. Discussion
5.1 Utility of screening
During a ten week period, we screened all eligible patients aged over 65 at a single time point
during their visit to the preoperative consultation hour with the aid of the MyDiagnostick for
the presence of atrial fibrillation (AF). A total of 505 patients participated in this study. 47
patients (9.3%) had previously been diagnosed with AF, either paroxysmal or permanent.
The screen-detected incidence for atrial fibrillation amounted 4.2%: 3.8% already known,
0.4% silent and 0% undertreated.
In a community hospital there seems to be a shift in distribution of AF: the proportion known
is bigger than expected, as where the proportions silent and undertreated are far smaller,
compared to screening in primary health care. Kaasenbrood et al found a screen-detected
incidence of 3.7%, with 30% cases of silent AF.5 This shift in distribution may be due to the
fact that patients being treated in the hospital generally have more comorbidities, thus are
exposed to more risk factors for the development of AF, and on the other hand the chance of
earlier detection or incidental findings of silent AF is enhanced when frequently being
examined in a hospital. These frequent visits with a doctor are plausibly also the reason for
the 100% correct anticoagulant prescription rate.
We did not screen all hospitalized patients, but the portion visiting the preoperative
consultation hour: those who were electively planned for hospital admission. However, we
believe that this group is a legitimate representation for the achievable yield. Patients admitted
via the emergency department or outpatient consultation often have an indication for an ECG,
and if AF is found, it would not contribute to the yield of screening.
As our study demonstrates, screening at single time point has the potential of missing
paroxysmal cases. Of the 30 patients with paroxysmal AF, only 2 showed the arrhythmia at
the time of screening. This fact has been confirmed by the STROKESTOP study, where the
single time prevalence of 0.5% increased to 3.0% with repeated recordings during 2 weeks.47
If one does decide to screen in a secondary health care setting, repeated recordings might
increase the screen-detected incidence.
The yield of screening a targeted population of >65 years in a community hospital in a
developed first world country as The Netherlands, where awareness already has been raised
concerning the extent of the morbidity and mortality of AF would presumably not be so high.
In ten weeks we detected two cases (0.4% of the total), from which only one would have
slipped through the net of routine pulse-palpation practice. If one was to extrapolate this
trend, five cases of silent AF might be singled out by targeted screening on a yearly basis. The
question that rises is whether this screening would be cost effective, a question we are not
able to answer based on this research.
5.2 Detection methods
When it comes to the comparison of the measuring instruments, we can state with certainty
that distinguishing AF from harmless cardiac rhythms by pulse palpation is hard. We found a
sensitivity of 42.9%, a much smaller number than what we expected from literature research.
The specificity (94.1%) is of the same magnitude as described in literature. A possible reason
as to why the sensitivity for manual pulse palpation is lower than expected, is the high
standard of rate control therapy in those who have AF. It is hard to detect AF when someone’s
22
pulse is reasonably calm and not significantly irregular. The MyDiagnostick does a better job,
proving itself to be 100% sensitive, and has a specificity of 95.2%.
As for targeted screening, it is important that the used measuring tool is utmost sensitive. The
crux of a screening programme is that it is designed to let as few undiagnosed but affected
people slip through the nets. In practice, the MyDiagnostick can guarantee that it would not
miss a single case of AF. The NPV is 100%, meaning a green light fully excludes the
diagnoses AF. With manual pulse palpation, AF cannot be safely rejected, even if it only
concerns a small fraction of false negatives. With the prevailing prevalence being 4.2%, it
results in a NPV of 97.4%. So when a regular pulse is detected, in 3.6% of the cases AF is
still present.
Besides a high sensitivity, it is preferable when a screening tool also has a high specificity.
This number for the MyDiagnostick and manual pulse palpation (95.2% and 94.1%
respectively) might not seem very different, but with a prevalence of 4.2%, it causes a
substantial difference in PPVs: 47.7% and 24.3%, for the MyDiagnostick and manual pulse
palpation respectively. Approximately one in two red lights from the MyDiagnostick indicates
AF, whereas with manual pulse palpation, it is circa one in four. This results in a low
credibility of irregular pulses, demonstrated by the ratio of ECGs made. In 21.6% (8 out of
37) the anaesthesiologist decided to take an ECG following irregular pulse. With regard to the
MyDiagnostick, this hassle can be avoided. Not only would the number false negatives be
less, no separate ECG needs to be made, for the MyDiagnostick stores it to its internal
memory, available for rhythm strip extraction and verdict.
In this study we did not objectify any disadvantages of implementing the MyDiagnostick. We
had no reports of consultation hours running late because of the minute it took to screen each
patient. This is possibly due to the fact that screening can be done during the interview, while
the patient is sitting down. None of the physician assistant had difficulty handling the
MyDiagnostick after brief training, the instrument was well received and found to be user
friendly.
5.3 Study population
The total screened population amounted 505 participants, who all visited the preoperative
consultation hours of the anaesthesiologist. The 47 patients known to have atrial fibrillation
had an average CHA2DS2VASc score of 3.2, comparable to Kaasenbrood et al (mean 3.4)
Svennberg et al. (mean 3.5 – 3.9).5,47
In our study, usage of NOAC was relatively high (43.5%
of total anticoagulant therapy), whereas in primary care this was only 9.1%.5 This might be
due to the fact that in secondary health care, anticoagulant therapy is initiated by a
cardiologist, who until January 2017 were the only group of doctors authorized to prescribe
NOACs.
The differences between the groups with and without AF we found are similar to those in
primary care, described in literature.5 The hospital population appears to be a fitted
representation of the normal population. People with AF were significantly older, were more
frequently men, and had more comorbidities, such as hypertension, heart failure and had more
often undergone aortic valve surgery. We also saw more ischaemic strokes and TIAs in
people bearing AF, but these differences were not significant. Unlike other studies, we found
that diabetes mellitus occurred more in those without AF, for no apparent reason.
Individual predictors for the presence of AF turned out to be heart failure (OR 7.3), male
23
gender (OR 2.9), BMI > 35 (OR 2.64) and increasing age (OR 1.07 per year). These findings
are concordant to prior studies.
5.4 Limitations
We did not evaluate the implementation of the screening in its entirety. We did not follow up
on time investment, costs, or patient’s opinion. For the variable ‘smoking’, we used patient’s
yes or no-answer to a questionnaire they had to fill in before seeing the anaesthesiologist.
Former smoking was therefore not included, while we know this also influences the chance of
developing AF.
5.5 Acknowledgements
First of all, I would like to thank my supervisor Dr. R. Tieleman for his guidance in the world
of science, his encouragements and his unceasing enthusiasm, despite the roadblocks that
sometimes were thrown up. I also would like to thank all the participating physician
assistants, who without a complaint were open-hearted towards me and open-minded about a
new screening procedure. I thank P. Wijtvliet for teaching me a not only a great deal about
atrial fibrillation, but more importantly how to transfer this knowledge to a patient’s
understanding. Last but not least I thank all the patients for their contribution to this study.
24
6. Conclusions
Atrial Fibrillation is well-known and treated properly in the Martini Hospital. Targeted
screening of patients over 65 years in secondary health care will increase the detection rate of
actionable AF, but probably so little, that the cost effectiveness could be debated. The greatest
yield of screening for actionable AF would be in primary care. Single time point screening
might underestimate the magnitude of the issue.
When detecting AF, the MyDiagnostick is superior to manual pulse palpation. It not only
generates a high sensitivity and specificity (100 and 95.2% respectively), guaranteeing not
missing a single undiagnosed case, it is also quick, non-invasive and easy to use. Besides, for
diagnosing AF, no additional ECG has to be made.
25
7. References
1: Fuster V, Ryden LE, Cannom DS, Crijns HJ, Curtis AB, Ellenbogen KA, et al.
ACC/AHA/ESC 2006 Guidelines for the Management of Patients With Atrial Fibrillation: A
Report of the American College of Cardiology/American Heart Association Task Force on
Practice Guidelines and the European Society of Cardiology Committee for Practice
Guidelines (Writing Committee to Revise the 2001 Guidelines for the Management of
Patients With Atrial Fibrillation): Developed in Collaboration With the European Heart
Rhythm Association and the Heart Rhythm Society. Circulation 2006 Aug 15,;114(7):e354.
2: Zoni-Berisso M, Lercari F, Carazza T, Domenicucci S. Epidemiology of atrial fibrillation:
European perspective. Clinical epidemiology 2014;6:213.
3: Friberg L, Bergfeldt L. Atrial fibrillation prevalence revisited. Journal of Internal Medicine
2013 Nov;274(5):461-468.
4: Chugh SS, Havmoeller R, Narayanan K, Singh D, Rienstra M, Benjamin EJ, et al.
Worldwide epidemiology of atrial fibrillation: a Global Burden of Disease 2010 Study.
Circulation 2014 Feb 25,;129(8):837-847.
5: Kaasenbrood Femke F. Yield of screening for atrial fibrillation in primary care with a
hand-held, single-lead electrocardiogram device during influenza vaccination. Europace 2016
-2-6.
6: Kim MH, Johnston SS, Chu B, Dalal MR, Schulman KL. Estimation of total incremental
health care costs in patients with atrial fibrillation in the United States. Circulation.
Cardiovascular quality and outcomes 2011 May;4(3):313.
7: Krahn AD, Manfreda J, Tate RB, Mathewson FAL, Cuddy TE. The natural history of atrial
fibrillation: Incidence, risk factors, and prognosis in the manitoba follow-up study. The
American Journal of Medicine 1995;98(5):476-484.
8: Crenshaw BS, Ward SR, Granger CB, Stebbins AL, Topol EJ, Califf RM. Atrial fibrillation
in the setting of acute myocardial infarction: the GUSTO-I experience. Global Utilization of
Streptokinase and TPA for Occluded Coronary Arteries. Journal of the American College of
Cardiology 1997 Aug 1,;30(2):406.
9: Santhanakrishnan R, Wang N, Larson M, Magnani J, McManus D, Lubitz S, et al. Atrial
Fibrillation Begets Heart Failure and Vice Versa: Temporal Associations and Differences in
Preserved Versus Reduced Ejection Fraction. Circulation 2016 Feb 2,;133(5):484-492.
10: Grigioni F, Avierinos J, Ling LH, Scott CG, Bailey KR, Tajik AJ, et al. Atrial fibrillation
complicating the course of degenerative mitral regurgitation: determinants and long-term
outcome. Journal of the American College of Cardiology 2002 Jul 3,;40(1):84.
11: Robinson K, Frenneaux MP, Stockins B, Karatasakis G, Poloniecki JD, McKenna WJ.
Atrial fibrillation in hypertrophic cardiomyopathy: a longitudinal study. Journal of the
American College of Cardiology 1990 May 1,;15(6):1279-1285.
26
12: Tedrow UB, Conen D, Ridker PM, Cook NR, Koplan BA, Manson JE, et al. The long-
and short-term impact of elevated body mass index on the risk of new atrial fibrillation the
WHS (women's health study). Journal of the American College of Cardiology 2010 May
25,;55(21):2319.
13: Fatemi O, Yuriditsky E, Tsioufis C, Tsachris D, Morgan T, Basile J, et al. Impact of
intensive glycemic control on the incidence of atrial fibrillation and associated cardiovascular
outcomes in patients with type 2 diabetes mellitus (from the Action to Control Cardiovascular
Risk in Diabetes Study). The American journal of cardiology 2014 Oct 15,;114(8):1217-1222.
14: Buch P, Friberg J, Scharling H, Lange P, Prescott E. Reduced lung function and risk of
atrial fibrillation in The Copenhagen City Heart Study. European Respiratory Journal 2003
Jun 1,;21(6):1012-1016.
15: CKD
16: Chamberlain AM, Agarwal SK, Folsom AR, Duval S, Soliman EZ, Ambrose M, et al.
Smoking and incidence of atrial fibrillation: Results from the Atherosclerosis Risk in
Communities (ARIC) Study. Heart Rhythm 2011;8(8):1160-1166.
17: Marott SCW, Nordestgaard BG, Zacho J, Friberg J, Jensen GB, Tybjaerg-Hansen A, et al.
Does elevated C-reactive protein increase atrial fibrillation risk? A Mendelian randomization
of 47,000 individuals from the general population. Journal of the American College of
Cardiology 2010 Aug 31,;56(10):789.
18: Larsson S, Drca N, Wolk A. Alcohol Consumption and Risk of Atrial Fibrillation A
Prospective Study and Dose-Response Meta-Analysis. JOURNAL OF THE AMERICAN
COLLEGE OF CARDIOLOGY 2014;64(3):282.
19: Soliman EZ, Goff DC. The Paradox of Racial Distribution of Atrial Fibrillation. Journal
of the National Medical Association 2008 Apr;100(4):447-448.
20: Allessie MA, Boyden PA, Camm AJ, Kleber AG, Lab MJ, Legato MJ, et al.
Pathophysiology and Prevention of Atrial Fibrillation. Circulation 2001 Feb 6,;103(5):769-
777.
21: Schotten U, Verheule S, Kirchhof P, Goette A. Pathophysiological mechanisms of atrial
fibrillation: a translational appraisal. Physiological reviews 2011 Jan;91(1):265-325.
22: Groot N, Houben R, Smeets J, Boersma E, Schotten U, Schalij MJ, et al.
Electropathological substrate of longstanding persistent atrial fibrillation in patients with
structural heart disease: Epicardial breakthrough. Circulation (Baltimore) 2010 Oct
26,;122(17):1674-1682.
23: Hall, John E, and Arthur C. Guyton. Guyton and Hall Textbook of Medical Physiology.
Philadelphia, PA: Saunders Elsevier, 2011.
24: Kirchhof P, Benussi S, Kotecha D, Ahlsson A, Atar D, Casadei B, et al. 2016 ESC
Guidelines for the management of atrial fibrillation developed in collaboration with EACTS.
European Heart Journal 2016 Aug 27,:ehw210.
27
25: Lim HS, Willoughby SR, Schultz C, Gan C, Alasady M, Lau DH, et al. Effect of atrial
fibrillation on atrial thrombogenesis in humans: impact of rate and rhythm. Journal of the
American College of Cardiology 2013 Feb 26,;61(8):852-860.
26: Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for
stroke: the Framingham Study. Stroke; a journal of cerebral circulation 1991 Aug;22(8):983-
988.
27: Lin H, Wolf PA, Kelly-Hayes M, Beiser AS, Kase CS, Benjamin EJ, et al. Stroke Severity
in Atrial Fibrillation: The Framingham Study. Stroke 1996 Oct 1,;27(10):1760-1764.
28: Lin H, Wolf PA, Benjamin EJ, Belanger AJ, D'Agostino RB. Newly Diagnosed Atrial
Fibrillation and Acute Stroke : The Framingham Study. Stroke 1995 Sep 1,;26(9):1527-1530.
29: Brambatti M, Connolly SJ, Gold MR, Morillo CA, Capucci A, Muto C, et al. Temporal
Relationship Between Subclinical Atrial Fibrillation and Embolic Events. Circulation 2014
May 27,;129(21):2094-2099.
30: Robert G. Hart, Lesly A. Pearce, Maria I. Aguilar. Meta-analysis: Antithrombotic Therapy
to Prevent Stroke in Patients Who Have Nonvalvular Atrial Fibrillation. Annals of Internal
Medicine 2007 Jun 19,;146(12):857-867.
31: Pisters R. A Novel User-Friendly Score (HAS-BLED) To Assess 1-Year Risk of Major
Bleeding in Patients With Atrial Fibrillation. Chest 2010 Nov;138(5):1093.
32: Apostolakis S, Sullivan RM, Olshansky B, Lip GYH. Factors affecting quality of
anticoagulation control among patients with atrial fibrillation on warfarin: the SAMe-TT₂R₂ score. Chest 2013 Nov;144(5):1555.
33: Ntaios G, Papavasileiou V, Diener H, Makaritsis K, Michel P. Nonvitamin-K-antagonist
oral anticoagulants in patients with atrial fibrillation and previous stroke or transient ischemic
attack: a systematic review and meta-analysis of randomized controlled trials. Stroke; a
journal of cerebral circulation 2012 Dec;43(12):3298.
34: Sjögren V, Grzymala-Lubanski B, Renlund H, Friberg L, Lip GYH, Svensson PJ, et al.
Safety and efficacy of well managed warfarin. A report from the Swedish quality register
Auricula. Thrombosis and haemostasis 2015 Jun;113(6):1370.
35: Kumana CR, Cheung BMY, Siu DCW, Tse H, Lauder IJ. Non‐vitamin K Oral
Anticoagulants Versus Warfarin for Patients with Atrial Fibrillation: Absolute Benefit and
Harm Assessments Yield Novel Insights. Cardiovascular Therapeutics 2016 Apr;34(2):100-
106.
36: Ruff CT, Giugliano RP, Braunwald E, Hoffman EB, Deenadayalu N, Ezekowitz MD, et
al. Comparison of the efficacy and safety of new oral anticoagulants with warfarin in patients
with atrial fibrillation: a meta-analysis of randomised trials. Lancet (London, England) 2014
Mar 15,;383(9921):955-962.
28
37: Connolly SJ, Ezekowitz MD, Yusuf S, Eikelboom J, Oldgren J, Parekh A, et al.
Dabigatran versus Warfarin in Patients with Atrial Fibrillation. The New England Journal of
Medicine 2009 Sep 17,;361(12):1139-1151.
38: Halsey C, Chugh A. Rate versus rhythm control for atrial fibrillation. Cardiology clinics
2014 Nov;32(4):521-531.
39: Nikolaidou T, Channer KS. Chronic atrial fibrillation: a systematic review of medical
heart rate control management. Postgraduate medical journal 2009 Jun;85(1004):303.
40: Friberg L, Rosenqvist M, Lindgren A, Terént A, Norrving B, Asplund K. High Prevalence
of Atrial Fibrillation Among Patients With Ischemic Stroke. Stroke 2014 Sep;45(9):2599-
2605.
41: Freedman B, Martinez C, Katholing A, Rietbrock S. Residual Risk of Stroke and Death in
Anticoagulant-Treated Patients With Atrial Fibrillation. JAMA Cardiology 2016 Jun
1,;1(3):366-368.
42: Pisters R, van Oostenbrugge RJ, Knottnerus ILH, de Vos CB, Boreas A, Lodder J, et al.
The likelihood of decreasing strokes in atrial fibrillation patients by strict application of
guidelines. Europace : European pacing, arrhythmias, and cardiac electrophysiology : journal
of the working groups on cardiac pacing, arrhythmias, and cardiac cellular electrophysiology
of the European Society of Cardiology 2010 Jun;12(6):779-784.
43: Tsang TS, Barnes ME, Pellikka PA, Gin K, Miyasaka Y, Seward JB, et al. 173 Silent
atrial fibrillation in olmsted county: A community-based study. Canadian Journal of
Cardiology 2011 Sep;27(5):S122.
44: Maxim LD, Niebo R, Utell MJ. Screening tests: a review with examples. Inhalation
Toxicology 2014 Nov;26(13):811-828.
45: Alberg A, Park J, Hager B, Brock M, Diener-West M. The use of “overall accuracy” to
evaluate the validity of screening or diagnostic tests. J GEN INTERN MED 2004
May;19(S5):460-465.
46: Taggar JS, Coleman T, Lewis S, Heneghan C, Jones M. Accuracy of methods for
detecting an irregular pulse and suspected atrial fibrillation: A systematic review and meta-
analysis. European journal of preventive cardiology 2016 Aug;23(12):1330-1338.
47: Svennberg E, Engdahl J, Al-Khalili F, Friberg L, Frykman V, Rosenqvist M. Mass
Screening for Untreated Atrial Fibrillation: The STROKESTOP Study. Circulation 2015 Jun
23,;131(25):2176-2184.
48: Lowres N, Neubeck L, Salkeld G, Krass I, McLachlan AJ, Redfern J, et al. Feasibility and
cost-effectiveness of stroke prevention through community screening for atrial fibrillation
using iPhone ECG in pharmacies. The SEARCH-AF study. Thrombosis and haemostasis
2014 Jun;111(6):1167.
29
49: Applied Biomedical Systems BV, MyDiagnostick. The added value of MyDiagnostick.
Available via https://www.mydiagnostick.com/professionals/mydiagnostick-en. Accessed at
13-02-2017.
50: Tieleman RG, Plantinga Y, Rinkes D, Bartels GL, Posma JL, Cator R, et al. Validation
and clinical use of a novel diagnostic device for screening of atrial fibrillation. Europace :
European pacing, arrhythmias, and cardiac electrophysiology : journal of the working groups
on cardiac pacing, arrhythmias, and cardiac cellular electrophysiology of the European
Society of Cardiology 2014 Sep;16(9):1291-1295.
51: Jacobs M, Kaasenbrood F, Postma M, Van Hulst M, Tieleman R. Abstract 11651: Cost-
effectiveness of Screening for Atrial Fibrillation in Primary Care With a Hand-held, Single-
lead ECG Device in the Netherlands. Circulation 2016 Nov 11,;134(Suppl_1 Suppl
1):A11651.
30
8. Appendix
I Informative letter sent to all patients
Screeningsonderzoek voor een hartritmestoornis
Het Martini Ziekenhuis is een topklinisch opleidingsziekenhuis dat zich inzet voor de beste
zorg. Om de zorg te kunnen verbeteren is het belangrijk dat er onderzoek plaatsvindt.
Op dit moment onderzoeken we of het nuttig is om alle patiënten ouder dan 65 jaar te
screenen op de aanwezigheid van boezemfibrilleren. Dit is een hartritmestoornis waarbij de
boezems van het hart te snel en onregelmatig samentrekken. Dit kan klachten geven als een
onregelmatige hartslag, duizeligheid of kortademigheid, maar sommige mensen ervaren
helemaal geen klachten. Deze hartritmestoornis is niet acuut levensgevaarlijk.
U bent reeds ingepland voor het preoperatieve spreekuur (POS) van de anesthesie op 03-01-
2017. Omdat u 65 jaar of ouder bent, wordt u tijdens een onderdeel van dit spreekuur
gevraagd deel te nemen aan dit onderzoek. We vragen u dan om voor één minuut een metalen
staafje vast te houden, de MyDiagnostick. Dit apparaat beoordeelt uw hartritme.
Voor het onderzoek zal een aantal gegevens worden verzameld, zoals uw leeftijd, diagnoses
en huidige medicatie. Deze gegevens zullen worden voorzien van een code en anoniem
worden verwerkt. Als u niet wilt deelnemen aan het onderzoek, dan heeft dit geen enkele
invloed op uw behandeling. U kunt dit dan in het ziekenhuis aan de onderzoeker doorgeven.
Namens het Martini-Onderzoeksteam bedanken wij u voor uw medewerking en tijd. Mocht u
vragen hebben, kunt u onderstaand nummer van de polikliniek Cardiologie bellen. Zij zorgen
dan dat u zo snel mogelijk antwoord krijgt op uw vraag. Mocht uw afspraak bij de POS reeds
verzet of geannuleerd zijn, kunt u deze brief als niet verzonden beschouwen
Het Martini-Onderzoeksteam
Dr. R. Tieleman, cardioloog
Mw. E. Roseboom, student-onderzoeker
050-524 5800
31
II Approval of the Medical Ethical Committee