[Downloaded free from http://www.ijamhrjournal.org on Friday, January 3, 2020, IP: 185.130.147.

68]

Original Article

Diagnostic Accuracy of Apex-Pulse Deficit for Detecting Atrial

Fibrillation
Anjali Rajkumar, Aditya Bhattacharjee, Raja J. Selvaraj
Department of Cardiology, Jawaharlal Institute of Postgraduate Medical Education and Research, Puducherry, India

Abstract
Background: Screening for asymptomatic atrial fibrillation (AF) can identify patients at risk of stroke and help initiate treatment. Apex-pulse
deficit, the difference between apex beat rate and peripheral pulse rate, has been described as a clinical sign to identify AF. However, the
accuracy of this measure to identify AF has not been studied before. Aims: The primary aim of this study was to determine the sensitivity
and specificity of apex-pulse deficit more than 10, measured over 1 min, to identify AF using 12‑lead electrocardiogram (ECG) as the gold
standard. Methods: This was a prospective cross‑sectional study. Subjects were those above 30 years of age with known AF (cases) or not
in AF (controls). Apex-pulse deficit was measured in each of them and correlated with rhythm detected in 12‑lead ECG. Results: A total of
70 patients were studied, 35 cases and 35 controls. Apex-pulse deficit was significantly larger for cases as compared to controls and was a
good discriminant to identify AF. Receiver operating characteristic curve analysis showed an area under the curve of 0.86. With a cutoff of 10,
sensitivity and specificity to identify AF were 62.8% and 85.7%, respectively. Using a cutoff of 5 increased the sensitivity to 80%. Counting
over 30 s was significantly less accurate than counting over one full minute. Conclusion: Apex-pulse deficit is a low‑cost method to identify
AF and may be useful for screening. A cutoff of 5 may enhance the sensitivity of measurement as compared to the traditional cutoff of 10.

Keywords: Atrial fibrillation, pulse deficit, screening

Introduction Various approaches have been used to screen for AF at a

Atrial fibrillation (AF) is the most common sustained disorder
of heart rhythm.[1] Normally, the atria are activated by regular
pacemaker impulses generated by the sinoatrial node. In AF,
there is rapid, chaotic activation of the atria. This results in an
irregular and often rapid heartbeat.
AF is present in 1%–4% of the population in the West[2]
and becomes more prevalent with increasing age. AF often
produces symptoms from the fast heart rate or heart failure,
resulting in the patients seeking treatment. However, AF can
also be asymptomatic. AF is associated with a risk of stroke
due to the formation of clots in the heart, and this risk is the
same in asymptomatic patients also. About 25% of stroke
patients have been diagnosed with asymptomatic AF. [3]
Although the evidence is not clear at present if screening
for asymptomatic patients with AF will provide a net
benefit, asymptomatic patients are also at risk of stroke. If these
asymptomatic patients are identified by screening, treatment
with antiplatelet or anticoagulant drugs, especially in patients
with additional risk factors can reduce the risk of stroke.
Access this article online
Quick Response Code:
Website:
www.ijamhrjournal.org
DOI:
10.4103/IJAMR.IJAMR_48_19
single time point. Devices have been developed to detect
AF by identifying pulse irregularity, and these include blood
pressure monitors with AF detection function.[4] Opportunistic
detection of AF using such blood pressure monitors
has been proved to have reasonable sensitivity and
specificity.[5] Smartphone applications with pulse detection
by finger photoplethysmography have also been used.[6]
These are, however, expensive, especially if required in large
numbers for community screening, and require training for
use. Low technology methods such as palpation of pulse by a
trained nurse to detect irregularity have been employed[7] but
require training and are subjective. AF can be detected in an
electrocardiogram (ECG) and this remains the gold standard.
Address for correspondence: Dr. Raja J. Selvaraj,
Department of Cardiology, Jawaharlal Institute of Postgraduate Medical
Education and Research, Dhanvantri Nagar, Puducherry ‑ 605 006, India.
E‑mail: rajajs@gmail.com
This is an open access journal, and articles are distributed under the terms of the Creative
Commons Attribution‑NonCommercial‑ShareAlike 4.0 License, which allows others to
remix, tweak, and build upon the work non‑commercially, as long as appropriate credit
is given and the new creations are licensed under the identical terms.
For reprints contact: reprints@medknow.com
How to cite this article: Rajkumar A, Bhattacharjee A, Selvaraj RJ.
Diagnostic accuracy of apex-pulse deficit for detecting atrial fibrillation.
Int J Adv Med Health Res 2019;6:52-5.
Received: 19‑04‑2019, Accepted: 02‑11‑2019, Published: 02-01-2020

52 © 2020 International Journal of Advanced Medical and Health Research | Published by Wolters Kluwer - Medknow
[Downloaded free from http://www.ijamhrjournal.org on Friday, January 3, 2020, IP: 185.130.147.68]

Rajkumar, et al.: Apex-pulse deficit in atrial fibrillation

For the purposes of screening, a 3‑lead or single‑lead ECG Analysis

has been tried; however, the need for equipment and trained The ECG was read by a cardiologist to identify the rhythm as
personnel for the interpretation has precluded its use for AF or not AF. The use of an apex-pulse deficit measurement
widespread community screening purposes in the background of more than 10 to identify AF was compared with this
of insufficient evidence for screening. gold standard method. Based on these results, sensitivity
An ideal screening technique must satisfy three important and specificity of the apex-pulse deficit to identify AF were
calculated. For the secondary endpoints,  (1) the apex-pulse
criteria – it must be rapid and preferably noninvasive, it must
deficit calculated by doubling the difference at 30 s was
be available at low cost, and minimal training is desirable
compared with the measurement at 1  min and  (2) receiver
with a view to cover large populations, including remote
operating characteristic (ROC) curve analysis was performed
areas. Apex-pulse deficit has been described as a clinical sign
to identify the optimal cut‑off of apex-pulse deficit to identify
in patients with AF. This is the difference between the heart
AF.
rate counted from heart sounds (in terms of apex beat) and
peripheral pulse palpated at the radial artery. The deficit is seen With a case‑to‑control ratio of 1:1, we estimated that a
in AF because some of the heartbeats which are audible do not minimum sample size of 66 patients will be required to provide
produce a palpable pulse owing to the irregularity. A deficit a power of 0.9 to detect an area under the curve of 0.7 using
of 10 or more has been described to indicate AF.[8] However, ROC curve analysis.
the sign has not been clinically validated, and the accuracy to
identify AF is not known. We designed this study to determine Results
the diagnostic accuracy of apex-pulse deficit to identify AF as
A total of 70 patients were studied, 35 in sinus rhythm and 35
compared to the gold standard of a 12‑lead ECG.
in AF. The baseline characteristics of the patients are shown
in Table 1.
Methods
The mean heart rate (counted from apex beat over 1 min) was
This was a prospective cross‑sectional study among patients 81 ± 15 bpm in cases and 90 ± 20 bpm in the controls. Among
attending the cardiology outpatient department in a tertiary the cases, eight patients had a heart rate of more than or equal
care center in southern India. The study was approved by to 100 bpm, while this was present in five patients from the
the Institute Ethics Committee (no: JIP/lEC/2018/0203) and control group.
informed consent was obtained from all participants.
Apex pulse deficit measured at 30 s and at 60 s was significantly
Inclusion criteria larger for cases as compared to controls. Irregular pulse by
1. Patients above the age of 30 years, with known AF (cases) palpation or auscultation was also more frequently noted in
2. Patients above the age of 30 years, attending the cardiology patients with AF [Table 2]. None of the patients in the control
outpatient clinic with conditions such as hypertension or group had premature atrial or ventricular beats on the ECG.
coronary artery disease but not in AF (controls).
The analysis of apex-pulse deficit measured at 60 s to identify
For each patient, informed consent was first obtained. AF showed an area under the ROC curve of 0.86. The ROC
Apex-pulse deficit was measured as described below by two
operators. Immediately following this, a 12‑lead ECG with a
rhythm strip was recorded. Care was taken to avoid a delay Table 1: Baseline characteristics of the patients
longer than 5 min between the calculation of apex-pulse deficit Characteristics Non‑AF (n=35) AF (n=35)
and recording of the ECG. Age, years 55.1±14.6 47.8±11
Male Sex, n (%) 32 (91.4) 18 (51.4)
Measurement of apex-pulse deficit Diabetes mellitus, n (%) 10 (28.6) 3 (8.6)
Two operators, both 3rd year medical students, performed the Hypertension, n (%) 11 (31.4) 6 (17.1)
measurements for all the patients after brief training. With Smoking, n (%) 9 (25.7) 2 (5.7)
the patient seated, operator 1 used a stethoscope to listen to Coronary artery disease, n (%) 22 (62.9) 4 (11.4)
the heart sounds, while operator 2 palpated the radial pulse. Rheumatic heart disease, n (%) 2 (5.7) 12 (34.3)
Once both the operators were ready, a timer was used to start AF=Atrial fibrillation
counting simultaneously. Heart rate counted by each operator
was recorded at the end of 30 s. The procedure was then
repeated and the heart rate was recorded at the end of 1 min. Table 2: Comparison of apex-pulse deficit in the control
The two operators interchanged their roles in different subjects group (sinus) versus atrial fibrillation
to avoid bias. Apex-pulse deficit was calculated as the difference Sinus (n=35) AF (n=35)
between the rates counted by the two operators at the end of Apex-pulse deficit at 30 s 0 (−2, 3) 14 (2, 25)
1 min. The apex-pulse deficit was also derived from the 30‑s (median, 25th and 75th percentiles)
recording by multiplying the difference by 2. In addition, each Apex-pulse deficit at 60 s 1 (0, 2.5) 16 (8, 26)
operator also designated the rhythm as regular or irregular. (median, 25th and 75th percentiles)

International Journal of Advanced Medical and Health Research  ¦  Volume 6  ¦  Issue 2  ¦  July‑December 2019 53
[Downloaded free from http://www.ijamhrjournal.org on Friday, January 3, 2020, IP: 185.130.147.68]
Rajkumar, et al.: Apex-pulse deficit in atrial fibrillation
Figure 1: Receiver operating characteristic curve of apex-pulse deficit
to identify atrial fibrillation
curve is shown in Figure 1. The analysis of apex-pulse deficit
measured at 30 s to identify AF showed an area under the
ROC curve of 0.75. The sensitivity and specificity for pulse
irregularity and different cutoffs of apex-pulse deficit are
listed in Tables 3 and 4 respectively. The positive predictive
value and negative predictive value of apex-pulse deficit of 10
or more to identify AF were 81.5% and 69.8%, respectively.
Discussion
In this study, we assessed the diagnostic accuracy of apex-pulse
deficit counted over a minute to identify AF. We found that an
apex-pulse deficit of 10 or more counted over 1 min identified
AF with reasonable accuracy. Although apex-pulse deficit has
been described as a clinical sign to identify AF, its diagnostic
accuracy has not been studied before. This assumes importance
in the current era where screening for asymptomatic AF is
assuming importance given the potential to reduce strokes. The
sensitivity of 62.8% and specificity of 85.7% suggest that it
is a reasonably accurate method to identify AF. This makes it
a promising approach for screening of AF in the community
since it requires minimal training, is cheap, and takes only a
couple of minutes to perform.
For a screening test, it is preferable to have a high sensitivity,
even at the cost of loss of some specificity to reduce false
negatives. Our ROC curve analysis suggests that a cutoff of
5 will improve the sensitivity to 80%, while the specificity
is still 80%. This cutoff may be preferable for the use for
screening.
We explored if counting for 30 s may be sufficient instead of for
a full minute as this will significantly reduce the time taken for
a measurement. However, this was less accurate to identify AF.
Apex-pulse deficit is likely to have reduced sensitivity to
identify AF when the heart rate is well controlled. In our
patients with AF, among the lowest tertile of heart rate
(heart rate ≤78 bpm), only 15.4% had an apex-pulse deficit
54 International Journal of Advanced Medical and Health Research  ¦  Volume 6  ¦  Issue 2  ¦  July‑December 2019
Table 3: Irregular pulse detection
Sinus (n=35), AF (n=35),
n (%) n (%)
Irregular pulse (palpation) 4 (11.4) 32 (91.4)
Irregular heart beat (auscultation) 4 (11.4) 33 (94.3)
Table 4: Sensitivity and specificity of apex-pulse deficit
and irregular pulse on auscultation and palpation
Measures Sensitivity (%) Specificity (%)
Apex-pulse deficit >10 62.8 85.7
Apex-pulse deficit >5 80 80
Irregular pulse (palpation) 91.4 88.6
Irregular heartbeat (auscultation) 94.3 88.6
more than 10, whereas in the highest tertile  (heart rate
>90 bpm), 100% had an apex-pulse deficit more than 10.
While this assumes significance in a hospital setting where the
patients are already diagnosed with AF and are on treatment
to control the ventricular rate, it may be less significant in a
community screening of previously undiagnosed patients. If
used for screening, an additional method of screening may be
considered in patients with a controlled ventricular rate.
The sensitivity and specificity of irregular pulse/heartbeat
detection seem to be better than the apex-pulse deficit. Similar
studies using trained nurses have shown sensitivity ranging
from 91% to 100%, while specificity ranged from 70% to
77%.[9] However, this is a subjective measure, and it is unclear
if the accuracy in the identification of irregular pulse achieved
in this study by medical students and in other studies by nurses
would apply to other health‑care workers.
Smartphone applications have been developed to screen for
AF. A smartphone application using photoplethysmography to
identify the pulse can help in screening for AF, according to
late‑breaking results from the DIGITAL‑AF study presented
at European Society of Cardiology (ESC) congress.[10] Blood
pressure monitors have also been used for opportunistic
detection and screening for AF and have generally shown
sensitivity >85% and specificity >90%.[4] Our study looked at
a low‑cost, low‑technology option that may be more applicable
for large‑scale screening.
In addition to evaluating apex-pulse deficit as a screening tool,
our study also has value in being the first time, this classically
described clinical sign in AF is evaluated in an evidence‑based
fashion. Our findings suggest that while this could be used as
a clinical sign to identify AF with reasonable accuracy, the
detection of an irregular pulse itself may have equal or better
accuracy.
Limitations
A significant limitation of the study is the use of a two‑gate
design where patients with previously diagnosed AF were used
as cases, while those without diagnosed AF were controls.
[Downloaded free from http://www.ijamhrjournal.org on Friday, January 3, 2020, IP: 185.130.147.68]
Rajkumar, et al.: Apex-pulse deficit in atrial fibrillation
Sensitivity and specificity measured from the study may not
accurately reflect the sensitivity and specificity of apex-pulse
deficit when used for screening in the community. However,
it does provide a preliminary estimate of the utility as a
diagnostic test.
Similarly, the prevalence of AF in the general population is
expected to be much lower than 50%. The positive predictive
value of the test in a population with a low prevalence of the
condition would be much lower. As with any screening test, it
would be preferable to use a prespecified population at higher
risk, older age, for example, to reduce false positives.
Conclusion
Apex-pulse deficit is a viable tool to identify AF. Using a
cutoff of 5 may improve the sensitivity of AF detection. This
raises the possibility of using apex-pulse deficit as a low‑cost
tool to screen for AF without the need for extensive training.
However, an irregular pulse on palpation may be more sensitive
and specific.
Financial support and sponsorship
One of the authors (A.R.) obtained funding from the Indian
Council of Medical Research for this study as a Short‑Term
Studentship project in the year 2018.
Conflicts of interest
There are no conflicts of interest.
International Journal of Advanced Medical and Health Research  ¦  Volume 6  ¦  Issue 2  ¦  July‑December 2019
References
1. Carlsson  J, Miketic  S, Windeler  J, Cuneo A, Haun  S, Micus  S, et al.
Randomized trial of rate‑control versus rhythm‑control in persistent
atrial fibrillation: The strategies of treatment of atrial fibrillation (STAF)
study. J Am Coll Cardiol 2003;41:1690‑6.
2. Friberg L, Bergfeldt L. Atrial fibrillation prevalence revisited. J Intern
Med 2013;274:461‑8.
3. Healey JS, Connolly SJ, Gold MR, Israel CW, Van Gelder IC, Capucci A,
et al. Subclinical atrial fibrillation and the risk of stroke. N Engl J Med
2012;366:120‑9.
4. Wiesel  J, Wiesel  D, Suri  R, Messineo  FC. The use of a modified
sphygmomanometer to detect atrial fibrillation in outpatients. Pacing
Clin Electrophysiol 2004;27:639‑43.
5. Kane  SA, Blake  JR, McArdle  FJ, Langley  P, Sims AJ. Opportunistic
detection of atrial fibrillation using blood pressure monitors:
A systematic review. Open Heart 2016;3:e000362.
6. Chan  PH, Wong  CK, Poh  YC, Pun  L, Leung  WW, Wong  YF, et al.
Diagnostic performance of a smartphone‑based photoplethysmographic
application for atrial fibrillation screening in a primary care setting.
J Am Heart Assoc 2016;5. pii: e003428.
7. Fitzmaurice DA, Hobbs FD, Jowett S, Mant J, Murray ET, Holder R,
et al. Screening versus routine practice in detection of atrial fibrillation
in patients aged 65 or over: cluster randomised controlled trial. Br Med
J (Clin Res Ed) 2007;335:383.
8. Longo D, Fauci A, Kasper D, Hauser S, Jameson JL, Loscalzo J.
Harrison’s Principles of Internal Medicine, 18th Edition. Published by
McGraw Hill Medical; 2011.
9. Cooke  G, Doust  J, Sanders  S. Is pulse palpation helpful in detecting
atrial fibrillation? A systematic review. J Fam Pract 2006;55:130‑4.
10. Digital AF, Vandervoort  PM. Late‑Breaking Science in Telemedicine.
Munich: European Society of Cardiology Congress; 2018. p. 25‑9.
55