Vous êtes sur la page 1sur 6

38

Journal of The Association of Physicians of India Vol. 64 August 2016

Original Article

Sensitivity and Specificity of Pulse Oximetry


and Ankle-Brachial Index for Screening
Asymptomatic Peripheral Vascular Diseases in
Type 2 Diabetes Mellitus
M Satheesh Kumar1, Ayush Lohiya2, Viviktha Ramesh2, Priyamadhaba Behera2,
Sarika Palepu 2, SA Rizwan3

Abstract

Editorial Viewpoint

Objectives: To compare pulse oximetry and Ankle-Brachial Index (ABI)


with duplex ultrasonography as reference standard to determine the
diagnostic accuracy for screening asymptomatic PVD in type 2 diabetes
mellitus.

Diabetic patients tend to


have a two to four fold
increase in the risk of
PVD.

Methods: This cross-sectional study was conducted in 2012 at tertiary


hospital in Madurai among diabetic patients attending the medicine
outpatient department (OPD). Type 2 Diabetes Mellitus patients,
asymptomatic with regards to symptoms and signs of PVD, aged above
40 years were included.
Pulse Oximetry was performed using a pulse oximeter and ABI using
sphygmomanometer cuffs and duplex ultrasonography of femoral,
popliteal, tibial, posterior tibial and dorsalis pedis arteries. A diagnosis
of PVD was based on: monophasic waveforms in any artery by duplex
ultrasonography, toe saturation being less than finger saturation by
>2% or if foot saturation decreased by >2% in an elevated position and
an ABI <0.9.
Results: Among 120 patients included in the study, prevalence of PVD
was 22.5% (95% CI: 15.9, 30.8). The PVD group had a higher proportion
of elderly, males, current smokers, long-standing diabetics and
comorbidities. The sensitivity, specificity, positive predictive value (PPV)
and negative predictive value (NPV) of pulse oximetry were 74.1% (95%
CI: 55.3, 86.8), 95.7% (89.4, 98.3), 83.3% (64.1, 93.3) and 92.7% (85.7, 96.4)
respectively, while those of ABI were 70.3% (51.5, 84.2), 87.1 (78.8, 92.5),
61.3% (43.8, 76.3) and 91.0% (83.3, 95.4) respectively. Parallel testing had
net sensitivity increased to 92.3% and net specificity decreased to 83.3%.
Performances did not differ across the subgroups.
Conclusion: Pulse oximetry was atleast as good as ABI for the screening
for asymptomatic PVD among diabetics.

Introduction

ndia, dubbed to be the diabetic


capital of the world, has the
second largest absolute number

Pulse oximetry is an easier


way to access PVD.
T h i s s t u d y s h o w s
pulse oximetry to be as
good as ankle-brachial
index for screening of
asymptomatic PVD.
of diabetic patients globally,
with a prevalence of about 8%
among adults next only to China.
Currently about 35 million Indians
are reported to suffer from
diabetes, a significant proportion
of whom are either undiagnosed
or diagnosed but under-treated
leading to poor glycemic
control. 1 This leads to accelerated
d e ve l o p m e n t o f c o m p l i c a t i o n s
like Peripheral Vascular Disease
(PVD). PVD, defined as a clinical
disorder in which there is stenosis
or occlusion in the arteries of the
limbs, 2 is a common complication
of long standing diabetes mellitus.
Individuals with diabetes tend to
have a two to four fold increase

Post-Graduate Trainee, Madurai Medical College, Madurai, Tamil Nadu; 2Post-Graduate Trainee, All India Institute
of Medical Sciences, New Delhi; 3Assistant Professor, Department of Community Medicine, Velammal Medical
College Hospital & Research Institute, Madurai, Tamil Nadu
Received: 25.06.2015; Revised: 21.07.2015; Accepted: 23.10.2015

Journal of The Association of Physicians of India Vol. 64 August 2016

Box 1: Eligibility criteria


Sr. Inclusion criteria
No.
1. Adults pre-diagnosed type 2 diabetes
mellitus, either physician diagnosed
or based on blood sugar records as
per American Diabetic Association
(ADA) criteria,16 irrespective of
control of blood sugar, duration of
diagnosis, treatment and presence of
other complications.
2. Not previously investigated for or
diagnosed as PVD and asymptomatic
with regard to symptoms of PVD
such as pain, swelling, ulcers,
previous amputations etc.
Sr. Exclusion criteria
No.
1. Age less than 40 years.
2. Patients suffering from
hypercoagulable states, congestive
heart failure, valvular heart disease,
suspected arteritis and collagen
vascular disease.
3. Patients who were unable to lie
supine for the period of testing.
4. Extremely sick patients who required
intensive care.

in the risk of PVD. 3 Prevalence of


PVD among overt diabetic patients
was estimated to be about 6 to 8%
in south Indians. 4,5 In diabetes,
PVD has certain peculiar features,
which make it more difficult
to treat; it is multi-segmental,
bilateral, extensive and involves
predominantly distal vessels
leading to poorer prognosis as
compared to non-diabetic PVD. A
large proportion of lower extremity
amputations among these patients
lead to significant disability and
economic burden to the individual
and the health system in developing
countries like India. 6 Stenosis in
PVD usually develops gradually
and is accompanied by formation of
extensive collaterals so most of the
patients are asymptomatic. Hence,
early detection and treatment of
PVD in patients with diabetes
mellitus can surely limit its
debilitating effects.
Various methods are available
for diagnosis of PVD. Duplex
ultrasonography, an ultrasoundbased technique, is a very accurate
me t h o d f o r d e t e c t i n g va s c u l a r
lesions. It is the combination
of real-time B mode scanning

in colour and Doppler spectral


analysis, but it is expensive and
hence not appropriate for screening
purposes. Other imaging modalities
like arteriography, Computerized
Tomography (CT) angiography and
Magnetic Resonance Angiography
(MRA) are also accurate though
expensive in clinical practice.
Among the methods used for
screening, Ankle-Brachial Index
(ABI) and pulse oximetry are
common. ABI, the ratio of blood
pressure (BP) in the legs to that of
in the arms was shown to have a
sensitivity of over 90% in detecting
P V D 7 , 8 b u t s o m e s t u d i e s we r e
done in overtly symptomatic cases
and therefore liable to spectrum
bias , for e.g ., a st udy done in
United States of America (USA)
reported that ABI had a sensitivity
of only 28.4% among asymptomatic
patients. 9
The pulse oximeter, which works
by combining spectrophotometry
and optical plethysmography
has a reported sensitivity of
about 80% and specificity of over
90%. 10 It is simple, non-invasive,
cheaper, more readily available
and can be performed without
any special training. However,
it performs poorly in cases of
dyshemoglobinemia, pulsating
veins and poor peripheral
perfusion. 11 Only a few studies
have assessed ABI and pulse
oximeter as a screening tool for
asymptomatic PVD in diabetic
patients in India and hence the need
for this study. 10,12 This study was
designed to assess the diagnostic
accuracy of pulse oximetry and
ABI as a screening tool to detect
asymptomatic PVD in patients with
type 2 diabetes mellitus against
duplex ultrasonography as the
reference standard and to compare
the performance of these tests
across different subgroups.

Material and Methods


Patients and setting

This cross-sectional study was


conducted between March and

39

November 2012 in the medicine


OPD of Government Rajaji
Hospital, Madurai, a tertiary care
hospital in southern India. Diabetic
patients who attended the OPD
were screened in a consecutive
manner for eligibility according to
predefined inclusion and exclusion
criteria using a screening checklist.
(Box 1) All eligible consenting
patients were then subjected to
all three tests. This process was
continued till the requisite sample
size was achieved.
Sample size

Using the formula, 13


4x[sensitivity(1-sensitivity)]
precision^2
and assuming a sensitivity of
90%, 7,8 relative precision of 20%,
alpha of 5% and beta of 20%,
we calculated that 11 patients of
PVD were required. Prevalence of
PVD among diabetic patients was
considered to be about 10%, 5 so the
number of diabetic patients needed
was 110. Applying a non-response
rate of 10%, the final number of
diabetic patients required for the
study was calculated as 122.
n=

Instruments

A handheld pulse oximeter was


used to measure SpO2 (peripheral
arterial oxygen saturation) of
fingers and toes. Systolic blood
pressures of the arms and legs were
measured with appropriate size
sphygmomanometer cuffs and a
handheld 8 MHz Doppler probe.
Duplex ultrasonography of lower
extremity arteries was performed
using a Doppler ultrasonogram
machine.
Procedures

Information on patient
demographics, duration of diabetes
and smoking history was collected
using a semi-structured, pre-tested
questionnaire. A comprehensive
physical examination was done
along with medical record
analysis to check for presence of
comorbidities. Patients medical
r e c o r d s we r e u s e d t o c o n f i r m
the presence of Coronary Heart
Disease (CHD), Cerebrovascular

40

Journal of The Association of Physicians of India Vol. 64 August 2016

Patients screened for


eligibility = 412
Not eligible = 274

Table 1: Distribution of participants


according to selected sociodemographic variables and
comorbidities
Variable

Patients found
eligible = 138
Refusal = 18
Total patients
included = 120

Fig. 1: Flow of participant


selection process

Disease (CVD) and Hypertension.


After this initial assessment,
one of the study investigators
performed ABI followed by pulse
oximetry on the same day. An
expert radiographer (not part of
the study team) who was blind to
the results of index tests, conducted
duplex ultrasonography within
one-week of the initial assessment.
Duplex ultrasonography of lower
extremities was done at the level
of femoral, popliteal, tibial,
posterior tibial and dorsalis pedis
arteries bilaterally. This test has
specificity of 97% when compared
with arteriography and has been
used as the reference standard in
many previous studies. 9,10 ABI was
calculated by measuring systolic
BP of the brachial artery at the
elbows and that of posterior tibial
artery (or dorsalis pedis artery
in case posterior tibial had no
signal) at the ankles, following
the usual precautions prescribed
by the seventh report of the Joint
National Committee on Prevention,
Detection, Evaluation, and
Treatment of High Blood Pressure
( JNC 7) for BP measurement. 14
A hand-held pulse oximeter was
used to measure SpO 2 to measure
the saturation of both index fingers
and both big toes. Toe saturation
was measured in two positions, one
in supine and the other at 12-inch
elevation from the horizontal plane.
Definitions used for PVD diagnosis

The presence of monophasic


waveforms in any one artery by

Age grp
(years)
Sex

Category Peripheral vascular


disease n (%)
Present, Absent,
n=27
n=93
41-50
7 (25.9) 40 (43.0)
51-60
15 (55.6) 43 (46.2)
61-70
Males
Females

5 (18.5)
22 (81.5)
5 (18.5)

10 (10.8)
67 (72.0)
26 (28.0)

Duration <10 years


of DM
10 years
Current smoker
Coronary heart dis.

5 (18.5)

55 (59.1)

22 (81.5)
20 (74.1)
18 (66.7)

38 (40.9)
29 (31.2)
37 (39.8)

Cerebrovascular dis.

4 (14.8)

6 (6.5)

Hypertension

22 (81.5)

55 (59.1)

duplex ultrasonography was taken


as confirmatory evidence for PVD.
B y p u l s e o x i m e t r y , P V D wa s
considered present if toe saturation
was less than finger saturation
by >2% or if the foot saturation
decreased by >2% in the elevated
position. 15 ABI values for each leg
were calculated by dividing the
ankle systolic BP by the elbow
pressure and a value of < 0.9 for
any leg was considered positive for
PVD. 16 A patient was considered to
be positive for PVD even if any one
leg had abnormal results.
Research ethics

Ethical clearance was obtained


from the Institutional Review Board
of the Government Rajaji Hospital,
Madurai. Written informed consent
was obtained from all participants
and confidentiality of information
was strictly maintained. Patients
who required treatment as per their
diagnosis were given appropriate
care.
Statistical analysis

Characteristics of the study


p o p u l a t i o n we r e d e s c r i b e d i n
percentages. Diagnostic accuracy
of the index tests was calculated
in terms of sensitivity, specificity,
PPV, and NPV with 95% Confidence
Intervals (CI). Each index test was
compared separately with the
reference standard and a combined
measure was also calculated for

a parallel testing scenario. When


calculating the net sensitivity
and net specificity for the parallel
testing scenario, positive by any
one test was considered positive
and negative by both tests was
considered negative. The parallel
testing formulae 17 used were:
Net sensitivity = [sensitivity
of test 1 + sensitivity of test 2
(sensitivity of test 1 sensitivity
of test 2)] and Net specificity =
[specificity of test 1 x specificity
o f t e s t 2 ] . We a l s o p e r f o r m e d
subgroup analysis according to
a few important covariates. Age,
sex, smoking status, diabetes
duration, cardiovascular disease
and hypertension, due to their
strong association with PVD were
considered as important covariates
for subgroup analysis.

Results
A total of 412 diabetic patients
were screened and 138 were found
eligible for inclusion. Eighteen
patients (13%) refused to participate
and finally a total of 120 patients
(against a calculated sample size of
122) with type 2 diabetes mellitus
were included in the study (Figure
1). Main reason for refusal was lack
of time to stay for the tests. The
prevalence of PVD in the sample
was 22.5% (95% CI: 15.9, 30.8).
Comparison of baseline parameters
showed that the PVD group, as
compared to non-PVD patients,
had a higher proportion of elderly
patients (18.5% vs. 10.8%), males
(81.5% vs. 72%), smokers (74.1%
vs. 31.2%), long-standing diabetics
(81.5% vs. 40.9%) and comorbidities
(Table 1).
The sensitivity and specificity
of pulse oximetry were 74.1% (95%
CI: 55.3, 86.8) and 95.7% (89.4,
98.3) respectively, while those
of ABI were 70.3% (51.5, 84.2)
and 87.1 (78.8, 92.5) respectively.
The PPV and NPV for pulse
oximetry were 83.3% (64.1, 93.3)
and 92.7% (85.7, 96.4) respectively
a n d t h o s e f o r A B I we r e 6 1 . 3 %
(43.8, 76.3) and 91.0% (83.3, 95.4)

Journal of The Association of Physicians of India Vol. 64 August 2016

41

Table 2: Sensitivity, specificity and predictive values of pulse oximetry and ABI with duplex ultrasonography as reference
standard
Index test

PVD (by duplex USG)


Present nos. Absent nos.

Pulse oximetry
Positive
Negative

20
7

4
89

Positive
Negative

19
8

12
81

Positive
Negative
Total

21
6
27

13
80
93

ABI

Combination
(parallel testing)

Sensitivity
% (95% CI)

Specificity
% (95% CI)

PPV
% (95% CI)

NPV
% (95% CI)

74.1 (55.3, 86.8)

95.7 (89.4, 98.3)

83.3 (64.1, 93.3)

92.7 (85.7, 96.4)

70.3 (51.5, 84.2)

87.1 (78.8, 92.5)

61.3 (43.8, 76.3)

91.0 (83.3, 95.4)

92.3 (86.4, 96.0)

83.3 (75.7, 88.9)

Table 3: Subgroup analysis of sensitivity, specificity and predictive values of pulse oximetry by selected covariates
Index
test

Validity
measure

Age
% (95% CI)
<55 years 55 years
Pulse
Sensitivity
66.6
80.0
oximetry
(35.4,
(51.3,
88.7)
94.6)
Specificity
94.9
97.0
(84.9,
(82.9,
99.8)
98.6)
PPV
72.7
92.3
(39.3,
(62.0,
92.6)
99.5)
NPV
93.3
91.6
(82.9,
(76.4,
97.8)
97.8)

Sex
Diabetes duration
% (95% CI)
% (95% CI)
Male
Female <10 years 10 years
72.7
80.0
60.0
77.2
(49.5,
(29.8,
(17.0,
(54.1,
88.3)
98.9)
92.7)
91.3)
94.7
95.5
96.1
96.3
(78.4,
(86.3,
(80.9,
(86.6,
98.8)
99.7)
99.3)
99.0)
60.0
89.4
84.2
80.0
(65.4,
(59.5,
(29.8,
(17.0,
98.9)
92.7)
98.1)
95.8)
91.4
96.1
96.3
87.8
(81.6,
(78.4,
(86.3,
(72.9,
96.4)
99.7)
99.3)
95.4)

Current smoker
% (95% CI)
Yes
No
70.0
85.7
(45.6,
(42.0,
87.1)
99.2)
96.5
95.3
(80.3,
(86.0,
99.8)
98.7)
66.6
93.3
(66.0,
(30.9,
99.6)
90.9)
82.3
98.3
(64.8,
(90.1,
92.6)
99.9)

CHD
% (95% CI)
Present Absent
77.7
66.6
(51.9,
(30.9,
92.6)
90.0)
97.2
94.6
(84.1,
(84.1,
98.6)
99.8)
93.3
66.6
(66.0,
(30.9,
99.6)
90.0)
90.0
94.6
(75.4,
(84.1,
96.7)
98.6)

Hypertension
% (95% CI)
Present Absent
72.7
80.0
(49.5,
(29.8,
88.3)
98.9)
98.1
92.1
(77.5,
(89.0,
99.9)
97.9)
94.1
57.1
(69.2,
(20.2,
88.1)
99.6)
90.0
97.2
(78.8,
(83.7,
95.8)
99.8)

Table 4: Subgroup analysis of sensitivity, specificity and predictive values of ABI by selected covariates
Index
test
ABI

Validity
measure

Age
% (95% CI)
<55 years 55 years
Sensitivity
66.6
73.3
(35.4,
(44.8,
88.7)
91.0)
Specificity
84.7
91.1
(72.5,
(75.1,
92.3)
97.6)
PPV
47.0
78.5
(23.8,
(48.8,
71.4)
94.2)
NPV
92.5
88.5
(81.2,
(72.3,
97.6)
96.2)

respectively. Therefore, all the


validity parameters were higher
for pulse oximetry as compared
to ABI but the limits of 95% CI
were overlapping, indicating no
significant statistical difference.
A combination of these two tests
in parallel testing mode showed
an increase in net sensitivity to
92% and a slight decline in net
specificity to 83%, as expected
(Table 2).
There was no statistically

Sex
Diabetes duration
% (95% CI)
% (95% CI)
Male
Female <10 years 10 years
68.1
80.0
60.0
72.7
(45.1,
(29.8,
(17.0,
(49.5,
85.2)
98.9)
92.7)
88.3)
78.9
89.5
80.7
92.7
(60.0,
(81.5,
(62.2,
(79.0,
95.3)
92.6)
97.6)
89.8)
42.8
66.6
68.1
44.4
(44.6,
(45.1,
(15.3,
(11.8,
85.2)
77.3)
79.7)
83.5)
89.5
95.4
96.2
83.3
(79.0,
(75.1,
(85.9,
(66.5,
95.3)
99.7)
99.3)
93.0)

Current smoker
% (95% CI)
Yes
No
65.0
85.7
(40.9,
(42.0,
83.6)
99.2)
75.8
92.1
(56.0,
(81.9,
88.9)
97.0)
65.0
54.5
(40.9,
(24.5,
83.6)
81.8)
75.8
98.3
(56.0,
(89.8,
88.9)
99.9)

significant difference (i.e.


overlapping CIs) in sensitivity,
specificity, PPV and NPV between
pulse oximetry and ABI across the
subgroups of age, sex, diabetes
duration, smoking status, presence
of CHD and hypertension (Tables 3,
4). In the parallel testing scenario,
sensitivity increased for all age
groups, male sex, non-smokers,
diabetes duration 10 years and
patients of CHD and hypertension.
The best performance of the parallel
testing strategy was seen in the

CHD
% (95% CI)
Present Absent
72.2
66.6
(46.4,
(30.9,
89.2)
90.9)
81.0
91.0
(64.2,
(79.6,
96.9)
91.4)
65.0
54.5
(40.9,
(24.5,
83.6)
81.8)
85.7
94.4
(68.9,
(83.6,
94.6)
98.5)

Hypertension
% (95% CI)
Present Absent
68.1
80.0
(45.1,
(29.8,
85.2)
98.9)
90.9
81.5
(79.2,
(65.1,
96.6)
91.6)
75.0
36.3
(50.5,
(12.3
,68.3)
90.4)
87.7
96.8
(75.7,
(82.0,
94.5)
99.8)

current non-smokers group with a


sensitivity of 98% and specificity of
88% (Table 5).

Discussion
We found that PVD was fairly
prevalent among the diabetics
included in this study. A few
community-based studies have
reported a lower prevalence
ranging from about 6% to 8%. 4,5
This could be due to the fact that
the study sample came from a

42

Journal of The Association of Physicians of India Vol. 64 August 2016

Table 5: Subgroup analysis of sensitivity, specificity and predictive values of combination of pulse oximetry and ABI by selected
covariates
Index test

Validity
measure

Age
% (95% CI)
<55 years 55 years
Combination Sensitivity
88.8
94.7
(parallel
(82.1,
(89.4,
testing)
94.1)
98.1)
Specificity
80.4
88.4
(71.7,
(81.2,
86.7)
93.4)

tertiary care hospital and hence the


prevalence was higher than in the
general community. In this sample
PVD was more common among
those aged 51-60 years, males and
smokers. Our unit of analyses was
patient and not limb as done by
other studies. 10,12 Since the factors
that are associated with PVD are
clustered within the individual and
PVD is a generalized process, it
was considered better to do patient
level analysis rather than extremity
level analysis.
We found that for asymptomatic
patients, pulse oximetry performed
atleast as good as ABI if not better
and the combination of both could
be used as a good screening tool for
the diagnosis of asymptomatic PVD
among diabetics. On combining
the two tests in parallel testing
scenario, sensitivity increased for
all age groups, among males, nonsmokers, for diabetes duration 10
years and among patients of CHD
and hypertension. Pulse oximetry
is quick, easily available, cheap,
requires minimal training, has low
inter-observer variability and its
ease of use even by paramedical
staff makes it a attractive option
for crowded OPD/physician office
settings. In terms of external
validity, although the study
was conducted in a tertiary care
hospital and the characteristics
of this group may be different
from those attending primary care
clinics, there are no strong reasons
to believe that the results would not
be applicable to them. This study
would be applicable to all such
diabetic patients who are currently
asymptomatic with regard to PVD.
Previous studies assessing the

Sex
Diabetes duration
% (95% CI)
% (95% CI)
Male
Female <10 years 10 years
91.2
96.0
84.0
93.8
(84.1,
(90.5,
(76.3,
(88.3,
95.3)
98.6)
90.1)
97.6)
77.6
89.3
74.7
85.5
(78.2,
(68.9,
(82.1,
(66.0,
91.5)
84.6)
94.1)
82.4)

Current smoker
% (95% CI)
Yes
No
89.5
98.0
(82.1,
(94.1,
94.1)
99.7)
73.2
87.8
(64.4,
(80.2,
80.9)
92.8)

utility of pulse oximetry as a


screening test for PVD have given
mixed results. In a study done in
patients having moderate PVD,
sensitivity of pulse oximetry was
r e p o r t e d t o b e o n l y 1 6 % . 15 I n
another study done in primary
care setting, the sensitivity of
pulse oximetry, ABI and their
combination was reported to be
77%, 63% and 86% respectively. 10
Specificity was 97% for both tests
and 92% for their combination.
Another study showing sensitivity
and specificity of pulse oximetry
to be 87% and 87% respectively
concluded that pulse oximetry
could be used as a simple, noninvasive screening device similar
to ABI in assessing PVD. 12 It has
been demonstrated in previous
studies that pulse oximetry was not
only able to detect asymptomatic
cases, but could also differentiate
between the various grades of
PVD, 1 4 alt houg h a few st udies
reported that ABI performs better
only in cases of severe disease, 9
whereas in our study it was found
to perform reasonably well among
asymptomatic patients.
The idea behind subgroup
analysis was to check if the test
performances were different across
covariates such as age, sex, diabetes
duration, smoking status, presence
of CHD and hypertension, which
are also known risk factors for
P V D . We f o u n d n o s i g n i f i c a n t
change in the performance of tests
and also sensitivity, specificity and
predictive values of pulse oximetry
and ABI in the subgroups were
comparable.
Patients with diabetes more
commonly develop symptomatic

CHD
% (95% CI)
Present Absent
93.8
88.8
(88.3,
(82.1,
97.6)
94.1)
78.7
86.1
(78.2,
(69.8,
85.3)
91.5)

Hypertension
% (95% CI)
Present Absent
91.2
96.0
(84.1,
(90.5,
95.3)
98.6)
89.2
75.1
(82.1,
(66.2,
94.1)
82.4)

forms of PVD, resulting in


large numbers of amputation
surgeries. Relative Risk for
lower extremity amputation in
patients with diabetes is 12 times
as compared with that of nondiabetic patients. 19 Early detection
and appropriate treatment can
prevent this significant disability.
Also, it has been shown that PVD
is correlated with cardiovascular
events such as MI and Stroke.
Therefore early detection and
control of PVD can serve as an
important preventive strategy for
reducing cardiovascular events
among diabetics. International
agencies such as the ADA have
recommended regular screening
for PVD, as a preventive measure
to reduce consequent disability. 2
Our study was carefully designed
to avoid commonly encountered
biases in diagnostic accuracy
studies. Firstly, we included only
asymptomatic patients with regard
to PVD, thereby avoiding spectrum
bias where test performances
are based on only severe cases.
Accuracy of these tests has been
previously established in severe
cases. This selection criterion also
tried to avoid to some extent, the
bias that occurs when the tests have
varying performances with regard
to disease duration. Secondly,
because the reference standard tests
were performed independently of
the index tests, we avoided the
diagnostic review bias. Thirdly,
since all patients underwent
all three tests, we avoided the
verification bias. Finally, although
the reference standard test has very
high sensitivity and specificity 20 is
not perfect and there is a possibility

Journal of The Association of Physicians of India Vol. 64 August 2016

of misclassification bias. However,


this misclassification was likely to
be non-differential affecting the
results only slightly.

Hospital, Madurai, for helping in


the conduct of the study.

There are a few limitations in


the study. Firstly, since a single
investigator performed both pulse
oximetry and ABI, the result of
one might have influenced that of
the other. Secondly, we could not
measure test reliability in terms of
inter-observer and intra-observer
variability. Thirdly, sample size
wa s p r o b a b l y i n a d e q u a t e f o r
subgroup analysis. Finally, we
could not study the effect of a few
covariates like occupation, control
of diabetes, extent of physical
activity and lipid profile, which
have a bearing on development of
PVD. Future research could study
the effect of these variables with a
larger sample size.

1. I nternational Diabetes Federation


[Internet]. IDF diabetes atlas. Sixth ed. 2013
[cited 2014 Jan 14]. Available from: www.
idf.org/diabetesatlas

References

2.

3. Newman AB, Siscovick DS, Manolio


TA, Polak J, Fried LP, Borhani NO,
et al. Ankle-arm index as a marker of
atherosclerosis in the Cardiovascular
Health Study. Cardiovascular Heart Study
(CHS) Collaborative Research Group.
Circulation 1993; 88:837845.
4.

Research Ethics

Ethical clearance was obtained


from the Institutional Review Board
of the Government Rajaji Hospital,
Madurai. Written informed consent
was obtained from all participants
and confidentiality of information
was strictly maintained. Patients
who required treatment as per their
diagnosis were given appropriate
care.
Acknowledgements

We are extremely thankful to all


the individuals who had helped
and participated in the study.
We would also like to thank staff
members of Government Rajaji

Premalatha G, Shanthirani S, Deepa R,


Markovitz J, Mohan V. Prevalence and risk
factors of peripheral vascular disease in
a selected South Indian population: the
Chennai Urban Population Study. Diabetes
Care 2000; 23:12951300.

5.

Pradeepa R, Chella S, Surendar J, Indulekha


K, Anjana RM, Mohan V. Prevalence of
peripheral vascular disease and its
association with carotid intima-media
thickness and arterial stiffness in type
2 diabetes: the Chennai Urban Rural
Epidemiology Study (CURES 111). Diabetes
Vasc Dis Res 2014; 11:190200.

6.

Cavanagh P, Attinger C, Abbas Z, Bal A,


Rojas N, Xu Z-R. Cost of treating diabetic
foot ulcers in five different countries.
Diabetes Metab Res Rev 2012; 28 Suppl
1:107111.

Conclusion

In conclusion, it can be stated


that pulse oximetry on account of
its high sensitivity will serve as
a good initial screening test for
asymptomatic PVD in patients with
type 2 diabetes; better still would
be a combination of pulse oximetry
and ABI, which has an even higher
sensitivity.

American Diabetes Association. Peripheral


arterial disease in people with diabetes.
Diabetes Care 2003; 26:33333341.

7. Fow k e s F G . Th e m e a s u re m e n t o f
atherosclerotic peripheral arterial disease
in epidemiological surveys. Int J Epidemiol
1988; 17:248254.
8.

9.

Yao ST, Hobbs JT, Irvine WT. Ankle systolic


pressure measurements in arterial disease
affecting the lower extremities. Br J Surg
1969; 56:676679.
Feigelson HS, Criqui MH, Fronek A, Langer
RD, Molgaard CA. Screening for peripheral
arterial disease: the sensitivity, specificity,
and predictive value of noninvasive tests in
a defined population. Am J Epidemiol 1994;
140:526534.

10. Parameswaran GI, Brand K, Dolan J. Pulse


oximetry as a potential screening tool
for lower extremity arterial disease in
asymptomatic patients with diabetes
mellitus. Arch Intern Med 2005; 165:442
446.

43

11. Mardirossian G, Schneider RE. Limitations


of pulse oximetry. Anesth Prog 1992;
39:194196.
12. Kwon J-N, Lee W-B. Utility of digital
pulse oximetry in the screening of lower
extremity arterial disease. J Korean Surg
Soc 2012; 82:94100.
13. Jones SR, Carley S, Harrison M. An
introduction to power and sample size
estimation. Emerg Med J 2003; 20:453458.
14. Chobanian AV, Bakris GL, Black HR,
Cushman WC, Green LA, Izzo JL, et al.
The Seventh Report of the Joint National
Committee on Prevention, Detection,
Evaluation, and Treatment of High Blood
Pressure: the JNC 7 report. JAMA 2003;
289:25602572.
15. Jawahar D, Rachamalla HR, Rafalowski A,
Ilkhani R, Bharathan T, Anandarao N. Pulse
oximetry in the evaluation of peripheral
vascular disease. Angiology 1997; 48:721
724.
16. Writing Committee to Develop Clinical Data
Standards for Peripheral Atherosclerotic
Vascular Disease, Creager MA, Belkin M,
Bluth EI, Casey DE, Chaturvedi S, et al. 2012
ACCF/AHA/ACR/SCAI/SIR/STS/SVM/SVN/
SVS key data elements and definitions
for peripheral atherosclerotic vascular
disease: a report of the American College
of Cardiology Foundation/American
Heart Association Task Force on Clinical
Data Standards ( Writing Committee
to Develop Clinical Data Standards for
Peripheral Atherosclerotic Vascular
Disease). Circulation 2012; 125:395467.
17. Haynes B, Sackett D, Guyatt G, Tugwell P.
Clinical Epidemiology: How to Do Clinical
Practice Research. 3rd edition. Philadelphia:
LWW; 2005. 480 p.
18. Bossuyt PM, Reitsma JB, Bruns DE, Gatsonis
CA, Glasziou PP, Irwig LM, et al. Towards
complete and accurate reporting of studies
of diagnostic accuracy: the STARD initiative.
BMJ 2003; 326:4144.
19. Centers for Disease Control and Prevention
(CDC). Diabetes-related amputations
of lower extremities in the Medicare
population--Minnesota, 1993-1995. Morb
Mortal Wkly Rep 1998; 47:649652.
20. Collins R, Burch J, Cranny G, AguiarIbez R, Craig D, Wright K, et al. Duplex
ultrasonography, magnetic resonance
angiography, and computed tomography
angiography for diagnosis and assessment
of symptomatic, lower limb peripheral
arterial disease: systematic review. BMJ
2007; 334:1257.