Blood Pressure Measurements In

Obese Persons
Comparison of Intra-arterial and Auscultatory Measurements*
KURT BERLINER, M.D., HERLEY FUJIY, M.D., DUK Ho LEE, M.D., MUHTAR YILDIZ, M.D.
and BERNARD GARNIER, M.D.

New York, New YORK

obese, thirty-nine were nonobese. The Ponderal

I
N RECENT YEARS, the relationship of obesity
and arterial hypertension has received con-
siderable attention. It is widely accepted that
arterial hypertension is more common among
the obese than among the lean.lm4 A technical
difficulty in measuring the blood pressure of
obese persons has been recognized ; conven-
tional sphygmomanometry may give falsely
high readings when the arm is big6 and may, on
occasion, result in an unwarranted diagnosis of
hypertension. Earlier investigations of this
problem, consisting of comparisons between
direct intra-arterial measurements and indirect
measurements by the cuff method, either
dealt with too small a number of subjects or
used technics which appear open to criticism
(e.g., measurements not simultaneous or not
made in the same arm). We, therefore, decided
to investigate the problemwith modernelectronic
equipment and a larger number of subjects.
MATERIAL AND METHOD
One hundred subjects were used for this study.
They were drawn from the wards and the outpatient
department of a large general hospital and were suf-
fering from the various diseases encountered in the
hospital population. Patients exhibiting atria1 fibril-
lation were excluded.
The subjects varied in age from sixteen to eighty-four
years (average, fifty-six years). Only three were
twenty-one years old or less; seven were seventy
years old or more. Forty-four were men, fifty-six
were women. Seventy patients were hypertensive,
thirty were normotensive. t Sixty-one patients were
t We regarded a person as hypertensive when his
systolic blood pressure, determined by intra-arterial
measurement, exceeded 150 and his diastolic pressure,
90 mm. Hg.
Index W/H, weight per inch of height, was used to
appraise body bulk and to delimit obesity. A Pon-
deral Index of 2.817 was arbitrarily chosen as the
dividing line. All subjects with a Ponderal Index
higher than 2.817 were classified as obese. The
obese group included fifty hypertensive subjects and
eleven normotensive subjects. The nonobese group
included nineteen hypertensive subjects and twenty
normotensive subjects.
Simultaneous direct and indirect arterial blood
pressure measurements were made in all 100 subjects.
The brachial artery was used for the intra-arterial
measurements, a Sanborn Electromanometer was
used for measuring the pressures and a direct writing
electrocardiograph, Sanborn Twin-Viso, was used for
recording them. Our technic has been described in
detail elsewhere.6 The indirect blood pressure meas-
urement (auscultatory method) was made on the
same arm by means of a commercial sphygmoma-
nometer (Baumanometer, wall type, cuff measurement
13 by 20 cm., with tapered cotton sleeve). The
highest and lowest systolic and diastolic pressures
recorded on each tracing were compared to the final
blood pressure reading obtained with the cuff
method.
RESULTS
Systolic Pressure: Table I is a summary of all
direct and indirect blood pressure measure-
ments. The subjects are arranged according to
their Ponderal Indices. The first thirty-nine
have a Ponderal Index ranging from 1.582 to
2.776; they constitute the nonobese group.
The remaining sixty-one subjects are obese.
Their Ponderal Indices range from 2.818 to
5.625. Discrepancies of varying magnitude are
shown to occur in both groups. Tables II and
III summarize the comparisons of systolic pres-

* From the Department of Medicine, New York Medical College, Metropolitan Medical Center, New York, New
York.

10 THE AMERICAN JOURNAL OF CARDIOLOGY

 Blood Pressure in Obese Persons 11

TABLE I
Summary of All Blood Pressure Determinations : Indirect and Direct Intra-Arterial
- - -

I I Blood Pressures (mm. Hg)

I Direct Intra-arterial
Higher
Indirect Method
1Ponderal Reading
Ci3.W Weight Method
NO. Age (lb.)
Height Index
(W/H)
(Cuff) I- Highest Lowest
Reading Reading

I sys- Dias- sys- Dias- sys- Dias- sys- Dias-

tolic tolic tolic tolic tolic tolic tolic tolic
_- .- --
I-
1 46 106 57 1.582 98 56 95 52 93 49 pff Cuff
2 39 103 53 1.634 96 68 93 47 92 49 cuff
3 21 108 5 51/zV 1.648 93 73 90 67 88 54 Ei Cuff
4 28 105 5lV 1.721 108 70 105 65 95 60 cuff Cuff
5 57 114 5 4 1.781 104 68 105 62 95 57 Direct Cuff
6 53 127 5 IO/* 1.801 126 78 118 63 112 58 Cuff Cuff
7 77 117 5 4 1.828 227 76 233 78 230 78 Direct Direct
8 21 114 52 1.838 128 71 124 85 120 83 cuff Direct
9 44 116/1 52 1.878 133 70 134 70 130 90 Direct Same
IO 37 132 510 1.885 122 90 123 77 117 74 Direct Cuff
11 16 126 56 1.916 127 73 165 78 153 78 Direct Direct
12 40 116 411/2 1.949 133 78 132 84 122 84 cuff Direct
13 84 140 5 9l// 2.015 152 92 187 83 172 80 Direct cuff
14 57 134 56 2.030 130 80 146 74 140 73 Direct cuff
15 60 140 5 8 2.058 172 72 152 63 145 66 cuff cuff
16 65 134 55 2.065 210 100 240 105 228 95 Direct Direct
17 56 135 55 2.076 130 80 132 72 118 72 Direct cuff
18 71 142 5 6 2.151 162 104 194 103 177 117 Direct Cuff
19 80 140 55 2.153 252 60 255 60 240 50 Direct Same
20 45 135 5121/z 2.160 130 80 130 78 120 72 Same Cuff
21 77 137 5 3 2.174 125 80 130 69 123 70 Direct cuff
22 58 124 481/1 2.194 188 90 182 83 176 83 cuff cuff
23 / 51 142 5 41/2b 2.201 164 58 155 50 145 54 cuff Cuff
24 63 148 57 2.208 146 62 173 76 168 79 Direct Direct
25 70 136 511/, 2.211 262 108 248 101 242 101 cuff cuff
26 48 133 50 2.216 245 135 270 130 225 118 Direct Cuff
27 65 145 53 2.301 260 100 260 100 245 95 Same Same
28 82 152 5 51/z 2.320 144 86 155 60 149 68 Direct cuff
29 60 137 4/91/p 2.382 260 109 223 87 215 85 Cuff Cuff
30 69 159 56 2.416 118 68 122 58 112 65 Direct CUE
31 63 162 56 2.454 117 73 112 64 105 63 CUE Cuff
32 64 159 5lN 2.606 160 94 163 78 150 68 Direct cuff
33 61 174 5 51/,# 2.656 152 74 158 72 150 69 Direct Cuff
34 56 187 59 2.710 118 70 132 70 115 70 Direct Same
35 55 185 58 2.720 126 74 138 70 123 70 Direct cuff
36 67 188 59 2.724 170 68 179 66 162 71 Direct Direct
37 60 169 51 2.759 123 68 118 65 111 58 cuff cuff
38 69 184 5 %1/Z 2.766 161 96 179 97 175 92 Direct Direct
39 74 193 5 91/, 2.776 142 78 182 82 158 83 Direct Direct
40 50 186 56 2.818 110 72 118 72 105 67 Direct Same
41 59 181 54 2.828 166 104 151 36 146 95 cuff Cuff
42 74 191 5 71/?.Y 2.829 189 62 165 65 162 65 cuff Direct
43 65 162 49 2.842 196 60 201 55 176 62 Direct cuff
44 50 185 54 2.872 200 130 198 100 170 90 Cuff cuff
45 50 171 411/2 2.873 202 100 190 88 186 89 cuff cuff
46 38 19B1/2 59 2.876 178 100 150 85 139 84 Cuff Cuff
47 68 201I? 510 2.878 111 71 97 63 90 59 Cuff cuff
48 64 190 56 2.878 131 76 153 83 145 67 Direct Direct
49 67 190 56 2.878 198 72 215 70 205 70 Direct cuff
50 48 183 4,91/?v 2.950 235 120 241 128 222 123 Direct Direct
- - - -I ,I
-
JULY 1961
12 Berliner et al.

TABLE I (Continued)
-
Blood Pressures (mm. Hg)

Direct Intra-arterial
Higher
Indirect Method
Ponderal Reading
Case We@ Method
Age Height Index
No. (lb.) (Cuff) Highest Lowest
(W/H)
Reading Reading

sys- Dias- sys- Dias- sys- Dias- sys- Dias-

tolic tolic tolic tolic tolic tolic tolic tolic

51 49 183 491/1# 2.950 210 114 212 108 209 107 Direct cuff
52 60 198 57 2.955 192 132 200 125 192 120 Direct CUff
53 54 190 5 4 2.968 103 72 120 58 112 55 Direct cuff
54 46 196 5 6 2.969 150 75 159 83 151 82 Direct Direct
55 57 200 57 2.985 166 110 168 112 159 106 Direct Direct
56 69 200 57 2.985 176 94 170 98 166 69 cuff Direct
57 50 212 510 3.028 140 78 143 72 132 70 Direct cuff
58 63 200 56 3.030 126 62 121 54 116 55 cuff cuff
59 51 176 410 3.034 182 90 195 100 185 95 Direct Direct
60 65 213 510 3.042 142 60 124 56 120 56 Cuff cuff
61 48 200 54 3.125 160 90 170 85 142 88 Direct Cuff
62 66 217 59 3.144 147 40 152 43 150 45 Direct Direct
63 69 1921/ 51 3.155 185 95 185 95 180 85 Same Same
64 65 193 51 3.163 206 130 195 113 181 110 Cuff Cuff
65 49 197 5 2 3.177 202 103 206 97 201 97 Direct cuff
66 68 183 4 9/* 3.182 128 90 150 68 135 64 Direct cuff
67 56 202 5 2 3.258 193 94 174 68 165 75 Cuff cuff
68 51 207 5 21/2X 3.312 186 108 200 122 190 120 Direct Direct
69 44 220 55 3.384 200 140 180 140 170 136 Cuff Same
70 36 220 55W 3.384 148 111 154 97 143 95 Direct cuff
71 60 210 5 2# 3.387 141 50 159 65 138 57 Direct Direct
72 60 225 5 l/s 3.488 222 126 223 110 214 111 Direct Cuff
73 59 232 56 3.515 122 88 126 80 120 80 Direct Cuff
74 55 239 57 3.567 192 115 198 110 195 110 Direct cuff
75 53 226 53 3.587 210 110 220 115 210 105 Direct Direct
76 59 216 50 3.600 154 90 180 88 168 83 Direct cuff
77 57 224 52 3.612 158 104 176 120 164 119 Direct Direct
78 73 212 410/1 3.623 146 92 156 101 136 91 Direct Direct
79 62 227 52 3.632 168 99 173 120 152 109 Direct Direct
80 50 221 411 3.745 254 142 242 123 229 117 cuff cuff
81 45 244 5 5 3.760 164 80 173 82 162 78 Direct Direct
82 34 248 5 5/2 3.786 133 77 137 73 122 73 Direct cuff
83 57 231 51 3.786 120 64 131 82 120 75 Direct Direct
84 58 238 50 3.966 188 108 180 96 170 95 Cuff Cuff
85 27 265 55 4.076 126 84 118 78 116 81 cuff Cuff
86 65 265 55 4.076 125 76 129 78 119 74 Direct Direct
87 51 264 53 4.190 160 90 176 83 166 83 Direct cuff
88 36 290 59R 4.202 184 123 201 122 185 111 Direct Cuff
89 71 237 48 4.232 188 80 205 87 190 82 Direct Direct
90 65 255 50 4.250 238 130 230 118 218 124 Cuff cuff
91 59 274 5 4 4.281 170 92 152 95 138 88 cuff Direct
92 38 270 5 3 4.285 156 102 159 94 154 97 Direct cuff
93 62 275 51 4.524 190 90 171 83 137 86 cuff cuff
94 63 275 51 4.524 196 94 152 66 134 61 cuff cuff
95 48 287 5 3 4.555 262 138 212 110 200 95 cuff cuff
96 38 278 4 11/2 4.672 178 100 150 85 139 84 cuff cuff
97 48 276 411 4.677 236 150 208 123 205 125 cuff cuff
98 58 315 5 ,I/%. 5.121 141 88 135 67 120 66 cuff cuff
99 49 350 57 5.223 250 150 199 115 178 104 cuff cuff
100 58 391 591/2ff 5 625 159 116 162 91 140 87 Direct cuff
I
THE AMERICAN JOURNAL OF CARDIOLOGY
 Blood Pressure in Obese Persons 13

TABLE II TABLE III

Systolic Blood Pressure Readings-Comparison of Direct Systolic Blood Pressure Readings-Extent of Discrep-
and Indirect Measurements ancies between Direct and Indirect Measurements

Discrepancy

Reading

A. Sirfy-One Obese Personr

? - -
Direct
ltlLZiSX-
j
ment
higher 10 9 3 0 36
mm. disregarded 33 23 i 5 61 Indirect
Dikerences of 1 or 2 i xneasure-
me,
B. 7hirty-Nine Nonobcse Pmom higher 1 10 4 3 24

Actual
pared
readings

Differences
com-

of 1 or 2
23 ~ 14 2 39
Readings
identical

El. Thirty-Nine
L
Nmobere Persons
I 1

mm. disregarded 19 13 7 39

sure in the obese and nonobese persons. Table

Direct
measure-
ment
higher 3 0
I
/ 1 14
II shows the frequency of the discrepancies, Indirect

Table III their magnitude. Most discrepancies XleaStY- I I I I I ; i

lent
in the obese persons were minor; in three sub- higher 4 3 4 23
Readings
jects, however, the indirect method grossly identical 2
underestimated systolic pressure by 22 to 26
mm., and in seven subjects the indirect method
grossly overestimated systolic pressure by 24, TABLE IV

28, 28, 28, 44, 50 and 51 mm., respectively. Diastolic Blood Pressure Readings-Comparison of
Direct and Indirect Measurements
In the nonobese persons, too, most of the dis-
crepancies were minor. However, gross over-
Comparison
estimation of systolic pressure also occasionally
occurred among these lean persons, in one in- Reading
Direct Indirect / Readigs
i 2:
stance amounting to 37 mm. Underestima- Reading Reading
tion of systolic pressure of lean persons by the Higher Higher i Identical
___=
indirect method was more common and more
A. Sixty-One Obese Pmons
marked, amounting to as much as 25 to 40 mm.
Diastotic Pwssure: In both groups systolic / I I
Actual readings com-
pressure was underestimated by the indirect pared 19 39 3 61
method more often that it was overestimated. Differences of 1 or 2
mm. disregarded 16 36 9 ! 61
The opposite was found when diastolic pressures
were compared. The indirect method over-
B. Thirty-Nine Nonobcre Persons
estimated diastolic pressure in the majority of
/ / ,
all cases, obese and nonobese. Tables IV and Actual readings corn
v summarize the comparisons of diastolic pres- pared 9 26 4 39
Differences of 1 or Z
sure in the obese and nonobese subjects, Table mm. disregarded 23 9 39
IV showing the frequency of the discrepancies,
Table v their magnitude. Among the obese
subjects, most of the discrepancies in diastolic pressure did not occur; gross overestimation
pressure were again minor. Gross underestima- occurred four times, amounting to 16 to 22 mm.
tion occurred only twice (by 16 and 18 mm.) Extreme Obesity us. Leanness: The similarity of
and gross overestimation much more frequently, findings in both groups at first seemed to indi-
twelve times (by 16 to 35 mm.) In the non- cate that obesity has little effect on the meas-
obese group gross underestimation of diastolic urement of blood pressure. Very different re-

JULY 1961
 Berliner et al.

TABLE v very stout subiects also was overestimated in

Diastolic Blood Pressure Readings-Extent of Discrep- nine out of ten cases, and by a substantial
ancies between Direct and Indirect Measurements
number of points. These findings were in
striking contrast to those in the ten leanest sub
jects. Here no significant overestimation or
underestimation of either systolic or diastolic
pressure occurred.
Lastly, we selected from both groups the twelve
cases which demonstrated the greatest over-

8
- estimation of systolic blood pressure by the
indirect method (Table VII). The discrepan-

T
Direct

cies ranged from 18 to 51 mm. Ten of the

measure
ment
higher 3 5 6 3 2 0 19 twelve subjects were obese and the five greatest
Indirect
Ul.?ZiSX. discrepancies of the entire series occurred in ex-
ment tremely obese persons. Here the effect of
higher 3 5 11 8 4 39
Readings obesity, falsely high readings, is most apparent.
identical . . . . . . . . . . 3
- COMMENT
B. Thirty-Nine Nonobcrc Persons
Effect of Obesity on Blood Pressure: From the out-
I I I I I I I set two separate problems presented themselves;
(1) the effect of obesity on blood pressure; (2)
errors of clinical blood pressure measurements
4 2 0 0 9
caused by the large arms of obese persons. As
to the first problem, any investigation into the
Inent
higher 3 6 10 3 3 1 26 effect of obesity on blood pressure obviously has
Readings
identical . 4
to find a true measure of obesity first. Meas-
I I urement of total body fatness can probably fur-
nish the best measure of obesity, but the
sults, however, were obtained when extreme methods are not yet practicable.8 Another
obesity was contrasted with marked leanness. method of assessment is to measure the circum-
Table VI limits the comparisons to the ten ference of the subjects arm. A recent study by
stoutest subjects (Ponderal Index 4.250 to Whyte,r however, clearly shows that circum-
5.625) and the ten leanest subjects (Ponderal ference of the arm and blood pressure are unre-
Index 1.582 to 1.885). In eight of the ten very lated. Blood pressure, measured in the ordi-
obese subjects, the indirect method overesti- nary way, was noted to be higher in subjects
mated the systolic pressure. The greatest dis- with larger arms, but the association lost
crepancies of the whole series were encountered significance when the influence of other factors
in this small group. Diastolic pressure in these such as age, weight and height was excluded.

TABLE VI
Ten Stoutest and Ten Leanest Subjects-Comparison of Direct and Indirect Blood Pressure Measurements

10 Stoutest Subjects 10 Leanest Subjects

Blood
Pressure Readings
Measure- Direct Indirect Direct Indirect Identical
ment Measurement Measurement Measurement Measurement
Higher Higher Higher Higher

Systolic 2 cases (each 8 cases (by 6, 18, 19, 4 cases (by 1, 1,l and 6 cases (by 3, 3, 3, 3, 4 0
oressure bv3mm.) 28,28,44,50 and 51 6 mm., respectively) and 8 mm., respec-
mm., respectively) tively)
9 cases (by 4,5,15,15, 2 cases (by 2 and 6 7 cases (by 4, 5, G, 6, 13, 1
21, 25, 28, 28 and mm., respectively) 15 and 19 mm., re-
35 mm.. resoec- spectively)
1 tively) A

THE AMERICAN JOURNAL OF CARDIOLOGY

 Blood Pressure in Obese Persons 15

TABLE VII pertension. This relationship is still better

Twelve Greatest Overestimations of Systolic shown if only the leanest and most stout are
Blood Pressure by the Indirect Method compared. Of the ten leanest patients in our
- group only one was hypertensive; nine were
T Systolic Pressure (mm. Hg)
normotensive. Of the ten stoutest patients, on
Cae.e vVeight Height I onderal the other hand, nine were hypertensive and only
NO. (lb.1 Index Systolic
W/H)
Direct Indirect
Pressure one was normotensive.
1%fC3SUX. MGASW
ment ment
Higher Errors in Blood Pressure Measurements in Obese
- (mm.)
.I-
Persons: The possible errors of clinical blood
60 213 510 3.042 124156 t42/60 18
pressure measurement in obese persons are, of
91 274 5 4 4.281 152/95 170/92 18
67 202 5 2 3.258 174160 193/94 19 course, an entirely different problem. These
15 140 5 8 2.058 152/63 172/72 20
69 220 5Y 3.384 180/140 200/l 40 20
errors have been widely publicized.7-1s Sev-
42 191 5 F/r* 2.829 165/65 189/62 24 eral authors5J7J8 have reported falsely high
96 278 41 l/P 4.672 150/85 178/100 28
97 276 411 4.677 208/123 236/l 50 28 readings in the obese arm and have given the
29 137 49 2.382 223/87 260/109 37 impression that an unwarranted diagnosis of
94 275 51 4.524 152/66 196/94 44
95 287 5 3 4.555 212/110 262/138 50 hypertension is a common occurrence. This
99 350 5 7U 5.223 199/115 250/150 51
problem seemed to us of enough importance to
- - - -
require further investigation. Does the obese
XOTE: Ten subjects were obese; two subjects were nonobese
arm regularly cause falsely high readings, or, if
not, how common are they? What is the extent
Still another way of judging obesity is by esti- of these errors, and are they proportionate to
mating the amount of subcutaneous fat. This the degree of obesity? Are many obese persons
is done by pinching up skin folds and measuring falsely classified as hypertensives? These were
their thickness. Whyte found that obesity thus the questions to which we sought the answers.
judged also has no apparent influence on blood We first reviewed the earlier investigations of
pressure execpt as it affects the total body the problem. Ragan and Bordley5 were the
weight. Whyte noted that the composition of a first to demonstrate the error incurred by taking
persons excess weight is immaterial ; it is the the blood pressure in obese persons. Using
over-all bulk which counts, be it muscle or fat. fifty-one subjects, they compared blood pres-
In our own investigation we, therefore, dispensed sure measurements made by the auscultatory
with measurements of circumference of the arm method in the right arm and simultaneous
and estimations of subcutaneous fat. Our task direct intra-arterial measurements, made with a
was to measure over-all bulk. Weight alone is Hamilton manometer, in the left arm. They
obviously not very informative about the bulk of found that in subjects with large arms the
the body, the bulk being dependent on height as auscultatory readings were usually too high, the
well as on other dimensions of the body. None errors amounting to between 10 and 20 mm. Hg.
of the available methods of measuring bulk is Trout, Bertrand and Williams*s made com-
entirely reliable. The Ponderal Index W/H, parisons on six extremely obese persons and
weight per inch of height, however, is in our eight nonobese persons (weights not stated).
opinion the simplest and best comparative Overestimation of systolic pressure by the in-
measure of body bulk and was employed in the direct method, varying from 32 to 102 mm. Hg,
present study. occurred in all six obese persons; in the non-
It has been shown by other investigators that obese persons there was a fair degree of agree-
obesity- is statistically associated with blood ment between the two methods. Observations
pressure levels above average, and that weight made by Irvin, quoted by Smirk,4 on eight
reduction by dieting may produce considerable patients with unequal arms showed slight
falls of arterial pressure when it is initially differences between the two on numerous
high.r0st5 Our series is relatively small and our repetitions, the higher reading being obtained
hospital patients do not represent a true cross on the arm with the larger circumference.
section of the entire population. Our figures, Smirk4 states : If an arm is artificially thickened
82 per cent hypertensive subjects in the obese by wrapping around it various materials such as
groups, 49 per cent hypertensive subjects in the cloth, cotton, wool, sponge rubber and meat
nonobese group, therefore, have no statistical From the butcher, it is found that these materials
significance although they effectively demon- have different effects upon the estimated blood
strate the relationship between obesity and hy- pressure. Cloth and meat of a rather compact

JULY 961
16 Berliner et al.

consistency had very little effect upon the esti- infrequently the sleeve is too short and part of
mated blood pressure, whereas cotton, wool and the cuff balloons out when it is inflated. After
sponge rubber increased the estimated blood readjusting the sleeve, markedly different
pressure. Apparently some materials when readings were obtained. Attempts to lengthen
interposed between the sphygmotnanometer the sleeve by adding a bandage caused even
cuff and the artery spread the pressure over a greater errors.
wider area so that, in fact, the pressure upon the We always bear in mind the possibility of
outside of the artery is less than the pressure falsely high readings when dealing with ex-
within the cuff. Hence the blood pressure is tremely stout persons and are, therefore, not
overestimated. The estimated pressure can be likely to make an unwarranted diagnosis of hy-
brought back to a pressure nearer to that ob- pertension. Moreover, most of these very obese
tained under standard conditions if the cuff is persons are true hypertensive subjects anyhow,
widened. This explanation of the falsely high as shown by intra-arterial measurement; there
readings is plausible and should be accepted is little danger of falsely classifying them as
without qualification. Pickering, Roberts and hypertensive subjects. In the rare cases in which
SowryzO reviewed the comparisons reported by there is doubt, intra-arterial measurement
Ragan and Bordley5 and published a list of should be taken.
adjustments for arm circumference to be made The large size of the arm, on the other hand,
to auscultatory pressure. The subtractions to does not always cause falsely high readings.
be made from systolic pressure readings in This was nicely shown by the stoutest subject
stout arms range from 5 to 25 mm. and those of our entire series, a woman weighing 391
from diastolic readings also from 5 to 25 mm. pounds. Her systolic pressure was underesti-
These authors admit that variations in circum- mated by the auscultatory method!
ference of the arm do not account for more than
a quarter of the differences between direct and SUMMARY

indirect readings and doubt the usefulness 1. Simultaneous direct intra-arterial blood
of applying a correction for arm circumference pressure measurements and indirect measure-
to an individual reading. ments with the auscultatory method were made
Our own findings lead us to the same con- in 100 subjects. The subjects were grouped
clusion. Besides, marked overestimation of according to their Ponderal Index W/H, weight
blood pressure by the auscultatory method, in per inch of height. Sixty-one subjects were
our series, was not at all limited to obese sub- obese, thirty-nine were nonobese.
jects, but occurred in lean persons also. Most 2. Nine out of the ten stoutest subjects were
noteworthy, however, is our finding of falsely hypertensive, whereas only one of the ten leanest
high readings in eight out of ten extremely subjects was hypertensive.
obese persons. The overestimations ranged 3. The indirect method generally under-
from 6 to 51 mm. and were greater than the estimates systolic and overestimates diastolic
general range of 10 to 30 mm. reported in the pressure. This holds for lean persons as well as
literature.5*17 Only Trout and associate@ for moderately obese persons.
found discrepancies greater than ours; they 4. The discrepancies between the direct and
ranged from 32 to 102 mm. Hg, with the two indirect methods were usually minor among the
greatest, 80 and 102 mm., occurring in the two lean and the moderately obese subjects and
stoutest arms. went in both directions. Falsely high blood
The accuracy of clinical blood pressure pressure readings in lean and moderately obese
measurement in general depends on many persons are not common.
factors such as age of the subject, height of true 5. In extremely obese persons the indirect
blood pressure, rigidity of arterial wall, contour method is subject to considerable error. In
of pulse wave,22 and sphygmomanometric eight out of ten such persons the systolic pres-
technic6 Obesity is but one such factor and sure was overestimated, and in nine out of ten
only in extremely stout subjects can it become such persons the diastolic pressure was over-
decisive. It is the large size of the arm which is estimated.
responsible for the error. In such persons we 6. Overestimation of systolic pressure in very
found that the difficulty of applying the cuff also obese persons can be much greater than was
plays a part. Often it is difficult to wrap the heretofore known; it amounted to 44, 50 and 51
cotton sleeve evenly around the big arm; not mm. in three of our stoutest subjects.

THE AMERICAN JOURNAL OF CARDIOLOGY

 Blood Pressure in Obese Persons 17

7. Falsely high readings in extremely obese 8. KEYS, A. and BROZEK, J. Body fat in adult man.
Physiol. Rev., 33: 245, 1953.
subjects are caused by the large size of their
9. DUBLIN, L. J. Relation of obesity to longevity.
arms. If this possibility is borne in mind, an New England J. Med., 248: 971, 1953.
unwarranted diagnosis of hypertension is not 10. ADLERSBERG, D., COIZR, A. R. and LAVAL, J.
likely to be made. In dubious cases, intra- Effect of weight reduction on course of arterial
arterial measurement should be resorted to. hypertension. J. Mt. Sinai Hosp., 12 : 984, 1946.
Il. NUZUM, F. R. and ELLIOT, A. H. An analysis of
8. Many factors other than arm size may
500 instances of arterial hypertension. Am. J.
cause inaccuracies of the indirect method, and M. SC., 181: 630, 1931.
the large arm does not regularly cause falsely 12. ROBINSON,S. C., BRUCES,M. and MASS, J. Hyper-
high readings. It is not feasible to apply a cor- tension and obesity. J. Lab. & Clin. Med., 25:
807, 1940.
rection to an individual reading.
13. TERRY, A. H., JR. Obesity and hypertension. J. A.
M. A., 81: 1283, 1923.
REFERENCES 14. PREBLE, W. E. Obesity: observations in one thou-
I. WHYTE, H. M. Blood pressure and obesity. Cir- sand cases. Boston tied. Surg. J., 188: 617, 1923.
culation, 19: 511, 1959. 15. FLETCHER, A. P. The effect of weight reduction
2. WHYTE, H. M., GRAHAM,J. A. D. and DE WOLFE, upon the blood pressure of obese hypertensive
M. S. Body fat, blood pressure, and serum women. Quart. J. Med., 23: 331, 1954.
cholesterol of Australian men. Australasian Ann. 16. BOE, J., HUMMERFELT, S. and WEDERVANG, F.
Med., 7: 328, 1958. The blood pressure in a population. Acta med.
3. WHYTP, H. M. Body fat and blood pressure of scandinav., suppl., 321, 1957.
natives in New Guinea. Australasian Ann. Med., 17. PERERA, G. A. Primary hypertension. Cmulation,
7: 36, 1958. 13: 321, 1956.
4. SMIRK, F. H. High Arterial Blood Pressure, p. 7. 18. TROUT, K. W., BERTRAND, C. A. and WILI.IAMS,
Springfield, Illinois, 1957. Charles C Thomas. M. H. Measurement of blood pressure in obese
5. RAGAN, C. and BORDLEY, J., III. The accuracy of persons. J. A. M. A., 162: 970, 1956.
clinical measurements of arterial blood pressure. 19. PICKERING, G. W. High blood pressure, p. 20.
Bull. Johns Hopktns Hosp., 69: 504, 1941. New York, 1955. Grune & Stratton.
6. BERLINER, K., FUJIY, H., LEE, D. H., YILDIZ, M. 20. PICKERING,G. W., ROBERTS, J. A. F. and SOWRY,
and GARNIER, B. The accuracy of blood pressure G. S. C. The effect of correcting for arm circum-
determinations. A comparison of dir&t and ference on the growth rate of arterial pressure
indirect measurements. Cardiolrxia.
u ,
37: 118. with age. C/in. SC., 13 : 267, 1954.
1960. 21. LACHMA~N, H. Abhaengigkeit der Blucdruckhoehe
7. BORDLEY, J., III, CONNOR, C. A., HAMILTON, W. F., von der Haerte der Manschette. Deutsches Gemnd-
KERR, W. J. and WIGGERS, C. J. Recommenda- heitswesen, 13: 59, 1958.
tions for human blood pressure determinations 22. THOMPSON,A. E. and DOUPE, J. Causes of error in
by sphygmomanometers. Circulation, 4: 503, auscultatory blood pressure measurements. Rev.
1951. canadienne biologic, 8 : 337, 1949.

JULY 1961