ORIGINAL ARTICLE doi: 10.1111/j.1463–1326.2004.00406.

Abdominal diameter index: a more powerful anthropometric

measure for prevalent coronary heart disease risk in adult males

D. A. Smith1,2, E. M. Ness3, R. Herbert1, C. B. Schechter4, R. A. Phillips5, J. A. Diamond6

and P. J. Landrigan1
1
Community and Preventive Medicine, Mount Sinai School of Medicine, New York, NY, USA
2
The Zena and Michael A. Wiener Cardiovascular Institute, Mount Sinai Medical Center, New York, NY, USA
3
Department of Medicine, Vanderbilt University Medical Center, Nashville, TN, USA
4
Department of Family Medicine and Community Health, Albert Einstein College of Medicine, Bronx, NY, USA
5
Department of Medicine, Lenox Hill Hospital, New York, NY, USA
6
Harris Chasanoff Heart Institute, Long Island Jewish Medical Center, New Hyde Park, NY, USA

Aim: The authors wished to compare the strength of association of several anthropometric measures of body size and
fat distribution among themselves and in comparison with other known risk factors for prevalent coronary heart
disease (CHD).
Methods: Prevalent CHD was assessed in 466 middle-aged, male, multiracial Triborough Bridge and Tunnel
Authority officers in New York City by verified history, electrocardiogram or exercise stress test. Anthropometric
measures included body mass index, waist, hip and thigh circumferences, waist-hip ratio, waist-thigh ratio, sagittal
abdominal diameter and abdominal diameter index (sagittal abdominal diameter/thigh circumference). Results were
compared with other CHD risk factors measured simultaneously (history of diabetes, smoking, blood pressure, lipid
profile, apolipoproteins A and B, lipoprotein (a), homocysteine, fibrinogen, urinary microalbumin, serum vitamin E
and ferritin) and a calculated 10-year CHD risk using a Framingham algorithm (10-year Framingham CHD risk).
Results: CHD was found in 29 individuals. Of the six anthropometric measures, abdominal diameter index gave the
largest and most significant standardized odds ratio (OR) for CHD [1.80, 95% confidence interval (CI) 1.20, 2.71],
equivalent to 10-year Framingham CHD risk. Men in the highest compared with the lowest tertile of abdominal
diameter index had a univariate OR of 5.47 (95% CI 1.55, 19.28) which was the only anthropometric measure that
remained significant after adjusting for 10-year Framingham CHD risk.
Conclusions: For middle-aged American men, abdominal diameter index may be the most powerful anthropometric
measure of risk for prevalent CHD.
Keywords: abdominal diameter index, anthropometry, body weight, coronary heart disease, risk assessment, visceral obesity
Received 15 January 2004; returned for revision 20 April 2004; revised version accepted 12 May 2004

Introduction
Increased abdominal adiposity is clearly associated with
increased risk for cardiovascular disease and premature
death [1–11]. Intra-abdominal obesity is also associated
with the metabolic abnormalities of the insulin resis-
tance syndrome [11–19]. Central obesity may be linked
through this pathway to elevated risk of coronary heart
disease (CHD).

Correspondence:
Donald A. Smith, MD, MPH, Mount Sinai Medical Center – Box 1014, 1 Gustave Levy Place, New York, NY 10029-6574, USA.
E-mail:
donald.smith@mssm.edu

370 Diabetes, Obesity and Metabolism, 7, 2005, 370–380 # 2004 Blackwell Publishing Ltd
D. A. Smith et al.
The best measurement of intra-abdominal adiposity is
attained by computed axial tomographic scanning or
magnetic resonance imaging [20–23]. However, the
high costs of these tests make them clinically impracti-
cal, and their size and lack of portability make them
unavailable for epidemiologic studies in many parts of
the world. The waist size or waist-to-hip ratio (WHR),
simple anthropometric measures, may be the most
widely used clinical estimates of central adiposity. Ele-
vation of WHR has been clearly associated with cardio-
vascular disease risk [1–8,12] and the insulin resistance
syndrome [14]. Other simple anthropometric measure-
ments, such as sagittal abdominal diameter and the ratio
of sagittal abdominal diameter to height, have also
shown similar or superior correlations with metabolic
cardiovascular risk profiles than WHR [14,24] and can
predict CHD mortality [11]. Kahn [25] has proposed that
the abdominal diameter index, the ratio of supine
sagittal abdominal diameter to mid-thigh circumference
may be more effective at estimating visceral adipose
tissue and, hence, a better predictor of CHD than WHR.
To determine which simple anthropometric measures
of total body size or of regional fat distribution are most
strongly associated with prevalent CHD and to compare
the associative strength of these measures to other
known and postulated risk factors for CHD, we utilized
data from a cross-sectional epidemiologic study in a
cohort of male bridge and tunnel officers in New York
City in 1994 whose primary purpose was examining the
relationship of work history and prevalent CHD. We
have previously reported [26] in this population
(which in that analysis included women) that CHD risk
was positively associated with total years of work as an
officer [odds ratio (OR) ¼ 1.98, 95% confidence intervals
(CI) 1.49, 2.71], and that risk increased for each decade
of employment adjusted for non-occupational CHD risk
factors.
Methods
All bridge and tunnel officers working for the New York
City Triborough Bridge and Tunnel Authority and those
who had retired since 1985 were invited to participate in
a comprehensive two-step medical evaluation empha-
sizing cardiovascular assessment approved by the Insti-
tutional Review Board of Mount Sinai Medical School.
Subjects gave written informed consent. A self-administered
questionnaire containing demographic information,
family history, occupational and medical history was
completed by each participant prior to the baseline
evaluation. At the baseline evaluation, a standard ques-
tionnaire was administered by the examining physician,
# 2004 Blackwell Publishing Ltd
Abdominal diameter index and CHO OA
which asked about prior evaluation or treatment of CHD
and other cardiovascular disease. The Rose question-
naire [27] for exertional angina, heart attack, stroke
and peripheral vascular disease was administered.
Participants were also questioned about history of
physician-diagnosed hypercholesterolemia, diabetes and
hypertension.
A standardized comprehensive physical examination,
including three serial, random-zero, sitting blood pres-
sures averaging the last two, was performed. Male bald-
ness pattern was coded according to the Hamilton Index
[28], which has previously been used to show that vertex
baldness, but not frontal baldness, is a risk factor for
CHD. Height and weight were measured without shoes
in a medical gown, height using a stadiometer and
weight using a standard physician’s balanced scale.
Abdominal obesity was measured in several ways. A
supine WHR was measured with the patient lying flat on
the examining table, the waist circumference measured
halfway between the bottom rib and the iliac crest in the
mid axillary line and hips measured at the level of the
greater trochanters. Sagittal abdominal diameter was
measured as the height of the abdomen from the table
in the supine position (cm) measured at the waist, first
using a chest caliper used by radiologists to determine
anterior-posterior chest thickness (S & S X-ray Products,
Inc., Brooklyn, NY, USA) and then when it became
available, using a portable sliding-beam caliper (Holtain
Ltd, Dyfed, Wales, UK) with the patient supine and after
gentle expiration. The caliper’s upper arm was brought
down to an abdominal mark made midway between the
iliac crests, which approximates the level of the L4-L5
interspace (figure 1) [10]. In a subset of 48 subjects, sagit-
tal abdominal diameter was measured with both the
chest caliper and the Holtain portable sliding-beam cali-
per. Abdominal diameter index was calculated by divid-
ing the sagittal abdominal diameter by the thigh
circumference measured 10 cm proximal to the superior
patellar border.
A resting electrocardiogram (ECG) was performed by a
technician trained by the Minnesota ECG Coding Center
and was sent to that centre for standardized coding [29].
After a minimum of four hours of fasting, blood was
drawn for baseline tests; the blood was centrifuged and
sent to MetPath Diagnostic Laboratory, Teterboro, NJ,
USA, for laboratory analyses. Baseline tests included
lipid profile [total, high-density lipoprotein (HDL) chol-
esterol and low-density lipoprotein (LDL) cholesterol,
triacylglycerol], lipoprotein (a), homocysteine, fibrino-
gen, apolipoproteins A and B, urinary microalbumin,
fibrinogen and vitamin E. All lipids were done enzyma-
tically using an automated blood analyser (Hitachi,
Diabetes, Obesity and Metabolism, 7, 2005, 370–380 371
OA Abdominal diameter index and CHO
Fig. 1. Measurement of the sagittal abdominal diameter in a
supine subject. (Permission to use this figure must be
obtained from the original article: Kahn HS et al. Simple
anthropometric indices associated with ischemic heart dis-
ease. J Clin Epidemiol 1996; 49: 1017–1024).
model 736). HDL cholesterol was determined by phos-
photungstic precipitation with enzymatic quantification
by spectrophotometry. LDL cholesterol was calculated
using the Friedewald equation [30]. Apolipoproteins
A and B, lipoprotein (a), fibrinogen and urinary micro-
albumin were measured by a nephelometric immuno-
Table 1 Classification of definite coronary heart disease (CHD) cases
Verifiable history of treatment for myocardial infarction (MI)
History of CABG1
History of PTCA2
Abnormal coronary angiogram
Definite MI on baseline ECG
Positive exercise stress test (EST) with:
Verifiable history of positive thallium EST
Previous hospitalization for possible MI
Previous history suspicious angina by Rose questionnaire
Previous history possible MI by Rose questionnaire
Asymptomatic, negative CHD history
Total
1
CABG, coronary artery bypass graft.
2
PTCA, percutaneous transluminal coronary angioplasty.
372 Diabetes, Obesity and Metabolism, 7, 2005, 370–380
D. A. Smith et al.
assay. Serum homocysteine and vitamin E were
performed using high-pressure liquid chromatography.
Ferritin measurement was assessed by a chemiluminescent
immunoassay method.
A cardiovascular risk assessment [31] was calculated
for each subject utilizing an algorithm derived from the
Framingham Study. A 10-year probability of developing
CHD (10-year Framingham CHD risk) was derived from a
logistic regression equation using the following vari-
ables: age, sex, systolic blood pressure, total cholesterol,
HDL cholesterol, history of diabetes and history of cur-
rent cigarette smoking. Consultation with a cardiologist
and exercise stress testing were offered to all stage 1
subjects who were found to have a suggestive history of
possible atherosclerotic cardiovascular disease or chest
pain or to have three or more standard cardiovascular
risk factors. Standardized treadmill exercise protocols
were performed under supervision of the program’s car-
diologist, using Bruce or modified Bruce protocols [32]
and were interpreted by one of two cardiologists. Def-
inite CHD was defined as in table 1. Because there were
no prevalent cases of CHD in the female population
(n ¼ 60), only the male population was used in these
analyses (n ¼ 466).
Participants filled out an occupational history which
included average number of hours worked per week in
either bridge or tunnel location per year for each year
they had worked up until that time. Full-time equivalent
work years were calculated for each participant by
n Definition
1 Diagnostic electrocardiogram (ECG), or
classic chest pain and diagnostic
enzymes, or classic chest pain and
equivocal enzymes and equivocal ECG.
2
2
4
50% occlusion of 1 vessel or more.
4 Presence of significant Q waves
(Minnesota Codes: 1-1-x or 1-2-x except
1-2-6 or 1-2-8).
Horizontal or downsloping ST segment
1 mm
at the J point or upsloping ST segment
depression
2 mm at 80 ms after the J point.
1
1
4
1
9
29
# 2004 Blackwell Publishing Ltd
D. A. Smith et al. Abdominal diameter index and CHO OA

dividing total work hours by 1920 h per full-time equiva-
lent year (40 h/week  48 weeks ¼ 1920 h).
Demographic, biochemical and anthropometric vari-
ables were compared between those with CHD (n ¼ 29)
and those without (n ¼ 437). In order to compare ORs for
CHD among many different risk variables, each variable
was divided by the standard deviation of that risk vari-
able to give a z-value for each variable for each partici-
pant. Z-values were then entered into a logistic
regression model for predicting prevalent CHD to give a
univariate OR for each variable. In order to see if certain
variables added any additional risk to those standard
variables used in the 10-year Framingham CHD risk
equation, each variable was entered along with the
10-year Framingham CHD risk variable to give a
Framingham-score-adjusted OR. Logarithmic transfor-
mations were performed with re-analyses for skewed
variables such as 10-year Framingham CHD risk, full-
time-equivalent working years, triacylglycerol, lipopro-
tein (a), homocysteine, fibrinogen and microalbumin
concentrations. The re-analyses confirmed and did not
appreciably differ from the untransformed analyses
which are the ones presented.
In order to further compare the predictive power of the
anthropometric variables of body mass index (BMI),
waist, WHR, waist-thigh ratio, sagittal abdominal dia-
meter, abdominal diameter index and thigh, each were
divided into tertiles and ORs (both univariate and
adjusted for 10-year Framingham CHD risk) were calcu-
lated by logistic regression for prevalent CHD for each of
the two upper tertiles relative to the lowest tertile.
In order to explore the contribution of single variables
in the Framingham CHD risk equation and other vari-
ables associated with insulin resistance but not included
in the risk equation to the predictive power of abdominal
diameter index, logistic regression was used to calculate
an OR of the highest compared to the lowest tertile of
abdominal diameter index for predicting prevalent CHD.
This ratio was then adjusted for age, BMI as a measure of
total body fat and current smoking. Four variables asso-
ciated with insulin resistance (serum triacylglycerol,
history of diabetes, HDL cholesterol and average systolic
blood pressure) were then competitively forced into the
model in a forward stepwise manner to see which vari-
ables contributed most to the overall OR for abdominal
diameter index and CHD. SPSS for Windows, Version 8.0
(SPSS Inc., Chicago, IL, USA) was used for all analyses.
Results
A total of 466 male subjects participated in the study, 29
with CHD as defined in table 1: one age 30–39, 15 age
40–49, seven age 50–59 and six age
60 years. Demo-
graphics and prevalence of hypertension, diabetes and
smoking status in the entire group and in those with and
without CHD are given in table 2. The mean age of this
ethnically mixed population was 44.8 years with a his-
tory of hypertension in 29%, of diabetes in 6.7% and of

Table 2 Demographics

CHD1 (n ¼ 29) No CHD (n ¼ 437) Total population (n ¼ 466)

n % n % n %

Mean age (years) 50.7 (8.8)2,3 44.4 (9.6)2 44.8 (9.6)2

Mean work years (full-time equivalents) 22.3 (11.3)2,3 14.8 (9.2)2 15.3 (9.5)2
Ethnic group
White 26 904 291 67 317 68
Black 2 7 90 21 92 20
Hispanic 1 3 47 11 48 10
Other 0 0 9 2 9 2
History of hypertension 15 524 118 27 133 29
History of diabetes 7 245 24 5.5 31 6.7
Smoking status
Never 10 35 166 38 176 38
Current 6 21 109 25 115 25
Former 13 45 162 37 175 38

1
Coronary heart disease.
2
Standard deviation.
Significantly different from those with no CHD, 3p < 0.001.
4
p < 0.05.
5
p < 0.01.

# 2004 Blackwell Publishing Ltd Diabetes, Obesity and Metabolism, 7, 2005, 370–380 373
OA Abdominal diameter index and CHO D. A. Smith et al.

current smoking in 25%. Those with CHD were older,
more Caucasian, had longer work histories as bridge and
tunnel officers and had a higher prevalence of diabetes
and hypertension. Only 2.8% of this working male
population was taking lipid-altering medications in
1994 and the percentage taking them with CHD was no
different from those without CHD (6.9% vs. 2.5%,
p ¼ 0.17).
The correlation between duplicate measurements of
sagittal abdominal diameter on the 48 subjects using
both chest and Holtain calipers was 0.99 and of abdom-
inal diameter index derived from the two sagittal
abdominal diameters divided by the one measure of
thigh circumference was 0.96. The coefficient of variation
of the two measures was 2.5% for sagittal abdominal
diameter and 2.6% for the abdominal diameter index.
These coefficients of variation are identical to those
found in a study of the precision of recumbent
anthropometry and are lower than the coefficients of
variation for waist and hip circumferences (4.3 and
5.9%) found in that study [33].
Table 3 gives differences in the anthropometric and
biochemical measurements between those with and
without CHD. Those with CHD had significantly higher
values for each of the following: WHR (0.980 vs. 0.955),
waist-thigh ratio (2.43 vs. 2.34), abdominal diameter
index (0.587 vs. 0.555), 10-year Framingham CHD risk
of CHD (0.144 vs. 0.089), fibrinogen (296 vs. 270 mg/dl),
urinary microalbumin (33.9 vs. 10.5 mg/dl) and serum
vitamin E concentration (19.4 vs. 15.8 mg/dl).
Table 4 gives the standardized ORs for CHD using
z-values of the anthropometric and biochemical risk fac-
tors both as univariate predictors and after adjusting for
10-year Framingham CHD risk. The 10-year Framingham

Table 3 Biochemical and anthropometric measurements

CHD1 (n ¼ 29) No CHD (n ¼ 437) Total population (n ¼ 466)

Risk factor Mean Mean Mean SD

2 3
10 years Framingham CHD risk 0.144 0.089 0.092 0.076
Body mass index (kg/m2) 30.7 30.4 30.4 5.3
Waist (cm) 101.0 98.5 98.7 13.1
Waist-hip ratio 0.9804 0.955 0.96 0.06
Waist-thigh ratio 2.434 2.34 2.35 0.19
Sagittal abdominal diameter (cm) 24.6 23.4 23.4 3.7
Abdominal diameter index 0.5875 0.555 0.56 0.06
Thigh (cm) 41.8 42.0 420 4.5
Baldness [n (%)]
None 12 (44) 225 (54) 237 (53)
Frontal 2 (7) 43 (10) 45 (10)
Vertex 13 (48) 151 (36) 164 (37)
Total cholesterol (mmol/l) 5.8 5.5 5.5 1.1
Triacylglycerol (mmol/l) 2.6 2.0 2.0 1.6
HDL cholesterol (mmol/l) 1.1 1.1 1.1 0.3
LDL cholesterol (mmol/l) 3.5 3.5 3.5 1.0
TC/HDL cholesterol ratio 5.7 5.3 5.3 1.5
Apolipoprotein A (g/l) 1.48 1.45 1.45 0.23
Apolipoprotein B (g/l) 1.38 1.32 1.32 0.31
Lipoprotein (a) (mg/dl)6 17.8 20.2 20.0 20.8
Homocysteine (umol/l) 15.3 16.8 16.7 6.6
Fibrinogen (g/l) 2.964 2.70 2.72 0.64
Urinary microalbumin (mg/l) 33.93 10.5 11.9 33.7
Vitamin E (mg/l) 19.45 15.8 16.0 6.5
Ferritin (ug/l) 266 230 232 292
(n ¼ 17) (n ¼ 330) (n ¼ 347)

1
Coronary heart disease.
2
10 years probability of developing CHD based on Framingham algorithm.
Significantly different from those with no CHD, 3p < 0.001.
4
p < 0.05.
5
p < 0.01.
6
umol/l ¼ mg/dl  0.0357.

374 Diabetes, Obesity and Metabolism, 7, 2005, 370–380 # 2004 Blackwell Publishing Ltd
D. A. Smith et al. Abdominal diameter index and CHO OA

Table 4 Standardized odds ratios for coronary heart disease

Univariate risk Framingham-score-adjusted risk

Risk factor OR 95% CI OR 95% CI

1 2
10 years Framingham CHD risk 1.72 1.29, 2.30
Work years (full-time equivalent) 1.812 1.34, 2.44 1.552 1.09, 2.23
Body mass index 1.05 0.72, 1.53 1.03 0.69, 1.54
Waist 1.20 0.83, 1.73 1.14 0.77, 1.69
Waist-hip ratio 1.673 1.06, 2.61 1.47 0.92, 2.35
Waist-thigh ratio 1.653 1.07, 2.54 1.36 0.93, 2.00
Sagittal abdominal diameter 1.37 0.95, 1.98 1.19 0.81, 1.75
Abdominal diameter index 1.804 1.20, 2.71 1.49 0.96, 2.30
Thigh 0.95 0.64, 1.42 0.95 0.64, 1.41
Baldness (type vs. none)
None 1.0
Frontal 0.87 0.19, 4.04 0.91 0.19, 4.26
Vertex 1.61 0.72, 3.63 1.38 0.61, 3.17
Total cholesterol 1.24 0.86, 1.79 1.02 0.68, 1.52
Triacylglycerol 1.27 0.98, 1.65 1.17 0.88, 1.55
HDL cholesterol 0.87 0.58, 1.30 1.07 0.72, 1.60
LDL cholesterol 1.13 0.82, 1.55 1.05 0.76, 1.46
Total/HDL cholesterol ratio 1.27 0.89, 1.79 0.93 0.63, 1.39
Apolipoprotein A 1.14 0.78, 1.68 1.27 0.84, 1.83
Apolipoprotein B 1.24 0.85, 1.81 0.91 0.59, 1.41
Lipoprotein (a) 0.88 0.57, 1.35 0.80 0.52, 1.23
Homocysteine 0.74 0.45, 1.23 0.70 0.40, 1.21
Fibrinogen 1.383 1.00, 1.89 1.18 0.82, 1.70
Microalbumin 1.324 1.08, 1.63 1.21 0.99, 1.50
Vitamin E 1.534 1.14, 2.06 1.35 0.99, 1.84
Ferritin 1.403 1.00, 1.97 1.37 0.97, 1.93

CHD, coronary heart disease; CI, confidence interval, HDL, high-density lipoprotein; LDL, low-density lipoprotein; OR, odds ratio.
1
10 years probability of developing CHD based on Framingham algorithm.
Significantly greater than 1.0 by logistic regression analysis, 2p < 0.001.
3
p < 0.05.
4
p < 0.01.

CHD risk gives a univariate risk of 1.72 per z-value.
Other variables showing significant increased risk on a
univariate basis include full-time equivalent years of
work (OR ¼ 1.81), abdominal diameter index (OR ¼ 1.80),
WHR (OR ¼ 1.67), waist-thigh ratio (OR ¼ 1.65), fibrino-
gen (OR ¼ 1.38), urinary microalbumin (OR ¼ 1.32),
serum vitamin E concentration (OR ¼ 1.53) and serum
ferritin (OR ¼ 1.40). After adjusting for 10-year Framing-
ham CHD risk, the only variable that retains its elevated
OR for prevalent CHD is full-time equivalent years of work
(OR ¼ 1.55). Abdominal diameter index (OR ¼ 1.49, 95%
CI 0.96, 2.30), urinary microalbumin (OR ¼ 1.21, 95% CI
0.99, 1.50), vitamin E concentration (OR ¼ 1.35, 95% CI
0.99, 1.84), and ferritin (OR ¼ 1.37, 95% CI 0.97, 1.93)
retain only borderline significance.
ORs for each of the upper two vs. the lowest tertile of
several anthropometric measures of body fat and its dis-
tribution for predicting CHD are given in table 5. In
univariate analyses, the ORs are significantly increased
in the highest tertile of WHR (OR ¼ 3.75), waist-thigh
ratio (OR ¼ 3.46) and abdominal diameter index
OR ¼ 5.47). After adjusting for 10-year Framingham
CHD risk, only abdominal diameter index remains sig-
nificant (OR ¼ 3.79).
When four components of the insulin resistance syn-
drome were forced into the model comparing the highest
to the lowest tertile of abdominal diameter index for
CHD (table 6), serum triacylglycerol and history of dia-
betes accounted for most of the predictive power with
HDL cholesterol and average systolic blood pressure
adding little more. No residual predictive power for
abdominal diameter index remains after adjusting for
serum triacylglycerol (not part of the American Heart
Association risk equation) and history of diabetes.

# 2004 Blackwell Publishing Ltd Diabetes, Obesity and Metabolism, 7, 2005, 370–380 375
OA Abdominal diameter index and CHO D. A. Smith et al.

Table 5 Odds ratios for coronary heart disease by tertile of body size measurements

Framingham-score-
Variable Tertile Inclusive values Univariate risk 95% CI adjusted risk 95% CI
2
Body mass index (kg/m ) 1 19.6–27.7 1.00 1.00
2 27.8–32.0 0.79 0.30, 2.06 0.66 0.25, 1.77
3 32.1–50.5 1.11 0.46, 2.69 0.92 0.37, 2.27
Waist (cm) 1 69.6–92.2 1.00 1.00
2 92.3–102.7 1.69 0.60, 4.77 1.45 0.51, 4.17
3 102.8–152.0 2.07 0.76, 5.67 1.66 0.59, 4.63
Waist-hip ratio 1 0.80–0.93 1.00 1.00
2 0.94–0.98 2.61 0.80, 8.50 2.28 0.69, 7.53
3 0.99–1.14 3.751 1.21, 11.68 2.92 0.92, 9.24
Waist-thigh ratio 1 1.76–2.27 1.00 1.00
2 2.28–2.42 1.42 0.44, 4.57 1.20 0.37, 3.91
3 2.43–2.91 3.461 1.23, 9.70 2.55 0.89, 7.34
Sagittal abdominal diameter (cm) 1 15.3–21.5 1.00 1.00
2 21.6–24.5 1.16 0.41, 3.28 1.00 0.35, 2.88
3 24.6–40.2 1.95 0.76, 5.03 1.55 0.59, 4.08
Abdominal diameter index 1 0.40–0.53 1.00 1.00
2 0.54–0.58 3.52 0.95, 13.03 2.82 0.75, 10.63
3 0.59–0 .73 5.472 1.55, 19.28 3.791 1.04, 13.82
Thigh (cm) 1 32.0–39.8 1.00 1.00
2 39.9–43.3 0.63 0.24, 1.67 0.60 0.22, 1.62
3 43.4–56.5 0.89 0.37, 2.16 1.01 0.41, 2.50

CI, confidence interval.

Significantly different from 1.0 by logistic regression analysis compared with lowest tertile, 1p < 0.05.
2
p < 0.01.

index is the superior measure, and the only anthropo-

Discussion
In this cross-sectional study of New York City bridge and
tunnel officers, we compared the predictive value of six
simple measures of total and abdominal fat for prevalent
CHD. We found that abdominal diameter index and
other measures of central obesity had more powerful
associations than BMI, which is a height-adjusted meas-
ure of total body weight and is highly correlated with
total body fat [20]. In our study, waist and sagittal
abdominal diameter are not univariately associated
with increased risk, but dividing each by hip or thigh
circumference make each significantly discriminatory
on univariate analysis for CHD risk. Thus, all three meas-
ures most associated with prevalent CHD risk in this
study – abdominal diameter index (sagittal abdominal
diameter/thigh circumference), waist-thigh ratio and
WHR – require a correction for relative body size differ-
ences in muscle, superficial fat and bone mass as estimated
by hip or thigh circumferences. Some studies have found
possible protective effects for CHD from thigh fat [34,35] or
hip fat (circumference) [36]; ours did not.
The 10-year-Framingham-score-adjusted tertile logis-
tic regression analysis suggests that abdominal diameter
metric measure which adds predictive information to
the 10-year Framingham CHD risk score. Kahn [25] has
predicted the theoretical superiority of the abdominal
diameter index, using mid-thigh circumference, as a pre-
dictor of the risk of intra-abdominal fat for coronary risk,
and in two previous case-control studies [10,37], he and
others have shown it to be superior to waist-mid thigh
ratio and WHR in predicting ischaemic heart disease and
sudden coronary death. Our study is the first cross-
sectional study to demonstrate this direct relationship
between abdominal diameter index and CHD and to
explore the risk factors through which that relationship
might occur.
The hypothesis for the increased CHD risk associated
with increased abdominal diameter index is that it may
be the best anthropometric measure of visceral obesity
which increases insulin resistance and its attendant car-
diovascular risk factors. Many studies, not measuring
abdominal diameter index, have shown the superiority
of sagittal abdominal diameter to other anthropometric
measures as a measure of visceral obesity and CHD risk.
Pouliot et al. [14] have shown that waist circumference
and computerized axial tomography-scan-measured

376 Diabetes, Obesity and Metabolism, 7, 2005, 370–380 # 2004 Blackwell Publishing Ltd
D. A. Smith et al.
Table 6 Odds ratio of highest vs. lowest tertile of abdominal
diameter index for prevalent coronary heart disease
OR 95% CI
Univariate 5.39 1.53, 19.02
Model 1: Includes age, BMI, 4.63 1.21, 17.67
current smoking
Model 2: Model 1 plus hx diabetes 4.10 1.05, 16.03
Model 3: Model 2 plus triacylglycerol 3.32 0.84, 13.1
BMI, body mass index; CI, confidence interval; OR, odds ratio.
(CAT-scan-measured) supine sagittal abdominal dia-
meter are more strongly correlated with abdominal visc-
eral adipose tissue on CAT scan than WHR. They suggest
that values for waist circumference greater than 100 cm
or sagittal abdominal diameter greater than 25 cm would
most likely be associated with a more atherogenic
metabolic profile. Likewise, using total body computed
tomography, Kvist et al. [21] have shown that CAT-scan-
measured standing sagittal abdominal diameter in men
is more highly correlated with visceral adipose tissue
than waist circumference or WHR (r ¼ 0.92, 0.88 and
0.39, respectively).
The Bogalusa Heart Study [38], a community-based
sample of 409 Blacks and 1011 Whites aged 20–38
years, found that sagittal abdominal diameter was more
strongly correlated with the coronary disease risk-factor
variables as a group than were other obesity measures. In
a Swedish study [39] of 885 men and women employees
in a Swedish telephone company participating in a
health survey, sagittal abdominal diameter showed a
stronger correlation to cardiovascular risk factors than
waist circumference, WHR or BMI.
Our findings are consistent with the hypothesis that
the CHD risk produced by increased abdominal diameter
index is mediated through increased insulin resistance.
Of the four physiologic parameters measured which are
associated with the insulin resistance syndrome – low-
ered HDL cholesterol, increased systolic blood pressure,
increased serum triacylglycerol and hyperglycaemia
(history of diabetes in this study) – the latter two were
chosen by the multivariate model as the most powerful.
Once diagnosis of diabetes and serum triacylglycerol
concentration were forced into the multivariate model,
the OR for abdominal diameter index dropped to an
insignificant level leaving no residual predictive value
for prevalent CHD. Seidell et al. [11] have shown in 981
males in the Baltimore Longitudinal Study on Ageing
that sagittal abdominal diameter is more highly correl-
ated with fasting triacylglycerol and glucose concentra-
tions than with cholesterol, systolic and diastolic blood
# 2004 Blackwell Publishing Ltd
Abdominal diameter index and CHO OA
pressure levels. This same study showed that baseline
sagittal abdominal diameter prospectively predicted
death from CHD in the next 15–20 years adding to the
p evidence for a causal, as opposed to just an innocent,
<0.01 association of visceral adipose tissue and CHD [40].
0.025 The standardized univariate CHD risk for abdominal
diameter index is 1.80 which is quite comparable to the
0.043
composite 10-year Framingham CHD risk of 1.72, which
0.087
includes age, gender, systolic blood pressure, history of
diabetes and smoking, total and HDL cholesterol con-
centrations. Other risk factors associated with equiva-
lent univariate increases in CHD risk in this study
include full-time equivalent years of work as a bridge
or tunnel worker, serum fibrinogen and vitamin E and
urinary microalbumin. Reasons postulated for the occu-
pational exposure risk include previous exposure to car-
bon monoxide in vehicular exhaust before the tollbooths
were adequately ventilated [41], job strain, physical
inactivity on the job and possibly other toxic compon-
ents of vehicular exhaust including particulates.
Fibrinogen was found in the Framingham Study [42]
to be an independent risk factor for CHD and increased
significantly with impaired glucose tolerance, smoking,
age and relative weight. The univariate predictive power
of fibrinogen in our study was eliminated when adjusted
for the 10-year Framingham risk score that included all
of these variables. The univariate association of urinary
microalbumin and CHD (OR ¼ 1.32) remained signifi-
cant when adjusted for history of diabetes (OR ¼ 1.24),
a finding duplicated in the Heart Outcomes Prevention
Evaluation trial [43], which found that any degree of
albuminuria was a risk factor for CVD events over 4.5
years, with or without diabetes. The positive association
of serum vitamin E concentration and CHD in this cross-
sectional study is explained by selection bias, i.e. the
preferential taking of vitamin E in those who have CHD
vs. those without CHD (48.3% vs. 24.6%, p < 0.01) and
becomes insignificant when adjusted for taking or not
taking vitamin E supplements.
It has been shown by magnetic resonance imaging [44]
that weight loss favours a larger percentage weight
decrease in abdominal adipose tissue than in subcutan-
eous adipose tissue, more evident in men than women.
Weight loss, increased exercise and the two combined
have been shown to decrease visceral fat and to improve
the adverse metabolic factors associated with insulin
resistance [44–48] and even to prevent the onset of type
2 diabetes in persons with insulin resistance [49]. The
specific occupational causes of increased CHD risk in
this population are not totally clear. While further efforts
to identify and remediate the workplace factor(s) respon-
sible for the increased CHD risk associated with
Diabetes, Obesity and Metabolism, 7, 2005, 370–380 377
OA Abdominal diameter index and CHO
occupation remain a public health imperative, the
demonstrated beneficial effects of weight loss on decreasing
CHD risk factors suggest that preventive health efforts
within this working population might also include
periodic efforts at weight loss.
Limitations of this study include the small number of
CHD cases, some of which were diagnosed by classic
cardiac stress testing criteria alone. Such a small number
of cases, however, would reduce the power to produce
significant associations, and thus the importance of
abdominal diameter index becomes even more impress-
ive. There were not enough women in this occupational
sample to allow a separate analysis, making these
findings relevant only to a multiracial group of males.
Insulin levels, as a measure of insulin resistance, CT and
MRI measurements of abdominal fat were not measured
because the principal aim of this study was to explore
the relationship of occupational exposure and CHD.
In conclusion, this cross-sectional study of 466 male
tunnel and bridge officers in New York City confirms the
well-demonstrated finding that central obesity conveys a
significant risk for prevalent CHD and suggests that
abdominal diameter index is its most discriminate
anthropometric measure. The risk, comparable to that
associated with a calculated 10 years risk of CHD using
a Framingham algorithm and with years worked as a
bridge and tunnel officer, was most likely mediated
through increased insulin resistance. Although abdom-
inal diameter index is simple and inexpensive to meas-
ure, it is unlikely that it will replace the tape measure
and abdominal and hip circumferences used in clinical
practice. Because it may be more precise than waist and
hip circumferences and is more powerfully associated
with prevalent CHD than other anthropometric variables
in this study, and because it adds information beyond
that provided by a 10-year Framingham CHD risk score,
it should be an important variable measured in epidemi-
ologic and clinical investigations of visceral obesity and
CHD.
Acknowledgements
The authors gratefully acknowledge the many people at
both the Bridge and Tunnel Officers Benevolent Asso-
ciation and the Triborough Bridge and Tunnel Authority
who supported this project and provided invaluable
assistance with its many logistical challenges. Addition-
ally, we thank Victor Herbert, MD, for his advice and
support, Raymond Gambino, MD, Executive Vice Presi-
dent and Chief Medical Officer and Patricia Maloney,
Director of Corporate Medical Services, MetPath, Inc., for
378 Diabetes, Obesity and Metabolism, 7, 2005, 370–380
D. A. Smith et al.
their generosity in providing the chemistry tests and re-
sults and Dr Sylvan Wallenstein for his statistical advice.
References
1 Terry RB, Page WF, Haskell WL. Waist/hip ratio, body
mass index and premature cardiovascular disease mor-
tality in US army veterans during a twenty-three year
follow up study. Int J Obes 1992; 16: 417–423.
2 Larsson B. Regional obesity as a health hazard in men –
prospective studies. Acta Med Scand Suppl 1988; 723:
45–51.
3 Lapidus L, Bengtsson C, Larsson B et al. Distribution of
adipose tissue and risk of cardiovascular disease and
death: a 12 year follow up of participants in the popula-
tion study of women in Gothenburg, Sweden. Br Med J
1984; 289: 1257–1261.
4 Prineas RJ, Folsom AR, Kaye SA. Central adiposity and
increased risk of coronary artery disease mortality in
older women. Ann Epidemiol 1993; 3: 35–41.
5 Thompson CJ, Ryu J, Craven TE, Kahl FR, Crouse JR.
Central adipose distribution is related to coronary
atherosclerosis. Arterioscler Thromb 1991; 11: 327–333.
6 Hodgson JM, Wahlquist ML, Balazs NDH, Boxall JA.
Coronary atherosclerosis in relation to body fatness and
its distribution. Int J Obes 1994; 18: 41–46.
7 Folsom AR, Kaye SA, Sellers TA et al. Body fat distribu-
tion and 5-year risk of death in older women. JAMA
1993; 269: 483–487.
8 Folsom AR, Burke GL, Byers CL et al. Implications of
obesity for cardiovascular disease in blacks: the CARDIA
and ARIC studies. Am J Clin Nutr 1991; 53: 1604S–1611S.
9 Curb JD, Marcus EB. Body fat, coronary heart disease,
and stroke in Japanese men. Am J Clin Nutr 1991; 53:
1612S–1615S.
10 Kahn HS, Austin H, Williamson DF, Arensberg D. Sim-
ple anthropometric indices associated with ischemic
heart disease. J Clin Epidemiol 1996; 49: 1017–1024.
11 Seidell JC, Andres R, Sorkin JD, Muller DC. The sagittal
waist diameter and mortality in men: the Baltimore
Longitudinal Study on Aging. Int J Obes Relat Metab
Disord 1994; 18: 61–67.
12 Reaven GM. Role of insulin resistance in human disease,
Banting Lecture 1988. Diabetes 1988; 37: 1595–1607.
13 Hunter GR, Kekes-Szabo T, Treuth MD et al. Intra-
abdominal adipose tissue, physical activity and cardiovas-
cular risk in pre-and post-menopausal women. Int J Obes
Relat Metab Disord 1996; 20: 860–865.
14 Pouliot M-C, Despres J-P, Lemieux S et al. Waist circum-
ference and abdominal sagittal diameter: best simple
anthropometric indexes of abdominal viseral adipose
tissue accumulation and related cardiovascular risk in
men and women. Am J Cardiol 1994; 73: 460–468.
15 Hunter GR, Kekes-Szabo T, Synder SW et al. Fat distri-
bution, physical activity, and cardiovascular risk factors.
Med Sports Exerc 1997; 29: 362–369.
# 2004 Blackwell Publishing Ltd
D. A. Smith et al.
16 Kaplan NM. The deadly quartet. Upper body obesity,
glucose intolerance, hypertriglyceridemia, and hyper-
tension. Arch Intern Med 1989; 149: 1514–1520.
17 Lemieux S, Despres JP, Moorjani S et al. Are gender
differences in cardiovascular disease risk factors
explained by the level of viseral adipose tissue? Dia-
betologia 1994; 37: 757–764.
18 Bjorntorp P. Metabolic implications of body fat distribu-
tion. Diabetes Care 1991; 14: 1132–1143.
19 Despres J-P, Moorjani S, Lupien PJ et al. Regional dis-
tribution of body fat, plasma lipoproteins, and cardio-
vascular disease. Arteriosclerosis 1990; 10: 497–511.
20 Ashwell M, Cole TJ, Dixon AK. Obesity: new insight into
the anthropometric classification of fat distribution
by computed tomography. Br Med J 1985; 290:
1692–1694.
21 Kvist H, Chowdhury B, Grangard U, Tylen U, Sjostrom L.
Total and visceral adipose-tissue volumes derived from
measurements with computed tomography in adult men
and women: predictive equations. Am J Clin Nutr 1998;
48: 1351–1361.
22 Staten MA, Totty WG, Kohrt WM. Measurement of
fat distribution by magnetic resonance imaging. Invest
Radiol 1989; 24 (5): 345–349.
23 McNeill G, Fowler PA, Maughan RJ et al. Body fat in lean
and overweight women estimated by six methods. Br J
Nutr 1991; 65 (2): 95–103.
24 Richelsen B, Pedersen SB. Associations between differ-
ent anthropometric risk parameters in non-obese,
middle-aged men. Int J Obes 1995; 19: 169–174.
25 Kahn H. Choosing an index for abdominal obesity: an
opportunity for epidemiologic clarification. J Clin Epi-
demiol 1993; 46: 491–494.
26 Herbert R, Schechter C. Smith DA et al. Occupational
coronary heart disease among bridge and tunnel officers.
Arch Environ Health 2000; 55: 152–163.
27 Rose GA. The diagnosis of ischaemic heart pain and
intermittent claudication in field surveys. Bull World
Health Org 1962; 27: 645–658.
28 Lesko SM, Rosenberg L, Shapiro S. A case-control study
of baldness in relation to myocardial infarction in men.
JAMA 1993; 269: 998–1003.
29 Prineas RJ, Crow RS, Blackburn H. The Minnesota Code
Manual of Electrocardiographic Findings. Standards and
Procedures for Measurement and Classification. Boston,
MA: John Wright, 1982.
30 Friedewald WT, Levy RI, Fredrickson DS. Estimation of
the concentration of low-density lipoprotein cholesterol
in plasma, without use of the preparative ultracentrifuge.
Clin Chem 1972; 18: 499–502.
31 Anderson KM, Wilson PWF, Odell PM, Kannel WB. An
updated coronary risk profile: a statement for health
professionals. Circulation 1991; 83: 356–362.
32 Bruce RA. Exercise testing of patients with coronary
artery disease. Ann Clin Res 1971; 3: 323–332.
# 2004 Blackwell Publishing Ltd
Abdominal diameter index and CHO OA
33 Williamson DF, Kahn HS, Worthman CM, Burnette JC,
Russell CM. Precision of Recumbent Anthropometry.
Am J Human Biol 1993; 15: 159–187.
34 Terry RB, Stefanick ML, Haskell WL, Wood PD. Contri-
butions of regional adipose tissue depots to plasma lipo-
protein concentrations in overweight men and women:
possible protective effects of thigh fat. Metabolism 1991;
40: 733–740.
35 Sjostrom CD, Hakangard AC, Lissner L, Sjostrom L. Body
compartment and subcutaneous adipose tissue distribu-
tion – risk factor patterns in obese subjects. Obes Res
1995; 3: 9–22.
36 Lissner L, Bjorkelund C, Heitmann BL, Seidell JC,
Bengtsson C. Larger hip circumference independently
predicts health and longevity in a Swedish female
cohort. Obes Res 2001; 9: 644–646.
37 Kahn HS, Simoes EJ, Kaponen M, Hanzlick R. The
abdominal diameter index and sudden coronary death
in men. Am J Cardiol 1996; 78: 961–964.
38 Gustat J, Elkasabany A, Srinivasan S, Berenson GS. Rela-
tion of abdominal height to cardiovascular risk factors in
young adults: the Bogalusa heart study. Am J Epidemiol
2000; 151: 885–891.
39 Ohrvall M, Berglund L, Vessby B. Sagittal abdominal
diameter compared with other anthropometric measure-
ments in relation to cardiovascular risk. Int J Obes Relat
Metab Disord 2000; 24: 497–501.
40 Seidell JC, Bouchard C. Visceral fat in relation to health;
is it a major culprit or simply an innocent bystander? Int
J Obes 1997; 21: 626–631.
41 Stern FB, Halperin WE, Hornung RW, Ringenburg VL,
McCammon CS. Heart disease mortality among bridge
and tunnel officers exposed to carbon monoxide. Am J
Epidemiol 1988; 128: 1276–1288.
42 Kannel WB, Wolf PA, Castelli WP, D’Agostino RB. Fibri-
nogen and risk of cardiovascular disease, The Framing-
ham Study. JAMA 1987; 258: 1183–1186.
43 Gerstein HC, Mann JF, Yi Q et al. Albuminuria and risk
of cardiovascular events, death, and heart failure in dia-
betic and nondiabetic individuals. JAMA 2001; 286:
421–426.
44 van der Kooy K, Leenen R, Seidell JC, Deurenberg AD,
Bakker CJG. Waist-hip ratio is a poor predictor of changes
in visceral fat. Am J Clin Nutr 1993; 57: 327–333.
45 Canai H, Tokunaga K, Shigenori F et al. Decrease in
intra-abdominal visceral fat may reduce blood pressure
in obese hypertensive women. Hypertension 1996; 27:
125–129.
46 Lehmann R, Vokac A, Niedermann K, Agosti K, Spinas GA.
Loss of abdominal fat and improvement of the cardio-
vascular risk profile by regular moderate exercise
training in patients with NIDDM. Diabetologia 1995; 38:
1313–1319.
47 Kopelman G. The effects of weight loss treatment on
upper and lower body fat. Int J Obes 1997; 21: 619–625.
Diabetes, Obesity and Metabolism, 7, 2005, 370–380 379
OA Abdominal diameter index and CHO D. A. Smith et al.

48 Katzel LI, Bleecker ER, Colman EG et al. Effects of weight 49 Knowler WC, Barrett-Connor E, Fowler SE et al. Reduction
loss vs aerobic exercise training on risk factors for coronary in the incidence of type 2 diabetes with lifestyle
disease in healthy, obese, middle-aged and older men. intervention or metformin. N Engl J Med 2002; 346:
A randomized controlled trial. JAMA 1995; 274: 1915–1921. 393–403.

380 Diabetes, Obesity and Metabolism, 7, 2005, 370–380 # 2004 Blackwell Publishing Ltd