International Journal of Obesity (2001) 25, 17941799

2001 Nature Publishing Group All rights reserved 03070565/01 $15.00

www.nature.com/ijo

PAPER
Anthropometric indexes in the prediction of type 2
diabetes mellitus, hypertension and dyslipidaemia in
a Mexican population
A Berber1*, R Gomez-Santos1, G Fanghanel1 and L Sanchez-Reyes1
1

Servicio de Endocrinologa del Hospital General de Mexico, Organo Descentralizado, Mexico City, Mexico

OBJECTIVE: To determine values of simple anthropometric measurements which are associated with the presence of type 2
diabetes mellitus, hypertension and dyslipidaemia and to assess anthropometric cut-off values for predicting the likelihood of
these chronic conditions in a Mexican population.
DESIGN AND SETTING: The data were obtained from PRIT (Prevalence of Cardiovascular Risk Factors in General Hospital
Workers) surveys from 1994 to 2000 adjusted to the structure of the overall Mexican population.
SUBJECTS: A total of 2426 men and 5939 women aged 38.99
7.11 and 39.11
14.25 y, respectively.
MEASUREMENTS: The optimal sensitivity and specificity of using various cut-off values of BMI (body mass index), WHR (waistto-hip ratio), WC (waist circumference) and WTH (waist-to-height ratio) to predict type 2 diabetes mellitus (DM), hypertension
(HT), or dyslipidaemia were examined by receiver operating characteristic curve (ROC) analysis. The likelihood ratios for having
diabetes, hypertension and dyslipidaemia in subjects with various cut-off values of BMI, WHR, WC and WTH were calculated.
Multiple step-wise logistic regression analysis was used to examine the independent relationship between the anthropometric
indexes, age and smoking, and the odds ratio of having chronic conditions.
RESULTS: The BMI cut-off to predict DM, HT, or dyslipidaemia varied from 25.2 to 26.6 kg=m2 in both men and women. The
optimal WC cut-offs were 90 cm in men and 85 cm in women. The WHR cut-off was about 0.90 in men and 0.85 in women, and
the optimal WTH cut-off was 52.5 in men and varied from 53 to 53.5 in women. The cut-off levels for WC, WHR and WTH
corresponded to the inflexion points in the likelihood ratio graphs. In the case of BMI likelihood ratio graphs, we found a
significant increase in the risk for chronic conditions from 22 to 23 BMI levels in both genders. Logistic regression analyses
disclosed that only BMI and age were included in all the models as well as the influence of smoking in DM and dyslipidaemia in
men.
CONCLUSION: Although these results may not be readily applied to the rest of the Mexican population or to other Hispanic
populations, they point to the necessity of similar studies with large randomized samples to find the cut-off levels for chronic
conditions in different populations.
International Journal of Obesity (2001) 25, 1794 1799
Keywords: anthropometric indexes; type 2 diabetes mellitus; hypertension; dyslipidaemia

Introduction
Obesity is associated with a higher prevalence of hypertension
(HT), type 2 diabetes mellitus (DM), and dyslipidaemia.1 3
Obesity anthropometric indexes such as body mass index

*Correspondence: A Berber, Cruz Galvez 269 CP 02800, Mexico City DF,

Mexico.
E-mail: arturoberber@aol.com
Received 6 March 2001; revised 8 May 2001;
accepted 12 June 2001

(BMI), waist circumference (WC), waist hip ratio (WHR),

and waist-to-height ratio (WTH) are all useful anthropometric measurements to provide important information on
cardiovascular risks.
Apart from the association between obesity indexes and
cardiovascular risks, it is also important to define the cut-off
values of an individual index to allow effective screening.
WHO and NIH have defined BMI, WC and WHR cut-off
levels for white, black and Hispanic American adults;4,5
however, these definitions cannot be readily applied to
other populations.6

Prediction of type 2 DM, hypertension and dyslipidaemia

A Berber et al

1795
The WHO MONICA survey found different WC distributions in different populations.7 These WC differences are
related to different cut-off levels as screening tools to detect
hypertension in African and Caribbean population.8
Ko found that cut-off values of anthropometric indexes to
define obesity used in Caucasians may not be applicable to the
Chinese. For instance, BMI cut-off values for chronic conditions are from 23 to 24.3 kg=m2 in men and from 23.2 to
24.3 kg=m2 in women, while likelihood ratios show significant
risk for chronic conditions from 23 kg=m2 BMI levels.9
It is beneficial to healthcare to assess what values of
simple anthropometric measurements are associated with
the presence of chronic conditions such as diabetes, hypertension or dyslipidaemia in different populations.

Materials and methods

Data were collected from 1994 to 2000 during the yearly
PRIT (Prevalence of Cardiovascular Risk Factors in General
Hospital Workers) surveys with only the data of new participants being evaluated. In these surveys all the workers of
Hospital General de Me xico were invited to participate.
Informed consent was obtained from all the participants.
Participants considered were 20 y or older.
Table 1 Anthropometry and prevalence of chronic conditions in the
studied population
Men
2426

Women
5939

38.99
7.11
1.67
0.08
69.82
11.74
87.26
7.11
25.01
3.87
0.91
0.05
52.33
4.76
47.3 (1148=2426)
29.6 (719=2426)
13.2 (321=2426)
6.1 (147=2426)
7.1 (172=2426)
16.7 (406=2426)
43.5 (1055=2426)

39.11
14.25
1.55
0.06
61.24
10.72
81.92
8.18
25.51
4.50
0.84
0.07
52.92
5.85
49.8 (2955=5939)
33.2 (1969=5939)
17.2 (1023=5939)
3.4 (203=5939)
5.8 (346=5939)
17.9 (1063=5939)
41.2 (2447=5939)

0.693
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
0.040
0.002
< 0.001
< 0.001
0.028
0.201
0.056

n
Age (y)
Height (m)
Weight (kg)
WC (cm)
BMI
WHR
WTH
BMI > 25
BMI > 27
BMI > 30
IGT
DM
HT
Dyslipidaemia

After an overnight fast, all subjects were seen in the

Endocrinology service. Demographic data were documented
and height and weight (to the nearest 0.1 kg) were measured
with the subject in light clothing without shoes. Body mass
index was calculated as the weight (kg) divided by the square
of the height (m). Waist circumference (WC) was taken as the
minimum circumference between the umbilicus and xiphoid
process and measured to the nearest 0.5 cm. Hip circumference was measured as the maximum circumference around
the buttocks posteriorly and the symphysis pubis anteriorly
and measured to the nearest 0.5 cm. Waist hip ratio (WHR)
and waist-to-height ratio (WTH) were then calculated. After
sitting for at least 5 min, blood pressure (BP) was measured in
the right arm by the same research nurse using a standard
mercury sphygmomanometer, and a large cuff was used for
the obese patients. The Korotkoff sound V was taken as the
diastolic BP.
Blood was taken after a 12 h fast for measurement
of plasma glucose (PG), total cholesterol (TC), fasting triglyceride (TG), and high-density lipoprotein cholesterol (HDL).
In the present analysis, the American Diabetes Association
criteria were used to diagnose diabetes.10 Diabetes was
defined as a fasting PG of 7.0 mmol=l and=or 2 h PG
11.1 mmol=l.10 Hypertension was defined as a systolic BP of
140 mmHg and=or diastolic BP of 90 mmHg.11 Dyslipidaemia
was considered to be present if plasma total cholesterol was
5.16 mmol=l and=or a fasting triglyceride of 2.26 mmol=l
and=or HDLC < 0.90 mmol=l (these patients also had LDLC

Table 3 Cut-off values of BMI, WC, WHR and WTH to predict DM, HT
and dyslipidaemia based on ROC analysis
Index
Men
WC

WHR

BMI

WTH

Table 2 Age-adjusted partial correlation coefficient among the

anthropometric indexes in the studied population

Men
WTH
WHR
WC

BMI

WC

WHR

WTH

0.094
0.250
0.043

0.859
0.328

0.232

0.328

0.232
0.859

Women
WC

WHR

BMI

Women
WTH
WHR
WC

0.252
0.080
0.200

All the coefficients have P < 0.05.

0.928
0.402

0.353

0.402

0.353
0.928

WTH

Condition

Cut-off level

Sensitivity and specificity

DM
HT
Dyslipidaemia
DM
HT
Dyslipidaemia
DM
HT
Dyslipidaemia
DM
HT
Dyslipidaemia

90.cm
90.cm
90.cm
0.90
0.91
0.90
25.3 kg=m2
26.2 kg=m2
24.9 kg=m2
52.5
52.5
52.5

47%
47%
44%
57%
70%
67%
57%
71%
65%
50%
51%
45%

DM
HT
Dyslipidaemia
DM
HT
Dyslipidaemia
DM
HT
Dyslipidaemia
DM
HT
Dyslipidaemia

85.cm
85.cm
85.cm
0.86
0.85
0.84
25.4 kg=m2
26.6 kg=m2
2
25.2 kg=m
53.5
53.5
53.0

53%
51%
47%
62%
54%
50%
55%
72%
64%
53%
56%
49%

International Journal of Obesity

Prediction of type 2 DM, hypertension and dyslipidaemia

A Berber et al

1796

> 5.16 mmol=l).12 Data were adjusted to the distribution of

the general population of Mexico in the 2000 census.13
Statistical analysis was performed using the Statistical
Package for Social Sciences (version 6.0) software on an
IBM compatible computer. All results are expressed as
mean
s.d. or percentage where appropriate. The optimal
sensitivity and specificity of using various cut-off values of
BMI, WHR, WC and WTH to predict diabetes, hypertension
or dyslipidaemia were examined by the receiver operating
characteristic curve (ROC) analysis.14,15 ROC curves were
plotted using measures of sensitivity and specificity based
on various anthropometric cut-off values. The ROC curve
analysis allows visual evaluation of the trade-offs between
sensitivity and specificity associated with different values of
the test result.14,15
With the ROC technique, comparison of sensitivity with
the specificity rate was made over the entire range of BMI,
WHR, WC and WTH. The BMI, WHR, WC and WTH cut-off
points were determined by interpolation from the point of
intersection of the lines of specificity and sensitivity (where
sensitivity equaled specificity). The point of intersection
between lines of specificity and sensitivity identified the
highest numbers of subjects with and without a given
chronic condition.16
Sensitivity was defined as the percentage of the total
number of subjects with the chronic condition who were
correctly identified by the anthropometric indexes cut-off
points with specificity defined as the percentage of the total
number of subjects without the chronic condition who were
correctly identified by anthropometric indexes cut-off
points.
The overall performance of the ROC test was quantified by
computing area under the curve (AUC). An AUC of 1 indicated perfect performance, while 0.5 indicated a performance that was not different from chance.15
The likelihood ratio (LR) was calculated to estimate the
odds of having diabetes, hypertension or dyslipidaemia in
subjects with various cut-off values of the anthropometric
indexes.17 LR is defined as sensitivity=(1 7 specificity). The
sensitivity and specificity of having chronic conditions at
various anthropometric cut-off values were calculated using
the adjusted population classified with reference to each cutoff level and hence the corresponding LR was derived. Since
LR refers to actual test results before disease status is known,
it is more immediately useful to clinicians than sensitivity
and specificity.17
Multiple step-wise logistic regression analysis was used to
examine the independent relationship between the four
anthropometric indexes of age and smoking and the odds
ratio of having DM, HT or dyslipidaemia.

Results
The adjusted population consisted of 2426 men and 5939
women. Their anthropometric indexes and prevalence of
glucose impaired tolerance (GIT), type 2 diabetes mellitus
International Journal of Obesity

(DM), blood hypertension (HT) and dyslipidaemia are summarized in Table 1. Men had a greater height, WC, WHR and
prevalence of DM and dyslipidaemia than women.
Except the significant correlation between WTH and WC,
there were no close associations amongst the anthropometric indexes, as shown by the age-adjusted partial correlation coefficients (see Table 2).
Table 3 summarizes the cut-off values of various anthropometric indexes to predict DM, HT, or dyslipidaemia using
the ROC analysis. The BMI cut-off to predict DM, HT or
dyslipidaemia varied between 25.2 and 26.6 kg=m2 in both
men and women. The optimal WC cut-off was 90 cm in men
and 85 cm in women. The WHR cut-off was approximately
0.90 men and 0.85 in women and the optimal WTH cut-off
was 52.5 in men and varied between 53 and 53.5 in women.
Figures 1 and 2 show the likelihood ratios of having DM,
HT, or dyslipidaemia in subjects with different values for
these four anthropometric indexes. The cut-off levels for
WC, WHR and WTH corresponded to the inflexion points
in the likelihood ratio graphs. In the case of BMI likelihood
ratio graphs, we found a significant increase in the risks for
chronic conditions from 22 23 BMI levels in both genders.

Table 4 Multiple logistic regression analysis using BMI, WC, WHR,

WTH, age and smoking as independent variables for the risk of
developing chronic conditions
Independent
variables
Men
DM mellitus (2315 cases in the analysis)
Smoking
(Nagelkerke, r2 0.104)
Age
BMI
HT (2315 cases in the analysis)
2
(Nagelkerke, r 0.130)
BMI
WHR
Age
Dyslipidaemia (2315 cases in the analysis)
2
WHR
(Nagelkerke, r 0.140)
BMI
Smoking
Age
WC
Women
DM (5608 cases in the analysis)
2
(Nagelkerke, r 0.031)

Odds ratio (95% CI)

1.295 (1.056, 1.589)

1.291 (1.215, 1.372)
1.216 (1.063, 1.390)
1.958 (1.772, 2.163)
1.363 (1.225, 1.517)
1.136 (1.085, 1.188)
1.476
1.457
1.271
1.141
0.884

(1.336,
(1.345,
(1.127,
(1.103,
(0.853,

1.631)
1.577)
1.433)
1.179)
0.916)

WHR
BMI
Age

1.472 (1.346, 1.609)

1.171 (1.123, 1.221)
1.124 (1.033, 1.222)

BMI
Age
WC
Dyslipidaemia0 (5608 cases in the analysis)
BMI
(Nagelkerke, r2 0.142)
Age
WTH
WHR
WC

1.967 (1.845, 2.097)

1.445 (1.397, 1.494)
1.030 (1.001, 1.060)

HT (5608 cases in the analysis)

(Nagelkerke, r2 0.206)

Only variables in the model are represented.

1.444
1.253
1.119
0.943
0.888

(1.376,
(1.224,
(1.049,
(0.898,
(0.850,

1.515)
1.283)
1.193)
0.990)
0.928)

Prediction of type 2 DM, hypertension and dyslipidaemia

A Berber et al

1797

Figure 1 Likelihood ratio of having DM, HT and dyslipidaemia at various BMI, WC, WHR and WTH ratio cut-offs in men.

Table 4 summarizes the independent relationships

between these four anthropometric indexes of age and
smoking and the relative risk of having DM, HT or dyslipidaemia using logistic regression analysis. BMI and age were
the parameters included in all models, however, in men with
DM and dyslipidaemia, smoking were included too.
The proportion of variance (Nagelkerke, r2) of the models
accounting for DM, HT, or dyslipidaemia was relatively
small up to 14% in men and 20% in women. This is
compatible with the fact that these diseases are heterogeneous and multifactorial with obesity only part of the underlying causes.

Discussion
PRIT surveys are an effort to describe the prevalence and
incidence of chronic conditions in the population of workers
of Hospital General de Me xico.18,19 The included population
may be consider as representative of middle class urban
population in Mexico City; so the obtained results may not
be readily applied to the rest of the Mexican population, for
instance countryside populations, or populations from other
regions of the country.

In the present study the adjusted prevalence of being

overweight in men and women (BMI > 25) and obesity in
men (BMI > 30) was lower than that reported by Arroyo in
the Mexican National Chronic Conditions Survey (60.7 and
25.1% in men and 56.3 and 14.9% in women).20
Similarly, the Mexican National Chronic Conditions
Survey disclosed a higher prevalence of DM and HT (7.2
and 26.6% in both genders).21 This may be ascribed to the
inclusion of populations from the north region of Mexico
with higher BMI and risks in the National Survey.
The studied population had a higher prevalence of being
overweight but a lower prevalence of obesity with regard to
the presence of chronic conditions when compared with the
Third National Health Examination Survey (overweight
39.4% in men and 24.7% in women; obesity 19.9% in men
and 24.9% in women).22 The occurrence of DM and HT was
also lower than that found in the American survey (DM 7.8%
in both genders,23 hypertension 25.9% in men and 22.2% in
women,24 probably because different age groups were not
considered.
According to WHO recommendations for Caucasian
populations, the thresholds for increased risk of comorbidities are BMI 25 kg=m2 men and women, WC 94 and 80 cm in
International Journal of Obesity

Prediction of type 2 DM, hypertension and dyslipidaemia

A Berber et al

1798

Figure 2

Likelihood ratio of having DM, HT and dyslipidaemia at various BMI, WC, WHR and WTH ratio cut-offs in women.

men and women, respectively, and WHIR 1.0 and 0.85 in

men and women, respectively.4 On the other hand, according to the recommendations for Asians, the thresholds for
increased risk of comorbidities are BMI 23 kg=m2 in men and
women, and WC 90 and 80 cm in men and women, respectively.25 The cut-off levels found in this study were lower
than those in the Caucasian population and similar to those
of the Asians.
WHR cut-offs in men and women (0.90 and 0.85 respectively) in this study are similar to those recommended
by WHO guidelines (0.90 and 0.80 in men and women
respectively).4
The cut-off levels for WC, WHR and WTH corresponded
to the inflexion points in the likelihood ratio graphs. In the
case of BMI likelihood ratio graphs, we found a significant
increase in the risk of chronic conditions with low BMI
levels. This is similar to the BMI likelihood ratios reported
by Ko.9 These findings may be ascribed to the short stature in
both populations.25
Logistic regression analyses disclosed that only BMI and
age were included in all the models to predict chronic
conditions. WC did not have significant influence in the
variability of the studied diseases. The proportion of variance
International Journal of Obesity

of the models accounting for the chronic conditions is very

small. Chronic diseases are heterogeneous and multifactorial
and besides anthropometric measurements other factors,
such as hereditary factors and menopausal state must be
considered.
The validity of these cut-off values must be assessed
prospectively in new participants coming from the same
population.
These results are important to implement public health
policies to detect chronic conditions.

Conclusion
Although these results may not be readily applied to the rest
of the Mexican population or to other Hispanic populations,
they point to the necessity for similar studies with large
randomized samples to find the cut-off levels for chronic
conditions in different populations.

References
1 Pi-Sunyer FX. Medical hazards of obesity. Ann Intern Med 1993;
119(7 Pt 2): 655 660.

Prediction of type 2 DM, hypertension and dyslipidaemia

A Berber et al
2 Higgins M, Kannel W, Garrison R, Pinsky J, Stokes J. Hazards of
obesity the Framingham experience. Acta Med Scand 1988; 723
(Suppl): S23 S36.
3 Must A, Spadano J, Coakley EH, Field AE, Colditz G, Dietz WH.
The disease burden associated with overweight and obesity. JAMA
1999; 282: 1523 1529.
4 World Health Organization. Obesity. Preventing and managing the
global epidemic. Report of a WHO Consultation on Obesity,
Geneva, 3 5 June. World Health Organization: Geneva; 1998.
5 US National Institutes of Health. Clinical guidelines for the identification, evaluation, and treatment of overweight and obesity in adults.
National Institutes of Health: Bethesda, MD; 1998.
6 Okosun IS, Liao Y, Rotimi CN, Choi S, Cooper RS. Predictive
values of waist circumference for dyslipidaemia, type 2 diabetes
and hypertension in overweight White, Black, and Hispanic
American adults. J Clin Epidemiol 2000; 53: 401 408.
7 Molarius A, Seidell JC, Sans S, Tuomilehto J, Kuulasmaa K. Varying sensitivity of waist action levels to identify subjects with
overweight or obesity in 19 populations of the WHO MONICA
Project. J Clin Epidemiol 1999; 52: 1213 1224.
8 Okosun IS, Rotimi CN, Forrester TE, Fraser H, Osotimehin B,
Muna WF, Cooper RS. Predictive value of abdominal obesity
cutoff points for hypertension in blacks from west African and
Caribbean island nations. Int J Obes Relat Metab Disord 2000; 24:
180 186.
9 Ko GT, Chan JC, Cockram CS, Woo J. Prediction of hypertension,
diabetes, dyslipidaemia or albuminuria using simple anthropometric indexes in Hong Kong Chinese. Int J Obes Relat Metab
Disord 1999; 23: 1136 1142.
10 Report of the Expert Committee on the Diagnosis and
Classification of Diabetes Mellitus. Diabetes Care 1997; 20:
1183 1197.
11 American Heart Association. Science advisory guide to primary
prevention of cardiovascular diseases: a statement for healthcare
professionals from the Task Force on Risk Reduction. Circulation
1997; 95: 2329 2331.
12 Summary of the second report of the National Cholesterol
Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult
treatment Panel II). JAMA 1993; 269: 3015 3023.
13 www.inegi.gob.mx/

14 Van der Schouw YT, Verbeek AL, Ruijs JH. ROC curves for the
initial assessment of new diagnostic tests. Family Pract 1992; 9:
506 511.
15 Altman DG. Diagnostic tests. In: Altman DG, Machin D, Bryant
TN, Gardner MJ (eds). Statistics with confidence, 2nd edn. BMJ
Books: London; 2000. pp 105 119.
16 Han TS, van Leer EM, Seidell JC, Lean ME. Waist circumference as a
screening tool for cardiovascular risk factors: evaluation of receiver
operating characteristics (ROC). Obes Res 1996; 4: 533 547.
17 Simel DL, Samsa GP, Matchar DB. Likelihood ratios with confidence: sample size estimation for diagnostic test studies. J Clin
Epidemiol 1991; 44: 763 770.
o S,
18 Fangha nel-Salmo n G, Sa nchez-Reyes L, Arellano-Montan
Va ldez-Liaz E, Chavira Lo pez J, Rasco n-Pacheco RA. Prevalencia
de factores de Riesgo de enfermedad coronoaria en trabajadores
del Hospital General de Me xico. Salud Publica (Mex) 1997; 39:
427 432.
19 Fangha nel G, Sa nchez-Reyes L, Berber A, Go mez-Santos R. Evolution of the prevalence of obesity in the workers of a General
Hospital in Mexico. Obes Res 2001; 9 268 273.
20 Arroyo P, Loria A, Ferna ndez V, Flegal KM, Kuri-Morales P, Olaiz
G, Tapia-Conyer R. Prevalence of pre-obesity and obesity in urban
adult Mexicans in comparison with other large surveys. Obes Res
2000; 8: 179 185.
21 Encuesta Nacional de Enfermedades Cro nicas. Epidemiologa SSA
1993.
22 Flegal KM, Carroll MD, Kuczmarski RJ, Johnson CL. Overweight
and obesity in the United States: prevalence and trends, 1960
1994. Int J Obes Relat Metab Disord 1998; 22: 39 47.
23 Harris MI, Flegal KM, Cowie CC, Eberhardt MS, Goldstein DE,
Little RR, Wiedmeyer HM, Byrd-Holt DD. Prevalence of diabetes,
impaired fasting glucose, and impaired glucose tolerance in U.S.
adults. Third National Health and Nutrition Examination Survey,
1988 1994. Diabetes Care 1998; 21: 518 524.
24 Burt VL, Whelton P, Roccella EJ, Brown C, Cutler JA, Higgins M,
Horan MJ, Labarthe D. Prevalence of hypertension in the US
adult population. Results from the Third National Health and
Nutrition Examination Survey, 1988 1991. Hypertension 1995;
25: 305 313.
25 World Health Organization. The Asia-Pacific perspective: redefining
obesity. World Health Organization: Geneva; 2000.

1799

International Journal of Obesity