nature publishing group articles

Epidemiology

Body Adiposity and Type 2 Diabetes:

Increased Risk With a High Body Fat
Percentage Even Having a Normal BMI
Javier Gómez-Ambrosi1,2, Camilo Silva2,3, Juan C. Galofré3, Javier Escalada3, Silvia Santos3,
María J. Gil2,4, Victor Valentí5, Fernando Rotellar2,5, Beatriz Ramírez1,2, Javier Salvador2,3
and Gema Frühbeck1–3

Obesity is the major risk factor for the development of prediabetes and type 2 diabetes. BMI is widely used as a
surrogate measure of obesity, but underestimates the prevalence of obesity, defined as an excess of body fat. We
assessed the presence of impaired glucose tolerance or impaired fasting glucose (both considered together as
prediabetes) or type 2 diabetes in relation to the criteria used for the diagnosis of obesity using BMI as compared to
body fat percentage (BF%). We performed a cross-sectional study including 4,828 (587 lean, 1,320 overweight, and
2,921 obese classified according to BMI) white subjects (66% females), aged 18–80 years. BMI, BF% determined by
air-displacement plethysmography (ADP) and conventional blood markers of glucose metabolism and lipid profile
were measured. We found a higher than expected number of subjects with prediabetes or type 2 diabetes in the
obese category according to BF% when the sample was globally analyzed (P < 0.0001) and in the lean BMI-classified
subjects (P < 0.0001), but not in the overweight or obese-classified individuals. Importantly, BF% was significantly
higher in lean (by BMI) women with prediabetes or type 2 diabetes as compared to those with normoglycemia (NG)
(35.5 ± 7.0 vs. 30.3 ± 7.7%, P < 0.0001), whereas no differences were observed for BMI. Similarly, increased BF%
was found in lean BMI-classified men with prediabetes or type 2 diabetes (25.2 ± 9.0 vs. 19.9 ± 8.0%, P = 0.008),
exhibiting no differences in BMI or waist circumference. In conclusion, assessing BF% may help to diagnose
disturbed glucose tolerance beyond information provided by BMI and waist circumference in particular in male
subjects with BMI <25 kg/m2 and over the age of 40.

Obesity (2011) 19, 1439–1444. doi:10.1038/oby.2011.36

Introduction circumference or the waist-to-hip ratio being better estimators

The prevalence of obesity is increasing dramatically worldwide
(1). Obesity is defined medically as a state of increased adi-
pose tissue of sufficient magnitude to produce adverse health
consequences and is associated with increased morbidity and
mortality. In this sense, obesity increases the risk, among other
diseases, of type 2 diabetes, cardiovascular disease, fatty liver,
sleep-breathing disorders, and certain forms of cancer (2).
BMI has been traditionally used as a surrogate measure of
adiposity and is the most frequently used diagnostic tool in
the current classification system for obesity. However, it is
well known that body fat distribution exerts a strong influ-
ence on the development of glucose intolerance and type 2
diabetes (3,4), with anthropometric measures such as the waist
of the obesity-associated type 2 diabetes risk than BMI (5).
Epidemiological studies analyzing the impact of the body fat
percentage (BF%) on the levels of metabolic risk factors are
scarce (6). It has been suggested that BF% is a better indicator
of other obesity comorbidities such as coronary heart disease
risk than waist circumference (7). Furthermore, BF% has been
associated with all-cause and cardiovascular mortality (8,9)
with high fat mass being more strongly associated with mor-
tality risk than BMI (10). Therefore, it may be crucial to dis-
criminate the clinical usefulness of measuring BF% to estimate
the obesity-associated type 2 diabetes risk.
The aim of this study was to analyze the impact of body
adiposity on the presence of prediabetes and type 2 diabetes.

1
Metabolic Research Laboratory, Clínica Universidad de Navarra, Pamplona, Spain; 2CIBER Fisiopatología de la Obesidad y Nutrición (CIBERobn), Instituto
de Salud Carlos III (ISCIII), Pamplona, Spain; 3Department of Endocrinology, Clínica Universidad de Navarra, Pamplona, Spain; 4Department of Biochemistry,
Clínica Universidad de Navarra, Pamplona, Spain; 5Department of Surgery, Clínica Universidad de Navarra, Pamplona, Spain. Correspondence: Gema Frühbeck
(gfruhbeck@unav.es)
Received 17 August 2010; accepted 23 January 2011; published online 10 March 2011. doi:10.1038/oby.2011.36

obesity | VOLUME 19 NUMBER 7 | july 2011 1439

articles
Epidemiology
Therefore, we conducted a cross-sectional study evaluating the
relation between the criteria used for the diagnosis of obesity
and the presence of prediabetes or type 2 diabetes and com-
paring BMI, BF%, and waist circumference between normo-
glycemic subjects and individuals with prediabetes or type 2
diabetes.
Methods and Procedures
Study sample
We conducted a cross-sectional analysis of 4,828 white subjects (3,186
females/1,642 males), aged 18–80 years including patients visiting the
Departments of Endocrinology and Surgery of the University Clinic of
Navarra. The study was performed to evaluate the presence of predia-
betes or type 2 diabetes in relation to the criteria used for the diagnosis
of obesity. Patients with chronic renal or liver disease, signs of acute
inflammation, or taking antidiabetic medication or any drug poten-
tially influencing insulin production were excluded. The experimental
design was approved, from an ethical and scientific standpoint, by the
Hospital’s Ethical Committee responsible for research and informed
consent was obtained from all subjects.
Anthropometric and body composition measurements
BMI was calculated as weight in kg divided by the square of height
in meters (kg/m2). Waist circumference was measured at the mid-
point between the iliac crest and the rib cage. Body density was esti-
mated by air-displacement plethysmography (ADP; Bod-Pod; Life
Measurements, Concord, CA) with percentage of body fat calculated
using the Siri equation. ADP is a validated and reproducible alternative
to the gold standard hydrodensitometry (11). It uses the pressure–vol-
ume relationship to estimate volume and density and has been shown
to predict fat mass and fat-free mass more accurately than dual-energy
X-ray absorptiometry and bioimpedance using hydrodensitometry
as reference method (11–14). The most frequently used cutoff points
for BF% defining overweight (20.1–24.9% for men and 30.1–34.9%
for women) and obesity (≥25% for men and ≥35% for women) were
applied (15–17).
Laboratory procedures
Blood samples were collected after an overnight fast. Plasma determi-
nations were analyzed as previously described (18). In 1,834 patients
(648 male and 1,186 female) plasma glucose concentrations after a 75-g
oral glucose tolerance test were available. Normoglycemia (NG) was
defined as having a glucose level <7.8 mmol/l 2 h after the oral glucose
tolerance test or a fasting plasma glucose concentration <5.6 mmol/l.
Prediabetes was defined as exhibiting a glucose concentration between
7.8 and 11.1 mmol/l 2 h after the oral glucose tolerance test (impaired
glucose tolerance) or exhibiting a fasting plasma glucose between 5.6
and 6.9 mmol/l (impaired fasting glucose). Type 2 diabetes was defined
as having a glycemia ≥11.2 mmol/l 2 h after the oral glucose tolerance
Table 1 Distribution of subjects with normoglycemia, prediabetes, and type 2 diabetes according to BMI and BF%
All Lean (BMI)
NG PreDiab T2D NG PreDiab
Body fat% (n = 4,828) (n = 587)
Lean 274 (92) 16 (5) 8 (3) 227 (95) 8 (3)
Overweight 310 (84) 47 (13) 14 (4) 140 (88) 14 (9)
Obese 2858 (69) 931 (22) 370 (9) 149 (79) 33 (18)
P <0.0001 <0.0001
Data presented as number of subjects (%). Cutoffs for defining overweight and obesity were ≥25 and 30 kg/m2, respectively, for BMI, and ≥20 and 25% in men and 30
and 35% in women, for BF%.
NG, nomoglycemia; PreDiab, prediabetes; T2D, type 2 diabetes.
a
Fisher–Freeman–Halton exact P value; P values are χ2 based.
1440
test or a fasting plasma glucose ≥7.0 mmol/l. This classification was
defined following the criteria of the ADA (19). Plasma glucose was ana-
lyzed by an automated analyzer (Roche/HitachiModular P800; Roche
Diagnostics, Basel, Switzerland) as previously described (20,21). Insulin
was measured by means of an enzyme-amplified chemiluminescence
assay (IMMULITE; Diagnostic Products, Los Angeles, CA). An indirect
measure of insulin resistance was calculated by using the homeostatic
model assessment (HOMA) (22). Total cholesterol and triglyceride con-
centrations were determined by enzymatic spectrophotometric meth-
ods (Roche, Basel, Switzerland). High-density lipoprotein-cholesterol
was quantified by a colorimetric method in a Beckman Synchron CX
analyzer (Beckman Instruments, Bucks, UK). Low-density lipoprotein-
cholesterol was calculated by the Friedewald formula.
Statistical analysis
Data are presented as mean ± s.d. Differences between groups were
analyzed by two-tailed unpaired Student’s t-tests or ANOVA followed
by Scheffé’s tests as appropriate. Differences in proportions or number
of subjects with NG, impaired glucose tolerance, and type 2 diabetes
by body fat category, globally and segregated by BMI category were
assessed by using contingency tests (χ2 tests). The calculations were
performed using the SPSS 15.0.1 (SPSS, Chicago, IL). A P value <0.05
was considered statistically significant.
Results
With the aim of evaluating the relation between the criteria
used for the diagnosis of obesity and the presence of predia-
betes or type 2 diabetes, we compared anthropometric vari-
ables in 1,642 males and 3,186 females (Table 1). We found a
higher than expected number of subjects with prediabetes or
type 2 diabetes in the obese category according to BF% when
the sample was globally analyzed (P < 0.0001) or in the lean
BMI-classified subjects (P < 0.0001), but not in the overweight
or obese BMI-classified individuals. In the global analysis, 370
individuals with type 2 diabetes exhibited a BF% within the
obese range (the expected number was 338), whereas only 8
subjects had a BF% within the lean range (the expected number
was 24). Similarly, 931 individuals with prediabetes exhibited a
BF% within the obese range (while the expected number being
856), whereas only 16 subjects had a BF% within the lean range
(the expected number was 61). A comparable distribution was
obtained when only lean BMI-classified individuals were con-
sidered, but not when only overweight or only obese BMI-
classified patients were included (Table 1).
The contingency table reported on Table 1 gives additional
and interesting information regarding the misclassification of
Overweight (BMI) Obese (BMI)
T2D NG PreDiab T2D NG PreDiab T2D
(n = 1,320) (n = 2,921)
5 (2) 44 (80) 8 (15) 3 (5) 3 (100) 0 (0) 0 (0)
5 (3) 152 (82) 26 (14) 8 (4) 18 (69) 7 (27) 1 (4)
6 (3) 837 (78) 182 (17) 60 (6) 1872 (65) 716 (25) 304 (11)
0.779 0.780a
VOLUME 19 NUMBER 7 | july 2011 | www.obesityjournal.org
 Table 2 Comparison of clinical characteristics of subjects with normoglycemia, prediabetes, and type 2 diabetes according to BMI and gender
Lean Overweight Obese
Men NG PreDiab T2D P NG PreDiab T2D P NG PreDiab T2D P
n 67 11 11 275 100 54 646 312 166
Age (years) 37.8 ± 12.8 48.4 ± 19.1 58.3 ± 12.7c <0.001 42.9 ± 12.6 53.2 ± 11.5c 56.3 ± 10.0c <0.001 40.7 ± 12.7 50.0 ± 12.8c 55.2 ± 10.2c,f <0.001
2
BMI (kg/m ) 22.4 ± 2.4 23.3 ± 1.6 23.4 ± 1.7 0.238 27.8 ± 1.4 28.1 ± 1.4 27.9 ± 1.5 0.274 36.4 ± 5.7 37.3 ± 6.3 37.2 ± 6.3 0.053
Body fat (%) 19.9 ± 8.0 26.8 ± 9.4a 23.6 ± 6.2 0.020 28.5 ± 6.4 29.2 ± 5.8 29.2 ± 5.0 0.487 38.5 ± 7.3 40.0 ± 7.4b 39.8 ± 6.1 0.004
b a b
Waist (cm) 85 ± 7 87 ± 9 89 ± 6 0.300 97 ± 6 99 ± 6 99 ± 6 0.005 116 ± 12 119 ± 12 120 ± 14 <0.001
SBP (mm Hg) 112 ± 12 109 ± 8 120 ± 16 0.240 117 ± 13 126 ± 14c 131 ± 14c <0.001 125 ± 14 130 ± 15c 134 ± 17c,d <0.001
b a e
DBP (mm Hg) 69 ± 7 69 ± 7 75 ± 9 0.080 74 ± 8 77 ± 8 77 ± 7 <0.001 79 ± 10 81 ± 10 82 ± 10 0.001
Glucose (mmol/l) 4.80 ± 0.51 5.68 ± 0.42a 8.77 ± 2.74c,f <0.001 4.99 ± 0.46 5.88 ± 0.58c 8.70 ± 2.60c,f <0.001 5.05 ± 0.50 5.84 ± 0.62c 8.67 ± 2.94c,f <0.001

obesity | VOLUME 19 NUMBER 7 | july 2011

Insulin (pmol/l) 45 ± 28 61 ± 13 87 ± 28a 0.016 69 ± 45 78 ± 38 99 ± 89a 0.012 104 ± 71 138 ± 84c 164 ± 173c,d <0.001
c,e c,f c c,f
HOMA 1.40 ± 0.92 2.07 ± 0.36 4.99 ± 2.32 <0.001 2.26 ± 1.72 2.84 ± 1.40 5.31 ± 5.37 <0.001 3.39 ± 2.45 5.11 ± 3.38 8.60 ± 11.19 <0.001
Triglycerides (mmol/l) 1.02 ± 0.53 1.43 ± 0.87 1.51 ± 1.19 0.058 1.20 ± 0.57 1.43 ± 0.77 2.39 ± 3.90c,e <0.001 1.46 ± 0.87 1.76 ± 1.44b 2.02 ± 1.91c <0.001
Cholesterol (mmol/l) 4.4 ± 0.9 4.6 ± 1.0 5.2 ± 1.4 0.056 4.9 ± 1.0 4.7 ± 1.0 5.0 ± 1.7 0.227 4.9 ± 1.1 5.1 ± 1.1 5.1 ± 1.2 0.270
LDL chol. (mmol/l) 2.7 ± 0.7 2.6 ± 0.7 3.3 ± 1.0 0.100 3.1 ± 0.9 2.9 ± 1.0 2.8 ± 1.0 0.030 3.2 ± 1.0 3.2 ± 1.0 3.0 ± 1.0 0.365
HDL chol. (mmol/l) 1.3 ± 0.3 1.5 ± 0.8 1.2 ± 0.4 0.425 1.2 ± 0.3 1.2 ± 0.3 1.2 ± 0.3 0.831 1.1 ± 0.3 1.1 ± 0.3 1.1 ± 0.3 0.582
Lean Overweight Obese
Women NG PreDiab T2D P NG PreDiab T2D P NG PreDiab T2D P
n 449 44 5 758 116 17 1,247 411 139
Age (years) 37.4 ± 13.7 52.3 ± 12.7c 47.0 ± 6.1 <0.001 41.7 ± 13.6 49.2 ± 11.6c 60.2 ± 9.2c,e <0.001 42.9 ± 13.4 50.0 ± 12.5c 55.2 ± 10.5c,f <0.001
2 c c,f
BMI (kg/m ) 21.9 ± 2.6 22.6 ± 2.0 20.8 ± 4.8 0.157 27.5 ± 1.4 27.7 ± 1.4 27.6 ± 1.0 0.545 36.5 ± 5.6 38.7 ± 6.8 40.9 ± 8.5 <0.001
Body fat (%) 30.3 ± 7.7 35.9 ± 6.4c 31.5 ± 11.6 <0.001 39.9 ± 4.9 41.3 ± 4.7b 41.9 ± 4.5 0.005 49.1 ± 5.8 50.9 ± 5.6c 51.7 ± 5.8c <0.001
c a b c c,f
Waist (cm) 75 ± 8 81 ± 9 78 ± 12 <0.001 88 ± 8 90 ± 6 95 ± 9 <0.001 107 ± 12 111 ± 13 117 ± 15 <0.001
SBP (mm Hg) 105 ± 14 113 ± 15b 110 ± 22 0.008 111 ± 13 116 ± 13b 126 ± 16c,e <0.001 120 ± 15 128 ± 17c 136 ± 21c,f <0.001
b a c c,d
DBP (mm Hg) 66 ± 8 69 ± 8 67 ± 11 0.097 69 ± 8 73 ± 8 75 ± 9 <0.001 75 ± 9 79 ± 11 82 ± 11 <0.001
c c,f c c,f c c,f
Glucose (mmol/l) 4.75 ± 0.45 5.64 ± 0.69 10.30 ± 6.87 <0.001 4.83 ± 0.45 5.67 ± 0.52 7.20 ± 1.73 <0.001 4.94 ± 0.49 5.78 ± 0.56 8.19 ± 2.58 <0.001
Insulin (pmol/l) 43 ± 26 52 ± 39 38 ± 12 0.292 59 ± 63 67 ± 37 111 ± 87b 0.009 88 ± 59 115 ± 69c 143 ± 88c,f <0.001
b a c,f c c,f
HOMA 1.32 ± 0.82 1.86 ± 1.63 1.70 ± 0.57 0.023 1.83 ± 1.94 2.42 ± 1.44 5.33 ± 5.00 <0.001 2.80 ± 1.97 4.27 ± 2.74 7.27 ± 5.53 <0.001
Triglycerides (mmol/l) 0.79 ± 0.30 0.98 ± 0.40c 0.81 ± 0.21 <0.001 0.94 ± 0.58 1.28 ± 0.65c 1.72 ± 1.08c,d <0.001 1.15 ± 0.51 1.40 ± 0.66a 2.79 ± 6.34c,f <0.001
c c c c,e
Cholesterol (mmol/l) 4.8 ± 0.9 5.3 ± 1.1 5.0 ± 1.8 0.003 4.9 ± 1.1 5.3 ± 1.3 5.3 ± 1.2 0.001 4.9 ± 1.1 5.3 ± 1.1 5.7 ± 1.8 <0.001
LDL chol. (mmol/l) 2.8 ± 0.9 3.2 ± 0.9b 3.3 ± 1.5 0.013 2.9 ± 1.0 3.2 ± 1.1a 3.0 ± 0.8 0.096 3.1 ± 0.9 3.4 ± 0.9c 3.4 ± 0.9c <0.001
b,f
HDL chol. (mmol/l) 1.6 ± 0.4 1.7 ± 0.5 1.4 ± 0.3 0.284 1.5 ± 0.4 1.6 ± 0.4 1.5 ± 0.6 0.628 1.3 ± 0.4 1.3 ± 0.3 1.2 ± 0.3 0.003
Data presented as mean ± s.d. Differences between groups were analyzed by ANOVA followed by Scheffé’s tests.
chol., cholesterol; DBP, diastolic blood pressure; HOMA, homeostatic model assessment; NG, normoglycemia; PreDiab, prediabetes; SBP, systolic blood pressure; T2D, type 2 diabetes.
a
P < 0.05, bP < 0.01, and cP < 0.001 vs. NG. dP < 0.05, eP < 0.01 and fP < 0.001 vs. PreDiab.

1441
Epidemiology
articles
articles
Epidemiology
Men
BMI
30 50
P = 0.091
40
25
30
kg/m 2
20
%
20
15
10
10 0 60
NG PreDiab/T2D
Women
BMI
30 60
P = 0.216
20 40
kg/m 2
%
10 20
0 0
NG PreDiab/T2D
Figure 1  Comparison of BMI, body fat percentage, and waist circumference between normoglycemic subjects and individuals with prediabetes or
type 2 diabetes. Only subjects with a BMI classified as lean (<25 kg/m2) were included in the analysis. Data presented as mean ± s.d. Differences
between groups were analyzed by Student’s t-tests. NG, normoglycemic; PreDiab, prediabetes; T2D, type 2 diabetes.
obesity when BMI instead of BF% is used. This misclassifica-
tion is evident finding that 32% (188 out of 587) of lean and
82% (1,079 from 1,320) of overweight individuals according to
BMI are in fact obese according to BF%.
In lean men (according to BMI) BF% was significantly differ-
ent between normoglycemic, prediabetic, and type 2 diabetic
groups (P = 0.020), with a significant increase in the predia-
betic group as compared to the normoglycemic one (26.8 ±
9.4 vs. 19.9 ± 8.0%) not reaching statistical significance in the
type 2 diabetic group although a tendency toward an increase
was evident (23.6 ± 6.2%). No differences were observed for
BMI (P = 0.238) or waist circumference values (P = 0.300)
(Table 2). In BMI-overweight men no differences were found
for BF% (P = 0.487) or BMI (P = 0.274), but waist circumfer-
ence was significantly different between groups (P = 0.005). In
obese men (according to BMI), BF% (P = 0.004) and waist (P <
0.001) were significantly different between groups, whereas no
differences were observed for BMI (P = 0.053). In lean women
(according to BMI) the BF%, as well as the waist circumfer-
ence, were higher in patients with prediabetes or type 2 diabe-
tes (P < 0.001 for both), whereas no differences for BMI were
observed (P = 0.157). In BMI-overweight women statistically
significant differences were observed for BF% (P = 0.005) and
waist circumference (P < 0.001), whereas no differences were
observed as regards BMI (P = 0.545). Finally, in obese women
(according to BMI) BMI, BF%, and waist were significantly
different (P < 0.001 for all) between groups (Table 2).
Differences between normoglycemic, prediabetic, and type
2 diabetic groups in other relevant cardiometabolic variables,
such as blood pressure or lipid profile, were more patent in
overweight and obese subjects than in lean individuals accord-
ing to BMI. The differences were also more evident in women
than in men, in particular regarding the lipid profile (Table 2).
1442
Body fat Waist
110
P = 0.008 P = 0.139
100
90
cm
80
70
NG PreDiab/T2D NG PreDiab/T2D
Body fat Waist
110
P < 0.0001 P < 0.0001
100
90
cm
80
70
60
50
NG PreDiab/T2D NG PreDiab/T2D
Finally, in order to get a clearer effect lean-BMI classified
patients with prediabetes and type 2 diabetes were grouped
together and BMI, BF%, and waist circumference were com-
pared to the values of the subjects with NG. BF% (P < 0.0001)
and waist circumference (P < 0.0001) were significantly higher
in lean (by BMI) women with prediabetes or type 2 diabetes
as compared to those with NG, whereas no differences were
observed for BMI (P = 0.216). Similarly, increased BF% (P =
0.008) was found in lean BMI-classified men with prediabetes
or type 2 diabetes, exhibiting no differences in BMI (P = 0.091)
or waist circumference (P = 0.139) (Figure 1).
A consistent finding is the statistically significant (Table 2)
older age of type 2 diabetes as compared to normoglycemic
individuals across all weight categories for both men and
women. As regards prediabetes this is also the case except for
lean men. In this sense, all lean-BMI classified male subjects
with type 2 diabetes and over 92% of patients of any other
group including both males and females were over 40 years
old (with the exception of lean-BMI classified female subjects
where 80% of patients were over 40). Since age is a well-known
risk factor for the development of insulin resistance, we per-
formed a further analysis adjusting for this factor. This analysis
showed that after adjusting for age no differences regarding
BMI, BF%, or waist circumference were observed in lean-BMI
classified men between prediabetes or type 2 diabetes as com-
pared to those with NG, whereas only BF% was significantly
different in lean-BMI classified women (P < 0.05).
Discussion
Obesity is the major risk factor for the development of predia-
betes and type 2 diabetes and represents an excess of body fat,
with the amount of this excess correlating with comorbidity
development (23). The present study provides evidence that
VOLUME 19 NUMBER 7 | july 2011 | www.obesityjournal.org

BF% should be measured when estimating the risk of prediabe-
tes and type 2 diabetes even in subjects with a BMI <25 kg/m2,
in particular over 40 years of age and especially in men.
In spite of the fact that excess adiposity underlies the increased
risk of type 2 diabetes observed in obesity, large epidemiological
studies analyzing the relation between BF% and prediabetes and
type 2 diabetes are scarce. The few studies analyzing the influ-
ence of BF% on metabolic risk factor levels have focused more
on the influence of body fat distribution than on the effect of
increased adiposity itself (7,17,24) or have been performed only
in volunteers with BMIs within the lean range (9). The present
study emphasizes that anthropometric indicators of body fat
distribution such as waist circumference or waist-to-hip ratio
are important when evaluating the risk of a subject regardless
of its body weight (Table 2). Particularly, in overweight or obese
subjects according to BMI where waist circumference is higher
in the prediabetes/type 2 diabetes groups. However, our data
further indicate that the actual body fat amount is also playing
a key role in the development of insulin resistance. This find-
ing has an important clinical implication given the fact that the
ADA recommends screening for prediabetes by health-care pro-
viders at 3-year interval if individuals are >45 years old and have
a BMI ≥25 kg/m2 while recommending to take screening into
consideration for age <45 years if BMI is ≥25 kg/m2 and another
risk factor is present (25). Nonetheless, in everyday clinical prac-
tice this recommended screening is infrequent (26). Actually,
our data shows that in lean individuals (especially in men)
waist circumference values can remain within the normal range
while a statistically significant effect of the increased adiposity
on prediabetes/type 2 diabetes prevalence is evident. Therefore,
our study provides scientific explanation to the finding that
body composition may help to understand the metabolic risk in
­“normal-weight” subjects with high ­adiposity (6,27).
Although in some studies waist circumference or waist-to-
hip ratio better predict diabetes risk than BMI or BF% (7,24),
others find that both overall and abdominal adiposity strongly
and similarly predict risk of type 2 diabetes (28,29). Very
recently, Biggs et al. (30) have shown that among older adults,
overall and central adiposity, as well as weight gain during
middle age and after the age of 65 years are associated with risk
of diabetes, with this observation being found even though in
their analysis fat mass in kg rather than BF% was used. Our
data show that waist circumference may be an important fac-
tor in obese-BMI classified subjects, but in lean-BMI classified
individuals BF% may be more determinant than BMI and even
than waist circumference. A plausible explanation for our find-
ings is that waist circumference may not be a good surrogate of
visceral fat in these individuals having been reported that the
waist measurement is more highly correlated to subcutaneous
than to visceral adipose tissue (31). Moreover, adipose tissue
from both depots has been associated with markers of inflam-
mation and oxidative stress (32,33). It is possible that in lean-
BMI classified subjects adipose tissue amount (BF%) is a better
marker of increased obesity-associated risk than indicators of
adiposity distribution such as waist circumference, especially
in men. Furthermore, it has been reported that equivalent
obesity | VOLUME 19 NUMBER 7 | july 2011
articles
Epidemiology
increases in body fat had lesser impact on insulin sensitivity in
women than in men (34), explaining the significant differences
in waist circumference between groups observed in lean BMI-
classified women, needing a higher central obese phenotype to
develop prediabetes.
Age is one of the major risk factors for the development of
insulin resistance and type 2 diabetes (35). The observation
that after adjusting for age differences regarding BMI, BF%, or
waist circumference between prediabetes or type 2 diabetes as
compared to those with NG were only observed in lean-BMI
classified women, confirms the detrimental effect of increased
BF% that takes place with ageing (36).
Another message that can be concluded from our study is
that there is a high degree of misclassification in the diagno-
sis of obesity in clinical practice, which results in the under-
diagnosis of obese patients at risk and, therefore, missed
opportunities to treat this life-threatening condition. This mis-
classification is evident in Table 1 finding that 32% of lean and
82% of overweight subjects are in fact obese according to BF%.
This observation suggests that current BMI cutoffs for diagno-
sis of obesity may need to be reconsidered.
Regarding the potential limitations of the present study
the body fat cutoffs used are those most frequently applied
for whites (15,16,37–40). Although the World Health
Organization establishes that obesity is considered an excess
of BF% it does not clearly state body fat cutoffs because these
values vary according to ethnicity. This study was conducted in
white subjects and would need to be extended to other popu-
lations to determine whether race differences yield different
observations (41–43). The validity of BF% measurement by
ADP according to weight status should be also considered as a
limitation worth of analysis. In this sense, ADP allows to study
patients encompassing a broad spectrum of BMIs from lean to
morbidly obese individuals. As opposed to bioimpedance and
dual-energy X-ray absorptiometry, which are know to under-
estimate in part actual BF%, body density can be accurately
measured by ADP in overweight and obese individuals (13).
In summary, BF% may be more determinant than BMI
and even than waist circumference for prediabetes and type
2 diabetes development, especially in lean subjects classified
by BMI and in males in particular. There is a high degree of
misclassification in the diagnosis of obesity when BMI instead
of BF% is used, which results in the underdiagnosis of obese
patients at risk. It is of particular clinical importance to assess
BF% to diagnose disturbed glucose tolerance beyond infor-
mation provided by BMI and waist circumference and try to
prevent the development of prediabetes and type 2 diabetes
especially in male subjects with BMI <25 kg/m2 and over the
age of 40 years.
Acknowledgments
This study was supported by grants from the FIS-ISCIII (PS09/02330 and
PI09/91029) and the Departments of Health (20/2005 and 31/2009) and
Education of the Gobierno de Navarra. CIBER de Fisiopatología de la
Obesidad y Nutrición (CIBERobn) is an initiative of the ISCIII, Spain. We
thank all the members of the Nutrition Unit, for their technical support in
body composition analysis.
1443
articles
Epidemiology
Disclosure
The authors declared no conflict of interest.
© 2011 The Obesity Society
REFERENCES
1. James WP. The epidemiology of obesity: the size of the problem. J Intern
Med 2008;263:336–352.
2. Haslam DW, James WP. Obesity. Lancet 2005;366:1197–1209.
3. Wahrenberg H, Hertel K, Leijonhufvud BM et al. Use of waist circumference
to predict insulin resistance: retrospective study. BMJ 2005;330:1363–1364.
4. Rodríguez A, Catalán V, Gómez-Ambrosi J, Frühbeck G. Visceral and
subcutaneous adiposity: are both potential therapeutic targets for tackling
the metabolic syndrome? Curr Pharm Des 2007;13:2169–2175.
5. Janiszewski PM, Janssen I, Ross R. Does waist circumference predict
diabetes and cardiovascular disease beyond commonly evaluated
cardiometabolic risk factors? Diabetes Care 2007;30:3105–3109.
6. Segal KR, Dunaif A, Gutin B et al. Body composition, not body weight, is
related to cardiovascular disease risk factors and sex hormone levels in men.
J Clin Invest 1987;80:1050–1055.
7. Dervaux N, Wubuli M, Megnien JL, Chironi G, Simon A. Comparative
associations of adiposity measures with cardiometabolic risk burden in
asymptomatic subjects. Atherosclerosis 2008;201:413–417.
8. Lahmann PH, Lissner L, Gullberg B, Berglund G. A prospective study of
adiposity and all-cause mortality: the Malmö Diet and Cancer Study. Obes
Res 2002;10:361–369.
9. Romero-Corral A, Somers VK, Sierra-Johnson J et al. Normal weight
obesity: a risk factor for cardiometabolic dysregulation and cardiovascular
mortality. Eur Heart J 2010;31:737–746.
10. Heitmann BL, Erikson H, Ellsinger BM, Mikkelsen KL, Larsson B. Mortality
associated with body fat, fat-free mass and body mass index among
60-year-old swedish men-a 22-year follow-up. The study of men born in
1913. Int J Obes Relat Metab Disord 2000;24:33–37.
11. Fields DA, Goran MI, McCrory MA. Body-composition assessment via air-
displacement plethysmography in adults and children: a review. Am J Clin
Nutr 2002;75:453–467.
12. Biaggi RR, Vollman MW, Nies MA et al. Comparison of air-displacement
plethysmography with hydrostatic weighing and bioelectrical impedance
analysis for the assessment of body composition in healthy adults. Am J Clin
Nutr 1999;69:898–903.
13. Ginde SR, Geliebter A, Rubiano F et al. Air displacement plethysmography:
validation in overweight and obese subjects. Obes Res 2005;13:1232–1237.
14. Das SK. Body composition measurement in severe obesity. Curr Opin Clin
Nutr Metab Care 2005;8:602–606.
15. Romero-Corral A, Somers VK, Sierra-Johnson J et al. Diagnostic
performance of body mass index to detect obesity in patients with coronary
artery disease. Eur Heart J 2007;28:2087–2093.
16. Romero-Corral A, Somers VK, Sierra-Johnson J et al. Accuracy of body
mass index in diagnosing obesity in the adult general population. Int J Obes
(Lond) 2008;32:959–966.
17. Bosy-Westphal A, Geisler C, Onur S et al. Value of body fat mass vs
anthropometric obesity indices in the assessment of metabolic risk factors.
Int J Obes (Lond) 2006;30:475–483.
18. Gómez-Ambrosi J, Salvador J, Silva C et al. Increased cardiovascular risk
markers in obesity are associated with body adiposity: role of leptin. Thromb
Haemost 2006;95:991–996.
19. American Diabetes Association. Diagnosis and classification of diabetes
mellitus. Diabetes Care 2010;33:S62–S69.
20. Gómez-Ambrosi J, Salvador J, Rotellar F et al. Increased serum amyloid A
concentrations in morbid obesity decrease after gastric bypass. Obes Surg
2006;16:262–269.
21. Gómez-Ambrosi J, Catalán V, Ramírez B et al. Plasma osteopontin levels
and expression in adipose tissue are increased in obesity. J Clin Endocrinol
Metab 2007;92:3719–3727.
1444
22. Matthews DR, Hosker JP, Rudenski AS et al. Homeostasis model
assessment: insulin resistance and beta-cell function from fasting plasma
glucose and insulin concentrations in man. Diabetologia 1985;28:412–419.
23. Prentice AM, Jebb SA. Beyond body mass index. Obes Rev 2001;2:
141–147.
24. Shen W, Punyanitya M, Chen J et al. Waist circumference correlates with
metabolic syndrome indicators better than percentage fat. Obesity (Silver
Spring) 2006;14:727–736.
25. ADA. Screening for type 2 diabetes. Diabetes Care 2004;27 Suppl 1:
S11–S14.
26. Ealovega MW, Tabaei BP, Brandle M, Burke R, Herman WH. Opportunistic
screening for diabetes in routine clinical practice. Diabetes Care 2004;
27:9–12.
27. Wildman RP, Muntner P, Reynolds K et al. The obese without
cardiometabolic risk factor clustering and the normal weight with
cardiometabolic risk factor clustering: prevalence and correlates of 2
phenotypes among the US population (NHANES 1999-2004). Arch Intern
Med 2008;168:1617–1624.
28. Wang Y, Rimm EB, Stampfer MJ, Willett WC, Hu FB. Comparison of
abdominal adiposity and overall obesity in predicting risk of type 2 diabetes
among men. Am J Clin Nutr 2005;81:555–563.
29. MacKay MF, Haffner SM, Wagenknecht LE, D’Agostino RB Jr, Hanley AJ.
Prediction of type 2 diabetes using alternate anthropometric measures in
a multi-ethnic cohort: the insulin resistance atherosclerosis study. Diabetes
Care 2009;32:956–958.
30. Biggs ML, Mukamal KJ, Luchsinger JA et al. Association between
adiposity in midlife and older age and risk of diabetes in older adults. JAMA
2010;303:2504–2512.
31. Fox CS, Massaro JM, Hoffmann U et al. Abdominal visceral and
subcutaneous adipose tissue compartments: association with metabolic risk
factors in the Framingham Heart Study. Circulation 2007;116:39–48.
32. Pou KM, Massaro JM, Hoffmann U et al. Visceral and subcutaneous
adipose tissue volumes are cross-sectionally related to markers of
inflammation and oxidative stress: the Framingham Heart Study. Circulation
2007;116:1234–1241.
33. Ferrante AW Jr. Obesity-induced inflammation: a metabolic dialogue in the
language of inflammation. J Intern Med 2007;262:408–414.
34. Sierra-Johnson J, Johnson BD, Bailey KR, Turner ST. Relationships between
insulin sensitivity and measures of body fat in asymptomatic men and
women. Obes Res 2004;12:2070–2077.
35. Amati F, Dubé JJ, Coen PM et al. Physical inactivity and obesity underlie the
insulin resistance of aging. Diabetes Care 2009;32:1547–1549.
36. St-Onge MP. Relationship between body composition changes and changes
in physical function and metabolic risk factors in aging. Curr Opin Clin Nutr
Metab Care 2005;8:523–528.
37. Wellens RI, Roche AF, Khamis HJ et al. Relationships between the Body
Mass Index and body composition. Obes Res 1996;4:35–44.
38. Deurenberg P, Andreoli A, Borg P et al. The validity of predicted body fat
percentage from body mass index and from impedance in samples of five
European populations. Eur J Clin Nutr 2001;55:973–979.
39. De Lorenzo A, Deurenberg P, Pietrantuono M et al. How fat is obese? Acta
Diabetol 2003;40 Suppl 1:S254–S257.
40. Agarwal R, Bills JE, Light RP. Diagnosing obesity by body mass index
in chronic kidney disease: an explanation for the “obesity paradox?”.
Hypertension 2010;56:893–900.
41. Deurenberg P, Yap M, van Staveren WA. Body mass index and percent body
fat: a meta analysis among different ethnic groups. Int J Obes Relat Metab
Disord 1998;22:1164–1171.
42. Jackson AS, Stanforth PR, Gagnon J et al. The effect of sex, age and race
on estimating percentage body fat from body mass index: The Heritage
Family Study. Int J Obes Relat Metab Disord 2002;26:789–796.
43. WHO Expert Consultation. Appropriate body-mass index for Asian
populations and its implications for policy and intervention strategies. Lancet
2004;363:157–163.
VOLUME 19 NUMBER 7 | july 2011 | www.obesityjournal.org