Clinical Nutrition xxx (2016) 1e6

Contents lists available at ScienceDirect

Clinical Nutrition
journal homepage: http://www.elsevier.com/locate/clnu

Original article

Associations of serum fibroblast growth factor 23 levels with obesity

and visceral fat accumulation
Xiang Hu a, b, c, d, e, 1, Xiaojing Ma a, b, c, d, e, 1, Yuqi Luo a, b, c, d, e,
Yiting Xu a, b, c, d, e, Qin Xiong a, b, c, d, e, Xiaoping Pan a, b, c, d, e,
Yunfeng Xiao f, Yuqian Bao a, b, c, d, e, *, Weiping Jia a, b, c, d, e
a
Department of Endocrinology and Metabolism, Shanghai Jiao Tong University Affiliated Sixth People's Hospital, Shanghai 200233, China
b
Shanghai Clinical Center for Diabetes, Shanghai 200233, China
c
Shanghai Key Clinical Center for Metabolic Disease, Shanghai 200233, China
d
Shanghai Diabetes Institute, Shanghai 200233, China
e
Shanghai Key Laboratory of Diabetes Mellitus, Shanghai 200233, China
f
Department of Radiology, Shanghai Jiao Tong University Affiliated Sixth People's Hospital, Shanghai 200233, China

a r t i c l e i n f o s u m m a r y

Article history: Background & aims: Recent findings raise the possibility that fibroblast growth factor (FGF) 23 may be
Received 24 August 2016 related with fat contents and distribution. The present study aimed to elucidate the associations of serum
Accepted 11 December 2016 FGF23 levels with the fat contents and distribution.
Methods: A total of 1599 normoglycemic individuals with preserved kidney function were enrolled.
Keywords: Serum levels of intact FGF23 were detected using a sandwich enzyme-linked immunosorbent assay.
Fibroblast growth factor 23
Overweight/obesity was defined by a body mass index of !25.0 kg/m2. Visceral fat area (VFA) was
Abdominal obesity
assessed by magnetic resonance imaging, and abdominal obesity was defined as VFA !80 cm2.
Visceral fat area
Results: Among the study population of 597 men, 411 premenopausal women, and 591 postmenopausal
women, serum FGF23 levels were significantly increased in participants with overweight/obesity and
abdominal obesity, respectively (both P < 0.001). Each one-unit increase in VFA was associated with a
0.028-pg/mL increase in serum FGF23 levels in men (P ¼ 0.001) and a 0.051-pg/mL increase in serum
FGF23 levels in postmenopausal women (P < 0.001). Whether in the presence of overweight/obesity or
not, the presence of abdominal obesity was independently associated with the increase in serum FGF23
levels in men and postmenopausal women (all P < 0.05).
Conclusions: Serum FGF23 levels are elevated in obese individuals, especially those with abdominal
obesity. The independent associations between the presence of abdominal obesity and the increase in
serum FGF23 levels in specific groups suggest that serum FGF23 levels may indicate the risk of metabolic
and cardiovascular disease in men and postmenopausal women.
© 2016 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.

1. Introduction
Abbreviations: BMI, body mass index; Ca, serum calcium; CRP, C-reactive pro-
tein; CVD, cardiovascular disease; DBP, diastolic blood pressure; FGF, fibroblast
Obesity can occur when the imbalance of energy intake and
growth factor; FINS, fasting serum insulin; FPG, fasting plasma glucose; HbA1c,
glycated hemoglobin A1c; HDL-c, high density lipoprotein cholesterol; HOMA-IR, expenditure persists for a prolonged period. With the imbalance in
homeostasis model assessment-insulin resistance; LDL-c, low density lipoprotein energy metabolism, excess body fat accumulates and further pre-
cholesterol; MRI, magnetic resonance imaging; SBP, systolic blood pressure; Scr, disposes individuals to developing metabolism disturbance and
serum creatinine; SFA, subcutaneous fat area; TC, total cholesterol; TG, triglyceride;
cardiovascular disease (CVD) [1]. Through the secretion of a variety
T2DM, type 2 diabetes mellitus; VFA, visceral fat area; W, waist circumference;
2hPG, 2-hour plasma glucose.
of adipokines, adipose tissue contributes to metabolic alterations,
* Corresponding author. Department of Endocrinology and Metabolism, Shanghai cardiovascular morbidity, and bone health. The functions of adipose
Jiao Tong University Affiliated Sixth People's Hospital, 600 Yishan Road, Shanghai tissue vary with the fat distribution. The accumulation of visceral
200233, China. Fax: þ86 21 64368031. fat, instead of subcutaneous fat, is largely responsible for the effects
E-mail address: yqbao@sjtu.edu.cn (Y. Bao).
1 of excess adipose tissue on CVD, diabetes, and mortality [2].
These two authors contributed equally to this work.

http://dx.doi.org/10.1016/j.clnu.2016.12.010
0261-5614/© 2016 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.

Please cite this article in press as: Hu X, et al., Associations of serum fibroblast growth factor 23 levels with obesity and visceral fat accumulation,
Clinical Nutrition (2016), http://dx.doi.org/10.1016/j.clnu.2016.12.010
2 X. Hu et al. / Clinical Nutrition xxx (2016) 1e6
Emerging evidence indicates that the skeleton is involved in ho-
meostatic processes of energy balance and adipose metabolism,
and the adipokines and bone-derived factors compose part of a
bone-adipose axis, leading to interaction between skeleton and fat
to establish a homeostatic feedback system [1,3].
As a bone-derived factor, fibroblast growth factor (FGF) 23 has
similar structure and biological function to the other members of
endocrine FGF subfamily, such as FGF19 and FGF21 [4]. Previous
studies have provided evidence supporting the associations of
FGF19 and FGF21 with overall and central obesity [5,6]. Functioning
as a circulating hormone, FGF23 is generally acknowledged as a
central regulator in mineral homeostasis [4]. Recent findings raise
the possibility that FGF23 also may be related with fat contents and
distribution [7e11]. Animal studies have revealed FGF23-deficient
mice are characterized by reduced fat content and alterations in
fat distribution [7]. Additionally, several clinical studies have re-
ported that circulating FGF23 levels are elevated in the obese
population and associated with waist circumference (W) [8e11].
Although W is a simple and recognized index used to identify
central obesity, it is not an accurate indicator of visceral fat accu-
mulation and distribution. The present study aimed to elucidate the
relationships between serum FGF23 levels and the fat contents and
distribution, with the precise indexes of abdominal fat (visceral fat
area [VFA] and subcutaneous fat area [SFA]) assessed by magnetic
resonance imaging (MRI). In order to eliminate the influences of
hyperglycemia and kidney function on serum FGF23 levels, the
study population included only normoglycemic individuals with
preserved kidney function. Based on known gender differences in
fat content and distribution [12], we divided the participants into
three subgroups of men, premenopausal women, and post-
menopausal women for analysis.
2. Materials and methods
2.1. Subjects
Normoglycemic individuals with normal renal function (esti-
mated glomerular filtration rate !60 mL/min/1.73 m2) were
selected from the Shanghai Obesity Study cohort [3]. Individuals
with impaired glucose regulation or diabetes, diagnosed on the
basis of the 1999 World Health Organization criteria [13], were
excluded. Further exclusion of individuals with hepatic dysfunc-
tion, hyperthyroidism or hypothyroidism, hypercalcemia or hypo-
calcemia, acute infection, a history of tumors, psychiatric disease, or
cardiovascular disease as well as those being treated for dyslipi-
demia or hypertension, undergoing replacement therapy with
systemic corticosteroids or thyroxine, and with a history of taking
medication known to influence bone, gave a final sample size of
1599. The participants provided complete clinical data and finished
a standardized questionnaire covering their disease history, medi-
cation usage, and menopause status.
The present study was performed in consistent with the
Declaration of Helsinki. The Ethics Committee of Shanghai Jiao Tong
University Affiliated Sixth People's Hospital preapproved this study.
Participants provided written informed consent before their
participation.
2.2. Anthropometric and biochemical assessments
Body mass index (BMI, calculated as weight [kg]/height2 [m2]),
W, and resting blood pressure (BP) were measured using standard
techniques [3]. Participants were examined in the morning after an
overnight fast to collect fasting blood samples. After them, each
participant underwent a 75-g oral glucose tolerance test.
Biochemical variables were detected as previously described [3],
Please cite this article in press as: Hu X, et al., Associations of serum fibroblast growth factor 23 levels with obesity and visceral fat accumulation,
Clinical Nutrition (2016), http://dx.doi.org/10.1016/j.clnu.2016.12.010
including fasting plasma glucose (FPG), 2-hour plasma glucose
(2hPG), glycated hemoglobin (HbA1c), fasting serum insulin (FINS),
lipid profiles (total cholesterol [TC], triglycerides [TG], low-density
lipoprotein cholesterol [LDL-c], high-density lipoprotein choles-
terol [HDL-c]), C-reactive protein (CRP), and serum calcium (Ca). A
sandwich enzyme-linked immunosorbent assay kit (Kainos kit,
Kainos Laboratories, Inc., Tokyo, Japan) was used to assess serum
intact FGF23 levels. The intra-assay and inter-assay coefficients of
variations for the assay were 5.6% and 8.2%, respectively. Insulin
resistance index (homeostasis model assessment-insulin resistance
[HOMA-IR]) was calculated as previously described [3].
2.3. Measurement of abdominal fat distribution
A clinical MRI scanner (Archiva 3.0T; Philips Medical Systems,
Amsterdam, The Netherlands) was applied to detecting visceral and
subcutaneous fat. Average values for the VFA and SFA were calcu-
lated with Image analysis software (slice-O-matic, version 4.2;
Tomovision Inc., Montreal, Quebec, Canada) according to a protocol
described in our previous study [14].
2.4. Definitions of obesity
Based on the 1998 World Health Organization criteria [15],
overweight/obesity was defined as a BMI of !25.0 kg/m2. A VFA of
80 cm2 was applied as the cutoff for abdominal obesity [16]. The
study population was divided into the following four groups on the
basis of fat distribution: neither overweight/obesity (normal body
weight) nor abdominal obesity (BMI <25.0 kg/m2 and VFA
<80 cm2), isolated abdominal obesity (BMI <25.0 kg/m2 and VFA
!80 cm2), isolated overweight/obesity (BMI !25.0 kg/m2 and VFA
<80 cm2), and overweight/obesity and abdominal obesity (BMI
!25.0 kg/m2 and VFA !80 cm2).
2.5. Statistical analysis
Statistical analyses were conducted using the SPSS 16.0 statis-
tical software package (SPSS Inc., Chicago, IL, USA). Data with
normal or skewed distribution, which were determined by the one-
sample KolmogoroveSmirnov test, were presented as
mean ± standard deviation or median with interquartile range. To
carry out inter-group comparison, the unpaired Student's t-test was
used for continuous data with normal distribution, ManneWhitney
U-test was use for those with skewed distribution, and the Chi-
squared test was used for categorical variables. Spearman correla-
tion analyses and linear regression analyses were applied to
exploring the associations of serum FGF23 levels and indexes of
obesity. The multiple logistic stepwise regression analysis was
conducted to identify factors independently correlated with the
presence of abdominal obesity. P values were reported as two-
tailed, which were considered significant statistically when they
were less than 0.05.
3. Results
3.1. Clinical characteristics of participants
The 1599 included study participants had a median age of 52.78
(45.98e57.84) years (overall age range: 20e78 years) and included
597 men, 411 premenopausal women, and 591 postmenopausal
women. Among normal weight participants, age, BMI, W, systolic BP,
diastolic BP, FPG, 2hPG, FINS, HOMA-IR, TC (P ¼ 0.01), TG, LDL-c, CRP,
VFA, and SFA were significantly greater (all P < 0.01) for those with
abdominal obesity compared with those without abdominal obesity.
In addition, the proportion of women and the HDL-c were lower
 X. Hu et al. / Clinical Nutrition xxx (2016) 1e6
(both P < 0.001) in normal weight participants with abdominal
obesity than those in subjects without abdominal obesity. For par-
ticipants with overweight/obesity, age, BMI (P ¼ 0.012), W, 2hPG,
FINS, HOMA-IR, TG, CRP, and VFA were significantly greater for those
with abdominal obesity compared with those without abdominal
obesity (all P < 0.01), and the proportion of women and HDL-c again
were lower (both P < 0.001) in individuals with abdominal obesity
than in subjects without abdominal obesity. No differences in the
other variables were observed between participants with and
without abdominal obesity in the corresponding category of over-
weight/obesity (all P > 0.05; Table 1).
3.2. Comparison of serum FGF23 levels
Serum FGF23 level for the entire study population was 28.30
(22.90e34.20) pg/mL, and levels of serum FGF23 were significantly
greater in participants with overweight/obesity and in those with
abdominal obesity, respectively (both P < 0.001). Even among
normal weight participants, those with isolated abdominal obesity
exhibited higher serum levels of FGF23 (P < 0.001).
Serum FGF23 levels were not different between men and
women (P ¼ 0.705). However, upon considering a difference in
menopause status in women, serum FGF23 levels were higher in
men and postmenopausal women than premenopausal women
(both P < 0.01). Serum FGF23 levels did not differ significantly
between men and postmenopausal women (P ¼ 0.128).
No matter with or without overweight/obesity, men with
abdominal obesity had higher serum FGF23 levels compared to
those without abdominal obesity (both P < 0.05). However, no
difference in levels of serum FGF23 was found between participants
with and without overweight/obesity, regardless of the presence of
abdominal obesity (both P > 0.05; Fig. 1a). In premenopausal
women, serum FGF23 levels were not higher in those with over-
weight/obesity and/or abdominal obesity (all P > 0.05; Fig. 1b). In
postmenopausal women, serum FGF23 levels were increased in
participants with overweight/obesity and/or abdominal obesity (all
P < 0.05), whereas those with overweight/obesity and abdominal
obesity did not show elevated serum FGF23 levels compared to
those with isolated overweight/obesity or isolated abdominal
obesity (both P > 0.05; Fig. 1c).
3.3. Associations of serum FGF23 levels with indexes of obesity
In men, serum FGF23 levels were associated positively with BMI,
W, VFA, and SFA (all P < 0.05). Each one-unit increase in VFA and
Fig. 1. Comparisons of serum FGF23 levels. *P < 0.05 versus those without overweight/obesity or abdominal obesity; #P < 0.05 versus those with overweight/obesity and abdominal
obesity.
Please cite this article in press as: Hu X, et al., Associations of serum fibroblast growth factor 23 levels with obesity and visceral fat accumulation,
Clinical Nutrition (2016), http://dx.doi.org/10.1016/j.clnu.2016.12.010
3
SFA was associated with a 0.028-pg/mL and a 0.016-pg/mL increase
in serum FGF23 levels, respectively (P ¼ 0.001 and 0.011, respec-
tively). After calculation, a 10% increase in VFA and SFA translated
into a 0.71% and 0.94% increase in the serum FGF23 levels,
respectively.
There was neither relationship of serum FGF23 levels with the
applied indexes of obesity (all P > 0.05) nor an increase in VFA or
SFA with higher serum FGF23 levels (P > 0.05) in premenopausal
women.
In postmenopausal women, serum FGF23 levels were correlated
positively with BMI, W, and VFA (all P < 0.05), but not SFA
(P ¼ 0.274). Each one-unit increase in VFA was associated with a
0.051-pg/mL increase in serum FGF23 levels (P < 0.001), which
translated into a 10% increase in the VFA with a 0.94% increase in
serum FGF23 levels. The SFA did not increase with increasing serum
FGF23 levels (P ¼ 0.152).
3.4. Multivariate stepwise regression analyses on serum FGF23
levels
Multivariate stepwise regression analyses defined serum FGF23
levels as a dependent variable and the subgroup of subjects without
overweight/obesity or abdominal obesity as reference. After
adjustment of age, systolic BP, diastolic BP, HbA1c, HOMA-IR, TG,
HDL-c, LDL-c, and CRP, the results showed that whether in the
presence of overweight/obesity or not, the presence of abdominal
obesity is associated with the increase in serum FGF23 levels in
men and postmenopausal women (all P < 0.05; Table 2). Addi-
tionally, in postmenopausal women, the presence of isolated
overweight/obesity, which were not combined with abdominal
obesity, was also correlated with the increase in serum FGF23 levels
(P ¼ 0.010).
4. Discussion
The present study revealed that for men and postmenopausal
women, serum FGF23 levels were increased significantly in those
with abdominal obesity. For men, the presence of abdominal
obesity contributed independently to an increase in serum FGF23
levels regardless of the presence of overweight/obesity. For post-
menopausal women, however, the presence of overweight/obesity
and abdominal obesity, separately or jointly, contributed inde-
pendently to an increase in serum FGF23 levels. Neither the pres-
ence of overweight/obesity nor abdominal obesity was associated
with serum FGF23 levels in premenopausal women.
4 X. Hu et al. / Clinical Nutrition xxx (2016) 1e6

Table 1
Characteristics of the study participants.

Variables Normal weight Overweight/obesity

VFA <80 cm2 VFA! 80 cm2 VFA <80 cm2 VFA !80 cm2
b
Women (%) 664 (72.73) 149 (46.13) 66 (73.33) 123 (45.05)b
Age (years) 51.92 (44.40e57.07) 55.04 (49.15e60.05)b 48.51 (42.56e56.55) 53.70 (46.93e58.50)b
BMI (kg/m2) 21.33 ± 1.91 23.16 ± 1.32b 26.58 ± 1.61 27.16 ± 1.97a
W (cm) 74.00 (70.00e78.25) 83.00 (79.00e86.00)b 84.75 (80.00e88.00) 91.00 (87.00e95.00)b
SBP (mmHg) 117.33 (107.33e123.33) 120.00 (110.00e129.33)b 120.00 (111.25e129.33) 123.33 (115.33e133.33)
DBP (mmHg) 74.00 (69.33e80.00) 77.33 (70.00e80.67)b 78.67 (70.67e80.17) 79.33 (71.33e83.33)
FPG (mmol/L) 5.06 ± 0.40 5.19 ± 0.39b 5.20 ± 0.41 5.24 ± 0.39
2hPG (mmol/L) 5.80 (5.01e6.56) 6.18 (5.42e6.86)b 5.71 (5.11e6.57) 6.35 (5.37e7.04)b
HbA1c (%) 5.5 (5.3e5.7) 5.5 (5.3e5.7) 5.5 (5.2e5.7) 5.5 (5.3e5.7)
FINS (mU/L) 5.62 (4.08e7.54) 6.98 (5.44e9.65)b 8.02 (5.72e11.16) 9.49 (6.89e12.62)b
HOMA-IR 1.26 (0.89e1.72) 1.60 (1.22e2.27)b 1.79 (1.32e2.57) 2.17 (1.58e3.05)b
TC (mmol/L) 5.01 (4.41e5.73) 5.14 (4.62e5.48)a 4.92 (4.18e5.63) 5.02 (4.38e5.74)
TG (mmol/L) 0.97 (0.72e1.36) 1.48 (1.05e2.06)b 1.10 (0.79e1.39) 1.50 (1.09e2.07)b
HDL-c (mmol/L) 1.57 (1.34e1.80) 1.30 (1.15e1.50)b 1.43 (1.20e1.67) 1.27 (1.09e1.50)b
LDL-c (mmol/L) 3.00 (2.48e3.55) 3.23 (2.74e3.74)b 3.09 (2.57e3.63) 3.20 (2.70e3.70)
CRP (mg/L) 0.41 (0.20e0.79) 0.78 (0.45e1.57)b 0.64 (0.32e1.25) 0.93 (0.55e1.97)b
Ca (mmol/L) 2.33 ± 0.09 2.34 ± 0.09 2.32 ± 0.09 2.33 ± 0.11
VFA (cm2) 50.97 (37.84e63.61) 96.72 (86.72e111.27)b 66.41 (56.76e74.23) 111.92(93.11e135.66)b
SFA (cm2) 155.44 (114.24e189.92) 162.77 (133.21e203.55)b 223.84 (181.96e270.07) 220.28 (178.00e269.14)

Abbreviations: VFA, visceral fat area; BMI, body mass index; W, waist circumference; SBP, systolic blood pressure; DBP, diastolic blood pressure; FPG, fasting plasma glucose;
2hPG, 2-hour plasma glucose; HbA1c, glycated hemoglobin A1c; FINS, fasting serum insulin; HOMA-IR, Homeostasis model assessment-insulin resistance; TC, total choles-
terol; TG, triglyceride; HDL-c, high density lipoprotein cholesterol; LDL-c, low density lipoprotein cholesterol; CRP, C-reactive protein; Ca, Serum calcium; SFA, subcutaneous
fat area.
Data are means ± SD, median (interquartile range).
a
P < 0.05 versus subjects without abdominal obesity (VFA <80 cm2) in the corresponding group of normal weight or overweight/obesity.
b
P < 0.01 versus subjects without abdominal obesity (VFA <80 cm2) in the corresponding group of normal weight or overweight/obesity.

Table 2
Multivariate stepwise regression analyses of serum FGF23 levels.

Subgroups Men Premenopausal women Postmenopausal women

b S.E Standardized b P b S.E Standardized b P b S.E Standardized b P

2
Normal weight VFA< 80 cm Reference Reference Reference
VFA !80 cm2 2.584 0.876 0.139 0.003 0.361 1.546 0.012 0.816 1.787 0.896 0.088 0.047
Overweight/obesity VFA <80 cm2 $1.725 1.801 $0.040 0.339 1.867 1.331 0.071 0.161 4.272 1.654 0.108 0.010
VFA !80 cm2 2.969 1.036 0.153 0.004 1.995 1.463 0.071 0.174 2.437 1.066 0.107 0.023

Abbreviations: FGF23, fibroblast growth factor 23; VFA, visceral fat area.
Independent variables originally included: age, SBP, DBP, HbA1c, HOMA-IR, TG, HDL-c, LDL-c, and CRP.

On the basis of two independent elderly cohorts, a Swedish
study discovered that a 10% increase in the serum FGF23 levels was
accompanied by a 3% increase in weight and BMI as well as in body
fat. Individuals in the highest FGF23 tertile were prone to over-
weight than those in the lowest FGF23 tertile [8]. Another
community-based study of 1261 men with preserved kidney
function revealed an upward trend in BMI from the lowest to the
highest FGF23 quartiles [9]. Individuals with obesity (BMI !30 kg/
m2) displayed significantly higher circulating concentrations of
FGF23 compared with those with a BMI of <23 kg/m2 [9]. A study
conducted in normotensive adolescents without chronic kidney
disease (CKD) also showed elevated circulating levels of FGF23 in
an obese population [10]. Additionally, two studies with small
samples, one performed in 20 obese women selected for bariatric
surgery and the other one performed in 48 obese perimenopausal
women, revealed that participants with obesity displayed higher
circulating FGF23 levels than their corresponding controls [11,17].
Moreover, the studies described above observed a positive cor-
relation between circulating FGF23 levels and central obesity
identified by W and waist-to-hip ratio [8,10]. Mirza et al. [8] re-
ported that serum FGF23 levels could explain 3% of the variation in
trunk fat mass, suggesting that as a circulating protein, FGF23 was
involved in the regulation of body fat content and distribution.
However, Hanks et al. [18] discovered that the association of FGF23
with both BMI and W was modified by the presence of CKD, as
evidenced by the presence of this association in participants
without CKD but not in those with CKD. Another study carried out
in elderly people aged more than 65 years also failed to find a
relationship between elevated circulating levels of FGF23 and the
presence of obesity [19]. Thus, there has been no consensus yet on
the clinical findings.
By selecting normoglycemic individuals with preserved kidney
function from communities, the present study mitigated the im-
pacts of glucose metabolism and kidney function on the circu-
lating levels of FGF23. The results discovered that levels of serum
FGF23 were increased significantly in subjects with overweight/
obesity as well as in those with abdominal obesity. Furthermore, in
those with normal weight (to eliminate the influence of over-
weight/obesity), elevated serum levels of FGF23 were observed in
the presence of isolated abdominal obesity. In line with most
findings described above, the present study suggested that serum
FGF23 levels were not only involved in the increase in total body
fat content but also contributed to ectopic fat deposition in the
visceral area.

Please cite this article in press as: Hu X, et al., Associations of serum fibroblast growth factor 23 levels with obesity and visceral fat accumulation,
Clinical Nutrition (2016), http://dx.doi.org/10.1016/j.clnu.2016.12.010
 X. Hu et al. / Clinical Nutrition xxx (2016) 1e6
Notably, despite of the lack of evidence supporting a gender
difference in serum FGF23 levels, Mirza et al. [8] demonstrated a
consistently stronger association between serum FGF23 levels and
fat mass in men than in women. Similarly, neither the previous
study nor the present study revealed that serum FGF23 levels were
different between men and women. In consideration of the
menopause status, however, the present study showed that
compared to premenopausal women, men and postmenopausal
women showed higher levels of serum FGF23. Thus, only in men
and postmenopausal women can FGF23 be used as a serum
biomarker indicating the presence of abdominal obesity. Addi-
tionally, serum FGF23 levels contributed predominantly to visceral
fat accumulation, and less so to the fat deposition over the total
body fat content in men and the subcutaneous fat content in
postmenopausal women.
The potential mechanism by which circulating FGF23 levels
were associated with the increase in body fat content, particularly
in the visceral area, remains to be determined. Streicher et al. [7]
revealed that mice in which FGF23 was ablated featured distinct
alterations in body composition (increased ash and protein content
and reduced fat content), whereas body weight and body compo-
sition in FGF23/vitamin D receptor double-mutant mice were
comparable with those of wild-type mice. This basic research
suggested that a vitamin D receptor-dependent mechanism un-
derlies the regulation of FGF23 in the fat accumulation and distri-
bution. In addition, animal studies also demonstrated the
stimulatory effects of adipokines on FGF23 expression in bone [20],
suggesting a feedback effect of adipose tissue on serum FGF23
levels.
It remains elusive why differences exist in the serum FGF23
levels and their association with fat content and distribution be-
tween men, premenopausal women, and postmenopausal women.
Given that FGF23 is believed to be a bone-derived hormone, the
regulators of bone metabolism such as estrogens, which balance
bone remodeling directly and indirectly, may exert impacts upon
FGF23 synthesis [4]. Therefore, the differences in estrogen levels
between men and women as well as those between women before
and after menopause could serve as possible explanations. More-
over, the function and deposition of adipose tissue differed by
gender and menopause status. Men were more likely to gather fat
in visceral depots, whereas women accrued more fat in the sub-
cutaneous depot before menopause. After menopause, the fat
deposition shifted to the visceral depot progressively [12]. This
variation in fat distribution may influence the strength of the cor-
relation between serum FGF23 levels and VFA.
Limited by the cross-sectional design, this study was unable to
clarify the causal relationships of serum FGF23 levels with fat
content and distribution. Perspective studies should be conducted
to generalize these findings and make them more comprehensive.
In conclusion, serum FGF23 levels were increased significantly
in an obese population, especially in those with abdominal obesity.
The relationships of serum FGF23 levels with body content and
distribution differed by gender and menopause status. The inde-
pendent associations between the presence of abdominal obesity
and the increase in serum FGF23 levels in specific groups suggest
that FGF23 could serve as a serum biomarker, indicating the
negative effects of visceral fat accumulation on metabolic and
cardiovascular disease in men and postmenopausal women.
Funding sources
This work was funded by 973 Program of China
(2013CB530606), Grant from Shanghai Health and Family Planning
Commission (2013ZYJB1001), and Translational Medicine
Please cite this article in press as: Hu X, et al., Associations of serum fibroblast growth factor 23 levels with obesity and visceral fat accumulation,
Clinical Nutrition (2016), http://dx.doi.org/10.1016/j.clnu.2016.12.010
5
Innovation Foundation of School of Medicine Shanghai Jiao Tong
University (15ZH2010 and 15ZH4006).
Conflict of interest statement
The authors have nothing to disclose.
Statement of authorship
The present study was designed by Yuqian Bao and Weiping Jia.
Xiaojing Ma, Xiang Hu, Yuqi Luo, Yiting Xu, and Qin Xiong collected
the data. Xiang Hu carried out the statistical analyses and
composed the paper. Xiaoping Pan conducted the FGF23 assess-
ment. Yunfeng Xiao assessed the abdominal fat by magnetic reso-
nance imaging. Xiang Hu, Xiaojing Ma, Yuqian Bao, and Weiping Jia
revised the paper and made contributions to the discussion. Xiang
Hu and Xiaojing Ma had equal contribution to this paper and were
the guarantors.
Acknowledgments
We appreciate the helps from all the staff of Baoshan, Gong-
hexin, Tianmuxi, and Daning communities. We are also much
grateful to all participants for their dedication in data collection and
laboratory assessments.
References
[1] Greco EA, Lenzi A, Migliaccio S. The obesity of bone. Ther Adv Endocrinol
Metab 2015;6:273e86.
[2] Sheu Y, Cauley JA. The role of bone marrow and visceral fat on bone meta-
bolism. Curr Osteoporos Rep 2011;9:67e75.
[3] Bao Y, Ma X, Yang R, Wang F, Hao Y, Dou J, et al. Inverse relationship between
serum osteocalcin levels and visceral fat area in Chinese men. J Clin Endocrinol
Metab 2013;98:345e51.
[4] Martin A, David V, Quarles LD. Regulation and function of the FGF23/klotho
endocrine pathways. Physiol Rev 2012;92:131e55.
[5] Go!mez-Ambrosi J, Gallego-Escuredo JM, Catala !n V, Rodríguez A, Domingo P,
Moncada R, et al. FGF19 and FGF21 serum concentrations in human obesity
and type 2 diabetes behave differently after diet- or surgically-induced weight
loss. Clin Nutr 2016. May 4. pii: S0261-5614(16)30076-0.
[6] Mai K, Schwarz F, Bobbert T, Andres J, Assmann A, Pfeiffer AF, et al. Relation
between fibroblast growth factor-21, adiposity, metabolism, and weight
reduction. Metabolism 2011;60:306e11.
[7] Streicher C, Zeitz U, Andrukhova O, Rupprecht A, Pohl E, Larsson TE, et al.
Long-term Fgf23 deficiency does not influence aging, glucose homeostasis, or
fat metabolism in mice with a nonfunctioning vitamin D receptor. Endocri-
nology 2012;153:1795e805.
[8] Mirza MA, Alsio € J, Hammarstedt A, Erben RG, Michae €lsson K, Tivesten A, et al.
Circulating fibroblast growth Factor-23 is associated with fat mass and dys-
lipidemia in two independent cohorts of elderly individuals. Arterioscler
Thromb Vasc Biol 2011;31:219e27.
[9] Gutie!rrez OM, Wolf M, Taylor EN. Fibroblast growth factor 23, cardiovascular
disease risk factors, and phosphorus intake in the health professionals follow-
up study. Clin J Am Soc Nephrol 2011;6:2871e8.
[10] Ali FN, Falkner B, Gidding SS, Price HE, Keith SW, Langman CB. Fibroblast
growth factor-23 in obese, normotensive adolescents is associated with
adverse cardiac structure. J Pediatr 2014;165:738e43.
[11] Grethen E, Hill KM, Jones R, Cacucci BM, Gupta CE, Acton A, et al. Serum leptin,
parathyroid hormone, 1,25-dihydroxyvitamin D, fibroblast growth factor 23,
bone alkaline phosphatase, and sclerostin relationships in obesity. J Clin
Endocrinol Metab 2012;97:1655e62.
[12] Palmer BF, Clegg DJ. The sexual dimorphism of obesity. Mol Cell Endocrinol
2015;402:113e9.
[13] World Health Organization. Department of Noncommunicable Disease Sur-
veillance. Definition, diagnosis and classification of diabetes mellitus and its
complication. Report of a WHO consultation. Part 1: diagnosis and classifi-
cation of diabetes mellitus. Geneva: World Health Organization; 1999.
[14] Wang Y, Ma X, Zhou M, Zong W, Zhang L, Hao Y, et al. Contribution of visceral
fat accumulation to carotid intima-media thickness in a Chinese population.
Int J Obes (Lond) 2012;36:1203e8.
[15] World Health Organization. Obesity: preventing and managing the global
epidemic. Report of a WHO consultation. World Health Organ Tech Rep Ser
2000;894:1e253.
[16] Bao Y, Lu J, Wang C, Yang M, Li H, Zhang X, et al. Optimal waist circumference
cutoffs for abdominal obesity in Chinese. Atherosclerosis 2008;201:378e84.
6 X. Hu et al. / Clinical Nutrition xxx (2016) 1e6

[17] Holecki M, Chudek J, Wie˛ cek A, Titz-Bober M, Duława J. The serum level of
fibroblast growth factor-23 and calcium-phosphate homeostasis in obese
perimenopausal women. Int J Endocrinol 2011;2011:707126.
[18] Hanks LJ, Casazza K, Judd SE, Jenny N, Gutie
!rrez OM. Associations of fibroblast
growth factor-23 with markers of inflammation, insulin resistance and
obesity in adults. PLoS One 2015;10:e0122885.
[19] Holecki M, Chudek J, Owczarek A, Olszanecka-Glinianowicz M, Bozentowicz-
_
Wikarek M, Duława J, et al. Inflammation but not obesity or insulin resistance
is associated with increased plasma fibroblast growth factor 23 concentration
in the elderly. Clin Endocrinol (Oxf) 2015;82:900e9.
[20] Tsuji K, Maeda T, Kawane T, Matsunuma A, Horiuchi N. Leptin stimulates
fibroblast growth factor 23 expression in bone and suppresses renal
1alpha,25-dihydroxyvitamin D3 synthesis in leptin-deficient mice. J Bone Min
Res 2010;25:1711e23.

Please cite this article in press as: Hu X, et al., Associations of serum fibroblast growth factor 23 levels with obesity and visceral fat accumulation,
Clinical Nutrition (2016), http://dx.doi.org/10.1016/j.clnu.2016.12.010