Cancer Causes Control

DOI 10.1007/s10552-015-0709-y

REVIEW ARTICLE

Comparison of anthropometric measurements of adiposity

in relation to cancer risk: a systematic review of prospective
studies
Josefine De Ridder1 • Cristina Julián-Almárcegui2,3 • Amy Mullee3 •
Sabina Rinaldi4 • Koen Van Herck1 • German Vicente-Rodrı́guez2,5 •
Inge Huybrechts1,3

Received: 4 June 2015 / Accepted: 22 December 2015

Ó Springer International Publishing Switzerland 2016

Abstract Results Thirteen articles were identified, with two

Purpose In epidemiology, the relationship between
increased adiposity and cancer risk has long been recognized.
However, whether the association is the same for measures of
abdominal or whole body adiposity is unclear. The aim of this
systematic review is to compare cancer risk, associated with
body mass index (BMI), an indicator of whole body adiposity,
with indicators of abdominal adiposity in studies in which
these indicators have been directly measured.
Methods We conducted a systematic search from 1974
(EMBASE) and 1988 (PubMed) to September 2015 with
keywords related to adiposity and cancer. Included studies
were limited to cohort studies reporting directly measured
anthropometry and performing mutually adjusted analyses.
Josefine De Ridder and Cristina Julián-Almárcegui have contributed
equally to this work.
Electronic supplementary material The online version of this
article (doi:10.1007/s10552-015-0709-y) contains supplementary
material, which is available to authorized users.
& Inge Huybrechts
huybrechtsi@iarc.fr
1
Department of Public Health, Faculty of Medicine and Health
Sciences, Ghent University, Ghent, Belgium
2
GENUD ‘‘Growth, Exercise, Nutrition and Development’’
Research Group, University of Zaragoza, Zaragoza, Spain
3
Dietary Exposure Assessment Group (DEX), International
Agency for Research on Cancer (IARC), 150 Cours Albert
Thomas, 69372 Lyon Cedex 08, France
4
Biomarkers Group, International Agency for Research on
Cancer, 150 Cours Albert Thomas, 69372 Lyon Cedex 08,
France
5
Department of Physiatrist and Nursing, Faculty of Health and
Sport Sciences, University of Zaragoza, Huesca, Spain
reporting on breast cancer, three on colorectal cancer, three
on endometrial cancer, two on gastro-oesophageal cancer,
two on renal cancer, one on ovarian cancer, one on bladder
cancer, one on liver and biliary tract cancer and one on
leukaemia. Evidence suggests that abdominal adiposity is a
stronger predictor than whole body adiposity for gastro-
oesophageal, leukaemia and liver and biliary tract cancer in
men and women and for renal cancer in women. Abdom-
inal adiposity was a stronger predictor for bladder and
colorectal cancer in women, while only BMI was a pre-
dictor in men. In contrast, BMI appears to be a stronger
predictor for ovarian cancer. For breast and endometrial
cancer, both measures were predictors for cancer risk in
postmenopausal women.
Conclusions Only few studies used mutually adjusted and
measured anthropometric indicators when studying adi-
posity–cancer associations. Further research investigating
cancer risk and adiposity should include more accurate
non-invasive indicators of body fat deposition and focus on
the understudied cancer types, namely leukaemia, ovarian,
bladder and liver and biliary tract cancer.
Keywords Anthropometry
Adiposity
Neoplasm

Cancer
Introduction
There is strong evidence that the risk of renal, gall bladder,
liver, advanced prostate cancer, ovarian, endometrial,
pancreatic, colorectal, postmenopausal breast and oeso-
phageal cancer increases with greater adiposity [1, 2], and
it is estimated that 481,000 of all new cancer cases in adults

123

worldwide in 2012 were attributable to high body mass
index (BMI; defined as 25 kg/m2 or greater) [3].
BMI intends to be a measure of whole body adiposity,
calculated by dividing the weight in kilograms by the
square of the height in metres, and the World Health
Organisation (WHO) defines overweight as a BMI greater
than or equal to 25 k/m2, and obesity as a BMI greater
than or equal to 30 kg/m2 [4]. However, BMI does not
distinguish between lean and fat mass, or indeed permit
assessment of the distribution of fat mass. Therefore, other
anthropometric indicators (waist circumference (WC), hip
circumference (HC), waist-to-hip ratio (WHR) and waist-
to-height ratio (WtHR)) have been used to identify pat-
terns of fat distribution and are primarily used to indicate
the extent of abdominal adiposity. Although obesity and
adiposity are synonymously used throughout the literature,
it is noteworthy that the association between the accu-
mulation of adipose tissue and cancer risk has been
observed in normal and overweight/obese individuals as
BMI does not fully capture the complex biology of adi-
posity. Therefore, the term ‘‘adiposity’’ will be used to
indicate the accumulation of adipose tissue throughout this
review article.
Systematic reviews have shown both BMI and measures
of abdominal adiposity to be cancer risk factors [1, 2, 5–7].
However, studies with self-reported (instead of objectively
measured) anthropometric measures were included in these
reviews and it is still unclear as to whether BMI, as an
indicator of whole body adiposity, or anthropometric
indicators of abdominal adiposity are better indicators of
adiposity-associated cancer risk.
The present systematic review aims to compare the
cancer risk of BMI and anthropometric indicators of
abdominal adiposity (WC, HC, WHR and WtHR) in
studies in which the anthropometric indicators had been
mutually adjusted and directly measured in adults, split by
cancer type.
Methods
Data sources and search strategy
A systematic search of PubMed and EMBASE from 1974
(EMBASE) and 1988 (PubMed) to September 2015 was
conducted. Our systematic search consisted of keywords
related to obesity and adiposity (‘‘obesity’’, ‘‘adiposity’’,
‘‘intra-abdominal fat’’, ‘‘body fat distribution’’, ‘‘abdominal
fat’’, ‘‘waist-hip ratio’’, ‘‘waist circumference’’, ‘‘body
shape’’ or ‘‘body fatness’’) combined with ‘‘cancer’’,
‘‘neoplasms’’ and ‘‘cohort studies’’. We used MeSH and
Emtree terms to build up a structured search (see Online
Resource 1). In addition, reference lists of other articles
123
Cancer Causes Control
and published systematic reviews on adiposity and cancer
were examined to identify any missing studies.
Inclusion and exclusion criteria
Only cohort or nested case control studies investigating the
difference between BMI and other anthropometric indica-
tors in relation to cancer risk of one or more cancer types
were included. Traditional case–control studies, which are
prone to recall and interviewer bias, were excluded [8]. All
studies were required to include BMI and at least one of the
following anthropometrical indicators of adiposity: WC,
HC, WHR, waist/thigh ratio or WtHR. Only studies in
which the anthropometry had been directly measured were
included, as self-reported height, weight and waist cir-
cumference data may be subject to bias [9, 10].
Studies were excluded that used adiposity measures as
mediators or confounders only; studies on cancer progno-
sis; studies on cancer treatment and survival; studies on
non-cancerous tumours; studies focusing on childhood
cancers; studies in the same cohort with overlapping
results; studies not including both BMI and other anthro-
pometric measurements and when they were not mutually
adjusted in the analyses; studies that were not published as
full-text reports; or studies where the protocol for assess-
ment of anthropometric data was missing. Excluded arti-
cles as well as reasons for exclusion are included in Online
Resource 2.
Data extraction and quality assessment
Quality assessment was based on that of Renehan et al. [11]
and assessed the following study components: potential
confounders taken into account and reporting of protocol
used for adiposity measurements (Online Resource 3).
Potential confounders used in the studies from this sys-
tematic review were compared with those cancer-specific
confounders proposed by IARC World Cancer Report 2014
[12]. Individual study data were extracted and tabulated in
a standardized format (by JDR) and checked for com-
pleteness and accuracy (by CJA). Studies were classified
by cancer type and summarized several study characteris-
tics (Online Resource 4).
Search summary
Two researchers (JDR and CJA) independently examined
titles and abstracts. Inter-reviewer disagreements were
solved by consensus. The electronic database search
retrieved 2,556 records in PubMed and 1,507 records in
EMBASE. After discarding 269 duplicates, a total of 3,794
potentially relevant records remained. Four articles were
identified from reference lists of articles and reviews. After
Cancer Causes Control
screening by title, the number of articles was reduced to
118. After assessing abstracts and full-text articles for
eligibility and performing the quality assessment, 13 arti-
cles were finally retained to be included in the systematic
review (see flow diagram in Fig. 1).
Results
We identified thirteen prospective cohort studies (fourteen
publications) that met our inclusion criteria [13–25].
Breast cancer
Two of the identified studies reported on breast cancer risk
in postmenopausal women [13, 14] (Table 1). No studies
on premenopausal women met the inclusion criteria. When
comparing the highest versus lowest quintile, Kabat et al.
[13] observed a similar significant positive association for
both BMI and WC for postmenopausal women that never
used hormone replacement therapy (HRT). Mellemkjaer
et al. [14] observed no significant association for BMI, WC
and HC in postmenopausal women who had never used
HRT, but observed a 15 % increased risk of breast cancer
per 5 cm increase in HC in women who had ever used
HRT.
Records identified through database
searching
(n=4063)
Records screened by title
(n=3798)
Abstracts and full-text articles assessed for
eligibility
(n=118)
Articles included in quality assessment
(n=13)
Articles included in the systematic review
(n=13)
Fig. 1 Flowchart of study selection
Colorectal cancer
Three of the identified studies reported on colorectal cancer
risk [13, 15, 16] (Table 2). One study reported a significant
positive association between WHR and colon cancer risk in
women only [16]. For women, colon cancer risk was 46 %
higher when comparing the highest and the lowest quintile
of WHR; this association was not significant in men. The
analysis on rectal cancer risk was not mutually adjusted
[16] and was therefore excluded. WC and BMI were both
positively associated with overall colorectal cancer risk in
postmenopausal women, although WC was shown to have
stronger association than BMI (90 vs. 40 % increased risk,
respectively) [13]. Moore et al. [15] found that BMI was
positively associated with colon cancer risk in men only;
the association was not significant in women.
Endometrial cancer
Three of the identified studies reported on measured
anthropometric indicators and endometrial cancer risk in
postmenopausal women [13, 17, 18] (Table 3). In all
studies, BMI showed a strong positive association with
endometrial cancer risk in postmenopausal women [13, 17,
18]. The European Prospective Investigation into cancer
(EPIC) study and the Women’s Health Initiative (WHI)
Additional records identified
through reference lists
(n=4)
Duplicates excluded
(n=269)
Records excluded
(n=3680)
Articles excluded
(n=105)
123
 Cancer Causes Control

Table 1 Summary risks of breast cancer in relation to BMI and adiposity measures after mutual adjustment
Authors Menopausal BMI WC HC
status: HRT use

Kabat et al. [13] PostM F HR (95 % CI) 1.41 (1.31–1.53) § 1.40 (1.29–1.52) §
PostM F: never HR (95 % CI) 1.70 (1.50–1.94) § 1.67 (1.46–1.90) §
HRT users
Mellemkjaer et al. [14] PostM F: never IRR (95 % CI) 1.13 (0.84–1.51) per Non-significant (NR) 1.08 (0.94–1.24) per
HRT users 4 kg/m2 increase 5 cm increase
PostM F: ever IRR (95 % CI) Non-significant Non-significant (NR) 1.15 (1.03–1.28) per
HRT users (NR) 5 cm increase
Significant values in bold letters
NR data not reported, HR hazard ratio, IRR incidence rate ratio, CI confidence interval, F female, PostM postmenopausal, HRT hormone
replacement therapy, WC waist circumference, HC hip circumference, WHR waist/hip ratio, BMI body mass index, § highest versus lowest
quintile

Table 2 Summary risks of colorectal cancer in relation to BMI and adiposity measures after mutual adjustment
Authors Cancer BMI WC WHR
type

Kabat CRC HR 1.44 (1.23–1.68) § in F 1.90 (1.61–2.25) § in F

et al. (95 %
[13] CI)
Moore CC RR 3.0 (1.1–8.3) ‡ 30 kg/m2 versus < 25 kg/ 3.0 (0.86–10.3) Largest versus smallest
et al. (95 % m2 in M; 1.6 (0.67–3.8) C 30 kg/m2 waist category in M; 2.1 (0.63–6.7)
[15] CI) versus \ 25 kg/m2 in F largest versus smallest waist category in
F
Pischon CC RR Non-significant (NR) C 29.4 kg/m2 1.19 (0.80–1.77)
et al. (95 % versus \ 23.6 kg/m2 in M; non-significant § in M; 1.46
[16] CI) (NR) C 29.4 kg/m2 versus \ 23.6 kg/m2 (1.06–2.01) § in
in F F
Significant values in bold letters
NR data not reported, CRC colorectal cancer, CC colon cancer, RC rectal cancer, RR risk ratio, HR hazard ratio, CI confidence interval, M male,
F female, WC waist circumference, HC hip circumference, WHR waist/hip ratio, BMI body mass index, § highest versus lowest quintile

Table 3 Summary risks of endometrial cancer in relation to BMI and adiposity measures after mutual adjustment
Authors Cancer type BMI WC WHR

Friedenreich et al. PreM F HR (95 % 1.55 (0.80–2.98) C 30 kg/m2 2.13

[17] CI) versus \ 25 kg/m2 (0.98–4.63) ¥
PostM F: never HRT HR (95 % 2.39 (1.62–3.53) ‡ 30 kg/m2 1.88
users CI) versus < 25 kg/m2 (1.11–3.19) ¥
Kabat et al. [13] PostM F HR (95 % 2.32 (1.93–2.80) § 2.20
CI) (1.82–2.65) §
PostM F: never HRT HR (95 % 5.84 (4.20–4.62) § 5.42
users CI) (3.86–7.61) §
Reeves et al. [18] PostM F HR (95 % 1.68 (1.33–2.13) ‡ 30 kg/m2 1.12
CI) versus < 25 kg/m2 (0.86–1.47) ¥
Significant values in bold letters
HR hazard ratio, CI confidence interval, F female, PostM postmenopausal, PreM premenopausal, WC waist circumference, HC hip circum-
ference, WHR waist/hip ratio, BMI body mass index, HRT hormone replacement therapy, ¥ highest versus lowest quartile, § highest versus lowest
quintile

123
Cancer Causes Control

study reported an increased risk among postmenopausal
women with a BMI greater than 30 kg/m2 (Table 3). Two
studies also reported a significant positive association of
cancer risk with WC [13] and WHR [17]. However, BMI
was more strongly positively associated with endometrial
cancer than WC or WHR [13, 17]. Neither WHR nor BMI
showed significant associations with endometrial cancer
risk in premenopausal women [17].
Gastro-oesophageal cancer
Two of the identified studies reported on measured
anthropometric indicators and gastro-oesophageal cancer
risk in men and women combined [19, 20] (Table 4).
Steffen et al. [19] studied two main subtypes of oesopha-
geal cancer: squamous cell cancer and adenocarcinoma.
Oesophageal squamous cell cancer risk was inversely
associated with BMI; higher BMI was associated with a
lower risk and positively associated with WC and WHR.
The analyses were further stratified by smoking status. In
smokers (RR 6.17; 1.78–21.3) and non-smokers (RR 7.23;
1.19–43.8), WC was positively associated with oesopha-
geal squamous cell cancer risk. The inverse association of
BMI with oesophageal squamous cell cancer risk was only
observed in non-smokers (RR 0.09; 0.03–0.29) [19]
(Table 4). Oesophageal adenocarcinoma risk was not
related to adiposity. Five years later, Steffen et al. [20]
repeated the same analyses with a larger study sample and
found that WHR was positively associated with the risk of
oesophageal adenocarcinoma, gastric cardia adenocarci-
noma and gastric non-cardia adenocarcinoma (Table 4).
Renal cancer
Two of the identified studies reported on measured
anthropometric indicators and cancer risk in women [13,
21] (Table 5). Luo et al. [21] found that WHR, but not
BMI, was positively associated with renal cancer risk. A
more recent study in the WHI cohort reported that both
BMI and WC were positively associated with renal cancer
risk, with WC showing a stronger association [13].
Ovarian cancer
One of the identified studies reported on measured
anthropometric indicators and ovarian cancer risk [22]
(Table 6). The prospective Shanghai Women’s Health
Study reported a significant positive association between
BMI and epithelial ovarian cancer risk. No associations
were found for WC, HC or WHR [22].
Bladder cancer
One of the identified studies reported on measured
anthropometric indicators and bladder cancer risk [25]
(Table 6). Per 5 cm increase in WC, the risk of bladder
cancer increased by 8 % among women. In men, the risk of
bladder cancer was positively associated with BMI and not

Table 4 Summary risks of gastro-oesophageal cancer in relation to BMI and adiposity measures after mutual adjustment
Authors Cancer BMI WC HC WHR
type

Steffen EAC RR Non-significant (NR) § in M and F 2.73 (0.91–8.19) § in 0.68 (0.26–1.78) § 1.66 (0.71–3.84) § in
et al. [19] (95 % M and F in M and F M and F
CI)
ESCC RR Significant inverse association 6.91 (2.54–18.8) § in 1.09 (0.42–2.87) § 3.12 (1.48–6.54) § in
(95 % (NR) § in M and F M and F in M and F M and F
CI)
Steffen EAC HR 1.19 (0.63–2.22) § in M and F 4.05 (1.85–8.87) § in
et al. [20] (95 % M and F
CI)
GCC HR 0.85 (0.51–1.42) § in M and F 1.95 (1.12–3.38) § in
(95 % M and F
CI)
GNCC HR 0.86 (0.56–1.34) § in M and F 2.05 (1.19–3.52) § in
(95 % M and F
CI)
Significant values in bold letters
NR data not reported, RR risk ratio, HR hazard ratio, CI confidence interval, F female, ESCC squamous cell cancer, EAC oesophageal
adenocarcinoma, GCC gastric cardia adenocarcinoma, GNCC gastric non-cardia adenocarcinoma, WC waist circumference, HC hip circum-
ference, WHR waist/hip ratio, BMI body mass index, § highest versus lowest quintile

123
 Cancer Causes Control

Table 5 Summary risks of renal cancer in relation to BMI and adiposity measures after mutual adjustment
Authors BMI WC WHR

Kabat et al. [13] HR (95 % CI) 1.65 (1.20–2.26) § in F 2.60 (1.87–3.63) § in F

Luo et al. [21] RR (95 % CI) Non-significant (NR) in F 1.6 (1.1–2.3) ¥ in F
Significant values in bold letters
NR data not reported, RR risk ratio, HR hazard ratio, CI confidence interval, WC waist circumference, HC hip circumference, WHR waist/hip
ratio, BMI body mass index, ¥ highest versus lowest quartile, § highest versus lowest quintile

Table 6 Summary risks of ovarian, bladder and liver and biliary tract cancer in relation to BMI and adiposity measures after mutual adjustment
Authors Cancer BMI WC HC WHR
type

Ma et al. Epithelial HR Significant positive association Non-significant (NR) ¥ in F Non- Non-

[22] ovarian (95 % (NR) ‡ 30.00 kg/m2 versus significant significant
cancer CI) 18.50–24.99 kg/m2 in F (NR) ¥ in (NR) ¥ in
F F
Roswall Bladder HR 1.05 (1.01–1.09) per 2 unit BMI 1.01 (0.95–1.08) per 5 cm WC
et al. [25] cancer (95 % increase in M; non-significant per 2 increase in M;
CI) unit BMI increase in F (NR) 1.08 (0.99–1.18) per 5 cm
WC increase in F
Schlesinger Liver and RR 1.04 (0.60–1.83) § in M and F 1.29 (1.13–1.47) per 5 cm
et al. [23] biliary (95 % increase in M and F
tract CI)
cancer
Significant values in bold letters
NR data not reported, RR risk ratio, HR hazard ratio, CI confidence interval, M male, F female, WC waist circumference, HC hip circumference,
WHR waist/hip ratio, BMI body mass index, ¥ highest versus lowest quartile, § highest versus lowest tertile

with WC; an increase in two units of BMI increased the Discussion

cancer risk by 5 % [25].
Liver and biliary tract cancer
One of the identified studies reported on measured
anthropometric indicators and liver and biliary tract cancer
risk [23] (Table 6). Analyses were carried out in men and
women combined, and liver and biliary tract cancer risk
was significantly associated with WC, while BMI was not.
Leukaemia
One of the identified studies reported on measured
anthropometric indicators and leukaemia risk [24]
(Table 7). Leukaemia risk was inversely associated with
BMI and positively associated with WC in males. Acute
myeloid leukaemia risk showed an inverse association with
BMI and a positive association with WC in males. WHR
was positively associated with acute myeloid leukaemia
risk in females.
To the best of our knowledge, this is the first comprehen-
sive systematic review comparing mutually adjusted and
measured anthropometric indicators of adiposity in relation
to cancer risk in all cancer types. Associations between
anthropometric indicators and cancer risk have been shown
to be gender, cancer type and histology (e.g. oesophageal
adenocarcinoma vs oesophageal squamous cell carcinoma)
specific. In summary, our systematic review suggests that
abdominal adiposity, as marked by WC and WHR, shows a
stronger association than whole body adiposity, as marked
by BMI, for gastro-oesophageal, leukaemia, liver and bil-
iary cancer risk in both males and females and for renal
cancer risk only in females (there were no data on men).
For colon and bladder cancer, abdominal adiposity showed
a stronger association for cancer in women and only BMI
was associated in men. In contrast, whole body adiposity
seems to show a stronger association for ovarian cancer
than abdominal adiposity. For breast and endometrial
cancer, both abdominal and whole body adiposity were

123
Cancer Causes Control

Table 7 Summary risks of leukaemia in relation to BMI and adiposity measures after mutual adjustment
Authors Cancer type Cancer subtypes BMI WC WHR

Saberi Lymphoid and All leukaemia RR 0.53 (0.33–0.87) ¥ in M; 1.92 (1.17–3.14) ¥ in M; 1.46 (1.00–2.12) ¥ in M;
et al. myeloid (95 % 1.21 (0.69–2.13) ¥ in F 1.09 (0.65–1.82) ¥ in F 0.93 (0.64–1.34) ¥ in F
[24] leukaemia CI)
Myeloid RR 0.54 (0.25–1.16) ¥ in M; 2.28 (1.03–5.02) ¥ in M; 1.58 (0.87–2.86) ¥ in M;
leukaemia (95 % 0.99 (0.44–2.26) ¥ in F 1.77 (0.83–3.77) ¥ in F 1.48 (0.84–2.60) ¥ in F
CI)
Lymphoid RR 0.55 (0.28–1.09) ¥ in M; 1.68 (0.85–3.29) ¥ in M; 1.37 (0.81–2.33) ¥ in M;
leukaemia (95 % 1.69 (0.74–3.84) ¥ in F 0.74 (0.35–1.59) ¥ in F 0.65 (0.38–1.10) ¥ in F
CI)
Acute myeloid RR 0.29 (0.10–0.87) ¥ in M; 3.22 (1.10–9.45) ¥ in M; 1.92 (0.85–4.33) ¥ in M;
leukaemia (95 % 0.63 (0.20–1.94) ¥ in F 2.52 (0.91–6.97) ¥ in F 2.56 (1.08–6.10) ¥ in F
CI)
Chronic RR 1.08 (0.22–5.37) ¥ in M; 1.06 (0.19–5.82) ¥ in M; 1.41 (0.81–2.44) ¥ in M;
myeloid (95 % 2.74 (0.51–14.82) ¥ in F 0.24 (0.05–1.25) ¥ in F 0.64 (0.36–1.12) ¥ in F
leukaemia CI)
Chronic RR 0.56 (0.28–1.14) ¥ in M; 1.72 (0.85–3.50) ¥ in M; 1.41 (0.81–2.44) ¥ in M;
lymphoid (95 % 1.76 (0.73–4.25) ¥ in F 0.68 (0.30–1.52) ¥ in F 0.64 (0.36–1.12) ¥ in F
leukaemia CI)
Significant values in bold letters
NR data not reported, RR risk ratio, HR hazard ratio, CI confidence interval, M male, F female, WC waist circumference, HC hip circumference,
WHR waist/hip ratio, BMI body mass index, ¥ highest versus lowest quartile

related to cancer risk in postmenopausal women. Hormone
replacement therapy use and smoking status were sug-
gested as possible effect modifiers that may influence
associations between adiposity and breast cancer and
oesophageal squamous cell cancer risk, respectively.
The mechanisms by which adiposity is thought to pro-
mote tumorigenesis are manifold, including both systemic
and local pathways [26–29]. The paracrine and autocrine
influence of adipose tissue on tumorigenesis is less well
investigated, whereas much research has focused on the
systemic effects of adipose tissue, including alterations in
adipokine production, insulin and insulin-like growth fac-
tor pathways, cancer cell signalling and inflammatory
pathways [26–30]. Several possible hypothesizes have
been put forward to explain why abdominal adiposity
indicators may be more strongly associated than BMI with
increased risk of several cancer. First, abdominal adiposity
measured by WC or WHR is mainly attributed to an
increase in visceral adipose tissue compared to BMI, which
is a measure of both lean and adipose tissue [31]. Those
individuals with higher BMI could therefore be those who
are more physically active and fit and present with higher
muscle tissue and not necessarily higher adipose tissue.
Second, adipose tissue in general, but more particularly
visceral fat, is metabolically active and exerts systemic
endocrine effects [29]. Persons who are overweight or
obese as categorized by BMI may have larger amounts of
subcutaneous fat in the hips and legs-fat linked to healthier
metabolic profiles than those with increased abdominal
adiposity [32–34]. However, previous reviews suggest that
further research is needed to determine the specific effect
of excess visceral adipose tissue on tumorigenesis [30].
It is noteworthy that for some cancer types, whole body
adiposity might still be stronger associated with increased
cancer risk than abdominal adiposity when considering the
theory of excess adipose tissue having a local (paracrine or
autocrine) toxic effect stimulating therefore local tumour
development [28]. This theory is increasingly relevant for
the development of breast cancer as the amount of fat in the
breast is proportional to the total body fat mass [28].
Abdominal adiposity, rather than whole body adiposity,
is shown as a strong risk factor of the development of
oesophageal adenocarcinoma, although the low number of
cases in women makes it difficult to evaluate gender dif-
ferences. As reported by Steffen et al. [20], one biological
explanation for this association is via gastro-oesophageal
reflux disease due to enhanced intra-abdominal pressure.
Another study found that those subjects with a higher waist
circumference and higher total abdominal fat quantified by
magnetic resonance imaging exhibited a greater lengthen-
ing of the cardiac mucosa that could lead to increased intra-
abdominal pressure and predispose individuals to adeno-
carcinoma [35].
Smoking status should be considered as possible effect
modifier when interpreting results on gastro-oesophageal
cancer risk. Steffen et al. [19] also reported that BMI was

123

inversely associated with cancer risk in non-smokers,
defined as lifelong non-smokers or former smokers who
had quit at least 10 years ago; however, results should be
interpreted cautiously as the evaluation of effect modifi-
cation by smoking was limited by the small number of
cases among non-smokers (n = 40).
For bladder and colorectal cancer, differences were
observed between males and females. Whole body adipos-
ity, as indicated by BMI, was observed to have a stronger
association than WC with bladder and colorectal cancer in
males; with the inverse observed in females. Greater body
weight is more closely related to abdominal adiposity than
to lower-body obesity in men, but more closely related to
gluteal femoral obesity in women [32]. Assuming that it is
primarily visceral and not subcutaneous or overall adipose
tissue that is involved in tumorigenic processes, body weight
and BMI may not accurately reflect the cancer risk that is
associated with abdominal adipose tissue accumulation, at
least in women. We also should consider that the study
reporting that BMI and not WC was related to colorectal
cancer risk in males [15] was carried out in a smaller sample
size (n = 7566) compared to the other two colorectal
studies (mean n = 256,089) (Online Resource 4).
For ovarian cancer, BMI seems to be a stronger predictor
of cancer risk than anthropometric measures of abdominal
adiposity. Information about menopausal status was not
included, and distinction cannot be made between pre- and
postmenopausal women. For endometrial and breast can-
cer, both whole body and abdominal adiposity are related to
cancer risk in postmenopausal women with BMI reported as
having a slightly stronger risk. Alterations in insulin [27]
and the metabolism of endogenous hormones, including sex
steroids like oestrogens, androgens and progesterone, have
been suggested to play a role in the relationship between
adiposity and both breast and endometrial cancer, respec-
tively [28, 36]. One hypothesis is that increases in circu-
lating oestrogens within adipose tissue are a result of the
enzyme aromatase that converts androgens to oestrogens
[14, 17, 37–40]. Moreover, other hormones such as insulin
and insulin-like growth factors (IGFs) and possibly adipo-
cyte-derived factors such as leptin and adiponectin may play
an important role [36]. However, the epidemiological evi-
dence to date for a role of endogenous hormones and the IGF
axis in cancer development is limited, and the biological
mechanisms may differ considerably with menopausal sta-
tus. In our systematic review, only one study on endometrial
cancer investigated the association between adiposity and
cancer risk in premenopausal women and authors did not
find any association with either abdominal or whole body
adiposity [17].
It is worth noting that comparison of highest versus
lowest results can be problematic as the categorizations or
the ranges of heights, weights and circumference
123
Cancer Causes Control
measurements can differ depending on the population and
depending on variables such as age, sex, reproductive
status and ethnicity [41]. Indeed, differences in the strength
of association between BMI and cancer risk have been
observed across age ranges [42]; however, the studies
included in our systematic review adjusted for several
confounders including age, sex and ethnicity (see online
Resource 3). Also, the possibility of reverse causality is a
limitation as exposure–outcome relation is biased by
weight loss due to pre-existing illness; this is especially
true in cancers that are not normally detected until an
advanced stage. However, anthropometric measures in the
studies included in this review were undertaken at baseline
when individuals were not presenting with disease and
before cancer diagnosis. Indeed, to avoid the possibility of
reverse causation, one study excluded individuals who
presented with cancer within four years of baseline mea-
surements [15]. Other anthropometric factors outside the
scope of this review paper have been associated with
cancer risk, as height and leg length [14, 23, 42, 43]. Data
on some of the cancer types were limited, with only one
study being identified per cancer type for ovarian, bladder,
liver and biliary tract and leukaemia. For renal cancer, data
were only available in women. Future studies are needed to
further clarify the possible associations between abdominal
and whole body adiposity and these cancer types.
Strengths include the broad range of measured anthro-
pometric indicators and the systematic information col-
lected. Only studies that included the protocol used for
adiposity measurements and adjusted for potential con-
founders were included. The majority of the included
studies evaluated multiple confounders including diet fac-
tors, physical activity and smoking. Large cohorts have
been included in the final number of studies, apart from one
(n = 7566), included more than 70,000 subjects. Our sys-
tematic review adds novelty to the current literature and
underlines the necessity of future studies to use the
appropriate methodology for investigating the relationship
between adiposity distribution and cancer risk. Only few
studies were using measured body fat indicators while
mutually adjusting at the same time. In 2007 and further in
2015, the WCRF and IARC elaborated a report showing
the current knowledge about adiposity and liver cancer
risk, including also studies using self-reported anthropo-
metrics [1, 12]. They found limited evidence about the
positive association with BMI and did not investigate the
role of other anthropometric measures [44]. In contrast, we
have found that WC seems to be stronger associated than
BMI with liver cancer risk. Our results agree with Larsson
et al. [7] meta-analyses and the report from WCRF [45]
who found that both abdominal adiposity and whole body
adiposity are associated with colon cancer risk, although
they did not compare the strength of the associations for
Cancer Causes Control
whole body and abdominal adiposity. We have observed
differences between sexes, being stronger the association
of WC than BMI with colon cancer risk in females. They
included self-reported measures in their meta-analyses, and
they also did find that rectal cancer risk was positively
associated with BMI, however, all the studies included in
the literature studying rectal cancer used self-reported
adiposity measures; therefore, we believe that future
studies on rectal cancer should include non-self-reported
adiposity measures to see whether the relationship between
rectal cancer and adiposity varies between whole body and
abdominal adiposity.
Our results are in disagreement with the systematic
review on breast cancer of Harvie et al. [6] who found that
WC and WHR were strongly associated with breast cancer
risk compared with BMI in premenopausal women. In
2010, the report from WCRF [46] observed that greater
BMI probably protected against premenopausal breast
cancer and was a risk for postmenopausal women. We have
not found enough evidence coming from studies using non-
self-reported indicators to conclude this finding in pre-
menopausal women. In contrast, we have found that both
whole body adiposity and abdominal adiposity are associ-
ated with breast cancer in postmenopausal women.
Conclusions
Although limited, the evidence suggests that abdominal
adiposity is a stronger predictor than whole body adiposity
for gastro-oesophageal, leukaemia and liver and biliary
tract cancer risk in both men and women and for renal
cancer risk in women. Abdominal adiposity, as indicated
by WC, was shown to be a stronger predictor for bladder
and colorectal cancer risk in women than BMI, while only
BMI was shown to be a predictor in men. In contrast, BMI
appears to be stronger predictor for ovarian cancer risk than
abdominal adiposity. For breast and endometrial cancer,
both abdominal adiposity and whole body adiposity are
related to cancer risk in postmenopausal women. As
already discussed, anthropometric measures are poor
approximations for understanding the distribution of sub-
cutaneous and visceral adipose tissue, making it hard to
identify the most important adiposity-related risk factors
for tumorigenesis and demonstrating the need for better
methods or techniques to quantify more accurately these
separate fat depots. Further research investigating cancer
risk and adiposity should therefore investigate possibilities
for including more accurate non-invasive indicators of
body fat deposition (e.g. imaging techniques such as
magnetic resonance imaging) and should focus on the
understudied cancer types, namely leukaemia, ovarian,
bladder and liver and biliary tract cancer.
Acknowledgments The contribution of JDR was undertaken during
a traineeship within the framework of the Leonardo da Vinci pro-
gramme (2012-LDV-PLM-214). The contribution of AM was
undertaken during the tenure of an IARC-Ireland Postdoctoral Fel-
lowship from the International Agency for Research on Cancer,
supported by the Irish Cancer Society (ICS).
References
1. World Cancer Research Fund/American Institute for Cancer
Research (2007) Food, nutrition, physical activity, and the pre-
vention of cancer: a global perspective. AICR, Washington
2. World Cancer Research Fund/American Institute for Cancer
Research (2015) World Cancer Research Fund (WCRF) Inter-
national continuous update project. WCRF International. http://
www.wcrf.org/int/research-we-fund/continuous-update-project-
findings-reports. Accessed 19 Nov 2015
3. Arnold M, Pandeya N, Byrnes G et al (2015) Global burden of
cancer attributable to high body-mass index in 2012: a popula-
tion-based study. Lancet Oncol 16:36–46
4. World Health Organization (WHO) (2000) Obesity: preventing
and managing the global epidemic report of WHO consultation.
WHO technical report series 894 World Health Organization
(ed). WHO, Geneva
5. MacInnis RJ, English DR (2006) Body size and composition and
prostate cancer risk: systematic review and meta-regression
analysis. Cancer Causes Control 17:989–1003
6. Harvie M, Hooper L, Howell AH (2003) Central obesity and
breast cancer risk: a systematic review. Obes Rev 4:157–173
7. Larsson SC, Wolk A (2007) Obesity and colon and rectal cancer
risk: a meta-analysis of prospective studies. Am J Clin Nutr
86:556–565
8. Schelsselman JJ (1982) Case control studies: design, conduct,
analysis. Oxford University Press, New York
9. Spencer EA, Appleby PN, Davey GK, Key TJ (2002) Validity of
self-reported height and weight in 4808 EPIC-Oxford partici-
pants. Public Health Nutr 5:561–565
10. Connor Gorber S, Tremblay M, Moher D, Gorber B (2007) A
comparison of direct vs. self-report measures for assessing height,
weight and body mass index: a systematic review. Obes Rev
8:307–326
11. Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen M (2008)
Body-mass index and incidence of cancer: a systematic review
and meta-analysis of prospective observational studies. Lancet
371:569–578
12. International Agency for Research on Cancer (IARC) 2014
World Cancer Report 2014. Stewart BW, Wild CP (eds). Lyon
13. Kabat GC, Xue X, Kamensky V et al (2015) Risk of breast,
endometrial, colorectal, and renal cancers in postmenopausal
women in association with a body shape index and other
anthropometric measures. Cancer Causes Control 26:219–229
14. Mellemkjaer L, Bigaard J, Tjonneland A et al (2006) Body
composition and breast cancer in postmenopausal women: a
Danish prospective cohort study. Obesity (Silver Spring)
14:1854–1862
15. Moore LL, Bradlee ML, Singer MR et al (2004) BMI and waist
circumference as predictors of lifetime colon cancer risk in Fram-
ingham Study adults. Int J Obes Relat Metab Disord 28:559–567
16. Pischon T, Lahmann PH, Boeing H et al (2006) Body size and
risk of colon and rectal cancer in the European Prospective
Investigation Into Cancer and Nutrition (EPIC). J Natl Cancer
Inst 98:920–931
123

17. Friedenreich C, Cust A, Lahmann PH et al (2007) Anthropo-
metric factors and risk of endometrial cancer: the European
prospective investigation into cancer and nutrition. Cancer Cau-
ses Control 18:399–413
18. Reeves KW, Carter GC, Rodabough RJ et al (2011) Obesity in
relation to endometrial cancer risk and disease characteristics in
the Women’s Health Initiative. Gynecol Oncol 121:376–382
19. Steffen A, Schulze MB, Pischon T et al (2009) Anthropometry
and esophageal cancer risk in the European prospective investi-
gation into cancer and nutrition. Cancer Epidemiol Biomarkers
Prev 18:2079–2089
20. Steffen A, Huerta JM, Weiderpass E et al (2015) General and
abdominal obesity and risk of esophageal and gastric adenocar-
cinoma in the European Prospective Investigation into Cancer
and Nutrition. Int J Cancer 137:646–657
21. Luo J, Margolis KL, Adami HO, Lopez AM, Lessin L, Ye W
(2007) Body size, weight cycling, and risk of renal cell carcinoma
among postmenopausal women: the Women’s Health Initiative
(United States). Am J Epidemiol 166:752–759
22. Ma X, Beeghly-Fadiel A, Shu XO et al (2013) Anthropometric
measures and epithelial ovarian cancer risk among Chinese
women: results from the Shanghai Women’s Health Study. Br J
Cancer 109:751–755
23. Schlesinger S, Aleksandrova K, Pischon T et al (2013) Abdom-
inal obesity, weight gain during adulthood and risk of liver and
biliary tract cancer in a European cohort. Int J Cancer
132:645–657
24. Saberi Hosnijeh F, Romieu I, Gallo V et al (2013) Anthropo-
metric characteristics and risk of lymphoid and myeloid leukemia
in the European Prospective Investigation into Cancer and
Nutrition (EPIC). Cancer Causes Control 24:427–438
25. Roswall N, Freisling H, Bueno-de-Mesquita HB et al (2014)
Anthropometric measures and bladder cancer risk: a prospective
study in the EPIC cohort. Int J Cancer 135:2918–2929
26. Louie SM, Roberts LS, Nomura DK (2013) Mechanisms linking
obesity and cancer. Biochim Biophys Acta 1831:1499–1508
27. Nieman KM, Romero IL, Van Houten B, Lengyel E (2013)
Adipose tissue and adipocytes support tumorigenesis and
metastasis. Biochim Biophys Acta (BBA) Mol Cell Biol Lipids
1831:1533–1541
28. Renehan AG, Zwahlen M, Egger M (2015) Adiposity and cancer
risk: new mechanistic insights from epidemiology. Nat Rev
Cancer 15:484–498
29. Doyle SL, Donohoe CL, Lysaght J, Reynolds JV (2012) Visceral
obesity, metabolic syndrome, insulin resistance and cancer. Proc
Nutr Soc 71:181–189
30. Donohoe CL, Doyle SL, Reynolds JV (2011) Visceral adiposity,
insulin resistance and cancer risk. Diabetol Metab Syndr 3:12
31. Hebert JR, Allison DB, Archer E, Lavie CJ, Blair SN (2013)
Scientific decision making, policy decisions, and the obesity
pandemic. Mayo Clin Proc 88:593–604
123
Cancer Causes Control
32. White UA, Tchoukalova YD (2014) Sex dimorphism and depot
differences in adipose tissue function. Biochim Biophys Acta
1842:377–392
33. Pischon T, Boeing H, Hoffmann K et al (2008) General and
abdominal adiposity and risk of death in Europe. N Engl J Med
359:2105–2120
34. Sahakyan KR, Somers VK, Rodriguez-Escudero JP et al (2015)
Normal-weight central obesity: implications for total and car-
diovascular mortality. Ann Intern Med 163:827–835
35. Robertson EV, Derakhshan MH, Wirz AA et al (2013) Central
obesity in asymptomatic volunteers is associated with increased
intrasphincteric acid reflux and lengthening of the cardiac
mucosa. Gastroenterology 145:730–739
36. Kahn BB, Flier JS (2000) Obesity and insulin resistance. J Clin
Invest 106:473–481
37. Ritte R, Lukanova A, Berrino F et al (2012) Adiposity, hormone
replacement therapy use and breast cancer risk by age and hor-
mone receptor status: a large prospective cohort study. Breast
Cancer Res 14:R76
38. Rinaldi S, Key TJ, Peeters PH et al (2006) Anthropometric
measures, endogenous sex steroids and breast cancer risk in
postmenopausal women: a study within the EPIC cohort. Int J
Cancer 118:2832–2839
39. Chionh F, Baglietto L, Krishnan K et al (2010) Physical activity,
body size and composition, and risk of ovarian cancer. Cancer
Causes Control 21:2183–2194
40. Li H, Yang G, Xiang YB et al (2013) Body weight, fat distri-
bution and colorectal cancer risk: a report from cohort studies of
134 255 Chinese men and women. Int J Obes (Lond) 37:783–789
41. World Health Organization (WHO) (2008) Waist circumference
and waist–hip ratio. WHO Expert Consultation, Geneva
42. Rinaldi S, Lise M, Clavel-Chapelon F et al (2012) Body size and
risk of differentiated thyroid carcinomas: findings from the EPIC
study. Int J Cancer 131:E1004–E1014
43. MacInnis RJ, English DR, Gertig DM, Hopper JL, Giles GG
(2004) Body size and composition and risk of postmenopausal
breast cancer. Cancer Epidemiol Biomarkers Prev 13:2117–2125
44. World Cancer Research Fund International/American Institute for
Cancer Research (2015) Continuous update project report: diet,
nutrition, physical activity and liver cancer
45. World Cancer Research Fund International/American Institute for
Cancer Research (2011) Continuous update project report: diet,
nutrition, physical activity and colorectal cancer
46. World Cancer Research Fund International/American Institute for
Cancer Research (2010) Continuous update project report: diet,
nutrition, physical activity and breast cancer