European Journal of Cancer (2015) 51, 705– 720
Available at www.sciencedirect.com
ScienceDirect journal homepage: www.ejcancer.com
Review
Obesity in breast cancer – What is the risk factor? F.R. James b,d, S. Wootton c, A. Jackson c, M. Wiseman c, E.R. Copson b,1, R.I. Cutress a,b,⇑,1 a
Southampton Breast Surgical Unit, University Hospitals Southampton, UK Cancer Sciences Division, University of Southampton, UK c Southampton NIHR Biomedical Research Centre, University Hospitals Southampton, UK d Jesus College, The University of Cambridge, UK b
Received 30 August 2014; received in revised form 11 January 2015; accepted 27 January 2015 Available online 3 March 2015
KEYWORDS Breast cancer Body fatness Risk factor Body composition Fat mass Fat free mass Prognosis Intervention
Abstract Environmental factors influence breast cancer incidence and progression. High body mass index (BMI) is associated with increased risk of post-menopausal breast cancer and with poorer outcome in those with a history of breast cancer. High BMI is generally interpreted as excess adiposity (overweight or obesity) and the World Cancer Research Fund judged that the associations between BMI and incidence of breast cancer were due to body fatness. Although BMI is the most common measure used to characterise body composition, it cannot distinguish lean mass from fat mass, or characterise body fat distribution, and so individuals with the same BMI can have different body composition. In particular, the relation between BMI and lean or fat mass may differ between people with or without disease. The question therefore arises as to what aspect or aspects of body composition are causally linked to the poorer outcome of breast cancer patients with high BMI. This question is not addressed in the literature. Most studies have used BMI, without discussion of its shortcomings as a marker of body composition, leading to potentially important misinterpretation. In this article we review the different measurements used to characterise body composition in the literature, and how they relate to breast cancer risk and prognosis. Further research is required to better characterise the relation of body composition to breast cancer. Ó 2015 Elsevier Ltd. All rights reserved.
⇑ Corresponding author at: Cancer Sciences Division, University of Southampton, Somers Building, MP 824, Tremona Road, Southampton SO16 6YD, UK. Tel.: +44 (0)7979 904339; fax: +44 (0)2380 795152. E-mail address: [email protected] (R.I. Cutress). 1
Both authors contributed equally as joint senior authors.
http://dx.doi.org/10.1016/j.ejca.2015.01.057 0959-8049/Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Breast cancer is the most common cancer in women worldwide. In the United Kingdom (UK) during 2011 49,936 new cases of breast cancer were reported [1] and the incidence is still rising. Although mortality rates
706
F.R. James et al. / European Journal of Cancer 51 (2015) 705–720
are falling, over 11,000 British women still die of this disease each year [2], and thousands of long-term survivors undergo intensive and prolonged adjuvant therapies following the initial surgical management of their primary tumour. Inherited mutations in high susceptibility genes account for only a small percentage of these cases, and environmental factors play a major causative role in most adult cancers, potentially interacting with polygenic factors. Theoretically, if these environmental factors can be sufficiently characterised, interventions can be developed, leading to a reduced incidence and improved treatments for breast cancer. The World Health Organisation (WHO) and many national authorities define obesity (a state of excess body fatness) in terms of body mass index (BMI). The current rate of obesity, (BMI >30 kg/m2), is estimated as 26% in British women and predicted to increase up to 43% by 2030 [3]. A recent updated comprehensive review of the evidence by the World Cancer Research Fund (WCRF) and American Institute for Cancer Research (AICR) [4] concluded that body fatness is an established and important risk factor for many cancers as well as other diseases (Fig. 1). The WCRF conclusion was an interpretation of evidence using body mass index (BMI) as a surrogate marker of body composition. The term ‘body fatness’ was used to describe the factor(s) underlying this increased risk, based on epidemiological BMI data, and evidence of mechanisms underpinning excess adiposity to cancer risk. However, BMI was not designed to describe body fatness [5], and does not describe body composition as it cannot distinguish lean from fat. Therefore, although there is now significant evidence that a high BMI is associated with an increased
• • • • • • • • • • • • • • • • • • • •
all-cause mortality ischaemic heart disease stroke diabetes mellitus chronic obstrucve pulmonary disease metabolic syndrome non-alcoholic fay liver disease benign prostate hyperplasia pulmonary embolism deep vein thrombosis gout gallstones reproducve disorders polycysc ovary syndrome osteoarthris lower back pain psychiatric disorders complicaons in pregnancy complicaons in surgery Cancers of the: postmenopausal breast, colorectum, pancreas, kidney, endometrium and gallbladder
Fig. 1. Diseases in which increased weight/obesity lead to increased risk – Dynamo-HIA project and [4].
risk of breast cancer in post-menopausal women, and with poorer clinical outcomes in all ages [4], the exact nature of the exposure remains uncertain (Fig. 2). This uncertainty is reflected in the range of different approaches taken to characterise or describe body composition in the literature: BMI, body weight, body composition, metabolic state and nutritional state. Most studies have used the relatively simple measure of BMI. Given the importance of this risk factor, not only in influencing breast cancer incidence, but also its association with poorer clinical outcomes after diagnosis, it is crucial to characterise more precisely the aspects of body composition, or associated metabolic or physiological factors, that underlie the observed association. It is then important to identify how it can be assessed clinically, whether it can be influenced and whether this can reduce risk and improve outcome. The aim of this review was to examine evidence, in terms of body composition, relating to the nature of the risk factor described as ‘body fatness’ in relation to breast cancer risk and outcome. To review the
Diet:
• •
Quality Quanty
Exercise: • Aerobic training • Weight training • General acvity
Genec make-up
Risk factor
Nutrional state
Metabolic state
Body composion
Measured by: • Weight • BMI • • • • • •
DXA BIA CT/MRI scanning Deuterium diluon test Under water weighing Total body plethysmography
Consists of: • Body fat mass •
• • •
Fat distribuon
Lean mass Water Bone
Fig. 2. Risk factor for breast cancer. There are several contributing factors (in blue) and interchangeable names (in red). In this review we specifically look at body composition. This is made up of various components (green) and can be measured by many methods (in black). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
F.R. James et al. / European Journal of Cancer 51 (2015) 705–720
literature the PUBMED database was searched using the appropriate terms to achieve the aims of the review, including; breast cancer, body composition, fat mass, lean mass, sarcopenia, dual energy X-ray absorptiometry (DEXA), bioelectrical impedance assay (BIA), BMI, risk, prognosis, outcome and mortality. The latest search to identify primary research publications was performed in July 2014. Full details of search strategy and results used to generate Tables 2–4 are provided in Supplementary Fig. 1. Since multiple methodologies have previously been used to measure body composition and the topic has not been previously addressed in a consistent manner an individual search strategy was designed for each table as outlined. Abstracts from meetings not published as peer-reviewed articles were not included. In addition potentially important references known to the authors or cited within relevant papers were also screened for the review. 2. Body composition measures 2.1. BMI as a measure of body composition Body mass index (BMI) is a simple proxy for human adiposity based on an individual’s height and weight (Fig. 3), and is based on Quetelet’s epidemiological observation that in a normal stature population body weight is approximately proportional to the square of height [6]. It is simple to calculate, allows measurement in a clinical setting, and has been widely used over many
BMI formula: BMI = Weight (Kg)/ height2 (m)
Underweight Severe thinness Moderate thinness Mild thinness
25% in men and >35% in women), but excellent specificity and positive predictive value [5]. In simple terms whilst those with a BMI of over 30 will almost certainly have excess body fat, over half the population with excess body fat will have a BMI of less than 30 and so not be identified by BMI alone. Reducing the BMI cut-off value will not improve the accuracy due to the confounding effects of lean. This may, at least in part, explain why epidemiologically mortality differences are most clearly seen with a BMI of over 30. Finally, the relationship between BMI and body fat also varies with sex, ethnicity and age. At different ages the same BMI corresponds to different amounts of lean and adipose tissue. Studies show that older adults have, on average, a greater proportion of fat than younger adults at any BMI due to loss of muscle mass with age [7]. Together these imprecisions can result in substantial misclassification of obesity in terms of body fatness, and as a result, the sensitivity of BMI cut points with respect to body fatness decreases with age. As the incidence of many chronic diseases influenced by body composition increases with age, the true relation with body composition as a predictive factor in these chronic diseases may be wrongly described by using BMI.
2.2. Other measures of body composition Fig. 3. BMI Formula – The World Health Organisation International Classification of adult underweight, overweight and obesity according to BMI.
Body composition is a complex variable, and impossible to measure directly in a clinical setting. The only
708
F.R. James et al. / European Journal of Cancer 51 (2015) 705–720
direct measure of body composition relies on chemical analysis of cadavers, and the study performed by Widdowson et al. [8] remains the gold standard against which indirect methods are validated. The most widely used indirect measures of body composition are based on the four compartment (4C) method. This splits the body into the compartments: fat, protein, water and mineral mass and uses a variety of methods such as dual energy X-ray absorptiometry (DEXA), total body plethysmography and deuterium dilution measurements to measure each of these to assess body mass, total body volume, total body water (TBW) and bone mineral. From this the contribution of the 4 compartments to body composition can be calculated [9]. This is the standard against which all new measures of body composition are validated. However, it is difficult, time consuming and expensive and hence used very infrequently, and is not realistic for clinical settings. Further, all these methods are based on assumptions regarding the density of body tissues, concentrations of water and electrolytes, and interrelationships between body compartments among healthy individuals. Unfortunately it is unknown whether these hold true for obese persons or those with chronic disease, where physiological, anatomical, metabolic and hormonal alterations may affect the underlying assumptions. As more attention is focused on the importance of the accurate assessment of body composition in accurate and clinically viable ways, more methods, bringing different advantages and disadvantages, are emerging and being tested in many patient populations (Table 1) [10]. Anthropometric measurements of body composition such as waist to hip ratio (WHR) and waist circumference have been shown to have added predictive values for some diseases than BMI alone [11], however, it is hypothesised that still stronger associations using more precise measures of body composition can be found. Methods such as Bioelectrical impedance assay (BIA), Dual-energy X-ray absorptiometry (DEXA), total body water, hydrodensitometry and computerised tomography (CT) scan analysis software (sliceOmatice) [12] are being explored as more convenient precise measures of body composition [13]. With these developments there is scope for characterising the components of body composition that might be better measures for disease risk and outcome in individuals than simple BMI measurements alone.
measure, is measured indirectly by methods such as DEXA, BIA, hydrodensitometry and computerised tomography (CT) scans, and is closely associated with metabolic risk factors such as higher triglycerides, lower high-density lipoprotein, higher blood pressure, impaired fasting glucose, diabetes mellitus and metabolic syndrome [14]. Whilst these factors have been well investigated with respect to cardiovascular disease risk, their association to breast cancer has not been well characterised. A recent large scale prospective study of two hundred ninety thousand women found metabolic syndrome, blood pressure and plasma glucose levels were all associated with an increased risk of breast cancer mortality in women over 60 [15]. As well as total body fat, – the distribution of fat around the body has also been investigated in relation to metabolic risk factors and other diseases. Body fat distribution varies from person to person and with age, ethnicity and sex. Assessment by anthropometric measurements such as waist to hip ratio and callipers are largely inaccurate in assessing body fat distribution, but new techniques such as bioelectrical impedance spectroscopy (BIS), dual X-ray absorptiometry (DXA), CT scan software and MRI scan analysis are providing a clearer picture of this body composition factor and its effects on certain disease risks. Body fat is usually separated into central adiposity (fat mainly distributed around the trunk and upper body) and peripheral adiposity (fat mainly distributed around the hips and lower body). Central adiposity reflects both visceral (the fat that surrounds the abdominal organs) and abdominal subcutaneous fat. Central adiposity is more strongly linked to adverse cardiometabolic outcomes, thought to be related to the metabolic activity of visceral as compared to subcutaneous fat. Body fat distribution, as described by central adiposity versus peripheral adiposity, was found to be most predictive of cardiovascular disease, independent of and to a greater magnitude than BMI in a recent review [16]. Many factors that have been suggested to be associated with increased breast cancer risk and poorer prognosis due to increased BMI, such as hormonal and inflammatory factors, are more closely associated with more precise measures of body fat mass than BMI. This suggests that body fat mass and fat distribution may play important roles in breast cancer and be a useful measure of risk [17].
3. Constituents of body composition
3.2. Lean body mass
3.1. Body fatness and fat distribution
Fat free mass, comprising muscle, viscera and bone, is the other major component of body composition. Recent interest has centred on the importance of skeletal muscle mass for normal physiological function. The term sarcopenia describes low lean muscle mass [18]. This is seen particularly with ageing where the amount
Body fatness is a major component of body composition. WCRF interpreted BMI as reflecting this in relation to its association with many cancer types. True body fatness, as described as a ‘total body fat mass’
F.R. James et al. / European Journal of Cancer 51 (2015) 705–720
709
Table 1 Complex measures of body composition. Measure
Characteristics measured
Method
Advantages
Disadvantages
BIA [13]
TBW
Accurate, very quick, easy, validated in many populations, can be carried out by anyone
Relies on a set of formulae that are different in different populations/conditions. Does not work in cohorts with fluid or electrolyte abnormalities. It is altered by hydration state. Expensive equipment
DEXA [73]
FFM
Accurate, quick, single procedure
Relies of formula to work out other compartments, radiation risk, needs X-ray machines and members of staff
Total body plethysmography [74]
BFM
Accurate, quick, easy to use, single measurement
Relies on formulas to work out FFM, Relies on formula to work out other body compartments, expensive equipment, needs trained staff to operate, altered by state of person
Underwater weighing [74]
BFM
An electric current is sent around the body and the resistance to flow measured. This resistance is used to work out TBW. Equations are then used to estimate other body fractions. Different methods exist such as full body, segmental, single frequency or multi frequency spectroscopy Two X-ray beams of different energy levels are aimed at the subject and the absorbance due to FFM can be calculated. Equations are then used to estimate other body fractions Volume of the body is measured by subtracting volume of pod with person in from volume of pod alone. Volume and weight allow density to be calculated. BFM is worked out from density using equations Density of the body is worked out by underwater and on land weighing. BFM is worked out from density using equations
Accurate, can be carried out by anyone, no complex equipment
Deuterium dilution [75]
Total body water
Unpleasant for subject, noncompliance is an issue, Relies on formulas to work out FFM, Relies on formula to work out other body compartments, time consuming Relies on formulas to work out other body compartments, time consuming, subject needs to drink deuterium label, expensive equipment, needs trained staff to work
CT/MRI [12]
BFM, Body fat distribution
A known amount of D2O is drunk and given time to equilibrate in the body. A sample is taken and the concentration worked out allowing calculation of TBW. Other body compartments are worked out from this CT scans used to calculate fat mass and distribution
Accurate, easy single measurement
Very accurate, doesn’t rely on formulas
Time consuming, requires trained staff and very expensive equipment, radiation exposure, machines cannot accommodate very large people
Abbreviations: BIA, bioelectrical impedance assay; TBW, total body water; DXA, dual energy X-ray absorptiometry; FFM, fat free mass; BFM, body fat mass; CT, computerised tomography; MRI, Magnetic resonance imaging.
and strength of muscle declines and is associated with functional impairment and morbidity in the elderly. A recent study found sarcopenic obesity to be a strong predictor of cardiovascular disease [19]. However, muscle mass and muscle strength showed independent effects, highlighting the need to characterise the structural and functional aspects that account for the observed associations. The health cost of sarcopenia was estimated
at $18.5 billion in the US for the year 2000 and is expected to increase [20]. Cachexia is loss of lean mass that cannot be reversed nutritionally and always occurs secondary to underlying pathology, in contrast to sarcopenia, which also occurs in health with age. Cachexia is a common result of late-stage cancer but the causes, consequences and potential for intervention have been little studied.
710
F.R. James et al. / European Journal of Cancer 51 (2015) 705–720
4. Body composition and risk of developing breast cancer 4.1. BMI and risk of breast cancer Many studies have investigated the relation between obesity and breast cancer incidence. Most have used BMI as a proxy measure for body fat. The World Cancer Research Fund (WCRF) [21] review of the literature (43 cohort studies, more than 100 case-control studies, and two ecological studies) concluded: ‘There is abundant and consistent epidemiological evidence and a clear dose response, with robust evidence for mechanisms operating in humans that greater body fatness is a cause of postmenopausal breast cancer.’ An updated metaanalysis of 31 cohort studies reported a 12 percent increased risk per 5 kg/m2 [4]. As discussed, the use of BMI as the measure may not fully describe body fat and stronger associations may exist with a more precise characterisation of body composition. Potential aetiological factors accounting for this association have been reviewed many times and are hypothesised to include a direct effect on levels of many circulating hormones, for example increasing plasma oestrogens and decreasing levels of sex hormone binding globulin [22]. With increased body fat there is also a difference in other growth factors such as insulin, insulin-like growth factor and leptin [23]. It is felt that this creates an environment that encourages proliferation and discourages apoptosis hence promoting carcinogenesis. In addition, the inflammatory state that accompanies obesity and metabolic syndrome may also contribute to the development of cancer [24]. However, a clearer understanding of specific structural and functional factors is needed to investigate these mechanisms more accurately. In contrast in premenopausal women greater body fatness is associated with lower risk of developing breast cancer, but, the mechanism is still not well understood. Some potential factors hypothesised are reduced exposure to oestrogens by disruption of the normal menstrual cycle, protection by increased concentrations of progesterone in overweight women [25] and earlier differentiation of breast tissue decreasing the probability of malignant transformation [26].
4.2. Other measures of body composition and breast cancer risk Despite the data on BMI and breast cancer risk and general acceptance of ‘body fatness’ as a risk factor for the development of postmenopausal breast cancer, little has been done to assess more precisely the importance of the various components of body composition. A review including five cohort studies found that central obesity was related to increased breast cancer risk in pre-menopausal women, but in post-menopausal women associations were related to the association between
waist to hip ratio (WHR) and BMI [27]. The WCRF has more recently reviewed the evidence for associations between anthropometric measures of body fatness and breast cancer risk. Waist to hip ratio (WHR) was used in 16 studies, and despite some heterogeneity many studies showed a significant association with breast cancer risk. A meta-analysis of the cohort studies gave a summary effect estimate of 1.19 (95% confidence interval (CI) 1.10–1.28) per ratio increment of 0.1. This is higher than for that of an increase of BMI of 2 kg/m2 (1.03 (95% CI 1.01–1.04)) discussed in the same report. This shows some measurements of ‘body fatness’ do better describe the risk factor than BMI, although perhaps not perfectly with five of the 16 studies showing no significant change in risk for increasing WHR. More precise measures of body composition and breast cancer risk have been investigated in only six studies [28–33] (detailed in Table 2) by means of BIA, DXA and the 4C model. As detailed in Table 2 all but one study found significant evidence for the effect of body composition on breast cancer risk. Increases body fat mass correlated with increased risk of breast cancer with one study reporting a 2-fold risk difference in women with the highest adiposities versus the lowest (26). Fat free mass was negatively associated with breast cancer risk in several studies. Importantly, this was found in a study where only patients with normal BMI were included (27). The only study not to find any association between complex body composition measures and breast cancer risk also failed to find a significant association between BMI and breast cancer risk. This suggests that these measures of body composition add value to BMI in predicting post-menopausal breast cancer risk. Central obesity, measured as waist to hip ratio (WHR), was a better predictor than BMI of the promoter methylation of E-cadherin, p16, and RAR-b2 genes [34]. This methylation is associated with distant metastasis and oestrogen receptor positive (ER+) tumours and hence worse prognosis.
5. Body composition and clinical outcome from breast cancer 5.1. BMI at diagnosis and survival There is increasing evidence that BMI at diagnosis of breast cancer also affects survival. Although some inconsistencies remain a recent systematic review and metaanalysis [35] of 45 studies showed poorer survival among women with a BMI P25 kg/m2 compared with women with a BMI
Available at www.sciencedirect.com
ScienceDirect journal homepage: www.ejcancer.com
Review
Obesity in breast cancer – What is the risk factor? F.R. James b,d, S. Wootton c, A. Jackson c, M. Wiseman c, E.R. Copson b,1, R.I. Cutress a,b,⇑,1 a
Southampton Breast Surgical Unit, University Hospitals Southampton, UK Cancer Sciences Division, University of Southampton, UK c Southampton NIHR Biomedical Research Centre, University Hospitals Southampton, UK d Jesus College, The University of Cambridge, UK b
Received 30 August 2014; received in revised form 11 January 2015; accepted 27 January 2015 Available online 3 March 2015
KEYWORDS Breast cancer Body fatness Risk factor Body composition Fat mass Fat free mass Prognosis Intervention
Abstract Environmental factors influence breast cancer incidence and progression. High body mass index (BMI) is associated with increased risk of post-menopausal breast cancer and with poorer outcome in those with a history of breast cancer. High BMI is generally interpreted as excess adiposity (overweight or obesity) and the World Cancer Research Fund judged that the associations between BMI and incidence of breast cancer were due to body fatness. Although BMI is the most common measure used to characterise body composition, it cannot distinguish lean mass from fat mass, or characterise body fat distribution, and so individuals with the same BMI can have different body composition. In particular, the relation between BMI and lean or fat mass may differ between people with or without disease. The question therefore arises as to what aspect or aspects of body composition are causally linked to the poorer outcome of breast cancer patients with high BMI. This question is not addressed in the literature. Most studies have used BMI, without discussion of its shortcomings as a marker of body composition, leading to potentially important misinterpretation. In this article we review the different measurements used to characterise body composition in the literature, and how they relate to breast cancer risk and prognosis. Further research is required to better characterise the relation of body composition to breast cancer. Ó 2015 Elsevier Ltd. All rights reserved.
⇑ Corresponding author at: Cancer Sciences Division, University of Southampton, Somers Building, MP 824, Tremona Road, Southampton SO16 6YD, UK. Tel.: +44 (0)7979 904339; fax: +44 (0)2380 795152. E-mail address: [email protected] (R.I. Cutress). 1
Both authors contributed equally as joint senior authors.
http://dx.doi.org/10.1016/j.ejca.2015.01.057 0959-8049/Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Breast cancer is the most common cancer in women worldwide. In the United Kingdom (UK) during 2011 49,936 new cases of breast cancer were reported [1] and the incidence is still rising. Although mortality rates
706
F.R. James et al. / European Journal of Cancer 51 (2015) 705–720
are falling, over 11,000 British women still die of this disease each year [2], and thousands of long-term survivors undergo intensive and prolonged adjuvant therapies following the initial surgical management of their primary tumour. Inherited mutations in high susceptibility genes account for only a small percentage of these cases, and environmental factors play a major causative role in most adult cancers, potentially interacting with polygenic factors. Theoretically, if these environmental factors can be sufficiently characterised, interventions can be developed, leading to a reduced incidence and improved treatments for breast cancer. The World Health Organisation (WHO) and many national authorities define obesity (a state of excess body fatness) in terms of body mass index (BMI). The current rate of obesity, (BMI >30 kg/m2), is estimated as 26% in British women and predicted to increase up to 43% by 2030 [3]. A recent updated comprehensive review of the evidence by the World Cancer Research Fund (WCRF) and American Institute for Cancer Research (AICR) [4] concluded that body fatness is an established and important risk factor for many cancers as well as other diseases (Fig. 1). The WCRF conclusion was an interpretation of evidence using body mass index (BMI) as a surrogate marker of body composition. The term ‘body fatness’ was used to describe the factor(s) underlying this increased risk, based on epidemiological BMI data, and evidence of mechanisms underpinning excess adiposity to cancer risk. However, BMI was not designed to describe body fatness [5], and does not describe body composition as it cannot distinguish lean from fat. Therefore, although there is now significant evidence that a high BMI is associated with an increased
• • • • • • • • • • • • • • • • • • • •
all-cause mortality ischaemic heart disease stroke diabetes mellitus chronic obstrucve pulmonary disease metabolic syndrome non-alcoholic fay liver disease benign prostate hyperplasia pulmonary embolism deep vein thrombosis gout gallstones reproducve disorders polycysc ovary syndrome osteoarthris lower back pain psychiatric disorders complicaons in pregnancy complicaons in surgery Cancers of the: postmenopausal breast, colorectum, pancreas, kidney, endometrium and gallbladder
Fig. 1. Diseases in which increased weight/obesity lead to increased risk – Dynamo-HIA project and [4].
risk of breast cancer in post-menopausal women, and with poorer clinical outcomes in all ages [4], the exact nature of the exposure remains uncertain (Fig. 2). This uncertainty is reflected in the range of different approaches taken to characterise or describe body composition in the literature: BMI, body weight, body composition, metabolic state and nutritional state. Most studies have used the relatively simple measure of BMI. Given the importance of this risk factor, not only in influencing breast cancer incidence, but also its association with poorer clinical outcomes after diagnosis, it is crucial to characterise more precisely the aspects of body composition, or associated metabolic or physiological factors, that underlie the observed association. It is then important to identify how it can be assessed clinically, whether it can be influenced and whether this can reduce risk and improve outcome. The aim of this review was to examine evidence, in terms of body composition, relating to the nature of the risk factor described as ‘body fatness’ in relation to breast cancer risk and outcome. To review the
Diet:
• •
Quality Quanty
Exercise: • Aerobic training • Weight training • General acvity
Genec make-up
Risk factor
Nutrional state
Metabolic state
Body composion
Measured by: • Weight • BMI • • • • • •
DXA BIA CT/MRI scanning Deuterium diluon test Under water weighing Total body plethysmography
Consists of: • Body fat mass •
• • •
Fat distribuon
Lean mass Water Bone
Fig. 2. Risk factor for breast cancer. There are several contributing factors (in blue) and interchangeable names (in red). In this review we specifically look at body composition. This is made up of various components (green) and can be measured by many methods (in black). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
F.R. James et al. / European Journal of Cancer 51 (2015) 705–720
literature the PUBMED database was searched using the appropriate terms to achieve the aims of the review, including; breast cancer, body composition, fat mass, lean mass, sarcopenia, dual energy X-ray absorptiometry (DEXA), bioelectrical impedance assay (BIA), BMI, risk, prognosis, outcome and mortality. The latest search to identify primary research publications was performed in July 2014. Full details of search strategy and results used to generate Tables 2–4 are provided in Supplementary Fig. 1. Since multiple methodologies have previously been used to measure body composition and the topic has not been previously addressed in a consistent manner an individual search strategy was designed for each table as outlined. Abstracts from meetings not published as peer-reviewed articles were not included. In addition potentially important references known to the authors or cited within relevant papers were also screened for the review. 2. Body composition measures 2.1. BMI as a measure of body composition Body mass index (BMI) is a simple proxy for human adiposity based on an individual’s height and weight (Fig. 3), and is based on Quetelet’s epidemiological observation that in a normal stature population body weight is approximately proportional to the square of height [6]. It is simple to calculate, allows measurement in a clinical setting, and has been widely used over many
BMI formula: BMI = Weight (Kg)/ height2 (m)
Underweight Severe thinness Moderate thinness Mild thinness
25% in men and >35% in women), but excellent specificity and positive predictive value [5]. In simple terms whilst those with a BMI of over 30 will almost certainly have excess body fat, over half the population with excess body fat will have a BMI of less than 30 and so not be identified by BMI alone. Reducing the BMI cut-off value will not improve the accuracy due to the confounding effects of lean. This may, at least in part, explain why epidemiologically mortality differences are most clearly seen with a BMI of over 30. Finally, the relationship between BMI and body fat also varies with sex, ethnicity and age. At different ages the same BMI corresponds to different amounts of lean and adipose tissue. Studies show that older adults have, on average, a greater proportion of fat than younger adults at any BMI due to loss of muscle mass with age [7]. Together these imprecisions can result in substantial misclassification of obesity in terms of body fatness, and as a result, the sensitivity of BMI cut points with respect to body fatness decreases with age. As the incidence of many chronic diseases influenced by body composition increases with age, the true relation with body composition as a predictive factor in these chronic diseases may be wrongly described by using BMI.
2.2. Other measures of body composition Fig. 3. BMI Formula – The World Health Organisation International Classification of adult underweight, overweight and obesity according to BMI.
Body composition is a complex variable, and impossible to measure directly in a clinical setting. The only
708
F.R. James et al. / European Journal of Cancer 51 (2015) 705–720
direct measure of body composition relies on chemical analysis of cadavers, and the study performed by Widdowson et al. [8] remains the gold standard against which indirect methods are validated. The most widely used indirect measures of body composition are based on the four compartment (4C) method. This splits the body into the compartments: fat, protein, water and mineral mass and uses a variety of methods such as dual energy X-ray absorptiometry (DEXA), total body plethysmography and deuterium dilution measurements to measure each of these to assess body mass, total body volume, total body water (TBW) and bone mineral. From this the contribution of the 4 compartments to body composition can be calculated [9]. This is the standard against which all new measures of body composition are validated. However, it is difficult, time consuming and expensive and hence used very infrequently, and is not realistic for clinical settings. Further, all these methods are based on assumptions regarding the density of body tissues, concentrations of water and electrolytes, and interrelationships between body compartments among healthy individuals. Unfortunately it is unknown whether these hold true for obese persons or those with chronic disease, where physiological, anatomical, metabolic and hormonal alterations may affect the underlying assumptions. As more attention is focused on the importance of the accurate assessment of body composition in accurate and clinically viable ways, more methods, bringing different advantages and disadvantages, are emerging and being tested in many patient populations (Table 1) [10]. Anthropometric measurements of body composition such as waist to hip ratio (WHR) and waist circumference have been shown to have added predictive values for some diseases than BMI alone [11], however, it is hypothesised that still stronger associations using more precise measures of body composition can be found. Methods such as Bioelectrical impedance assay (BIA), Dual-energy X-ray absorptiometry (DEXA), total body water, hydrodensitometry and computerised tomography (CT) scan analysis software (sliceOmatice) [12] are being explored as more convenient precise measures of body composition [13]. With these developments there is scope for characterising the components of body composition that might be better measures for disease risk and outcome in individuals than simple BMI measurements alone.
measure, is measured indirectly by methods such as DEXA, BIA, hydrodensitometry and computerised tomography (CT) scans, and is closely associated with metabolic risk factors such as higher triglycerides, lower high-density lipoprotein, higher blood pressure, impaired fasting glucose, diabetes mellitus and metabolic syndrome [14]. Whilst these factors have been well investigated with respect to cardiovascular disease risk, their association to breast cancer has not been well characterised. A recent large scale prospective study of two hundred ninety thousand women found metabolic syndrome, blood pressure and plasma glucose levels were all associated with an increased risk of breast cancer mortality in women over 60 [15]. As well as total body fat, – the distribution of fat around the body has also been investigated in relation to metabolic risk factors and other diseases. Body fat distribution varies from person to person and with age, ethnicity and sex. Assessment by anthropometric measurements such as waist to hip ratio and callipers are largely inaccurate in assessing body fat distribution, but new techniques such as bioelectrical impedance spectroscopy (BIS), dual X-ray absorptiometry (DXA), CT scan software and MRI scan analysis are providing a clearer picture of this body composition factor and its effects on certain disease risks. Body fat is usually separated into central adiposity (fat mainly distributed around the trunk and upper body) and peripheral adiposity (fat mainly distributed around the hips and lower body). Central adiposity reflects both visceral (the fat that surrounds the abdominal organs) and abdominal subcutaneous fat. Central adiposity is more strongly linked to adverse cardiometabolic outcomes, thought to be related to the metabolic activity of visceral as compared to subcutaneous fat. Body fat distribution, as described by central adiposity versus peripheral adiposity, was found to be most predictive of cardiovascular disease, independent of and to a greater magnitude than BMI in a recent review [16]. Many factors that have been suggested to be associated with increased breast cancer risk and poorer prognosis due to increased BMI, such as hormonal and inflammatory factors, are more closely associated with more precise measures of body fat mass than BMI. This suggests that body fat mass and fat distribution may play important roles in breast cancer and be a useful measure of risk [17].
3. Constituents of body composition
3.2. Lean body mass
3.1. Body fatness and fat distribution
Fat free mass, comprising muscle, viscera and bone, is the other major component of body composition. Recent interest has centred on the importance of skeletal muscle mass for normal physiological function. The term sarcopenia describes low lean muscle mass [18]. This is seen particularly with ageing where the amount
Body fatness is a major component of body composition. WCRF interpreted BMI as reflecting this in relation to its association with many cancer types. True body fatness, as described as a ‘total body fat mass’
F.R. James et al. / European Journal of Cancer 51 (2015) 705–720
709
Table 1 Complex measures of body composition. Measure
Characteristics measured
Method
Advantages
Disadvantages
BIA [13]
TBW
Accurate, very quick, easy, validated in many populations, can be carried out by anyone
Relies on a set of formulae that are different in different populations/conditions. Does not work in cohorts with fluid or electrolyte abnormalities. It is altered by hydration state. Expensive equipment
DEXA [73]
FFM
Accurate, quick, single procedure
Relies of formula to work out other compartments, radiation risk, needs X-ray machines and members of staff
Total body plethysmography [74]
BFM
Accurate, quick, easy to use, single measurement
Relies on formulas to work out FFM, Relies on formula to work out other body compartments, expensive equipment, needs trained staff to operate, altered by state of person
Underwater weighing [74]
BFM
An electric current is sent around the body and the resistance to flow measured. This resistance is used to work out TBW. Equations are then used to estimate other body fractions. Different methods exist such as full body, segmental, single frequency or multi frequency spectroscopy Two X-ray beams of different energy levels are aimed at the subject and the absorbance due to FFM can be calculated. Equations are then used to estimate other body fractions Volume of the body is measured by subtracting volume of pod with person in from volume of pod alone. Volume and weight allow density to be calculated. BFM is worked out from density using equations Density of the body is worked out by underwater and on land weighing. BFM is worked out from density using equations
Accurate, can be carried out by anyone, no complex equipment
Deuterium dilution [75]
Total body water
Unpleasant for subject, noncompliance is an issue, Relies on formulas to work out FFM, Relies on formula to work out other body compartments, time consuming Relies on formulas to work out other body compartments, time consuming, subject needs to drink deuterium label, expensive equipment, needs trained staff to work
CT/MRI [12]
BFM, Body fat distribution
A known amount of D2O is drunk and given time to equilibrate in the body. A sample is taken and the concentration worked out allowing calculation of TBW. Other body compartments are worked out from this CT scans used to calculate fat mass and distribution
Accurate, easy single measurement
Very accurate, doesn’t rely on formulas
Time consuming, requires trained staff and very expensive equipment, radiation exposure, machines cannot accommodate very large people
Abbreviations: BIA, bioelectrical impedance assay; TBW, total body water; DXA, dual energy X-ray absorptiometry; FFM, fat free mass; BFM, body fat mass; CT, computerised tomography; MRI, Magnetic resonance imaging.
and strength of muscle declines and is associated with functional impairment and morbidity in the elderly. A recent study found sarcopenic obesity to be a strong predictor of cardiovascular disease [19]. However, muscle mass and muscle strength showed independent effects, highlighting the need to characterise the structural and functional aspects that account for the observed associations. The health cost of sarcopenia was estimated
at $18.5 billion in the US for the year 2000 and is expected to increase [20]. Cachexia is loss of lean mass that cannot be reversed nutritionally and always occurs secondary to underlying pathology, in contrast to sarcopenia, which also occurs in health with age. Cachexia is a common result of late-stage cancer but the causes, consequences and potential for intervention have been little studied.
710
F.R. James et al. / European Journal of Cancer 51 (2015) 705–720
4. Body composition and risk of developing breast cancer 4.1. BMI and risk of breast cancer Many studies have investigated the relation between obesity and breast cancer incidence. Most have used BMI as a proxy measure for body fat. The World Cancer Research Fund (WCRF) [21] review of the literature (43 cohort studies, more than 100 case-control studies, and two ecological studies) concluded: ‘There is abundant and consistent epidemiological evidence and a clear dose response, with robust evidence for mechanisms operating in humans that greater body fatness is a cause of postmenopausal breast cancer.’ An updated metaanalysis of 31 cohort studies reported a 12 percent increased risk per 5 kg/m2 [4]. As discussed, the use of BMI as the measure may not fully describe body fat and stronger associations may exist with a more precise characterisation of body composition. Potential aetiological factors accounting for this association have been reviewed many times and are hypothesised to include a direct effect on levels of many circulating hormones, for example increasing plasma oestrogens and decreasing levels of sex hormone binding globulin [22]. With increased body fat there is also a difference in other growth factors such as insulin, insulin-like growth factor and leptin [23]. It is felt that this creates an environment that encourages proliferation and discourages apoptosis hence promoting carcinogenesis. In addition, the inflammatory state that accompanies obesity and metabolic syndrome may also contribute to the development of cancer [24]. However, a clearer understanding of specific structural and functional factors is needed to investigate these mechanisms more accurately. In contrast in premenopausal women greater body fatness is associated with lower risk of developing breast cancer, but, the mechanism is still not well understood. Some potential factors hypothesised are reduced exposure to oestrogens by disruption of the normal menstrual cycle, protection by increased concentrations of progesterone in overweight women [25] and earlier differentiation of breast tissue decreasing the probability of malignant transformation [26].
4.2. Other measures of body composition and breast cancer risk Despite the data on BMI and breast cancer risk and general acceptance of ‘body fatness’ as a risk factor for the development of postmenopausal breast cancer, little has been done to assess more precisely the importance of the various components of body composition. A review including five cohort studies found that central obesity was related to increased breast cancer risk in pre-menopausal women, but in post-menopausal women associations were related to the association between
waist to hip ratio (WHR) and BMI [27]. The WCRF has more recently reviewed the evidence for associations between anthropometric measures of body fatness and breast cancer risk. Waist to hip ratio (WHR) was used in 16 studies, and despite some heterogeneity many studies showed a significant association with breast cancer risk. A meta-analysis of the cohort studies gave a summary effect estimate of 1.19 (95% confidence interval (CI) 1.10–1.28) per ratio increment of 0.1. This is higher than for that of an increase of BMI of 2 kg/m2 (1.03 (95% CI 1.01–1.04)) discussed in the same report. This shows some measurements of ‘body fatness’ do better describe the risk factor than BMI, although perhaps not perfectly with five of the 16 studies showing no significant change in risk for increasing WHR. More precise measures of body composition and breast cancer risk have been investigated in only six studies [28–33] (detailed in Table 2) by means of BIA, DXA and the 4C model. As detailed in Table 2 all but one study found significant evidence for the effect of body composition on breast cancer risk. Increases body fat mass correlated with increased risk of breast cancer with one study reporting a 2-fold risk difference in women with the highest adiposities versus the lowest (26). Fat free mass was negatively associated with breast cancer risk in several studies. Importantly, this was found in a study where only patients with normal BMI were included (27). The only study not to find any association between complex body composition measures and breast cancer risk also failed to find a significant association between BMI and breast cancer risk. This suggests that these measures of body composition add value to BMI in predicting post-menopausal breast cancer risk. Central obesity, measured as waist to hip ratio (WHR), was a better predictor than BMI of the promoter methylation of E-cadherin, p16, and RAR-b2 genes [34]. This methylation is associated with distant metastasis and oestrogen receptor positive (ER+) tumours and hence worse prognosis.
5. Body composition and clinical outcome from breast cancer 5.1. BMI at diagnosis and survival There is increasing evidence that BMI at diagnosis of breast cancer also affects survival. Although some inconsistencies remain a recent systematic review and metaanalysis [35] of 45 studies showed poorer survival among women with a BMI P25 kg/m2 compared with women with a BMI
