The prevalence of obesity has increased rapidly over the past two decades, reaching epidemic levels in many parts of the world.3 Currently worldwide, more than 1 billion adults are overweight and at least 300 million of them are obese.3 Present levels of obesity vary from less than 5% in China, Japan, and certain African nations, to over 75% in urban Samoa.3 Recent data from Europe found that the prevalence of obesity varied from 8 to 40% in men and from 5 to 53% in women, with high prevalence (>25%) in Spain, Greece, Ragusa and Naples (Italy), and the lowest prevalence (<10%) in France.4 Obesity is more common in Eastern Europe than in other parts of Europe, especially among women.5 Obesity prevalence remains at 1 to 5% in most populations in sub-Saharan Africa; however, it has risen to 8% in men and to 34% in women in some areas of South Africa and in neighboring countries.6 Obesity is becoming a problem in some urban areas of developing countries, especially in those undergoing economic transition,7 and the rates are almost 20% in some cities.3
The prevalence of obesity (BMI >30 kg/m2) in the U.S. adult population rose from 15% in 1976-1980 to 22.9% in 1988-1994 and 30.5% in 1999-2000, and excess weight (BMI > 25 kg/m2) rose from 55.9% in 1988-1994 to 64.5% in 1999-2000; the increases occurred for both men and women in all age groups and for non-Hispanic whites, non-Hispanic blacks, and Mexican Americans.8 A Canadian survey in 2000/2001, which used self-reported data, showed that about half of adult Canadians are overweight (56% in men and 39% in women) and 15% of them are obese (16% in men and 14% in women);9 however, the true rate could be higher because of a tendency for individuals to underreport their weight.
Childhood obesity is epidemic in some areas of the world and increasing rapidly in others.3,10 Worldwide, about 17.6 million children under 5 are estimated to be overweight.3 The highest prevalences of overweight among children are in the U.S., Ireland, Greece, and Portugal.11 The prevalence of overweight among U.S. children is increasing: in 1999-2000, 15.5% of 12- to 19-year-olds were overweight, 15.3% of 6- to 11-year-olds, and 10.4% of 2- to 5-year-olds, compared with 10.5, 11.3, and 7.2%, respectively, in 1988-1994.12 In Canada, between 1981 and 1996-1997, the excess weight rate increased from 11 to 33% in 7- to 13-year-old boys and from 13 to 27% in girls, and the obesity rate rose from 2 to 10% in boys and from 2 to 9% in girls.13 In a relatively short period of time
(1994-1995 to 1998-1999), the proportion of 2- to 11-year-olds who were overweight (including obese) rose from 34 to 37% and the obesity rate rose from 16 to 18%.14
Obesity is now known to be a major contributor to the global burden of disease.3 Obesity is associated not only with some nonfatal but debilitating health problems, such as respiratory difficulties, chronic musculoskeletal problems, skin problems, infertility, sleep apnea, and poor mental health, but also more life-threatening problems, including hypertension, type 2 diabetes, heart disease, stroke, some types of cancer.315 Obese people have higher mortality from all causes;16-19 more than 280,000 deaths annually have been attributed to obesity among U.S. adults.20 In Europe, this number is at least 279,000 deaths, of which 175,000 are attributable to obesity and 104,000 to overweight.21 Furthermore, a large decrease in life expectancy is observed among obese individuals, especially among younger adults.22,23 Severely obese children and adolescents have lower health-related quality of life24 and early-onset obesity has been suggested as a risk factor for morbidity and mortality later in life.25,26
Obesity is also associated significantly with excess medical costs.3,27,28 It is estimated that obesity accounts for 2 to 6% of total health care costs in several industrialized countries3 and for an even higher percentage in the U.S. where obesity-related health problems consume about 7 to 8% of the U.S. health care budget in direct medical costs.29,30 The total cost (direct and indirect) for obesity-related health problems is around $117 billion annually in the U.S.31 The total direct costs of obesity in Canada have been estimated at $1.8 billion, or 2.4% of total health care expenditures for all diseases in 1997.32
This chapter focuses on the epidemiological evidence of obesity as a risk factor for various cancers. Most studies used BMI as an indicator of overall obesity, and some had other measures as well, such as waist circumference or waist-to-hip ratio (WHR) (an indicator of body fat distribution) to examine whether central obesity is a risk factor for cancer. Different cut points for BMI, waist circumference, and WHR were used in various studies. We identified studies published between 1966 and 2004 via the MEDLINE database. Included in this chapter are published reviews and meta-analyses as well as individual studies published since the reviews. For individual studies, only those with at least 100 cases were included; however, all identified studies were included for less common cancers. When several articles were published from the same study, only the most recent is included.
Most of the cohort and case-control studies, included in three published reviews, as well as studies published subsequently33-48 consistently demonstrated a positive association between the risk of colorectal cancer and body fatness, as indicated by BMI or WHR or waist circumference. Most studies found nearly a doubling of risk in those with BMI >30.0 kg/m2 compared with those having BMI <23 kg/m2.33 A trend of increasing risk with increasing BMI across a wide range was also indicated in most investigations, with no clear evidence of a threshold effect.33 Only three studies49-51 showed no association of colorectal cancer and body fatness in both sexes.
The observed association between obesity and colorectal cancer is generally more consistent and stronger for men than for women and for colon cancer than for rectal cancer.33 Some studies found that WHR is a strong predictor (of the same strength as BMI) of colorectal cancer risk, especially for women.36,42,44,52-55 The association is also generally stronger and more consistent for cancer of the distal colon than the proximal colon.36,39,40,52,54-58 A meta-analysis of 13 studies on BMI and colon cancer found that the observed associations were of similar strength for cohort and case-control studies, and that the pooled estimates were statistically significant for both male and female but larger for men than for women.34 Excess weight early in life appears to be at least as important as recent excess weight in relation to colorectal cancer risk.40,45,54,56,58 One large population-based case-control study found that estrogen modifies the association between BMI and colon cancer risk; i.e., obese women who were estrogen-positive (pre-menopausal women or women who used hormone replacement therapy) experienced a greater than twofold increase in risk, whereas obese women who were estrogen-negative (postmenopausal women not taking hormone replacement therapy) did not.48
A positive association of similar magnitude to the association with colorectal cancer was also observed between body fatness and colorectal adenomas,50,53,59-66 which suggests that obesity may affect progression from adenoma to cancer. One study demonstrated the association in both sexes combined,61 others found it for men only,64,65 while another observed no association in either sex.67 Studies also identified a stronger effect of BMI, WHR, or waist circumference on large adenomas than on small adenomas or an effect restricted to large ade-
Although some studies suggest that weight gain between early adult ages and later adult ages increases the risks of developing colorectal cancer56,57 and col-orectal adenomas.60,61 the findings have not been substantiated by other stud-
The relationship between breast cancer risk and various measures of body size at different periods of life and adult weight gain has been extensively explored in many epidemiological studies. Because the associations differ between premenopausal and postmenopausal women, we reviewed the literature separately for the two groups.
In general, obesity does not relate to or is inversely associated with premenopausal breast cancer risk, as demonstrated in four reviews and studies published after the reviews.33-35,47,68-72 However, there were also studies that reported increased risk.73-77 Two of these were studies of non-Caucasian women.74,77
Both race and the distribution of body fat, rather than obesity per se may be important factors. Increased risk with higher WHR but not with higher BMI has been reported,78-80 as has decreased risk with higher BMI but not with higher WHR,81-83 decreased risk with higher BMI and increased risk with higher WHR among white women, but increased risk with higher WHR and no change with BMI among black women,84 and no change in risk with both higher BMI and higher WHR.85
Among the limited data on weight gain and premenopausal breast cancer risk, no association was observed in some case-control studies78,82,86 and a decreased risk was reported in one cohort study87 and several case-control stud-ies.76,88-90 One study found that adult weight gain was associated with increased risk among Hispanic white women but not among non-Hispanic white women.77
Convincing evidence exists for an association between obesity and postmenopausal breast cancer.33-35 The association remains after adjusting for various reproductive and lifestyle factors including physical activity. Most prospective cohort studies showed a significant increase in risk associated with recent obesity.33-35,68-70798591-95 Various degrees of increased risk (from 10% to more than twofold) were reported in many case-control studies.33-35,47,71,72,74,77,78,80,83,86,96-101 However, the increase is more modest in cohort studies than case-control studies. No increase in risk has also been found.33 34
Most studies suggested that obesity in young adulthood (at the age of 18 to 25 years) was either not associated with breast cancer risk76,77,91,96,99,100,102 or was associated with a 10 to 30% decrease in risk.69,84,86,90,93,95,103
Adult weight gain appeared to be the strongest and most consistent predictor of increased postmenopausal breast cancer risk, in both cohort studies 69,87,92,103,104
and case-control studies.33,74,77,78,86,90,96,98-100,102,105-108
Family history of breast cancer, hormone replacement therapy (HRT) use, estrogen receptor status, tumor morphology, and age at diagnosis may modify the association between obesity/weight gain and postmenopausal breast cancer risk. Studies have found a stronger association with obesity for women with a family history of breast cancer than for those without.75 95 96 Some studies also suggested that the association was restricted to non-HRT users73,87,91,93,100,108 or stronger among women who had never used HRT;7292 106 109 however, one large case-control study observed a stronger association among women who had used estrogen replacement therapy.76 Two studies have observed the association of breast cancer with weight gain among women who never used HRT but not among current HRT users.9193
A study that examined breast cancer risk by estrogen receptor (ER) and progesterone receptor (PR) status102 found that, among postmenopausal women, BMI-related risk was increased only among those with ER-/PR-positive tumors and not increased among those either with ER-/PR-negative tumors or among premenopausal women with any ER/PR subgroup. A Japanese study71 demonstrated that among postmenopausal women, obesity-related risk increase was stronger for ER-positive than ER-negative tumors and for PR-positive than PR-negative tumors. Another study indicated an increased risk of breast cancer with weight gain among Hispanic and postmenopausal non-Hispanic white women with ER-/PR-positive tumors, but not those with ER-/PR-negative tumors.77 However, one study suggested that the association between WHR and breast cancer is independent of ER status.80 Furthermore, one case-control study found that obesity was more strongly associated with elevated risk of invasive ductal carcinoma than invasive lobular carcinoma.97
Age at diagnosis appears to modify obesity-related breast cancer risk. Among studies that examined the association stratified by age at diagnosis, the risk appeared to increase with age at diagnosis, from 1.1 to 1.3 for women younger than 60 years to 1.6 to 2.9 for women older than 65 years.69,76,82,86,94,100,101 However, one cohort study of non-HRT users found that the effect of obesity was more pronounced among younger postmenopausal women (aged 50 to 69 years) than older women (aged 70 to 79 years).93
Several studies have shown obesity to be associated with an increased risk of a more advanced stage of breast cancer at diagnosis.110114 However, obese women are less likely to be screened with mammography,115 more likely to present late to a health-care professional,116 and have lower socioeconomic status.117,118 These factors are associated with delay in diagnosis of breast cancer and therefore may contribute to the more advanced stage at diagnosis for obese women.
The literature suggests a modest negative effect of greater body size on breast cancer prognosis in both pre- and postmenopausal women even after adjustment for other prognostic factors.94119121 However, some studies found no statistically significant difference in overall and disease-free survival between obese and non-obese women.119 122 123 Some studies also found that the influence of obesity on prognosis was limited to or was stronger among women with negative nodes, ER- and PR-positive status, and early disease stages (I and II).33
The evidence for a positive association between obesity and endometrial cancer risk is also convincing.33-35 Obesity has been found consistently to be associated with a two- to fivefold increase in endometrial cancer risk in both pre- and postmenopausal women in a majority of cohort and case-control studies conducted in various countries.33-35'37'41'4469124-131 Only two case-control studies observed no association.132 133 Obesity has been estimated to account for 39% of endometrial cancer in European women.35 Most studies observed a linear increase in risk with increasing BMI,33 and the association was independent of other known risk factors in most studies. The association for older women was generally similar to124,134,135 or somewhat stronger than for younger women.70 129 However, one study observed an association only among older women (>60 years old).136 One study identified a greater obesity-related risk increase among nonsmokers than among current smokers.130 Several studies that assessed WHR, waist-to-thigh circumference ratio (WTR), and skinfold thickness also indicated a positive association.129,137-140 Although some studies suggested that WHR or WTR influenced the risk independent of BMI,129,138 other studies found that the association disappeared once the results were adjusted for BMI.137,139,140 Large body weight among post-menopausal women was also found to increase the risk of hyperplasia of the endometrium.126
Body weight at early adulthood was not, or just weakly, associated with endometrial cancer risk.129,134,136,138,139,141-143 However, one study found a similar magnitude of association as at late adulthood.125 Weight gain during adult life substantially increased endometrial cancer risk in most studies that assessed adult weight gain125,129,134,136,138,139,141,142 with a linear dose-dependent pattern.33 However, studies that presented both crude and adjusted results showed that the association of risk with weight gain disappeared or became much weaker after adjusting for recent BMI or weight.125,134 One study showed no increase in unadjusted risk.143
Obesity has been consistently associated with an increased risk of renal cell cancer.33-35 Of all the cohort and case-control studies374146 47 and reviews3334144 examined, all but one study145 demonstrated a more than twofold increase in renal-cell cancer risk among obese men and women compared with those of normal weight, with a dose-response effect. However, no association has been observed for cancer of the renal pelvis.146 147 These studies were conducted on various populations in North America, northern and southern Europe, Asia, and Australia. The summary relative risk estimate from 14 studies on each sex in a meta-analysis was 1.07 (95% confidence interval 1.04 to 1.09) for men and 1.07 (95% confidence interval 1.05 to 1.09) for women per unit of increase of BMI.144 It is estimated that 27% of the renal cell cancer cases among American men and 29% among women could be related to overweight and obesity.144
Some studies reported a stronger association among women than men;148-153 however, the meta-analysis of 14 studies on each sex revealed no evidence of effect modification by sex, and BMI is equally strongly associated with the risk of renal cell cancer among men and women.144 The meta-analyses also found comparable strengths of association of obesity with renal cell cancer between cohort studies and case-control studies, between small and large studies, between studies conducted in the U.S. and in other countries, and between studies adjusted and not adjusted for smoking.144
The evidence of a link between prostate cancer incidence and obesity is insufficient: 13 cohort studies reported a small or modest positive association37,43,44,154-163 and 12 other cohort studies, 6 of which were included in the three reviews, did not.33-35,41,46,69,164-166 For case-control studies, only 4 studies observed a modest increase in prostate cancer risk associated with BMI47,167-169 and others found no association with BMI;33-35 one study found an inverse association with BMI.170 Among the studies reporting a positive association with obesity, the association is stronger for, or restricted to, fatal or more aggressive tumors or is stronger with mortality than with incidence.156-158,160,162 One meta-analysis of 6 studies of incident cases produced a modest relative risk of 1.01 (95% confidence interval 1.00 to 1.02) per unit of increase in BMI.35 Two studies suggested that the larger effect found in men aged 50 to 59 years than in other age groups154 and the increased risk only among those who never drank alcohol47 might partly explain the inconsistent findings among studies.
Only a few studies examined the influence of abdominal obesity on prostate cancer risk. One case-control study conducted in China found no association with usual adult BMI, but an excess risk with high levels of WHR171 and another study observed a nonsignificant increase in metastatic prostate cancer risk.172
Although epidemiological studies have not consistently shown that obesity is a risk factor for prostate cancer incidence, studies of mortality from prostate cancer have more clearly reported a positive association.37,155,157,160,161 Some studies also suggested that a higher BMI has been associated with higher-grade prostate tumors, non-organ-confined disease, disease progression,174 and poor prognosis.173
26.4.6 Adenocarcinoma of the Esophagus and Gastric Cardia
A limited number of cohort44175 and case-control studies176-182 have consistently linked obesity with increased risk of adenocarcinoma of the esophagus and gastric cardia. Two case-control studies reported no association,183 184 and one case-control study in Northern China suggested an inverse association between BMI
The association between BMI and risk of adenocarcinoma of the esophagus and cardia is strong, as indicated by the 2.7 relative risk summary estimate of a meta-analysis34 of seven studies.175-181 184 The observed association seems not to be explained by bias and confounding.33 The association also seems stronger for adenocarcinoma of the esophagus than of the cardia.177 178 181 However, no association with BMI was reported for esophageal squamous cell carcinoma.175 177 178 181
Results for the association between obesity and ovarian cancer risk are inconsistent.33 However, a systematic review of 13 hospital case-control studies, 11 population case-control studies, and 5 cohort studies published through 2001 concluded that the evidence supports a small to moderate positive association between high BMI and ovarian cancer risk.186 Seven cohort studies37,69,187-191 and five case-control studies47,192-195 were published later. Two cohort studies of mortality37,187 and one cohort study of incidence188 observed an increased risk with higher adult BMI. One cohort study identified a higher risk of ovarian cancer with higher WHR.189
The timing of overweight may be a risk modifier and may account for some of the inconsistent results. The Iowa Women's Health study found that higher BMI in early adulthood (but not current BMI) increased the risk of ovarian cancer.189 An increased risk associated with higher BMI in adolescence and young adulthood but not in older women was also indicated in a large Norwegian cohort study.190 A significantly increased risk associated with higher BMI at age 18 and for most of adult life, but not with change of weight between age 18 and adult life, was shown in one large case-control study with 1269 cases.192 Two cohort studies found no association between recent BMI and ovarian cancer risk.69,191 Although one case-control study indicated a positive association with recent BMI,47 two others found no association with recent BMI for the overall group193,194 but a significant positive association among premenopausal women193 with BMI at age 30194 and with higher WHR.194
High BMI was strongly related to risk of serous borderline ovarian cancer in three case-control studies,186,193,195 especially among premenopausal women.193 However, one nested case-control study suggested an inverse relation of risk with
Among studies that examined BMI-related risk of ovarian cancer by menopause separately, a positive association was found among premenopausal women only in two studies191,193 and among both groups in one study.186 One study of mortality found that the increased risk with obesity was limited to women who had never used postmenopausal estrogen and was not seen among ever users.187
Although there have been limited studies published on this issue, a meta-analysis of six case-control studies and eight cohort studies provided evidence that obesity may be weakly associated with the risk of pancreatic cancer.197 This meta-analysis observed a summary relative risk per unit increase in BMI of 1.02 (95% confidence interval 1.01 to 1.03). The estimates were slightly higher for studies that had adjusted for smoking and for case-control studies that had not used proxy respondents.197 Of the four studies that were not included in the meta-analysis, both cohort studies4144 and one case-control study47 suggested a positive association, whereas one case-control study conducted in China indicated a nonsignificant positive association.198
Although some studies found that the associations were of similar magnitude for men and women,4447 198-200 others suggested a larger effect of obesity on risk among men than among women.41201 One study observed a similar effect of obesity among white and black.200
Most cohort37,43,202-205 and case-control studies 47,206,207 have observed an inverse association between BMI and lung cancer risk. Three cohort studies41,44,46 found no association and one case-control study208 a positive relation. The inverse association may be explained by weight loss due to preclinical disease43,202,204 or a residual confounding effect of smoking.37,47,202,204,207 However, inconsistent results have been reported: the inverse relation with BMI was stronger among never-smokers in one cohort study205 and among never-smoking women in another study,207 and the positive association was found among both never-smokers and former-smokers.208 The issue of whether BMI is related to lung cancer risk is still controversial.
One cohort study and seven case-control studies, including those studies reviewed by the 1ARC Handbook, suggested that overweight or obesity is not significantly associated with testicular cancer risk.33,47,209,210 A significant positive association with obesity was observed for non-Hodgkin's lymphoma,37 41 47 multiple myeloma,3747 211 and leukemia.374447 Very few studies have been published on other cancer sites. There are too few studies on these cancer sites to allow a firm conclusion to be drawn.
The results from several studies suggest that overweight and obesity are associated with the risk of incidence41,46,47 and death37,212 from all types of cancer combined. Overweight and obesity may account for 7.7% of all cancer in Canada47 and 5% in Europe,35 and in the U.S., for as high as 14% of all deaths from cancer in men and 20% in women.37
Was this article helpful?