Obesity – incidence and mortality from gastrointestinal tract cancers

The prevalence of obesity is increasing at an alarming rate all over the world. This epidemic phenomenon that began in the early 1980's continues unabated and affects both adults and children. Obesity is one of the strongest emerging risk factors for many cancers in western countries, among them some gastrointestinal tract cancers. In this review, the presented evidence is based mainly on meta-analyses.

Epidemiological studies show that obesity is associated with an increased risk of esophageal adenocarcinoma, but not esophageal squamous cell carcinoma. The association between body mass index (BMI) and the risk of esophageal adenocarcinoma was recently systematically reviewed and meta-analysis of two cohort and 12 case-control studies including 2 488 cancers was performed [1]. Summary estimates indicated a statistically significant positive association between high BMI and risk of esophageal adenocarcinoma and the strength of the association increased with increasing BMI. The risk for overweight men (Summary relative risk, RR_Summary=1.80; 95% CI 1.5–2.2; P_{heterogeneity}=0.93) increased further for obese men (2.4; 95% CI 1.9–3.2; P_{heterogeneity}=0.35) when comparing to normal weight; corresponding risk estimates were similar among women (for overweight RR_Summary=1.5; 95% CI 1.1–2.2; P_{heterogeneity}=0.57 and for obesity 2.1; 95% CI 1.4–3.2; P_{heterogeneity}=0.94).

Although adenocarcinoma of the distal esophagus and gastric cardia share several characteristics and are difficult in clinical practice to be distinguished, these tumors should be regarded as separate diseases [2]. Results from meta-analysis indicated that overall associations with cardia cancer were heterogeneous, although stratification by study population provided homogenous results for studies from the USA or from Europe. A high BMI (>25 kg/m²) was positively associated with the risk of cardia adenocarcinoma (RR_Summary=1.5; 95% CI 1.3–1.8; P heterogeneity = 0.38) for men and women from the USA and Europe when comparing to BMI < 25 kg/m² [1]. Stratified analysis showed that BMI-adenocarcinoma association was strongest for cancers that were above 2 cm from the gastroesophageal junction and weakest in the gastric cardia. In a recent prospective study of 41 295 people from Australia, not included in the above-presented meta-analysis, BMI > 30 kg/m² as compared to BMI < 25 kg/m² was associated with RR = 3.7; 95% CI 1.1–12.4 for adenocarcinoma of combined 30 cases of lower third of esophagus and gastric cardia [3].

One of the proposed mechanisms is that high BMI is associated with higher risk for gastro esophageal reflux disease (GERD). In a recent meta-analysis of nine studies that examined the association between BMI and GERD symptoms, six found statistically significant association. In eight studies, there was a trend toward a dose-response relationship, with RR_Summary=1.4; 95% CI 1.2–1.8 for overweight and 1.9; 95% CI 1.5–2.6 for obesity [4]. Patients with long-standing frequent reflux symptoms were reported to have about 40-fold increased risk of esophageal adenocarcinoma [5]. Although the above results consistently show a positive link obesity-reflux-esophageal adenocarcinoma, there are still controversies regarding this issue [2]. Namely, the link between BMI and reflux is stronger among women and is further augmented by high estrogen exposure, so why is esophageal carcinoma predominant in men.

Mortality from esophageal cancer among men who never smoked was 1.8 (95% CI 1.1–2.9) for overweight and 1.9 (95% CI 0.9–4.0) for obesity, p _trend=0.04; corresponding risk estimates for all women were 1.5 (0.9–2.6) and 2.6 (1.4–5.1), p _trend=0.004.

Although the incidence and mortality of gastric cancer have fallen dramatically in the USA and other countries, this cancer site remains a major public health problem as a second leading cause of cancer death worldwide [6]. The two main tumour sites of gastric adenocarcinoma have shown diverging incidence trends during the last decades – decline of distal (non-cardia) tumors and increase of proximal (cardia) tumors, especially among men in western countries. Gastric cardia cancer shows similarities to esophageal cancer and was commented above. In contrast, incidence of gastric non-cardia cancer has not been shown to be positively associated with BMI. Mortality from total gastric cancer (not subdivided by subsites) among men was 1.2 (0.9–1.5) for BMI 30–35 kg/m² and 1.9 (1.2–3.1) for BMI > 35 kg/m², p _trend=0.03; corresponding risk estimates for women were 1.3 (1.0–1.7) and 1.1 (0.6–1.9), p _trend=0.46 [7].

There is accumulating evidence that obesity is positively associated with the risk of pancreatic cancer as was shown in a recent meta-analysis of six case-control and eight cohort studies involving 6 391 cases of pancreatic cancer [8]. The estimated increase in the risk corresponded to RR_Summary=1.2 (1.1–1.3) for obese people compared to people with normal body weight. Recently published results from a Swedish prospective cohort further supported this positive association; obese men and women had RR = 1.8 (1.04–3.2) compared to those with normal weight [9]. Results from an American cohort showed a statistically significant association with obesity at age 40 (RR = 1.7; 95% 1.1–2.6) but not with BMI at baseline (RR = 1.14) [10]. Mortality from pancreatic cancer among never smoking men was 1.2 (1.01–1.5) for BMI = 25–29.9 kg/m²: 1.3 (1.09–2.0) for BMI = 30–34.9 kg/m², and 2.6 (1.3–5.4) for BMI 35–40 kg/m², p _trend=0.005; corresponding risk estimates among all women were 1.1 (1.0–1.2), 1.3 (1.1–1.5), 1.4 (1.01–2.0) and 2.8 (1.7–4.4) for women with BMI > 40 kg/m²; p _trend<0.001 [7].

In a meta-analysis of BMI and colon cancer based on four cohort studies and two population-based case-control studies, an increment by one BMI unit was associated with statistically significant 3% risk (95% CI 2–4%). The dose-relationship corresponded to a 15% increase in the risk for an overweight person and 33% increased risk for an obese person [11]. Rectum cancer was analyzed separately in a fewer number of studies. According to estimates based on overweight and obesity prevalence in Europe in the late 1980s, about 11% of colon cancer in men and in women could be attributable to excess body weight [11]. In a large prospective cohort of over 900 000 adult Americans, mortality from colorectal cancer was associated positively with increasing excess body weight [7]. In men overweight was associated with RR = 1.2 (1.1–1.3), obesity 30–34.9 kg/m² with RR = 1.5 (1.3–1.7) and 35–40 kg/m² with RR = 1.8 (1.4–2.4) comparing to normal weight, p _trend<0.001: in women corresponding risk estimates were 1.1 (1.01–1.1), 1.3 (1.2–1.5), 1.4 (1.1–1.7) and for BMI > 40 kg/m² 1.5 (1.09–2.2), p _trend<0.001. In a London cohort of over 18 000 middle-aged men, cancer mortality in relation to overweight and obesity was evaluated separately for carcinoma of the colon and rectum [12]. Excess body weight was associated in a dose-response way with colon cancer mortality (p _trend=0.02) and risk ratio for obesity was 2.2 (1.3–3.8) comparing to normal weight; excluding the first 10 years of follow-up almost did not change risk estimates RR = 2.2 (1.2–4.0). Risk estimates for mortality from rectal cancer were based on smaller number of cases (104 rectum vs. 279 colon cancer); RR for being obese was 2.0 (0.8–4.9), p _trend=0.21. In a prospective cohort from Japan mortality from colon cancer was associated with excessive weight gain after the age of 20 in women but no such relationship was found in Japanese men [13].

In summary, the accumulated evidence indicates that excess body weight is positively associated with majority of gastrointestinal tract cancers. These findings call for measures to prevent obesity and to promote healthy lifestyle among the general population.