From January 1995 to July 2003, 2138 patients from a total population of 2765 patients with histologically confirmed adenocarcinoma of the colon were enrolled in this study; 14 nonmetastatic patients without available BMI data and 613 patients with metastatic disease were excluded. The included patients had regular follow-up visits with a healthcare professional until December 2008 (i.e., a postoperative follow-up period of at least 5 years) or until death.

The demographic and clinicopathologic data evaluated for each patient included age, gender, body height and weight, tumor stage (as defined according to the American Joint Committee on Cancer TNM staging system, sixth edition, New York: Springer-Verlag, 2002), tumor location, degree of tumor differentiation, timing of surgery (i.e., elective or emergent), medical illness, and preoperative laboratory data.

BMI was calculated as the weight in kilograms divided by the height in meters squared. The following BMI categories were established according to the classification of the Department of Health (DOH) of Taiwan:[8] underweight (BMI < 18.5 kg/m²), normal weight (BMI = 18.5-23.9 kg/m²), overweight (BMI = 24.0-26.9 kg/m²), and obese (BMI ≥ 27 kg/m²). For long-term outcome, BMI was also categorized according to the World Health Organization (WHO) classification[9] as follows: underweight (BMI < 18.5 kg/m²), normal weight (BMI = 18.5-24.9 kg/m²), overweight (BMI = 25-29.9 kg/m²), and obese (BMI ≥ 30 kg/m²) to enable comparison with the Taiwanese classification, since The Expert Consultation on BMI in Asian populations recommended that the current WHO BMI cut-off points be retained as the international classification[10].

Local recurrence or distant metastasis was confirmed histologically or radiographically, while postoperative mortality was defined as death within 30 d of the primary surgery. The patients were divided into three age groups: younger than 40 years (younger group), 40-75 years (middle age group), and older than 75 years (elderly group). The carcinoembryonic antigen (CEA) level was considered to be abnormal at > 5 ng/mL. Tumor location was categorized as right (from the cecum to the transverse colon) or left (from the splenic flexure to the sigmoid colon). Hypoalbuminemia was defined as a serum albumin level of < 35 g/L and apparent anemia was defined as a hemoglobin level of < 10 g/dL.

Survival curves were constructed using the Kaplan-Meier method and then compared using the log-rank test. OS was calculated as the number of years from primary surgery to the date of death. DFS was calculated as the number of years from primary surgery to either the first disease recurrence or death. CSS was calculated as the number of years from primary surgery to the first of disease recurrence. The two arms were compared by Pearson χ² test and independent-samples t tests to detect any differences in proportions and means. A Cox regression model was used for multivariate analysis. All P values were two-tailed, and they were considered to be statistically significant at < 0.05.

The study population comprised 2138 patients with colon cancer, of whom 1109 were males and 1029 were females. BMI in this study ranged from 12.2 kg/m² to 49.0 kg/m², with a mean of 23.4 kg/m². Of the 2138 patients, 164 (7.7%) were underweight, 1109 (51.9%) were normal weight, 550 (25.7%) were overweight, and 315 (14.7%) were obese.

The characteristics of the patients and tumors are presented stratified according to BMI category in Table 1. The age of the entire study population was 61.4 ± 13.9 years (mean ± SD); those of the underweight, normal-weight, overweight, and obese patients were 63.5 ± 17.3 years, 61.8 ± 14.0 years, 61.9 ± 12.4 years, and 61.7 ± 12.3 years, respectively. The mean age did not differ significantly with the BMI category (P = 0.828). However, the distribution of age groups differed significantly with the BMI category, with the proportions of younger and elderly patients being higher in underweight-patients group than in the other groups (P < 0.001). The gender distribution also differed significantly between the underweight patients and the other groups, with underweight patients being more likely to be female. The tumor location also differed significantly, with the proportion of right colon cancer decreasing as the BMI category increased (P < 0.001).

With respect to the distribution of tumor stage among the BMI categories, the number of stage I tumors was lower in the underweight and normal-weight patients than in the overweight and obese patients. The obese patients had the lowest number of stage III tumors.

The proportion of emergent operations did not differ significantly with the BMI category: 3.7% of underweight, 3.4% of normal-weight, 2.5% of overweight, and 2.2% of obese patients (P = 0.584).

Underweight patients were the most likely to exhibit apparent anemia (hemoglobin < 10 g/dL) and have hypoalbuminemia and body weight loss (P < 0.001). The proportion of patients with body weight loss, hypoalbuminemia, and apparent anemia decreased as the BMI category increased. The percentage of patients with an abnormal preoperative CEA level did not differ significantly with the BMI category.

Finally, with regard to associated medical illnesses, the percentage of patients with hypertension or diabetes mellitus increased with the BMI category. Obese patients were the most likely to have heart disease. The other associated comorbidities including previous stroke, asthma, peptic ulcer disease, chronic hepatitis, renal insufficiency, and others (e.g., gall stones and thyroid disease) did not differ significantly with the BMI category.

The postoperative morbidity, anastomostic leakage, and mortality rates did not differ significantly between the underweight, normal-weight, overweight, and obese patients: 12.8%, 12.9%, 13.3% and 15.6%, respectively (P = 0.667); 1.2%, 1.4%, 1.6%, and 1.9%, respectively (P = 0.880); and 3.7%, 1.3%, 1.1%, and 0%, respectively (P = 0.007).

Data from all patients who received surgery were used to analyze the long-term outcome. Univariate analysis was performed using the Kaplan-Meier method, along with the log-rank test and Cox proportional-hazards model. The results of our statistical analysis are presented in Table 2, and the survival curves are shown in Figure 1A (OS), Figure 1B (DFS), and Figure 1C (CCS).

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Overall survival curves (A) , disease-free survival curves (B) and cancer-specific survival curves (C) relative to body mass index category using the Kaplan-Meier method, and comparison using the log-rank test. BMI: Body mass index.

Univariate analysis revealed that OS (P < 0.001) and DFS (P = 0.001) were lowest in the underweight patients. The OS was marginally higher (P = 0.046) in overweight patients than in normal-weight patients, but the DFS (P = 0.103) did not differ significantly between the two groups. No significant difference in OS (P = 0.500) or DFS (P = 0.807) was observed between normal-weight and obese patients.

CSS did not differ significantly with the BMI category (P = 0.346). The results were similar when these patients were categorized according to the WHO BMI classifications.

Adjuvant chemotherapy was offered to 58.0%, 70.1%, 73.9%, and 70.8% of underweight, normal-weight, overweight, and obese patients with stage III tumors, respectively (P = 0.096). In total, 70.2% of patients with stage III tumors received adjuvant chemotherapy.

Since the compositions of patients and tumors differed with the BMI category, multivariate analysis was used to determine the effect of BMI along with other confounding factors on the long-term outcome. The variables of the Cox regression model included BMI, TNM stage, age group, gender, comorbidities (patients were divided into three groups: without, with one or two kinds, and with more than two kinds of comorbidity), CEA (normal vs abnormal), hemoglobin (< 10 g/L vs≥ 10 g/L), albumin (normal vs hypoalbuminemia), timing of surgery (elective vs emergent), postoperative morbidity (with vs without), tumor location (right vs left), histologic type (adenocarcinoma vs mucinous vs signet-ring type), and histologic grade (high vs moderate vs low differentiation). The hazard ratio (HR) of each BMI category was compared with that of the normal-weight patients. The results of multivariate analysis for OS, DFS, and CSS are listed in Table 3.

As documented in Table 4, underweight patients had a significantly worse OS [HR, 1.58; 95% confidence interval (CI), 1.23-2.05; P < 0.001] and DFS (HR, 1.49; 95% CI, 1.16-1.93; P = 0.002), but their CSS did not differ significantly (HR, 1.33; 95% CI, 0.94-1.87; P = 0.107) when compared with the normal-weight patients. OS, DFS, or CSS did not differ significantly between the normal-weight, overweight, and obese patients.

To determine whether the prognostic effect of BMI was related to patient gender, a multivariate analysis was applied to separated data from male and female patients. The results of this statistical analysis yielded the same findings between the genders: CSS did not differ significantly with the BMI category for all patients or for male or female patients.

The prevalence of obesity and the incidence of colon cancer continue to increase in Taiwan as in other developed countries. Studies that have investigated the influence of body mass index (BMI) on the outcomes of colon cancer patients have produced inconsistent findings. These previous reports were mostly from Western countries, and lacked data regarding Asian people. The authors therefore investigated the role of BMI in general colon cancer patients of Taiwan, in terms of the clinical characteristics and short- and long-term outcomes.

BMI is a simple index that is used to classify individuals as being underweight, overweight, or obese. The cut-off points of BMI for classifications vary in different countries. The World Health Organization BMI classifications were used to evaluate the long-term outcome of colon cancer survivors and were compared with those determined using the Taiwan BMI classification system.

The findings of the present study demonstrate that a low BMI could be a marker of weight loss, blood loss, nutrition depletion, and more-advanced disease. For the population of Taiwan, which represents both Chinese people and other Asians in general, BMI does not appear to significantly affect the colon-cancer-specific survival, although underweight patients do appear to have a higher postoperative mortality and worse overall survival, and are less likely to experience disease-free survival than patients in the other BMI categories.

BMI is a simple index that can be used to evaluate the likelihood of colon cancer patients developing weight loss, blood loss, nutrition depletion, and more-advanced disease.

BMI was calculated as the weight in kilograms divided by the height in meters squared. The classification of BMI was established to evaluate health risks such as type 2 diabetes and cardiovascular disease.

This study investigated the effect of body mass index on the characteristics and overall outcome of colon cancer in Taiwan. This manuscript addresses a highly relevant subject and could be important for the cancer field.