Background: Obesity is a well-established risk factor for colorectal cancer (CRC) incidence. However, in cancer survival, an inverse association between body mass index (BMI) and mortality—often referred to as the “obesity paradox”—has been reported. While most studies have focused on BMI measured at diagnosis, the prognostic significance of changes in body weight across the diagnostic period remains poorly understood. Because post-diagnostic weight loss or gain may reflect disease progression, treatment effects, and nutritional status, examining transitions in weight status over time may provide more clinically meaningful prognostic information than static BMI measures.
Objectives: To investigate the association between pre- to post-diagnostic body weight change and fatality among CRC patients, using transitions across BMI ranges as indicators of weight change, stratified by cancer subsite.
Methods: We conducted a retrospective cohort study using the K-CURE Colorectal Cancer Collaborative Staging and Cancer Public Library Database. Patients were stratified by cancer subsite (colon and rectum). Changes in body weight status were operationalized as transitions across BMI ranges (kg/m²: ≤20, 20–<23, 23–<25, ≥25) between the pre- and post-diagnostic periods. Patients who remained within the BMI range of 20–<23 kg/m² served as the reference group. Adjusted hazard ratios (aHRs) and 95% confidence intervals (CIs) for all-cause fatality were estimated using multivariable Cox proportional hazards models adjusted for age, sex, and AJCC 7th edition cancer stage.
Results: Among 12,935 patients with colon cancer, 5,030 with available pre- and post-diagnostic BMI data were included; most were men aged 60–69 years and diagnosed at stage I. Among 8,838 patients with rectal cancer, 3,195 eligible patients were identified. Associations between weight change and fatality differed by cancer subsite. In colon cancer, post-diagnostic weight loss was strongly associated with increased fatality. Compared with patients remaining within the BMI range of 20–<23 kg/m², transitions from BMI 23–<25 kg/m² to ≤20 kg/m² (aHR: 5.34 [95% CI: 2.60–10.98]) and from BMI ≥25 kg/m² to ≤20 kg/m² (10.96 [3.31–36.32]) were associated with elevated risks. Weight loss from BMI 20–<23 kg/m² to ≤20 kg/m² (2.34 [1.35–4.06]) and transitions to the BMI range of 20–<23 kg/m² from higher ranges (from 23–<25 kg/m²: 2.02 [1.18–3.46]; from ≥25 kg/m²: 4.62 [2.31–9.24]) were also associated with increased fatality. In rectal cancer, associations between weight change and fatality were weaker. A significantly increased risk was observed only for transitions from BMI 20–<23 kg/m² to ≤20 kg/m² (1.81 [1.05–3.11]), with no statistically significant associations for other weight-loss patterns.
Conclusions/Implications: Pre- to post-diagnostic body weight loss was differentially associated with fatality by CRC subsite. Substantial weight loss, particularly transitions to BMI ≤20 kg/m², was strongly associated with increased fatality in colon cancer but not in rectal cancer. These findings suggest that dynamic weight change, rather than BMI at diagnosis alone, provides important prognostic information. From a public health perspective, unintended weight loss among colon cancer patients may serve as an early marker of poor prognosis, underscoring the importance of timely nutritional assessment and survivorship care.