Early-onset colorectal cancer (EOCRC, defined here as diagnosis before 50 years) is increasing in many regions. The relationship between body mass index (BMI) at diagnosis and outcomes after curative-intent surgery in colorectal cancer (CRC) is complex. An established but paradoxical observation in CRC overall is that overweight patients sometimes show better outcomes despite obesity being a recognized risk factor. This study sought to evaluate the association between BMI at diagnosis and recurrence-free survival (RFS) in EOCRC and to compare that relationship with average-onset colorectal cancer (AOCRC, ≥50 years).
The analysis used a retrospective cohort from Sahlgrenska University Hospital comprising 1,459 patients who underwent curative-intent surgery for colorectal adenocarcinoma. The cohort was partitioned into EOCRC (n = 159) and AOCRC (n = 1,300). Patient data were obtained from electronic health records, de-identified after collection, and stored securely. All participants had provided informed consent.
The investigators applied Cox proportional hazards models to assess the relation between BMI and RFS, adjusting for tumour stage, tumour location, and tumour differentiation. BMI was considered both categorically and as a continuous variable modelled using restricted cubic splines to allow for non-linear associations. Model fit and significance of non-linear terms were evaluated with likelihood-ratio tests (LRT). An interaction analysis assessed whether the BMI–RFS relationship differed between EOCRC and AOCRC cohorts.
In the EOCRC subgroup, modelling BMI as a continuous variable revealed a statistically significant association with RFS (LRT p = 0.02). The spline model described a U-shaped relationship: the lowest hazard for recurrence was observed at approximately BMI 27, while hazards were higher at BMI values below 20 and above 30. This pattern suggests that both underweight and obese BMI ranges were associated with worse RFS compared with a mid-range BMI in this younger cohort.
By contrast, in the AOCRC cohort continuous BMI was not associated with RFS (LRT p = 0.15). The restricted cubic spline for BMI in this older cohort was effectively flat, indicating no detectable non-linear relationship between BMI at diagnosis and recurrence-free survival after adjustment for stage, location, and differentiation.
An interaction test comparing the BMI–RFS relationships between EOCRC and AOCRC showed a statistically significant difference (LRT p = 0.046). This indicates that the association between BMI and recurrence-free survival varied by age-at-onset group in this dataset.
The authors interpret the observed U-shaped BMI–RFS association in EOCRC as consistent with a version of the obesity paradox that may be specific to early-onset disease. Proposed explanations include differences in body composition between younger and older patients, distinct tumour biology in EOCRC, or variations in systemic metabolism that modify the prognostic impact of BMI. The authors also acknowledge that methodological biases could contribute to the observed pattern.
The report highlights the need for more granular measures beyond BMI; the authors recommend inclusion of biomarkers and refined body composition assessments in future research to elucidate mechanisms. The study is retrospective and observational, and the source notes potential methodological biases as an alternative explanation for the findings. Specific limitations such as detailed confounders beyond stage, location, and differentiation, follow-up duration, or absolute event counts were not detailed in the provided source text.
The study was conducted on Swedish patients with informed consent and was approved by the Regional Ethical Review Board in Gothenburg, Department 1 Medicine (approval no. 118-15; approved 3 June 2015). Data were de-identified after extraction from electronic health records. The authors declared no competing interests and reported adherence to relevant ethical guidelines.
In this single-institution retrospective cohort, BMI at diagnosis showed a U-shaped association with recurrence-free survival in patients with early-onset colorectal cancer, with the lowest recurrence hazard near BMI 27, whereas no association was observed in older patients. These findings suggest the prognostic implications of BMI may differ by age at CRC onset and support further study using biomarkers and body composition measures to determine whether observed patterns reflect biology or bias.