Page 94 - Risk quantification and modification in older patients with colorectal cancer

Chapter 5
BMI skeletal muscle mass < 41 cm2/m2. For low skeletal muscle density, body mass index (BMI) specific thresholds were used independently of gender: BMI< 25 kg/ m2 HU< 41 and BMI ≥25 kg/m2 HU < 33.
Secondly, we determined the sex-specific quartiles for skeletal muscle mass and density in our cohort. Patients were stratified into quartiles to allow comparison between the four groups and to compare patients with the lowest quartile of skeletal muscle mass and density with those with the highest skeletal muscle mass and density.
Physical performance
The preoperative use of a mobility aid was used as a marker of physical performance. Subgroup analysis was intended for patients with low muscle mass and low muscle density that also used a mobility aid preoperatively. Only patients with a Katz ADL score of ≥2 was considered functionally dependent.
Statistical analysis
OS was calculated from the day of surgery until death (all possible causes) or last follow-up. Continuous variables are reported mean with standard deviation (SD) and categorical variables with numbers and percentages. The Mann–Whitney U test or X2 test were used to compare groups. OS was depicted through the Kaplan–Meier method. To assess the association between skeletal muscle mass, skeletal muscle density and physical performance with postoperative outcomes, logistic regression models were used to estimate Odds Ratios (ORs) with their corresponding 95% Confidence Interval (CI). To study the effect of these variables on OS, univariable and multivariable Cox-proportional hazards model was used to estimate hazard ratios (HRs) with their corresponding 95% CI.
The following confounders were considered: age, gender, BMI (for skeletal muscle mass and density only)21 and tumour stage (for OS analysis). Comorbidity was considered a confounder for physical performance and outcomes.22 A sensitivity analysis was performed on colon cancer patients only and this did not change our findings. Therefore, analyses were performed on all patients.
For assessing the additional effect of physical performance in patients with a low muscle mass and low muscle density, the interaction between muscle mass/
92