Healthier pre-diagnosis lifestyle and long-term survival in patients after diagnosis of colorectal cancer: Evidence from the Singapore Chinese Health Study

Purpose To quantify survival differences between colorectal cancer (CRC) patients with healthier and those with less healthy pre-diagnosis lifestyles using restricted mean survival time (RMST) at 5 and 10 years, and to assess which lifestyle components contribute to these differences. Methods Incident CRC cases (n = 2124) were identified from the Singapore Chinese Health Study, a prospective cohort of 63 257 adults enrolled in 1993-1998. Cancer status and cause of death were obtained via linkage with the National Disease Registry through 31 December 2015. A pre-diagnosis lifestyle score (0-7) incorporated BMI and smoking, sleep duration, diet, and weekly physical activity at baseline. RMST differences comparing less healthy (scores 0-3) versus healthier (4-7) lifestyles were estimated at 5- and 10-years post-diagnosis, adjusted for CRC stage (early, late) and age at diagnosis, sex, education attainment, and the interval between lifestyle assessment and diagnosis. Results Over a median follow-up of 4.8 years (IQR 1.2-11.8) from CRC diagnosis, 1557 deaths occurred, including 1103 CRC-specific deaths. Late-stage diagnosis was associated with shorter survival than early stage by 1.53 years (95% CI 1.35 to 1.72) at 5 years and 3.33 years (2.93 to 3.72) at 10 years. The median interval between lifestyle assessment and diagnosis was 12.2 years (IQR 6.9 to 16.7). RMST for all-cause deaths was significantly improved by a healthier lifestyle but results for CRC-specific deaths were mixed. For all-cause mortality, adjusted RMST differences (less healthy - healthier) were −0.15 (95% CI −0.32 to 0.01) at 5 years and −0.37 (−0.71 to −0.02) at 10 years. CRC-specific mortality: At 5 years, the adjusted RMST difference (less healthy − healthier) was −0.15 years (−0.32 to 0.02), and at 10 years it was −0.34 years (−0.71 to 0.03). Among patients diagnosed at age ≤70 years, healthier pre-diagnosis lifestyles were associated with modestly longer survival, with a 10-year CRC-specific RMST gain of 0.58 years (0.03 to 1.13), while other estimates were borderline or not significant. Lifestyle components associated with improved CRC-specific survival included diet (lower vs higher AHEI category: 5-year RMST difference −0.20 (−0.38 to −0.02); 10-year −0.46 (−0.84 to −0.08)) and weekly physical activity (no vs yes: 5-year −0.19 (−0.37 to −0.02); 10-year −0.42 (−0.78 to −0.06)). Similar results were obtained for all-cause death. Smoking and sleep duration were not significantly associated with survival. Conclusions A healthier pre-diagnosis lifestyle was associated with improved long-term survival in CRC patients. Short-term differences were modest after accounting for CRC stage at diagnosis, underscoring the importance of early detection. Having maintained healthy behaviors before diagnosis may favorably influence long-term outcomes among CRC patients.