Chronological Age and Risk of Chemotherapy Nonfeasibility: A Real-Life Cohort Study of 153 Stage II or III Colorectal Cancer Patients Given Adjuvant-modified FOLFOX6.

OBJECTIVES: To assess nonfeasibility of adjuvant-modified FOLFOX6 chemotherapy in patients with stage II or III colorectal cancer.

METHODS: Consecutive patients managed between 2009 and 2013 in 2 teaching hospitals in the Paris urban area were included in the CORSAGE (COlorectal canceR, AGe, and chemotherapy fEasability study) cohort study. Nonfeasibility was defined by the frequencies of empirical first-cycle dose reduction (>15%), early discontinuation (<12 cycles), and low relative dose intensity (RDI) (<0.85). Risk factors for chemotherapy nonfeasibility were identified using multivariate logistic regression.

RESULTS: Among 153 patients, 56.2% were male (median age, 65.6 y; 35.3%≥70 y; 7.3% with performance status [PS]≥2). For 5-fluorouracil (5-FU), 20.9% of patients had first-cycle dose reduction and 28.1% early discontinuation; RDI was 0.91 (25th to 75th percentiles, 0.68 to 0.99). Factors independently associated with first-cycle 5-FU dose reduction were aged 65 to 69 years versus those younger than 65 years (adjusted odds ratio [aOR], 5.5; 95% confidence interval [CI], 1.5-19.9) but not age 70 years and older, PS≥2 (aOR, 6.02; 95% CI, 1.15-31.4), higher Charlson Comorbidity Index (aOR1-point increase, 1.4; 95% CI, 1.05-1.82), or larger number of medications (aOR 1-medication increase, 1.19; 95% CI, 1.00-1.42). Oxaliplatin dose reduction occurred in 52.3% of patients and early discontinuation in 62.7%; the latter was more common in the 70 years and older group (92.6% vs. 74.6% in the <65-y group; P=0.01); RDI was 0.7 (95% CI, 0.55-0.88).

CONCLUSIONS: In the real-world setting, compared with their younger and older counterparts, patients aged 65 to 69 years given modified FOLFOX6 for stage II or III colorectal cancer had higher frequencies of 5-FU nonfeasibility defined based on first-cycle dose reduction, early discontinuation, and RDI; and these differences were independent from PS, comorbidities, and number of medications.