Smoking Linked to Shorter Disease-free Survival in Stage III Colon Cancer 

Effect most evident in BRAF wild-type and KRAS mutant disease

Get Permission

This analysis ... provides evidence that the effects of smoking may extend beyond an adverse impact on colon cancer risk to also adversely impact outcomes after diagnosis.

—Amanda Phipps, PhD, MPH, and colleagues

According to an analysis reported by Amanda Phipps, PhD, MPH, Assistant Professor of Epidemiology at the University of Washington and Assistant Member at Fred Hutchinson Cancer Research Center, Seattle, and colleagues in the Journal of Clinical Oncology, patients with stage III colon cancer who ever smoked had significantly reduced disease-free survival compared with never-smokers.1 The association between smoking and reduced disease-free survival was most marked in patients with BRAF wild-type or KRAS mutant colon cancer. Prior studies have indicated that cigarette smoking is associated with a modestly increased risk for colon cancer, particularly after long durations or high levels of exposure.

Study Details

The study involved 1,968 patients with stage III disease from the North Central Cancer Treatment Group phase III trial N0147 examining FOLFOX (infusional fluorouracil, leucovorin, and oxaliplatin) vs FOLFOX plus cetuximab (Erbitux) as adjuvant therapy. Questionnaires regarding smoking and other risk factors were completed by patients prior to study randomization. Of the total of 2,686 patients enrolled, 718 did not complete questionnaires.

A number of significant differences were seen between never-smokers (n = 931) and ever-smokers (n = 1028) at baseline. Ever-smokers were more likely to be older (mean age = 59.4 vs 56.6 years, P < .001), be male (57% vs 47%, overall P < .001), have mismatch repair gene–deficient status (15% vs 10%, overall P = .001), have mutated BRAF (16% vs 11%, overall P = .006), have poorer ECOG performance status (0 in 74% vs 79%, and 1 in 25% vs 20%, overall P = .02), never engage in vigorous physical activity (71% vs 64%, overall P < .001), and be current alcohol consumers (18% vs 45% were never-consumers, and 45% vs 33% were current consumers, overall P < .001).

There was a borderline significant trend for never-smokers to have greater body mass index (≥ 30.0 kg/m2 for 38% vs 32%, overall P = .06). There were no significant differences between groups in allocation to adjuvant therapy study arm, tumor subsite, number of affected nodes, T stage, or KRAS mutation status.

Disease-free Survival in Ever-smokers

Median follow up in patients not experiencing events was 3.5 years. On unadjusted analysis, ever-smokers had a significant 21% increased risk of recurrence or death (3-year disease-free survival 70% vs 74%, hazard ratio [HR] = 1.21, P = .03). Associations with smoking status were assessed in multivariate analysis adjusting for tumor site, number of involved lymph nodes, T stage, mismatch repair gene status, performance status, physical activity, body mass index, alcohol consumption, age, and sex.

On multivariate analysis, the significantly increased risk for recurrence or death persisted for ever-smokers (HR = 1.23, P = .03). The association with poorer disease-free survival was somewhat stronger for current smokers (HR = 1.47, 95% confidence interval [CI] = 1.04–2.09) than for former smokers (HR = 1.20, 95% CI = 0.99–1.46).

With regard to associations of disease-free survival with smoking patterns, there were no statistically significant associations with patterns of numbers of cigarettes per day, time since smoking cessation, or duration of smoking interval, although the data suggested that disease-free survival was poorest among patients who smoked more than 30 cigarettes per day (HR = 1.35, 95% CI = 0.86–2.12, for those smoking 31–40 per day; HR = 1.40, 95% CI = 0.80–2.44, for those smoking > 40). There was a significant trend for greater risk for recurrence or death in patients who initiated smoking at age 20 or older (HR = 1.39, 95% CI = 1.08–1.80, P = .04 for trend across age at initiation).

BRAF Wild-type or KRAS Mutant Disease

Analyses stratified by patient and clinical characteristics, and adjusted for the same factors noted above, showed a significant interaction between smoking and BRAF mutation status in relation to disease-free survival (overall P = .03). Poorer disease-free survival in ever-smokers vs never-smokers was limited to patients with BRAF wild-type tumors (HR = 1.36, 95% CI = 1.11–1.66), with smokers with BRAF mutant tumors having nonsignificantly reduced risk (HR = 0.80, 95% CI = 0.50–1.29).

Although the interaction of KRAS mutation status and smoking status for disease-free survival only approached statistical significance (P = .07), there was evidence of poorer disease-free survival for smokers vs nonsmokers with KRAS mutant disease. Among patients with KRAS mutant disease, ever-smokers had a significant 50% increased risk for recurrence or death (HR = 1.50, 95% CI = 1.12–2.00), whereas risk was not significantly increased in ever-smokers with KRAS wild-type disease (HR = 1.09, 95% CI = 0.85–1.39).

The effects of BRAF and KRAS mutation status were most marked in comparisons of current smokers vs never-smokers, with significantly poorer disease-free survival observed for current vs never-smokers with BRAF wild-type disease (HR = 1.60, 95% CI = 1.10–2.32) and KRAS mutant disease (HR = 2.30, 95% CI = 1.40–3.77). Compared with never-smokers, risk in current smokers was nonsignificantly reduced among those with BRAF mutant disease (HR = 0.82, 95% CI = 0.29–2.35) and nonsignificantly increased in those with KRAS wild-type disease (HR = 1.08, 95% CI = 0.65–1.77).

There were no significant interactions among smoking status and disease-free survival across sex, age, T stage, number of involved nodes, tumor subsite, or mismatch repair gene status stratifications. However, disease-free survival was significantly poorer for ever-smokers vs never-smokers who were male (HR = 1.33, 95% CI = 1.02–1.73), aged 50 years or less (HR = 1.53, 95% CI = 1.03–2.26), had stage T3 disease (HR = 1.29, 95% CI = 1.04–1.60), had one to three involved nodes (HR = 1.52, 95% CI = 1.13–2.04), had disease of the distal colon (HR = 1.36, 95% CI = 1.01–1.82), and had mismatch repair gene–proficient status (HR = 1.25, 95% CI = 1.03–1.53).

Analyses of time to recurrence among ever-smokers and never-smokers produced results similar to the analyses of disease-free survival.

Mechanisms Underlying Effects?

The authors noted that the finding of significantly reduced disease-free survival in smokers with BRAF wild-type, KRAS mutant, or mismatch repair gene–proficient tumors—disease subtypes not previously associated with smoking in risk factor studies—suggests an impact of smoking on colon cancer progression or response to treatment via as yet undefined pathway-specific mechanisms. As an example of one of the many ways the numerous compounds in cigarette smoke could affect multiple pathways of tumor initiation and progression and treatment response, they cite the effect of nicotine in increasing proliferation and decreasing apoptosis, perhaps via activation of the PI3K/AKT pathway; this effect could be amplified in the presence of a KRAS mutation, since KRAS also activates PI3K.

The authors concluded:

Overall, smoking was significantly associated with shorter [disease-free survival] and time to recurrence in patients with colon cancer. These adverse relationships were most evident in patients with BRAF wild-type or KRAS mutated colon cancer…. This analysis contributes to a sparse literature [on the effects of smoking on outcomes of colon cancer] and provides evidence that the effects of smoking may extend beyond an adverse impact on colon cancer risk to also adversely impact outcomes after diagnosis. Further research is needed to confirm and better understand observed differences in the association between smoking and survival outcomes in patients with colon cancer according to BRAF and KRAS mutation status and the mechanisms responsible for these patterns of association. ■

Disclosure: Dr. Phipps reported no potential conflicts of interest.


1. Phipps AI, Shi Q, Newcomb PA, et al: Associations between cigarette smoking status and colon cancer prognosis among participants in North Central Cancer Treatment Group phase III trial N0147. J Clin Oncol. April 1, 2013 (early release online).