Nuts after stage III colon cancer: survival associations, not a nut trial
An 826-patient analysis linked nut intake to outcomes while leaving the dietary exposure observational.

Illustrative food image, not a photograph of the study intervention.
The chemotherapy trial did not randomize nuts
This analysis followed 826 eligible patients with stage III colon cancer who were enrolled in a randomized adjuvant chemotherapy trial. Their dietary intake was reported using food frequency questionnaires. Although the parent study randomized chemotherapy, the nut comparison itself was prospective and observational. The investigators used Cox proportional hazards regression to assess associations between nut intake, cancer recurrence, and mortality. Calling this a randomized nut trial would therefore misdescribe the evidence. Its particular contribution is a dietary follow-up analysis in a clearly defined cancer population, not an experiment testing nuts as a substitute for oncology treatment.
Fadelu T, Zhang S, Niedzwiecki D, Ye X et al.. Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 2018. PMID 29489429 · DOI
Two survival outcomes, two estimates
After a median follow-up of 6.5 years, participants reporting two or more servings of nuts per week had an adjusted hazard ratio of 0.58 for disease-free survival compared with participants who abstained from nuts. The 95% confidence interval was 0.37 to 0.92, and the trend P value was 0.03. For overall survival, the corresponding hazard ratio was 0.43, with a confidence interval of 0.25 to 0.74 and a trend P value of 0.01. These are separate outcome estimates from an observational comparison; neither number is an absolute survival probability or proof that adding nuts produces that result.
Fadelu T, Zhang S, Niedzwiecki D, Ye X et al.. Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 2018. PMID 29489429 · DOI
The tree-nut subgroup deserves its own label
In subgroup analysis, the apparent benefit was confined to tree nut intake. The reported hazard ratio for disease-free survival was 0.54, with a 95% confidence interval of 0.34 to 0.85 and a trend P value of 0.04. For overall survival, the tree-nut hazard ratio was 0.47, with a confidence interval of 0.27 to 0.82 and a trend P value of 0.04. The subgroup wording is important: the abstract does not present every nut category as equally associated with outcomes. Nor does a subgroup association establish the effect of any particular tree nut, preparation, or recipe.
Fadelu T, Zhang S, Niedzwiecki D, Ye X et al.. Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 2018. PMID 29489429 · DOI
A frequency comparison is not a prescription
The authors report that the association of total nut intake with improved outcomes persisted across other known or suspected risk factors for recurrence and mortality. They conclude that diets containing more nuts may be associated with reduced recurrence and death in this stage III colon cancer population. The cautious phrase is essential because intake was reported, not assigned. The two-or-more-servings-per-week category describes the study comparison rather than a tested treatment dose. This paper therefore supports a specific survival association worth studying further, while leaving unanswered whether deliberately changing nut intake would reproduce the observed differences in disease-free or overall survival.
Fadelu T, Zhang S, Niedzwiecki D, Ye X et al.. Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 2018. PMID 29489429 · DOI