Early-life fiber and breast cancer: timing, recall, and a cohort signal
A Nurses’ Health Study II analysis separated early-adult fiber intake from recalled high-school diet.

Illustrative food image, not a photograph of the study intervention.
Two windows of dietary information
The Nurses’ Health Study II analysis examined dietary fiber during adolescence and early adulthood in relation to breast cancer risk. Among 90,534 premenopausal women who completed a dietary questionnaire in 1991, the researchers documented 2,833 invasive breast cancer cases over 20 years. A subset of 44,263 women also completed a questionnaire in 1998 about their high-school diet; 1,118 breast cancer cases were documented among these women by the end of follow-up. The two dietary windows therefore did not have identical sample sizes or methods of reporting. In particular, the high-school information was supplied later rather than collected during high school.
Farvid MS, Eliassen AH, Cho E, Liao X et al.. Pediatrics, 2016. PMID 26908709 · DOI
The early-adult comparison
For total dietary fiber in early adulthood, the relative risk comparing the highest with the lowest intake quintile was 0.81, with a 95% confidence interval of 0.72 to 0.91. The trend P value was 0.002. The researchers used multivariable-adjusted Cox proportional hazards regression to model risks across fiber categories. This is a comparison of reported dietary intake within a cohort, not a trial assigning a high-fiber diet. The estimate describes relative breast cancer risk in those categories; it does not state an individual's absolute risk or show that a particular high-fiber food independently accounts for the association.
Farvid MS, Eliassen AH, Cho E, Liao X et al.. Pediatrics, 2016. PMID 26908709 · DOI
Soluble and insoluble fiber were both examined
The abstract reports inverse associations for both soluble and insoluble fiber in early adulthood. Comparing the highest with the lowest quintile, soluble fiber had a relative risk of 0.86, with a 95% confidence interval of 0.77 to 0.97 and a trend P value of 0.02. Insoluble fiber had a relative risk of 0.80, with a confidence interval of 0.71 to 0.90 and a trend P value below 0.001. These parallel estimates do not establish that one fiber type is a superior intervention. They describe two dietary exposure analyses rather than a randomized head-to-head comparison of fiber treatments.
Farvid MS, Eliassen AH, Cho E, Liao X et al.. Pediatrics, 2016. PMID 26908709 · DOI
The adolescent estimate was less precise
For adolescent total fiber intake, the highest-versus-lowest quintile relative risk was 0.84, with a 95% confidence interval of 0.70 to 1.01 and a trend P value of 0.04. The interval for that category comparison includes 1.00, even though the trend test was below 0.05. Averaging adolescent and early-adult fiber intake gave a relative risk of 0.75, with a confidence interval of 0.62 to 0.91. The authors interpret the findings as supporting an early-life fiber hypothesis. Recalled high-school diet and the absence of dietary randomization remain important boundaries when interpreting that hypothesis as evidence rather than a prevention guarantee.
Farvid MS, Eliassen AH, Cho E, Liao X et al.. Pediatrics, 2016. PMID 26908709 · DOI