November 12, 2011

Dietary fibre and cancer

Q: Did you see the BBC news article about dietary fibre and bowel cancer?

A: No, but I’ll go look. [reading noises]

Q: Well?

A: They definitely  get points for linking to the article in the British Medical Journal (and it’s even an open access article that anyone can read).

Q: Haven’t we beaten that issue into the ground yet?

A: Apparently not.

Q: So, should we all be eating more fibre?

A: If you believe the article, yes.   They combined the results of lots of cohort studies, where people were asked about their diets and then followed up to see if they got cancer.  The people who ate more whole-grain fibre had lower rates of bowel cancer.

Q: Why wouldn’t we believe it?

A: Well, neither the BBC article nor the scientific paper makes any mention of the randomized trials of dietary fibre supplementation.

Q: I suppose I’m supposed to ask “What randomised trials?”

A: Yes.

Q: Consider it asked.

A: There have been several trials where people were randomly assigned to additional dietary fibre or their normal diet, to see if the fibre would prevent colon polyps.

Q: Which are?

A: Pre-malignant tumors that get found and removed in colonoscopies.  They can progress to actual cancer if they are left alone.

Q: And did the additional fibre prevent colon polyps?

A: As you will have guessed by now, no.

Q: So it doesn’t work?

A: Well, it’s conceivable that it might work. Polyps aren’t cancers, and perhaps fibre makes them grow more slowly or something. But the randomized trials have to be discouraging, and they weren’t even mentioned.

Q: How can you get away with just ignoring them?

A: The scientific paper was about studies that measured actual colon cancers, not polyps. There’s only been one randomized diet trial large enough to do that, and it was primarily of a low-fat diet, not of high fibre. [ It didn’t prevent cancer, either.]

Q: Still seems a bit sneaky.

A: You might very well think that. I couldn’t possibly comment.

avatar

Thomas Lumley (@tslumley) is Professor of Biostatistics at the University of Auckland. His research interests include semiparametric models, survey sampling, statistical computing, foundations of statistics, and whatever methodological problems his medical collaborators come up with. He also blogs at Biased and Inefficient See all posts by Thomas Lumley »

Comments