Patrick Holford blogs cohort study which finds that “Multivitamin use was not related to total mortality”

I was interested to see Holford blogging about Pocobelli et al’s recent article on “Use of supplements of multivitamins, vitamin C, and vitamin E in relation to mortality”. Holford reports the study as finding that “Multivitamin use cuts heart disease risk”. However, I have a number of concerns about Holford’s interpretation of this study:

• This is a cohort study, not an intervention study. Correlation does not necessarily equate to causation: it may not be the multivitamins that are reducing heart disease risk. For example, as the paper’s authors make clear, “Some of the associations we observed were small and may have been due to unmeasured healthy behaviors that were more common in supplement users.”
• Holford argues that “these results appear to contradict the meta-analyses published last year”. However, assuming he means the Cochrane analysis of “Antioxidant supplements for prevention of mortality in healthy participants and patients with various diseases”, this was a meta-analysis of randomised trials. The Pocobelli et al paper is a cohort study, and I fail to see how its results falsify last year’s Cochrane meta-analysis.
• The article reports that “Multivitamin use was not related to total mortality”. While the study also found that “[i]n cause-specific analyses, use of multivitamins [was] associated with decreased risks of CVD mortality” – so Holford is technically correct to report a reduction in heart disease risk – if overall mortality is not changed then this may be of little comfort to many readers.

Holford promotes himself as interpreting the scientific evidence on nutrition for his non-expert followers. Here, he is doing a particularly bad job at this.