Misleading and erroneous papers on vitamin C: A list of critiques by Harri Hemilä

In the scientific literature, there is a lot of rubbish about vitamin C. Several randomized controlled trials have been analyzed incorrectly or reported in a misleading manner (#5, #6, #16, #22, #38, #51, #57, #61, #62, #63, #64, #73, #74, #75). For example, the subgroup analysis in the Karlowski (1975) trial was based on participants’ guesses regarding treatment allocation; however, the trial report did not disclose that 42% of participants were excluded from this subgroup analysis (#5, #6, #16); and three trials involving critically ill patients administered vitamin C for only 4 days, yet the investigators failed to consider the significant changes in mortality occurring after the abrupt discontinuation of supplementation (#51, #71, #73, #74). 
    Numerous systematic reviews and meta-analyses have been affected by errors in data extraction, inappropriate trial inclusion, and flawed analytical methods (#4, #7, #11, #12, #13, #14, #25, #33, #34, #39, #40, #41, #42, #43, #44, #46, #47, #48, #50, #52, #54, #58, #59, #60, #70, #72, #74, #76); notably, three of these erroneous meta-analyses were published as Cochrane reviews (#25, #33, #47, #54). For instance, in one of the most influential meta-analyses, Thomas Chalmers (1975) concluded that vitamin C reduced the duration of the common cold by only 0.11 days. Subsequent correction of the errors increased the estimated effect to 0.93 days (#4, #14). 
    A central objective of contemporary biomedical research is specificity of causal inference. The principal strength of randomized controlled trials lies in the ability to attribute observed differences between study groups specifically to the intervention under investigation. Nevertheless, several meta-analyses combined trials in which vitamin C was administered together with other substances (#44, #46, #47, #70, #72, #74), thereby confounding interpretation because observed effects—or the absence of them—could not be attributed exclusively to vitamin C.
    One particularly unusual publication in the vitamin C literature copied empirical data, including standard deviations with 4-digit accuracy, directly from a paper published 20 years earlier (#62). 
    The seriousness of these concerns is reflected by the fact that four vitamin C papers were subsequently retracted (#32, #43, #60) and one Cochrane review was formally withdrawn (#34).
    These critiques were written by Harri Hemilä to document problems in the specific publications and to caution researchers against accepting such studies uncritically when evaluations of vitamin C research.