Vitamin C Supplementation and Respiratory Infections: a Systematic Review

Military Medicine, Nov 2004 by Hemilä, Harri

The one-tailed p values are used In the tables and text because the explicit question in the present analysis is whether vitamin C supplementation decreases the incidence and severity of respiratory infections. There is no experimental or theoretical reason to expect that vitamin C supplementation would increase morbidity from respiratory infections. Confidence intervals are not calculated because the primary purpose of this review was to test the hypothesis that vitamin C affects respiratory infections. Trial settings and outcome definitions vary substantially, which limits the possibility to generalize the point estimates; therefore, no pooled estimates were calculated.

Trials with Military Personnel

Seven trials examined the effect of vitamin C supplementation on respiratory infections in military personnel. Table I shows the number of participants, the duration of the trial, the dosage of vitamin C, and the outcome values. There is substantial variation in the common cold incidence in the control groups. Pitt and Costrini113 observed 11 colds per person-year, whereas Dahlberg et al.17 observed only 0.4 colds per person-year, although both of these large trials examined military recruits.

Pitt and Costrini16 carried out a large-scale, randomized, double-blind, placebo-controlled trial with Marine recruits in a training camp in South Carolina in the wintertime. The participants were vaccinated against adenovirus and influenza, and they received penicillin or erythromycin as streptococcal prophylaxis. The study groups showed no difference in common cold incidence. Vitamin C caused a statistically significant, although clinically minor, reduction in common cold severity. Of the eight pneumonia cases observed, five were caused by pneumococcus but two of them had a coinfection with parainfluenza virus. The incidence of pneumonia was significantly lower in the vitamin C group.

Interestingly, Pitt and Costrini16 also found that a considerable proportion of their participants correctly inferred their treatment from subjective observations (for a comparable case, see Ref. 10). When asked which pill they thought they were taking, 53% (358) of all participants (674) stated they did not know. Of those who replied vitamin C or placebo (316), 27% (89) and 26% (89) answered correctly, whereas 20% (66) and 21% (72) answered incorrectly in the vitamin C and placebo groups, respectively. Consequently, in all, 178 participants answered correctly and 138 incorrectly. Assuming pure guessing, the expected number of correct and incorrect answers is 158, which allows the calculation of observed and expected difference to yield χ^sup 2^(1 df] = 5.0 and p[one-tail] = 0.013. Consequently, 40 participants (178 - 138) correctly inferred their treatment in this double-blind trial. Formulated from citric acid, the placebo tablets were indistinguishable in appearance and taste from the ascorbic acid tablets. Therefore, it is worth noting that 6% (40 of 674) of all participants inferred their treatment by subjective perception alone.