Padayatty · Annals of internal medicine 2004 · pharmacokinetic dose-concentration study and modeling · n=17

Vitamin C pharmacokinetics: implications for oral and intravenous use.

Cited 980 times in the scientific literature.

Level 3 - non-randomized controlled study

Non-randomized pharmacokinetic dose-concentration study and mathematical modeling in healthy volunteers.

PubMed 15068981 · doi:10.7326/0003-4819-140-7-200404060-00010 · record verified 2026-08-29

What was done

Seventeen healthy hospitalized volunteers received oral and intravenous vitamin C across a dose range of 0.015 to 1.25 g while plasma and urine concentrations were measured. Pharmacokinetic modeling was then used to calculate expected plasma and urine concentrations for doses ranging from 1 to 100 g.

What was found

Peak plasma vitamin C concentrations were significantly higher after intravenous versus oral doses (P < 0.001), with widening differences as dose increased. Oral administration of 1.25 g yielded a mean peak plasma concentration of 134.8 +/- 20.6 micromol/L compared with 885 +/- 201.2 micromol/L intravenously. Pharmacokinetic modeling predicted peak plasma concentrations of 220 micromol/L for the maximum tolerated oral dose (3 g every 4 hours) versus 13,400 micromol/L for a 50-g intravenous dose. Predicted peak urine concentrations from intravenous administration were 140-fold higher than those from maximum oral doses.

Why it matters

Prior clinical trials showing no benefit for vitamin C in cancer relied on oral administration, which cannot achieve cytotoxic plasma levels due to tight physiological saturation controls. This pharmacokinetic divergence demonstrates that oral trials cannot rule out the potential efficacy of high-dose intravenous vitamin C.

Limits

The study included only 17 healthy subjects rather than patients with cancer. Doses above 1.25 g (including the simulated 50-g to 100-g regimens) were predicted via pharmacokinetic modeling without empirical in vivo confirmation, and clinical antitumor outcomes were not directly tested.

Cited by