Phase I safety trial of intravenous ascorbic acid in patients with severe sepsis

Vitamin C — the thing in orange juice and the supplement on every drugstore shelf — measured at seventeen point nine micromolar in the blood of critically ill patients, while the normal range is three to four times that. Sepsis drains the body of ascorbic acid faster than it can be replaced. If that depletion contributes to mortality, and you can fix it with an intravenous drip, that changes everything about how cheap and accessible the treatment could be. Sepsis is one of intensive care medicine's most stubborn problems. Its incidence keeps rising, organ failure cascades are its primary killing mechanism, and despite more than fifteen thousand patients enrolled in prior trials and over a billion dollars spent on research, no targeted therapy has reliably reduced mortality. The pattern, as Fowler and colleagues note, is consistent: single-target approaches — whether statins, activated protein C, or other anti-inflammatory agents — have failed against a process that is fundamentally multicellular and multi-system. Into that frustrating landscape comes an observation that has largely gone unacted upon in the literature. Plasma ascorbic acid is almost universally depleted in severe sepsis. The Fowler team measured a mean entry level of seventeen point nine micromolar in their cohort, against a normal level of fifty to seventy micromolar. Those low levels do not just reflect a sick patient. Prior studies show they inversely correlate with the incidence of multiple organ failure and directly correlate with survival. The depletion happens rapidly, driven by at least three simultaneous processes: ascorbic acid is consumed reducing plasma free iron, consumed scavenging aqueous free radicals, and destroyed when its oxidized form — dehydroascorbic acid — cannot be recycled by cells under stress. Why does that matter? In healthy endothelium — the cells lining blood vessels — ascorbic acid is taken up in millimolar concentrations where it scavenges reactive oxygen species, restores a molecule called tetrahydrobiopterin that endothelial nitric oxide synthase requires to function, and can inhibit NF-kappa-B, the master switch of the inflammatory cytokine storm. In animal models, Armour and Wu and their respective colleagues showed that ascorbic acid infusion improved capillary blood flow, microvascular barrier function, and arteriolar responsiveness. Fowler's own preclinical work showed it attenuated vascular lung injury, enhanced alveolar fluid clearance, and prevented sepsis-induced coagulopathy. There was a plausible mechanism, a clear biological depletion, and an intravenous route that Padayatty and colleagues showed is the only way to reach pharmacologic plasma concentrations. The question was: is it safe to administer in that way to the sickest patients in the intensive care unit? That is exactly what this Phase I trial was designed to answer. Twenty-four medical intensive care unit patients with severe sepsis were randomized one-to-one-to-one: placebo, which was five percent dextrose in water, low-dose ascorbic acid at fifty milligrams per kilogram per day, or high-dose at two hundred milligrams per kilogram per day. Each daily dose was split into four equal infusions, each given over thirty minutes every six hours, for a total of ninety-six hours. Eight patients per arm. Because this was a Phase I study, the primary endpoint was safety, not efficacy. The team prespecified four adverse outcomes to monitor: hypotension defined as a twenty millimeter mercury drop in mean arterial pressure, tachycardia defined as a twenty beat-per-minute increase in heart rate, hypernatremia from the sodium load in the preparation, and nausea or vomiting. Vital signs were recorded every five minutes during each infusion and for forty-five minutes afterward. Secondary endpoints gave the trial its biological texture. Organ failure was tracked using the Sequential Organ Failure Assessment score — SOFA — a clinical tool that scores dysfunction across six organ systems, calculated at enrollment and at twenty-four, forty-eight, seventy-two, and ninety-six hours. The biomarker panel included high-sensitivity C-reactive protein, procalcitonin, and thrombomodulin — that last one measured by ELISA as a surrogate for endothelial injury because thrombomodulin is shed from the vessel wall when it is being damaged. The primary result is clear. No treatment-related adverse events led to withdrawal. No episodes of hypotension, tachycardia, hypernatremia, or nausea and vomiting were attributed to the infusions. One high-dose infusion was halted after the fourteenth dose — at eighty-four hours — due to a ventricular arrhythmia that was subsequently judged to be electrical artifact. That patient remained in the analysis. At the doses tested, intravenous ascorbic acid was safe and well tolerated in critically ill septic patients. What the infusions did to plasma levels is striking. Placebo patients, who were already depleted at entry with a mean of twenty point two micromolar, drifted further down to fifteen point six micromolar by day four. Low-dose patients went from sixteen point seven micromolar at baseline to three hundred thirty-one micromolar on day four — roughly a twenty-fold increase. High-dose patients climbed from seventeen micromolar to three thousand eighty-two micromolar. That's millimolar territory. Both treatment groups separated from placebo within twelve hours, and the high-dose group separated from the low-dose group at twelve hours as well. The intermittent every six hour schedule produced sustained, steady-state elevated concentrations across the entire ninety-six hours. With safety confirmed and plasma levels massively restored, the secondary signals become the focus. On SOFA scores, both ascorbic acid groups showed descending trajectories over the four days — slopes that were significantly different from zero — while placebo patients showed no such reduction. The high-dose group showed the faster improvement: the regression slope of the Delta daily total SOFA score was negative zero point zero four three for high-dose versus zero point zero zero three for placebo, a difference reaching a p-value below zero point zero one. Placebo patients, rather than improving, drifted upward in organ failure burden over the same window. The inflammatory biomarkers told a consistent story. C-reactive protein fell in ascorbic acid-treated patients and was significantly lower than both their own baseline and placebo by twenty-four hours. Procalcitonin in the high-dose arm trended downward from the start, becoming significantly lower than baseline by forty-eight hours, and falling by more than fifty percent over the full ninety-six hours. Placebo procalcitonin, by contrast, trended upward at twenty-four hours. Thrombomodulin is the subtler but mechanistically important signal. Placebo patients began showing an upward trend in plasma thrombomodulin beyond thirty-six hours — that's the vessel wall shedding this protein under ongoing injury. Ascorbic acid-treated patients did not show that rise. The between-group differences did not reach statistical significance — eight patients per arm is simply not enough for that — but the directional pattern was consistent: placebo patients accumulating a vascular injury signal and treated patients not. Fowler and colleagues interpret the convergence of the SOFA trajectory, the inflammatory marker declines, and the blunted thrombomodulin rise as aligned evidence that intravenous ascorbic acid attenuates both the systemic inflammatory response and the endothelial injury that characterizes severe sepsis. The authors are careful about what a Phase I trial of twenty-four patients can claim. It cannot prove efficacy. It was not powered to detect a mortality difference. The SOFA and biomarker findings are hypothesis-generating, not confirmatory. But that is the right framing for what the study accomplished: it established that the intervention is safe at both doses, demonstrated that plasma repletion is achievable and pharmacologically dramatic, and produced concordant biological signals across organ dysfunction scores and three independent biomarkers. In a Phase I study that wasn't designed to show those signals, their consistency across endpoints is what makes them worth taking seriously. The mechanistic story that Fowler and colleagues propose runs through the endothelium. Parenteral ascorbic acid reaches concentrations in the bloodstream high enough to be taken up by vascular endothelial cells, where it scavenges reactive oxygen species, restores nitric oxide bioavailability, and inhibits NF-kappa-B-driven inflammatory gene expression. The thrombomodulin signal suggests that the vessel wall itself is being protected. The SOFA improvement suggests that this protection translates, at least directionally, into less organ failure over four days. What makes this worth watching is the economics. Sepsis is among the most expensive conditions in intensive care, consuming enormous resources with mortality rates that have stubbornly refused to fall. The intervention tested here costs a small fraction of the drugs that have already failed in larger trials. A safe, biologically plausible treatment with early signals pointing in the right direction — that is precisely the kind of result that merits a Phase II trial. Fowler and colleagues stated so explicitly. The work that came after this study is a separate story, but this is where it started: a depleted vitamin, a twenty-fold repletion, and no harm done. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.