Abrupt termination of vitamin C from ICU patients may increase mortality: secondary analysis of the LOVIT trial

The study addresses whether the higher mortality observed in the vitamin C arm of the LOVIT sepsis trial could be due to abrupt termination of vitamin C rather than to harmful effects during administration. Prior evidence suggests vitamin C may benefit critically ill patients: meta-analyses report shorter ICU stay and reduced ventilation time, with greater effects in more severe illness. Experimental and clinical data indicate vitamin C may mitigate sepsis pathophysiology, including hypotension, lung injury, microcirculatory impairment, and cardiac stress; vitamin C deficiency increases pneumonia risk, a common cause of sepsis. Two randomized trials reported reduced sepsis mortality with vitamin C, and a re-analysis of the CRITIS-AL trial showed time-dependent effects limited to the administration period. In LOVIT, vitamin C was given for 4 days despite a median ICU stay of 6 days, raising concerns about a rebound effect if high-dose vitamin C is abruptly stopped in patients with low baseline levels. The authors hypothesized that harm in the vitamin C group might be attributable to the abrupt termination of vitamin C during ongoing critical illness, and analyzed the early time distribution of deaths to test this.

This is a secondary analysis of the randomized, double-blind LOVIT trial. The authors extracted the early survival data by digitizing steps from published survival curves (figure 2 in the LOVIT report) using a graphics program, focusing on deaths over the first 11 days (the 4-day administration period plus 7 days after cessation). They applied Cox regression models to estimate mortality rate ratios (RRs) comparing vitamin C versus placebo. Using R (cph procedure of the survival package) they estimated a single RR over days 0–11 and also fit two-period models with a split at varying days, employing survsplit to create time-dependent strata. Model fit between the single-RR and two-period models was compared via likelihood ratio tests. They reported RRs with 95% confidence intervals and p-values. Analyses and data extraction steps are detailed in a Supplementary file.

- Over the first 11 days, total deaths were 106 in the vitamin C group and 83 in the placebo group; a single-effect Cox model gave RR = 1.32 over days 0–11. - Allowing a separate effect starting on day 5 (i.e., after the 4-day administration) significantly improved model fit compared with a uniform effect model (likelihood ratio test p = 0.026). - During the 4-day vitamin C administration (days 1–4): 49 deaths (vitamin C) vs 51 (placebo); RR = 0.97 (95% CI 0.65–1.44) — no evidence of harm during administration. - During the week after termination (days 5–11): 57 deaths (vitamin C) vs 32 (placebo); RR = 1.9 (95% CI 1.2–2.9; p = 0.004) — strong evidence of excess mortality following abrupt cessation. - Early post-cessation (days 5–7): 33 vs 14 deaths; RR = 2.28 (95% CI 1.24–4.2; p = 0.008), indicating the greatest divergence immediately after stopping vitamin C.

Findings indicate that the excess mortality observed in LOVIT among patients randomized to vitamin C is temporally concentrated after the 4-day course ended, not during active administration. This supports the hypothesis that abrupt termination of high-dose vitamin C during ongoing critical illness may precipitate harm, potentially via a rebound effect where upregulated vitamin C metabolism and persistent increased utilization in sepsis lead to even lower systemic levels upon cessation. The pattern parallels prior time-dependent effects seen in other sepsis vitamin C trials, where benefits or differences were confined to the treatment window. The authors argue that, given elevated vitamin C consumption and low baseline levels in critically ill patients, administration should cover the entire period of critical illness rather than stop early. They also raise ethical concerns about not treating severe vitamin C deficiency (scurvy or near-scurvy levels) in randomized trials, noting that many LOVIT participants had very low baseline levels and that deficiency symptoms overlap with sepsis manifestations. Overall, the results challenge interpretations that vitamin C per se increases mortality and instead implicate the timing and abrupt discontinuation of therapy. The clinical relevance is that trial designs and practice should consider duration and weaning strategies for vitamin C in critically ill patients, and ensure deficiency is addressed.

The increased mortality in the vitamin C arm of the LOVIT trial is attributable to the abrupt termination of vitamin C, not to harm during its administration. The LOVIT findings should not be used as evidence against vitamin C therapy for critically ill patients. Future research should continue to evaluate vitamin C in critical illness, avoid abrupt cessation that may cause rebound effects, consider longer treatment durations aligned with the course of critical illness, and address ethical issues of randomizing patients with very low vitamin C levels to placebo without treating deficiency.

- Secondary analysis based on data extracted from published survival curves rather than individual patient-level data, which may introduce measurement or extraction error. - Focused on the first 11 days, potentially overlooking later effects beyond this window. - The analysis infers a rebound mechanism without direct longitudinal vitamin C level measurements post-randomization in LOVIT. - Potential inconsistencies or typographical errors in reported statistics within the source text highlight reporting limitations. - As a re-analysis, subject to the constraints and design of the original trial (e.g., treatment duration, timing of initiation).