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

**Background:** The LOVIT trial (n=862) previously reported that intravenous vitamin C increased the risk of death or persistent organ dysfunction at 28 days in adults with sepsis receiving vasopressor therapy. However, vitamin C was administered for only 4 days despite a median ICU stay of 6 days. The authors hypothesized that the observed harm might be due to a rebound effect from abrupt termination of high-dose vitamin C, a phenomenon previously documented in the literature. This secondary analysis aimed to determine whether the timing of deaths relative to vitamin C cessation could explain the trial's findings.

**Methods:** The authors extracted early survival data from the published LOVIT trial survival curves using a graphics program, focusing on the first 11 days (the 4-day administration period plus 1 week thereafter). Cox regression was used to estimate mortality rate ratios (RR) comparing vitamin C to placebo. The dataset was split using the survSplit procedure to allow separate RR estimates for early and late periods. A likelihood ratio test compared the fit of a single uniform RR model versus a two-period model.

**Key Results:** Over the 11-day follow-up, there were 106 deaths in the vitamin C group and 83 in the placebo group (RR=1.32, 95% CI 0.99-1.76). A two-period Cox regression model with a split at day 4 provided a significantly better fit than a uniform effect model (p=0.026). During the 4-day vitamin C administration period (days 1-4), there were 49 deaths in the vitamin C group and 51 in the placebo group (RR=0.97, 95% CI 0.65-1.44), showing no evidence of harm during active treatment. In the week after termination (days 5-11), there were 57 deaths in the vitamin C group versus 32 in the placebo group (RR=1.9, 95% CI 1.2-2.9; p=0.004). The harm was most pronounced immediately after cessation: during days 5-7, RR=2.28 (95% CI 1.24-4.2; p=0.008), while during days 8-11, RR=1.51 (95% CI 0.81-2.8), with overlapping confidence intervals. The authors note that after day 30, there was no difference in mortality between groups (vitamin C: 41/267; placebo: 48/289; RR=0.93, p=0.7).

**Clinical Implications:** The authors argue that the LOVIT trial's conclusion—that receiving intravenous vitamin C caused harm—is misleading. Instead, the harm occurred only after abrupt termination of vitamin C, consistent with a rebound effect where high-dose vitamin C increases metabolism, and sudden cessation leads to systemic levels even lower than baseline. This is particularly concerning given that 25% of LOVIT patients had baseline vitamin C levels below 5.37 µM (less than half the usual deficiency threshold of 11 µM), and nearly half had levels below 12.38 µM. The authors raise ethical concerns about randomizing patients with very low vitamin C levels to placebo without treating potential scurvy, noting that scurvy symptoms (hypotension, dyspnea, edema, fatigue) overlap with sepsis symptoms and can be life-threatening. They recommend that vitamin C should be administered throughout the entire period of critical illness, not just for 4 days, and that future research should not be discouraged by the LOVIT findings but should address ethical considerations around placebo groups for vitamin C-deficient patients.