Sepsis resuscitation generally focuses on hemodynamics. Rivers of ink have been spilled writing about oxygen delivery and fluid responsiveness. This is clearly important, but it's possible that our focus on easily observable phenomena has led us to ignore something of equal importance: metabolic resuscitation. We can deliver all the oxygen we want to the tissues, but if the mitochondria are failing it won't work.

Very brief overview of the scientific rationale

Vitamins become depleted during sepsis

Sepsis causes increased consumption of thiamine and Vitamin C:

Vitamin C levels invariably fall during sepsis, sometimes dropping below the level of detection. Vitamin C deficiency correlates with multiorgan failure and death (Wilson 2009).

Thiamine deficiency is common in sepsis, occurring in perhaps one-third of patients. This is associated with increased mortality (Manganese 2011).

Imagine what would happen if we routinely measured vitamin levels on all of our patients. We would be alarmed by our patients' low levels. However, in practice we don't check these, so profound derangements pass unnoticed.

Vitamin deficiency causes clinical disease

Vitamin C deficiency causes scurvy. Vitamin C is important for the maintenance of endothelial boundaries, with edema noted in scurvy. Vitamin C is also required for the synthesis of catecholamines and cortisol, so deficiency causes failure of the sympathetic nervous system.

Thiamine deficiency may cause delirium (Wernicke's encephalopathy) and cardiac dysfunction (Beriberi). Beriberi in particular can mimic sepsis, by causing distributive shock and lactic acidosis.

It is unclear precisely how important vitamin deficiencies may be to the pathogenesis of sepsis. However, it's not difficult to imagine that the simultaneous development of several vitamin deficiencies could synergize to promote organ failure. For example, a patient with high-output heart failure (from thiamine deficiency) and sympathetic nervous system failure (from vitamin C deficiency) would develop profound shock.

Clinical studies of IV vitamin C in critical illness

Intestinal absorption of vitamin C is saturable. Given the increased metabolic consumption of Vitamin C in critical illness, the only way to replete Vitamin C in this context is intravenously.

Tanaka H et al. 2000: Reduction of resuscitation fluid volumes in severely burned patients using ascorbic acid administration

This is a prospective RCT involving 37 patients with major burns (>30% body surface area). Patients were randomized regarding whether or not to receive an infusion of IV vitamin C, 66 mg/kg/hr for the first 24 hours of hospitalization.

Patients were resuscitated to an identical MAP and CVP. However, patients in the vitamin C group required less fluid resuscitation, had higher urine output, and developed less wound edema (figure below). As might be expected, this translated into improved oxygenation and less time on mechanical ventilation among the Vitamin C group (average of 12 vs. 21 days of ventilation, p=0.03).

Nathens AB et al. 2002: Randomized, prospective trial of antioxidant supplementation in critically ill surgical patients

These investigators postulated that prophylactic antioxidants provided to surgical ICU patients could reduce pulmonary complications. They randomized patients shortly after admission to the ICU to no therapy vs. a combination of enteral vitamin E plus IV vitamin C 1000 mg q8hr until ICU discharge.

The primary endpoint was a composite of pneumonia or ARDS. Although there were trends towards fewer pulmonary complications among patients treated with antioxidants, these didn't reach statistical significance (the study was underpowered due to low rates of respiratory complications). However, patients treated with vitamins E and C fared better on a variety of secondary endpoints including less time on the ventilator and less multiorgan failure.

Fowler A et al. 2014: Phase I safety trial of intravenous ascorbic acid in patients with severe sepsis

This was a prospective RCT involving patients with severe sepsis in a medical ICU. 24 patients were randomized to receive placebo, low-dose vitamin C (12.5 mg/kg IV q6hr), or high-dose vitamin C (50 mg/kg IV q6hr). The primary endpoint was safety and tolerability, with no adverse events noted.

Secondary endpoints included SOFA score and inflammatory markers. Given the low number of patients in each arm, comparisons were made to the patients' baseline scores. Patients treated with vitamin C experienced a dose-dependent improvement in SOFA score over time (below, right). Vitamin C also improved inflammatory markers (below, left).

Zabet MH et al 2016: Effect of high-dose ascorbic acid on vasopressor requirement in septic shock

Since vitamin C is required for catecholamine synthesis, many authors have hypothesized that it could expedite weaning off vasopressors (Carr 2015). These investigators tested this hypothesis among 24 patients from a surgical ICU with vasopressor-dependent septic shock (tables below, left). Patients were randomized to receive placebo vs. Vitamin C 25 mg/kg IV q6hr.

The primary endpoint was vasopressor dose and duration, which were significantly reduced in patients receiving vitamin C (tables below, right). There was also a reduction in mortality among patients treated with Vitamin C.

Clinical studies of IV thiamine in sepsis

Donning MW 2016: Randomized, double-blind, placebo-controlled trial of thiamine as a metabolic resuscitator in septic shock: A pilot study

This was a RCT investigating the effect of thiamine (200 mg IV q12hr) in 88 patients with septic shock. Thiamine had no effect on lactate levels or mortality within the entire population (figure below, left). This may be explained by the fact that only 35% of the patients had thiamine deficiency. Within this pre-specified subgroup, thiamine administration did reduce lactate levels and mortality (figure below, right).

Marik P 2017: Hydrocortisone, vitamin C, and thiamine for the treatment of severe sepsis and septic shock: A retrospective before-after study.

This is a before-after study investigating the impact of treating sepsis with a combination of thiamine 200 mg IV q12hr, Vitamin C 1.5 grams Q6hr, and hydrocortisone 50 mg IV q6hr. The rationale for combining steroid and Vitamin C is that they work synergistically, for example vitamin C may restore sensitivity to glucocorticoids (Marik 2016). In tissue cultures of endothelial cells, the combination of steroid and vitamin C (but neither drug in isolation) can preserve endothelial integrity against challenge from lipopolysaccharide.

This therapy was applied to patients in a medical ICU with sepsis and a procalcitonin >2 ng/ml. 47 patients were included in each group, with good matching between groups:

The primary outcome was mortality, which was substantially reduced in patients receiving vitamin C (p<0.001, figure below). Similar results were obtained when the data was analyzed via two alternative methods, using either a propensity-adjusted outcome or logistic multivariate analysis. The robustness of these analyses suggest that mortality differences reflect a true treatment effect, rather than statistical confounding.

Patients treated with Vitamin C were weaned off vasopressors much earlier than control patients, despite receiving on average a bit less fluid (figure below). This result is similar to findings discussed above from the RCT by Zabet 2016.

Patients treated with Vitamin C experienced more rapid decline in SOFA score and procalcitonin (figures below). These results are similar to results shown above from the RCT by Fowler 2014.

This study is limited by its single-center, before-after, non-blinded design. However, the patients were quite well-matched at baseline and the outcomes are objective. Two statistical analyses failed to detect confounding. The results show dramatic separation between the two groups, considerably larger differences than might be realistically explained by a Hawthorne effect.

Another limitation is that the study evaluated three simultaneous interventions. As such, it is impossible to determine which agent(s) are responsible for the clinical improvement seen. To attempt to reproduce these results clinically, this combination of three agents must be used.

This study was recently presented at the Critical Care Reviews conference:

CCR17 Cure for Sepsis from Critical Care Reviews on Vimeo.

How should we use this information?

Safety of Vitamin C and Thiamine (for a septic patient already on steroid)

The safety of IV thiamine 400 mg/day is universally accepted and not worth further discussion.

Interest in vitamin C has persisted for over half a century, so its safety is well established even at enormous doses. For example, none of the above studies reported any adverse event, despite the use of massive doses by Tanaka et al. (1.6 grams/kg over 24 hours).

One potential concern regarding vitamin C is that it may be metabolized into oxalic acid, leading to calcium oxalate nephropathy. This doesn't seem to be a significant problem using Marik's treatment regimen, for several reasons:

Oxalate formation is a dose-dependent toxicity. This has rarely been reported from short courses of IV vitamin C, but only at much higher doses (>40 grams/day; Buehner 2016).

Concurrent use of thiamine should reduce the conversion of vitamin C into oxalate.

In Marik's study, Vitamin C use correlated with improved renal outcomes (the rate of dialysis was reduced from 33% to 10%, p=0.02).

Another concern which has been raised is that vitamin C at extremely high doses may have a pro-oxidant effect. This was shown not to occur even at a dose of 7.5 grams IV daily (Muhlhofer 2004).

Ready for prime time?

This is a controversial question, which will undoubtedly leave practitioners with divided opinions. When exactly do we reach a tipping point, beyond which we feel that there is enough evidence to implement a therapy? This is almost more of a philosophical question than a scientific one, making it impossible to reach consensus. Hardcore evidence-based medicine disciples may be aghast at using a therapy without a large multi-center RCT, whereas more integrative, theoretically-minded clinicians may be willing to consider it.

It would certainly be nice to have a large, multicenter, placebo-controlled RCT. However, not every therapy requires one. The use of vitamin C to treat scurvy is based on a single-center open-label trial involving twelve sailors. There isn't any RCT proving that potassium should be repleted among patients with septic shock. Replacing an essential endogenous substance which is deficient doesn't necessarily mandate level-I evidence.

Although a single unassailable RCT is lacking, an extensive body of evidence does exist regarding vitamin C and thiamine. Over a half-century of experimentation attests to the safety of these vitamins. Several RCTs suggest that they are beneficial in critical illness, as reviewed above. Marik's results have similarities to prior RCTs, implying replicability. This clinical evidence is supported by robust basic science.

Clinical practice isn't scientifically perfect. Nearly all of our current therapies for sepsis lack level-I evidence. On the front lines of critical care, we are forced to treat dying patients based on the evidence that we have, not the evidence that we might wish for. In the context of this imperfect reality, treating septic patients with Vitamin C and thiamine may be a rational and evidence-based practice. Indeed, some authors recommended using IV vitamin C even before Marik's study was released (Honore 2016).

Septic patients are invariably deficient in Vitamin C, and frequently deficient in thiamine.

Deficiencies in Vitamin C and thiamine might explain many of the abnormalities seen in sepsis.

Vitamin C and thiamine have an outstanding track record of safety, proven over decades of experimentation and clinical experience.

Five RCTs have suggested benefit from Vitamin C or thiamine in critically ill patients, with no evidence of toxicity.

A recent before-after study found a substantial mortality benefit from the combination of stress-dose steroid, IV vitamin C, and IV thiamine. Although this isn't an RCT, the results are quite striking.

Further research is required, but in the interim this is a reasonable intervention given the excellent safety profile of these agents.

Expert Commentary

Below are some additional thoughts and clinical updates from Dr. Marik. As he reports ongoing success with growing numbers of patients, it is increasingly difficult to dismiss this as a statistical or methodological fluke. Dr. Marik also shared a video with me, which is shown below. I like this video because it places this therapy into a real-life context: this isn't a futuristic theoretical treatment, it's something that's being done right now as standard-of-care at several Virginia hospitals.

We have now treated over 150 patients with severe sepsis and septic shock. We have had only one patient die from sepsis, this being a complex surgical case who died in the immediate post-operative period. While a few of the treated patients have died, none died from progressive organ failure related to sepsis. All these patients were weaned off pressors/mechanical ventilation and died from their underlying disease. There can be no question of doubt that we have changed the natural history and disease progression of patients with sepsis… patients with sepsis simply just don’t develop progressive organ failure. The cocktail is without side effects; renal function has improved in all treated patients. On average patients receive 3-4 days of treatment with the cocktail which is then stopped on discharge from the ICU or day 4. Typically patients with septic shock who are admitted to our ICU are discharged on the third ICU day with minimal organ failure (low SOFA Score). The protocol is initiated by our residents in the Emergency Room at the same time that they start antibiotics. Our CEO and CMO has confirmed our results (shorter LOS and fewer deaths) and has requested that the protocol be instituted throughout our hospital system. It is noteworthy that all animals (except human and guinea pigs) produce Vitamin C during stress and this has been shown to be protective during infection. GULO knockout mice (Vitamin C knockout mice) are much more likely to die from sepsis than wild type mice… treatment of infected GULO mice with Vitamin C is largely protective against death.

Related

Image credit: Mitochondria