Background
A definitive cure for sepsis has eluded physicians and scientists since the advent of modern medicine. Over the years, numerous agents have undergone testing in clinical trials, but most have shown limited to no efficacy for the treatment of this deadly condition. The recombinant activated protein C, dtrotrecogin alfa (Xigris), was FDA approved for the treatment of sepsis in 2001, only to be pulled from the market in 2011 when a large post-marketing study failed to show significant efficacy. Other methods like hemofiltration have shown some promise, but remain mostly experimental and are not in wide use.
So far, the only effective and widely accepted treatment for sepsis remains early antibiotic administration and cardiovascular support with fluids and vasopressors. However, for this approach to be effective it must be applied early in the course of the disease, before severe septic shock sets in, which can quickly lead to organ damage and death. If the patient reaches septic shock, mortality remains very high.
In January of 2016, Dr. Paul Marik, a critical care doctor at Eastern Virginia Medical School (EVMS) was treating a septic patient in the ICU of Sentara Norfolk General Hospital in Norfolk, Virginia. The patient, a 48-year old woman with overwhelming sepsis, had failing renal and pulmonary function, and Marik thought she had little chance for survival. Marik had recently read a number journal articles on the beneficial effects of vitamin C in sepsis, so in a last-ditch effort to alter the course of her severely deteriorating condition, he decided to try a cocktail of IV vitamin C and steroids.
When he returned to the hospital the next day, instead of finding his patient on her death bed as expected, Marik found her showing significant signs of improvement. Soon she no longer required pressors, her renal and pulmonary function quickly improved, and three days later she was discharged from the ICU. Thinking this was a fluke, in the days that followed Marik administered the vitamin C cocktail to two other septic patients, with similar results.
Soon Marik started giving his cocktail to all his septic patients with the same positive results. A third ingredient, thiamine (vitamin B1) was later added to the mix for reasons described further below. To date, Marik claims to have treated 150 patients with his cocktail, dubbed the Norfolk or Marik cocktail, of whom only one died from sepsis, although others have died from other comorbidities. The treatment has presumably been so effective that the president of Norfolk General Hospital recently announced that its use is being rolled out to other ICUs in the health system to validate these positive findings. The ingredients of this cocktail are listed below:
| The Norfolk/Marik Sepsis Cocktail |
| Vitamin C 1.5 g IV every 6 hours x 4 days
Hydrocortisone 50 mg IV every 6 hours x 7 days then taper Thiamine 200 mg IV every 12 hours x 4 days or until ICU discharge |
Wishing to share his impressive results with the world, Marik, who’s no stranger to academic medicine having authored over 400 peer reviewed journals, 50 book chapters and 4 textbooks in critical care, according to his EVMS bio, decided to put together a study with his findings. His study was published in the December 2016 online issue of CHEST.
Vitamin C, thiamine and steroids in sepsis – Marik et al, CHEST 2016
This study compared outcomes in 47 consecutive patients with sepsis treated in the usual manner over a 7 month period before use of the vitamin C cocktail was instituted, with outcomes in 47 consecutive patients treated with the vitamin C cocktail over a 7 month period after institution of this treatment. Its primary outcome was all-cause in-hospital mortality. The patients were approximately 60 years of age and were well-matched in terms of presentation, organ function, biomarker levels, infection type and severity as assessed by SOFA and APACHE II and IV scores.
| Patient Characteristics | ||
| Treatment (n=47) | Control (n=47) | |
| Age (yrs) | 58.3 | 62.2 |
| Male | 57% | 49% |
| Mechanical ventilation | 47% | 55% |
| Vasopressors | 46% | 46% |
| WBC | 20.6 | 17.2 |
| Lactate (mmol/L) | 2.7 | 3.1 |
| Creatinine (mg/dL) | 1.9 | 1.9 |
| AKI (KDIGO ≥1) | 66% | 64% |
| PCT (ng/mL) | 25.8 | 15.2 |
| Blood cultures + | 28% | 28% |
| Vitamin C (μmol/L) | 14.1 | – |
| Pneumonia | 38% | 40% |
| Urosepsis | 23% | 21% |
| Primary bacteria | 15% | 15% |
| GI/biliary sepsis | 13% | 13% |
| Other source | 11% | 11% |
| SOFA day 1 | 8.3 | 8.7 |
| APACHE II | 22.1 | 22.6 |
| APACHE IV | 79.5 | 82.0 |
| Predicted mortality | 39.7 | 41.6 |
All-cause in-hospital mortality (primary outcome) was 40% in the control group and 9% (4 patients) in the treatment group (p<0.001). Notably, 3 of the 4 deaths in the treatment group did not occur in the ICU due to sepsis, but rather from accompanying comorbidities including severe Alzheimer’s, severe sarcoidosis, and severe cardiomyopathy, after discharge from the ICU. These 3 patients had their treatment withdrawn and did not die of sepsis or sepsis-induced organ failure, according to a presentation by Marik at a recent Critical Care Reviews conference.
Similar results were obtained when the data was analyzed using a logistic multivariate and a propensity-adjusted outcome analysis. On logistic multivariate analysis, the treatment intervention was the variable most highly predictive of mortality (F-value 17.33) followed by the APACHE IV score (F-value 13.29). On propensity adjusted outcome analysis, the odds ratio for death was 0.13 (95% CI 0.04-0.48, p=0.002) in favor of the treatment group. Mortality and other clinical outcomes are summarized in the table below.
| Outcomes | ||
| Treatment (n=47) | Control (n=47) | |
| Predicted mortality | 39.4% | 41.4% |
| Actual hospital mortality | 9% | 40% |
| Duration of vasopressors (hrs) | 18.3 | 54.9 |
| Change in SOFA at 72 hours | 4.8 | 0.9 |
| Procalcitonin clearance (72 hrs) | 86.4 | 33.9 |
| Need for RRT for AKI | 10% | 33% |
| Mean ICU LOS (days) | 4 | 4 |
On the surface, these results are extraordinary, but the question that everybody has been asking since their publication is are they real or the product of a statistical anomaly perhaps due to inferior study design?
Many observers have pointed out that Marik’s study is statistically weak because it is a small, single-center study that employed a retrospective observational and non-blinded before-after design. This is hardly the type of study that would definitively answer the question of efficacy of any treatment, they point out, and such a study is typically hypothesis generating at best. These observers note that only a larger blinded, randomized placebo controlled clinical trial can truly determine the efficacy of this vitamin C-based treatment. Given this, are its results at all meaningful and should they impact clinical care?
Generally speaking, it is true that RCTs are the statistical gold standard, and the retrospective, observational and non-blinded design of Marik’s study is certainly prone to bias and confounding variables, so it cannot meet the high standard of an RCT. However, a few things are worth noting that do strengthen its case. One, patients in both groups were consecutive, reducing bias from potential “cherrypicking” of patients. Secondly, the patients’ baseline characteristics appear well matched in most respects, and investigators ran additional analyses including propensity outcome adjustment to compensate for group differences, and multivariate analysis to compensate for confounders. Thirdly, operator bias cannot be avoided since the study was unblinded, however, the treatment effect was so large (absolute risk reduction of 31% and a relative risk reduction of 87% [OR=0.13]), that it is unlikely to be caused solely by this bias. No matter how perfectly the clinicians performed their jobs in the treatment group, a 9% mortality from sepsis over such a period of time (and only 2% if the non-sepsis related deaths are discounted), is a remarkable achievement. This of course assumes that mortality and other clinical events were correctly adjudicated. Marik is an accomplished academic and the author of a textbook titled Evidence Based Critical Care, so it would be very hard to believe he is unfamiliar with the methods of rigorous research and statistical analysis.
The biochemical basis for vitamin C in sepsis
But why would vitamin C have such a dramatic effect on the course of sepsis? It is well-known that vitamin C is a powerful anti-oxidant with beneficial effects on the immune system, but in recent years some of the more outlandish claims of vitamin C megadose use to cure various infections have been debunked. Why would this be any different?
In his recent Critical Care Reviews presentation, Marik provided some interesting and compelling arguments as to why this might be the case. We encourage the reader to watch this presentation and draw his or her own conclusions, but we will summarize the main points here.
First, according to Marik, vitamin C is both an anti-oxidant and a co-enzyme in many important reactions, and studies have shown that it is quickly and markedly depleted in high metabolic stress states such as sepsis. Most animals can make their own vitamin C to address this problem, but primates and Guinea pigs are the only animals that lack the necessary enzyme (L-gulono-lactate oxidase) for this, and thus must obtain it from outside sources. However, in humans, absorption of vitamin C in the gut is limited by the capacity of the sodium-dependent vitamin C transporter SVCT1 to about 500 mg per day. This is something Linus Pauling apparently did not know when he initially recommended large daily doses of oral vitamin C which were shown to have no effect above approx. 500 mg because they are largely excreted. For this reason, large doses of oral vitamin C are not effective in humans for fast replenishment in a highly depleted state, this requires IV administration, says Marik. The normal serum level of vitamin C is approx. 30 to 50 μmol/L and this level gets concentrated x100 fold in the cells by the SVCT2 vitamin C transporter, according to Marik.
Vitamin C plays numerous roles, among these being the production of catecholamines by the adrenal gland. According to Marik, the adrenal gland requires vitamin C to produce the catecholamines and cortisol that play crucial roles in the stress response, and if vitamin C levels are depleted, production of these stress hormones can seize. Vitamin C is also critical in endothelial function and nitric oxide (NO) production, adrenergic receptor activation, neurotransmitter production, tight junction maintenance, collagen production, and immune cell function. These wide functions explain the severe vascular, nervous and immune derangements seen with vitamin C depletion in scurvy (bleeding, ecchymosis, edema, neuropathy, behavioral changes, infections, etc), symptoms that overlap with those seen in sepsis. Sepsis seems therefore to be as much a metabolic derangement as it is a cardiovascular derangement, and administration of vitamin C may address this metabolic component.
If vitamin C alone is so important, why then the need to add steroids and thiamine? According to Marik, septic patients are also thiamine deficient, and even administering thiamine alone can reduce mortality in this population (Donnino et al, 2016). So adding thiamine adds an extra kick to the mix. The other important reason is that megadoses of vitamin C can lead to oxalate crystal deposition in the kidneys, something that can be exacerbated by thiamine deficiency. By administering thiamine, the risk of oxalate production in the setting of large vitamin C doses is minimized.
The reason for the use of hydrocortisone in Marik’s cocktail is the interesting synergistic relationship between vitamin C and glucocorticoids. In sepsis, the thiol groups of the glucocorticoid receptors (GR) are oxidized leading to an inability of the glucocorticoid to bind to the receptor and the DNA elements, according to Marik. Administration of vitamin C restores GR function, and thus glucocorticoid action. Conversely, dexamethasone works to upregulate expression of the SVCT2 transporter, leading to an enhancement of Vitamin C concentration in the cells. So vitamin C and glucocorticoids work together to potentiate and amplify each other’s effects – alone they don’t work well, but together they are a potent mix, says Marik.
So should critical care clinicians start to use Marik’s vitamin C cocktail on their septic patients? The evidence is compelling, but Marik’s study is too small and limited in design to warrant a wide-scale change in sepsis treatment recommendations based on its findings. Further study is required, preferably a randomized clinical trial, something that Marik himself admits. However, randomized clinical trials are very expensive and take years to design and implement, so it could be a long time before the results of such a study would be available. In the meantime, given the striking results of this preliminary study, the low adverse event risk profile of vitamin C and the other ingredient’s in Marik’s cocktail, as well as their wide availability and low cost, some clinicians may well be justified to try it in their septic patients after carefully explaining the justification, risks and potential benefits to their patients and their families.
We are cautiously optimistic that this will be a major breakthrough in sepsis treatment, but for now can make no recommendations based on the available data. We encourage our readers to examine the resources listed in this blog and make up their own minds. The results and concepts discussed here are explained in far greater detail in Marik’s recent talk at the Critical Care Reviews conference and we encourage all of our readers to view it online here. In addition, a fantastic analysis of this study as well as a review of the relevant literature pertaining to vitamin C use in sepsis was written by Josh Farkas of the PulmCrit blog, and is available here. Marik’s CHEST study can be accessed here.
If any of our practitioner users opt to try this treatment and wish to report their experience to us, we would very much welcome their comments and observations. Just email us at [email protected], and we will share your experience with our large critical care community.
If this treatment truly turns out to be as effective as Marik’s preliminary work suggests, it would be a medical milestone of tremendous impact with the potential to save millions of lives, and a lesson that sometimes the most effective solution is the simplest solution that has been staring us right in the face all along. We certainly are keeping our fingers crossed.
Daniel Nichita, MD
And the ESCAVO Team