In this episode, I’ll discuss the recent study of vitamin C, hydrocortisone, and thiamine for severe sepsis and septic shock.
Lead author: Paul E. Marik
Published in Chest December 2016
Although published 3 months ago, this paper is being heavily discussed this week in traditional and social media.
The trial is described as a “retrospective before-after clinical study.” The treatment protocol involves:
Vitamin C 1.5 G IV q6 hrs
Hydrocortisone 50 mg IV q6 hrs
Thiamine 200 mg IV q12 hrs
Initially, the treatment was given to 3 patients with fulminant sepsis who appeared certain to die. After all 3 of these patients survived, the author decided to use the treatment in all septic patients admitted to the ICU. After use in 47 additional patients, matching was done to find 47 historical controls with which to compare outcomes.
The hospital mortality was 8.5% (4 of 47) in the treatment group compared to 40.4% (19 of 47) in the control group (p < 0.001). No patients in the treatment group developed progressive organ failure. Vasopressors were weaned off all patients in the treatment group, a mean of 18.3 ± 9.8 hours after starting treatment with vitamin C protocol. The mean duration of vasopressor use was 54.9 ± 28.4 hours in the control group (p<0.001).
The authors concluded:
Our results suggest that the early use of intravenous vitamin C, together with corticosteroids and thiamine may prove to be effective in preventing progressive organ dysfunction including acute kidney injury and reducing the mortality of patients with severe sepsis and septic shock. Additional studies are required to confirm these preliminary findings.
I glanced over this article when it was published in December 2016, but I didn’t think much of it.
— Pharmacy Joe ?? (@PharmacyJoe) December 22, 2016
It looked interesting, but the retrospective single-center design made me think I would want the results confirmed before applying the protocol to patients. With the strength of the evidence so low, this article didn’t make the cut in my mind to add to my weekly literature digest. It was just one of 147 articles published that week in the journals I follow regularly. Now, the article has been picked up by the lay press as a potential cure for sepsis. EVMS, the primary author’s employer, is also promoting the therapy with several professionally produced videos meant for easy consumption by the general media. I have heard of family members requesting the treatment for patients:
Given the dramatic results, the authors of the study no longer have the clinical equipoise required to do a randomized controlled trial. In other words, they believe the treatment is significantly better than placebo and could not in good faith enroll a placebo arm at their institution.
The study has already made the rounds in the blogging and podcasting world, and I gave my opinion of how the results fit into practice this Monday to my Academy members.
At Pulmcrit.org, Josh Farkas has laid out a thorough explanation of the science behind how vitamin C could be responsible for such a dramatic reduction in mortality.
At Emcrit.org, Scott Weingart has interviewed the primary author Paul Marik.
At Fluidphysiology.org, Tom Woodcock has discussed the involvement of the lay press and issues related to studying this protocol in a randomized controlled trial.
The strength of the study lies in the underlying mechanisms that could explain such a dramatic effect. As explained in the full article by Marik, both hydrocortisone and vitamin C are necessary for the production of endogenous catecholamines. Vitamin C preserves endothelial function and microcirculatory flow. Vitamin C restores glucocorticoid function. Patients with sepsis have low vitamin C levels.
The weaknesses of the study are that it was observational, single-center, and had a small sample size. Inexplicably, the authors chose to pull controls from a different season of the year than the active treatment.
Advocates of the study cite the dramatic effect, two previous trials of vitamin C alone in sepsis, the lack of known side effects of the medications studied, and the low cost of the intervention.
The effect is certainly dramatic. In the professionally produced videos from EVMS, Dr. Marik claims he has not lost a patient due to sepsis-related mortality since he started the protocol.
Interestingly, of the two existing trials that examined vitamin C in patients with sepsis, the authors of the study referenced one trial only in terms of what dose of vitamin C to use, and the other trial as proof that vitamin C alone has “little clinical impact” on outcomes in sepsis.
The lack of side effects of vitamin C is supported by small studies in critically ill patients.
Hydrocortisone is not a drug that is without side effects. Recently in patients with severe sepsis (but not septic shock) hydrocortisone was shown to provide no benefit but increased the risk of secondary infection, hyperglycemia, and muscle weakness. More than half of the patients in the Marik study had severe sepsis, although a risk of secondary infection, hyperglycemia, and muscle weakness is certainly preferable to a risk of mortality.
The cost of the intervention, as best I can tell, has been miscalculated by a factor of 9. EVMS claims a total cost of therapy to be $60. However, a 25 g vial of IV vitamin C costs ~$90. Although this contains enough vitamin C for a 4-day course of therapy, the vial is preservative free, and the best-known stability data of vitamin C in admixture is only 24 hours. This means that unless multiple patients are being treated simultaneously, the cost of treatment is $360 just for the vitamin C.
The cost of hydrocortisone is $7.50 per dose (16 doses for a course of therapy = $120). The cost of thiamine is $6 per dose (8 doses for a course of therapy = $48).
This brings the total cost of therapy to $528. If the mortality effect in this observational trial is true, then the number needed to treat to save one life is 3. Despite the stated cost of a course of therapy being off by an order of magnitude, spending $1500 on any therapy to prevent a patient death is an incredibly low cost. The mortality benefit would have to be reduced to 0.5% before the cost to prevent the death of one patient exceeded $100,000.
Critics of the study cite the observational nature of the trial and the high mortality rate of 40% in the control group.
The fact that this trial was observational means the quality of the evidence is low. It is reasonable to want additional data before applying the results to every patient with sepsis.
The criticism of the high rate of mortality in the control group is unfair in this instance. This is usually an appropriate criticism to make in studies with an unusually high mortality rate in the control group. But in this case, the high mortality rate in the control group was dictated by the study design. The patients in the treatment group had very high APACHE-IV scores, with a predicted mortality of 39.7%. In order to find appropriately matched controls, the control group needed to have similarly high APACHE-IV scores, with a predicted mortality of 41.6%. Therefore the high mortality in the control group is completely appropriate.
Pharmacy Joe’s thoughts
I’m very surprised at the way the results of this study are being promoted.
There is not even a small randomized controlled trial of this protocol to support its use. The only other RCT that shows reduced mortality with vitamin C did not include steroids, and the authors make clear how important it is to include steroids in the treatment to achieve a beneficial effect.
I have seen several ad hominem attacks on skeptics of this study, suggesting that not accepting the results means that a skeptic would only be satisfied by a multi-center double-blind randomized controlled trial (more on this later).
In a talk at a conference, the lead author of the study compared himself to Ignac Semmelweis, a mid-1800s physician who was mocked and involuntarily committed for using an observational before/after study to suggest that doctors should wash their hands after touching cadavers before delivering newborns. This seems to assert that any argument that is not in favor of the study is a non-sequitur.
Then there is the B-roll video, and the sound bites, the NPR article, and the dramatic video testimonials of nurses and family members by EVMS. The slow translation of evidence to practice is a real problem that results in the unnecessary death of many patients. The videos and media attention certainly seem to be making an effective case for the translation of evidence to practice. But in this case, the evidence is a single observational study. I know if I thought I found the cure for sepsis I would be shouting about it from the rooftops. But I would expect this type of media push more so if the results were confirmed in some way (not necessarily an RCT).
I can’t help but compare this protocol to the VSE (vasopressin+steroids+epinephrine) protocol in cardiac arrest. The VSE protocol was subjected to a randomized controlled trial in 2009 and again in 2013. The effects of VSE were dramatic. Patients who developed post-resuscitation shock in the VSE group had a 30% chance of survival to discharge compared to 0% in the control group. The rate of survival with good neurological outcome was tripled. Results of this quality are unheard of in patients who experience cardiac arrest. The VSE protocol was given IIb recommendation in the 2015 ACLS guidelines. VSE has everything that the vitamin C + steroid protocol has – cheap drugs, dramatic effect, low risk of side effects. PLUS VSE has two RCTs. Yet VSE is not widely adopted.
If VSE had professional video testimonials and an NPR article would we be using it in every hospital?
IV Vitamin C is made by a single generic manufacturer.
With rapid adoption of the protocol, the drug is ripe for two things:
- A price spike
- A drug shortage due to increased demand
How should the results of this study by Marik et al be confirmed?
How should the vitamin C + steroids Marik protocol in sepsis be studied to validate results? — Pharmacy Joe ?? (@PharmacyJoe) March 28, 2017
- Other institutions might adopt this protocol and do before/after studies of their own.
- A sequential treatment strategy could be used in a randomized controlled trial. Tom Woodcock at fluidphysiology.org explains that this type of trial design is used to minimize patient exposure to placebo while maintaining statistical validity. Such a trial would give the vitamin C protocol to patients until 1 patient died, then would give placebo to patients until 1 died, then switch back to vitamin C, etc…
- A patient registry could be started where the APACHE-IV score and predicted vs actual mortality of patients who received the treatment is reported and shared with researchers.
- Animal studies could be done – perhaps in guinea pigs as Marik notes that apart from humans guinea pigs are the only other species to not produce endogenous vitamin C.
If you can think of a different way to confirm these results, post in the comment section below.
I hope that Paul Marik has discovered the cure for sepsis. Until I see the results confirmed in some way, I’ll facilitate providers who wish to use this protocol on patients, but I won’t actively recommend it on my own.
As I see no plans for an ongoing study to confirm these results, I am considering starting a patient registry. If you are currently using this protocol, and are willing to calculate and report APACHE-IV scores and mortality data on your patients to a registry that will be shared with others, please contact me at firstname.lastname@example.org.
If you like this post, check out my book – A Pharmacist’s Guide to Inpatient Medical Emergencies: How to respond to code blue, rapid response calls, and other medical emergencies.