BY PAUL MARIK, MBBCh
Below is a list of the most common questions that have been asked in response to our paper, "Hydrocortisone, Vitamin C, and Thiamine for the Treatment of Severe Sepsis and Septic Shock: A Retrospective Before-After Study" (Chest 2016), and our best answers.
Q1. Why was the mortality rate in your control group so high relative to world averages?
A. When we began using this "cocktail," we were unsure of its benefits (and risks), and we therefore only used the cocktail in patients at highest risk of dying and developing progressive organ failure. To receive the cocktail, patients must be admitted via our emergency department with severe sepsis or septic shock and an initial procalcitonin (PCT) >2. (We routinely measure PCT in all our patients suspected of having sepsis.)
The median procalcitonin in the treatment and control groups were 25 ng/ml and 15 ng/ml respectively. The median PCT in the MOSES study, the largest study to date to evaluate the time course of PCT in patients with severe sepsis and septic shock, was 5 ng/ml. (Crit Care Med 2017;45:781.) In the MOSES study, nearly 50 percent of patients had a PCT <2. PCT is a very good quantitative marker of sepsis severity. We therefore selected a subgroup of patients with sepsis at highest risk of complications and death, i.e., we include the sickest half of the patients with sepsis.
When we performed the retrospective, case-controlled study, we matched the control group patients based on PCT and severity of illness, meaning we used the same selection criteria for the treatment and control groups, which was indicated in the manuscript. After completing the study and once we analyzed the data for efficacy and safety, we changed our indications for use: We now treat every patient with sepsis who is sick enough to be admitted to the ICU with the cocktail, starting in the ED and at the same time as antibiotics. This may explain the progressive decline in sepsis mortality at our hospital from January 2016 to January 2017, which has been reported by an independent data analytics company. (See YouTube video at http://bit.ly/2rdozj9.)
Furthermore, the exact incidence and mortality of sepsis in the United States (or elsewhere) is difficult to determine. It depends on the selection criteria and whether the data are severity-adjusted. Administrative databases have been shown to be notoriously unreliable. In addition, one cannot extrapolate from the mortality reported in "large" clinical trials due to the large number of exclusion criteria. They do not represent real-world practice. It is noteworthy that in the Surviving Sepsis Campaign Database, which includes more than 130,000 patients, the overall mortality was 32.8 percent.
Q2. How do you mix and administer vitamin C, and what is its shelf life and cost?
A. Vitamin C is provided by the manufacturer as a 50 ml vial at a concentration of 500 mg/ml. We add 3 ml of vitamin C into a 50 ml bag of normal saline (1,500 mg vitamin C in 50 ml bag), which is then infused over one hour. The dosing schedule is 1,500 mg every six hours for four days or until discharge from the ICU.
Once the vial is open, it is only stable for six hours. The resulting product has given 24 hours of stability. When first ordered, four doses are made and sent to the nurse. For the next day, we will wait until one hour prior to make the next batch. The acquisition cost of IV vitamin C as sold by Mylan Institutional is $81 per vial. KRS Global Biotechnology in Boca Raton, FL, compounds a vitamin C formulation (Tapioca) 500 mg/ml 50ml at $20 per vial, which has a 90-day shelf life from the time made and a seven-day shelf life after opening the vial.
Q3. What about the use of vitamin C and spurious blood glucose levels using point-of-care glucose monitors?
A. Vitamin C and glucose have very similar molecular structures, both being six-carbon molecules with glucose-6-phospate being the precursor molecule of vitamin C. Spuriously elevated POC glucose levels have been reported in patients with burns, who have received large pharmacologic doses of vitamin C (more than 50 g/day). (Psych Res 1989;30:165; J Burn Care Res 2015;36:50.) This phenomenon has been reported with vitamin C and other compounds with POC glucose devices that incorporate the glucose-dehydrogenase-pyrroloquinoline quinone (PQQ) amperometric method of testing. The pharmacologic doses of vitamin C used in patients with burns (PsychRes 1989;30:165; J Burn Care Res 2015;36:50) and those with malignancy (Proc Natl Acad Sci U S A 2005;102:13604; PLoS One 2015;10:e0120228; Science 2015;350:1391) result in millimolar concentrations of vitamin C.
Our dosing strategy (1.5 g IV q 6 hourly) results in blood vitamin C levels that are in the 200 umol/l range, which should not cause significant cross-reaction with blood glucose concentrations that are in the millimolar range (6-11 mmol). To validate this, we measured blood glucose levels with POC testing (Accu-Chek Inform, Roche, Indianapolis) and simultaneously with our central laboratory at the end of the vitamin C infusion, and noted minimal differences in the measured levels. We, therefore, believe that in the dosage used in our study, the interaction between vitamin C and POC glucose testing is not a clinically significant problem.
Q4. Why did you apply a protocol that had not yet been tested?
A. As clinicians responsible for the care of patients who are at an exceedingly high risk of death, one "has to think outside the box" because our current approach is unsatisfactory. The three agents that we combined in our "cocktail" are widely available, safe, and cheap, and have been used individually to treat sepsis. All patients with severe sepsis/septic shock have low or undetectable levels of vitamin C. The vitamin C was dosed according to the indications and dosage of the approved package insert. (See below.) We combined these three agents on the presumption that the sum would be greater than the individual components.
Q5. If our intensivists begin using vitamin C for our sepsis patients, would the IRB need to oversee this as research? Did you obtain IRB approval for your study?
A. No, IRB approval is not required. As is clear from our paper, we performed a retrospective, before-after study approved by our IRB. This treatment approach has now become standard of care in our ICU and in many ICUs across the world, although the selection criteria may be different. The use of the cocktail in the clinical context is not research; the cocktail consists of three widely available and approved drugs with a long safety record.
Many of the drugs used in the ICU in the United States do not having supporting scientific data based on large RCTs, nor do they have specific FDA labeling for that indication. Haloperidol is a good example. The use of haloperidol for a patient with delirium in the ICU does not require IRB approval. It should be noted that haloperidol has significant risks and toxicity. When using "nonstandard" treatments or a drug for non-approved indications, however, the clinician must balance the potential benefits of the intervention with the consequences of the disease (untreated) and the safety of the proposed therapy.
It is our opinion based on an extensive review of the literature as well as our experience and that of others that vitamin C, thiamine, and hydrocortisone are exceedingly safe and have no known side effects in a disease without alternative treatments and an exceedingly high mortality. Indeed, we believe that clinicians have an ethical obligation to treat overwhelmingly septic patients with this cocktail. The balance of beneficence, nonmaleficence, and social justice clearly supports this treatment approach.
Q6. Are patients still consistently responding well to your therapy?
A. We have now treated more than 250 patients with this cocktail, and received the same reproducible response. During this time (over 16 months), only one patient has died while receiving this therapy. That patient was a complicated surgical patient with inadequate source control who died within hours of surgery. Our therapeutic cocktail has been used in hundreds of patients across the globe with strikingly similar positive results.
Here's a typical example: "An elderly man with ischemic cardiomyopathy and EF 15% pacer/AICD at baseline presented to outside hospital with shock and BP 60/30 mm Hg.He was given two-three liters of fluid, and immediately went into respiratory failure requiring intubation. He was transferred to us intubated on four vasopressors with AKI (Cr 3). Eventually he ended up growing Group B Strep from his blood with procalcitonin 43. He was started on the vitamin C cocktail, and within a day his pressor requirements melted away, and he was extubated. His kidneys have improved, and he is walking around the ICU. Tomorrow he will probably leave the ICU with no residual organ dysfunction, no volume overload, and no ICU complications."
Q7. After the publication of your work, how has your protocol been extended? What experience have you gathered from the other centers?
A. More than 30 medical centers across the United States and the world are now using this protocol in daily practice. As already indicated, the results are reproducible time and time again, and hundreds of lives have been saved.
Q8. Do you think adopting your treatment is a matter of patents and money?
A. The wonderful thing about our novel intervention is that nobody will become rich from this therapy, including me, because it is cheap and readily available. It is not possible to patent these drugs or the combination, as I understand it. This is important because sepsis is common in resource-limited countries that cannot afford expensive designer pharmaceutical drugs.
Q9. What do you think about the opinion of those who have criticized your protocol?
A. Criticism and skepticism are an essential part of scientific discourse, but it must be professional, scientific, and not personal. Our protocol was not "sucked out of thin air." There is an enormous body of scientific research to support using all three components; all we did was put them together.
Q10. From your critical care experience, do you think it is currently the best treatment in cases of sepsis or septic shock?
A. I believe this treatment is an important component for treating sepsis and septic shock, but it cannot be applied in isolation. The important components are:
- Early identification of sepsis
- Early prescription of the right antibiotic(s) in the right dose
- Adequate source control (very important)
- A physiologic, restrictive fluid strategy with the early use of norepinephrine ( very important)
- The "metabolic resuscitation protocol:" steroids, vitamin C, and thiamine
- State-of-the-art supportive care based on the best scientific evidence
- A multidisciplinary, team approach to patient care
Q11. Do you think it is important to know to what degree each protocol component works?
A. As indicated previously, each component has some benefits, but together the sum is more than the parts. This is no different from the approach oncologists use to treat cancer; almost every oncology protocol uses multiple drugs that target different pathways.
Q12. What about the safety of IV vitamin C, particularly in patients with renal impairment?
A. Vitamin C in the doses we recommend is extremely safe, even in patients with renal impairment. We have checked serum oxalate levels in patients with chronic renal failure who are not receiving hemodialysis (HD) at the end of treatment with the cocktail, and they are all in the safe range. Vitamin C is a small molecule that is freely dialyzable and safe in the doses we recommend in patients receiving renal replacement therapy (continuous renal replacement therapy or conventional HD). We have closely followed the serum creatinine (Cr) in our patients with acute kidney injury (AKI) treated with the cocktail, and the creatinine has fallen in all of them. (See graph.)
Dosages of 150 g (100 times the dose we use) have been safely given to patients with cancer. The only precaution with such high doses is chronic renal impairment that increases the risk of hyperoxalosis and worsening renal function. Hyperoxalosis and acute kidney injury have not been reported in patients receiving less than 10 g/day. These facts are well documented in the literature. The only other potential complication with these pharmacologic doses of vitamin C is hemolysis in patients with G6PD deficiency. (See Q14.)
Q13. Is there a risk of oxalate accumulation in patients with chronic renal failure (CRF) on dialysis? Is there a level of renal failure beyond which you would advise dose-reducing the vitamin C?
A. This is the one area that did concern us, and we have explored in much detail (also see answer to the previous question). In patients with CRF on HD, there is no issue with oxalate. AKI is common in patients with sepsis; we have found that the Cr comes down in all cases. (See graph above.) What's trickier is the situation in which patients have CRF and AKI, as well as septic shock. We had one such patient at the very beginning. We tried a dosage adjustment in him, but he failed to respond to the cocktail.
We have subsequently used the standard dose (1.5 g IV q 6) in all patients with CRF, and the Cr has fallen in all. We monitored oxalate levels in the first half-dozen or so of these patients; the levels were in the safe range. In patients at highest risk of kidney injury, I would suggest that the Cr be closely monitored, and if the Cr increases, it is not unreasonable to stop the vitamin C on day 3 and measure the serum oxalate level until further safety data are available on these patients. We believe, however, it is safe to continue treatment with vitamin C in patients receiving CRRT.
Q14. What about the risk of hemolysis in patients with G6PD deficiency?
A. Low-dose vitamin C is protective against hemolysis in patients with G6PD deficiency. It only causes hemolysis in very high doses. We have not had any issue with hemolysis or any other complications.
Q15. "Twenty-eight of 47 (59.6%) patients in the control group were treated with hydrocortisone.'' Was an analysis done on the mortality outcome for these 28 patients?
A. During the control period, hydrocortisone was used at the discretion of the treating physician. While the sample size is very small, the use of hydrocortisone compared with no hydrocortisone did not appear to affect any of the outcome measures.
Q16. It would have been interesting to have a prospective study comparing the effects of vitamin C and thiamine with a hydrocortisone group alone. I do believe this might be the variable that may have created such a gap between the primary outcome results.
A. As already indicated and as supported by our bench research currently in press (Chest 2017), we believe that vitamin C and hydrocortisone act synergistically on multiple metabolic pathways to limit and reverse the multiple derangements in cellular and organ dysfunctions that occur in sepsis. The added benefit of IV thiamine requires further study. Additional basic science and clinical studies are required to confirm our preliminary findings.
Q17. Our surgeons will not let us use this cocktail because they believe it impairs wound healing and increases the risks of secondary infections.
A. This is a common and absurd myth. There are multiple mechanistic reasons why the cocktail may actually promote wound healing and prevent secondary infections. Septic patients have critically low vitamin C levels. It is well known that vitamin C is essential for collagen formation and wound healing. In addition, vitamin C improves macrophage and T cell function, and prevents the development of post-sepsis immunosuppression. (Arch Surg 1997;132:129.) It is important to emphasize that the metabolic resuscitation protocol uses stress doses of hydrocortisone, which are equivalent to those produced by a normally functioning adrenal gland during stress; these are not immunosuppressive doses of corticosteroids that impair immune function or wound healing. On the contrary, this dose of hydrocortisone is likely to promote resolution of tissue inflammation and wound healing.
Schulze, et al., performed a randomized, double-blind, placebo-controlled trial that investigated the effects of acute, preoperative corticosteroid administration on cutaneous wound healing. (Arch Surg 1997;132:129.) The 24 patients in this study received a single dose of 30 mg/kg methylprednisolone (equivalent to 10,000 mg of hydrocortisone; 100 times the dose used in our protocol) or placebo intravenously 90 minutes prior to colon resection. There was no difference in wound healing between the two groups. Proline levels in the collagen and the amount of collagen accumulation within the wounds over 10 postoperative days were also evaluated, with no difference between the groups. These results suggest that acute, high-dose steroid administration does not significantly affect wound healing, as measured by both clinical and biochemical parameters. Audrey Wang, MD, and colleagues reviewed the literature on the use of corticosteroids in the perioperative period and wound healing. (Am J Surg 2013;206:410.) These authors concluded that "the preponderance of human literature found that high-dose corticosteroid administration for less than10 days has no clinically important effect on wound healing."
Q18. We believe that the findings from your study are spurious, that the results are biased, and the conclusion overstated. There is no high-quality evidence that any of the three interventions individually improves survival in patients with sepsis. The substantial methodological flaws of the study undermine its external validity and veracity. We highly recommend meticulous assessment of all patient-important benefits and harms of vitamin C, hydrocortisone, and thiamine prior to adopting this unproven strategy in clinical ICU practice. (Synopsis of letter to Chest by Moller, et al.)
A. First and foremost, it is important to state that we do not refute nor did we attempt to mask the nature of our study: retrospective, single-center, nonrandomized, and unblinded. We initiated this therapy after our review of small trials in similar populations. (J Transl Med 2014;12:32; Crit Care Med 2016;44:360.) We agree that the supporting data on efficacy were not robust, but felt that the available safety data on these particular interventions justified their introduction as salvage therapy in septic patients who were unlikely to survive. Our anecdotal experience was impressive, and led us to use this treatment in a number of consecutive patients.
Our decision to describe and publish our experience occurred afterwards; thus the methodological characteristics were unmodifiable. As a result of this study design, we understand why the results of our study have been met with some skepticism by the scientific community. We agree with Moller, et al., that additional trials should be performed to support or refute the findings of our study. As stated in our conclusion, "…additional studies are required to confirm our preliminary findings." Given our experience, however, we felt that publication of these results was necessary, and it's our ethical responsibility to do so.
Sepsis is common, debilitating, and often lethal. Despite exhaustive attempts at therapies to interrupt the mechanism of a dysregulated immune response and subsequent organ damage, we are currently limited to antibiotics and supportive care as our only consensus therapeutic measures. We recognize that the decision to use three readily available pharmacologic agents, each with limited supporting clinical data, can be viewed as unconventional, but therapeutic interventions in the absence of high-quality, randomized, controlled trials are commonplace in the ICU. When using such interventions, the clinician must balance the potential consequences of the disease with the safety of the proposed therapy. It is our opinion that the safety profile of vitamin C, thiamine, and hydrocortisone for a disease without alternative treatments and an exceedingly high mortality allows clinicians to use this therapeutic intervention to prevent death and limit the complications and long-term sequelae of this devastating disease. (J Transl Med 2014;12:32; Crit Care Med 2016;44:360; Ann Surg 2002;236:814; J Res Pharm Pract 2016;5:94; Crit Care Med 2008;36:1937.)
We continue to support the effort to investigate this therapy further in studies. The mortality reductions described in retrospective studies are often not reproduced in large, multicenter RCTs. A therapy that effectively targets pivotal pathways in patients with sepsis, however, could plausibly result in a large reduction in mortality. The associated reduction in the dose of vasopressors, qSOFA score, and procalcitonin clearance in our treated patients, all independent markers of the successful treatment of sepsis (Crit Care Med 2017;45:781) , suggests a true biologic effect. In addition, we have independent validation that the sepsis mortality in our hospital has been dramatically reduced since the introduction of this novel therapeutic intervention.
It has previously been suggested that "the best hope for therapeutic advances [in sepsis] will depend on broad-base targeting, in which multiple components are targeted at the same time." (Blood 2003;101:3765.) Such combination "chemotherapy" targeting multiple biological pathways is the standard approach in treating malignant diseases. The benefits of vitamin C, hydrocortisone, and thiamine alone are likely limited, but we believe that these medications act synergistically to effect the desired outcomes. Laboratory evidence of a synergistic interaction between hydrocortisone and vitamin C in preserving endothelial integrity supports this claim. (Chest 2017; in press.)
Read "Vitamin C for Sepsis Remains a Hypothesis" at http://bit.ly/2uDwzaR.
Dr. Marik is a professor of internal medicine and the chief of pulmonary and critical care medicine at Eastern Virginia Medical School in Norfolk.