Effect of Vitamin C, Thiamine, and Hydrocortisone on Ventilator- and Vasopressor-Free Days in Patients With Sepsis: The VICTAS Randomized Clinical Trial | Complementary and Alternative Medicine | JAMA | JAMA Network
[Skip to Navigation]
Sign In
Visual Abstract. Vitamin C, Thiamine, and Hydrocortisone and Ventilator- and Vasopressor-Free Days in Sepsis
Vitamin C, Thiamine, and Hydrocortisone and Ventilator- and Vasopressor-Free Days in Sepsis
Figure 1.  Participant Flow in the Vitamin C, Thiamine, and Steroids in Sepsis Trial
Participant Flow in the Vitamin C, Thiamine, and Steroids in Sepsis Trial

aCategories are not mutually exclusive; participants could have more than 1 reason.

Figure 2.  Survival in the Intervention and Control Groups at 180 Days
Survival in the Intervention and Control Groups at 180 Days

Cumulative deaths for the 180-day follow-up period are shown. Median time to death in the intervention group was 16 (interquartile range, 4-41) days and in the control group was 15 (interquartile range, 6-29) days. The proportionality P = .81, suggesting that the proportionality assumption was met.

Table 1.  Participant Characteristics
Participant Characteristics
Table 2.  Exploratory Outcomes in the Main and Per-Protocol Analyses
Exploratory Outcomes in the Main and Per-Protocol Analyses
1.
Singer  M, Deutschman  CS, Seymour  CW,  et al.  The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3).   JAMA. 2016;315(8):801-810. doi:10.1001/jama.2016.0287PubMedGoogle ScholarCrossref
2.
Rhee  C, Dantes  R, Epstein  L,  et al; CDC Prevention Epicenter Program.  Incidence and trends of sepsis in US hospitals using clinical vs claims data, 2009-2014.   JAMA. 2017;318(13):1241-1249. doi:10.1001/jama.2017.13836PubMedGoogle ScholarCrossref
3.
Torio  CM, Moore  BJ. National Inpatient Hospital Costs: The Most Expensive Conditions by Payer, 2013. Agency for Healthcare Research and Quality; 2016. Healthcare Cost and Utilization Project Statistical Brief No. 204.
4.
Buchman  TG, Simpson  SQ, Sciarretta  KL,  et al.  Sepsis among Medicare beneficiaries, 1: the burdens of sepsis, 2012-2018.   Crit Care Med. 2020;48(3):276-288. doi:10.1097/CCM.0000000000004224PubMedGoogle ScholarCrossref
5.
Yende  S, Austin  S, Rhodes  A,  et al.  Long-term quality of life among survivors of severe sepsis: analyses of two international trials.   Crit Care Med. 2016;44(8):1461-1467. doi:10.1097/CCM.0000000000001658PubMedGoogle ScholarCrossref
6.
Rhodes  A, Evans  LE, Alhazzani  W,  et al.  Surviving Sepsis Campaign: international guidelines for management of sepsis and septic shock: 2016.   Crit Care Med. 2017;45(3):486-552. doi:10.1097/CCM.0000000000002255PubMedGoogle ScholarCrossref
7.
Marshall  JC.  Why have clinical trials in sepsis failed?   Trends Mol Med. 2014;20(4):195-203. doi:10.1016/j.molmed.2014.01.007PubMedGoogle ScholarCrossref
8.
Spoelstra-de Man  AME, Elbers  PWG, Oudemans-Van Straaten  HM.  Vitamin C: should we supplement?   Curr Opin Crit Care. 2018;24(4):248-255. doi:10.1097/MCC.0000000000000510PubMedGoogle ScholarCrossref
9.
Wilson  JX.  Evaluation of vitamin C for adjuvant sepsis therapy.   Antioxid Redox Signal. 2013;19(17):2129-2140. doi:10.1089/ars.2013.5401PubMedGoogle ScholarCrossref
10.
Fujii  T, Luethi  N, Young  PJ,  et al; VITAMINS Trial Investigators.  Effect of vitamin C, hydrocortisone, and thiamine vs hydrocortisone alone on time alive and free of vasopressor support among patients with septic shock: the VITAMINS randomized clinical trial.   JAMA. 2020;323(5):423-431. doi:10.1001/jama.2019.22176PubMedGoogle ScholarCrossref
11.
Fowler  AA  III, Truwit  JD, Hite  RD,  et al.  Effect of vitamin C infusion on organ failure and biomarkers of inflammation and vascular injury in patients with sepsis and severe acute respiratory failure: the CITRIS-ALI randomized clinical trial.   JAMA. 2019;322(13):1261-1270. doi:10.1001/jama.2019.11825PubMedGoogle ScholarCrossref
12.
Putzu  A, Daems  AM, Lopez-Delgado  JC, Giordano  VF, Landoni  G.  The effect of vitamin C on clinical outcome in critically ill patients: a systematic review with meta-analysis of randomized controlled trials.   Crit Care Med. 2019;47(6):774-783. doi:10.1097/CCM.0000000000003700PubMedGoogle ScholarCrossref
13.
Barabutis  N, Khangoora  V, Marik  PE, Catravas  JD.  Hydrocortisone and ascorbic acid synergistically prevent and repair lipopolysaccharide-induced pulmonary endothelial barrier dysfunction.   Chest. 2017;152(5):954-962. doi:10.1016/j.chest.2017.07.014PubMedGoogle ScholarCrossref
14.
Marik  PE, Khangoora  V, Rivera  R, Hooper  MH, Catravas  J.  Hydrocortisone, vitamin C, and thiamine for the treatment of severe sepsis and septic shock: a retrospective before-after study.   Chest. 2017;151(6):1229-1238. doi:10.1016/j.chest.2016.11.036PubMedGoogle ScholarCrossref
15.
Hager  DN, Hooper  MH, Bernard  GR,  et al.  The Vitamin C, Thiamine and Steroids in Sepsis (VICTAS) protocol: a prospective, multi-center, double-blind, adaptive sample size, randomized, placebo-controlled, clinical trial.   Trials. 2019;20(1):197. doi:10.1186/s13063-019-3254-2PubMedGoogle ScholarCrossref
16.
Lindsell  CJ, McGlothlin  A, Nwosu  S,  et al.  Update to the Vitamin C, Thiamine and Steroids in Sepsis (VICTAS) protocol: statistical analysis plan for a prospective, multicenter, double-blind, adaptive sample size, randomized, placebo-controlled, clinical trial.   Trials. 2019;20(1):670. doi:10.1186/s13063-019-3775-8PubMedGoogle ScholarCrossref
17.
Hager  DN, Martin  GS, Sevransky  JE, Hooper  MH.  Glucometry when using vitamin C in sepsis: a note of caution.   Chest. 2018;154(1):228-229. doi:10.1016/j.chest.2018.03.018PubMedGoogle ScholarCrossref
18.
Vincent  JL, Moreno  R, Takala  J,  et al; Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine.  The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure.   Intensive Care Med. 1996;22(7):707-710. doi:10.1007/BF01709751PubMedGoogle ScholarCrossref
19.
Ely  EW, Margolin  R, Francis  J,  et al.  Evaluation of delirium in critically ill patients: validation of the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU).   Crit Care Med. 2001;29(7):1370-1379. doi:10.1097/00003246-200107000-00012PubMedGoogle ScholarCrossref
20.
Moskowitz  A, Huang  DT, Hou  PC,  et al; ACTS Clinical Trial Investigators.  Effect of ascorbic acid, corticosteroids, and thiamine on organ injury in septic shock: the ACTS randomized clinical trial.   JAMA. 2020;324(7):642-650. doi:10.1001/jama.2020.11946PubMedGoogle ScholarCrossref
21.
de Grooth  HJ, Elbers  PWG, Vincent  JL.  Vitamin C for sepsis and acute respiratory failure [letter].   JAMA. 2020;323(8):792. doi:10.1001/jama.2019.21981PubMedGoogle ScholarCrossref
22.
Fowler  AA  III, Fisher  BJ, Kashiouris  MG.  Vitamin C for sepsis and acute respiratory failure [letter reply].   JAMA. 2020;323(8):792-793. doi:10.1001/jama.2019.21987PubMedGoogle ScholarCrossref
Original Investigation
Caring for the Critically Ill Patient
February 23, 2021

Effect of Vitamin C, Thiamine, and Hydrocortisone on Ventilator- and Vasopressor-Free Days in Patients With Sepsis: The VICTAS Randomized Clinical Trial

Author Affiliations
  • 1Division of Pulmonary, Allergy, Critical Care, and Sleep, Emory University School of Medicine, Atlanta, Georgia
  • 2Emory Critical Care Center, Emory Healthcare, Atlanta, Georgia
  • 3Department of Emergency Medicine, Johns Hopkins University, Baltimore, Maryland
  • 4Division of Pulmonary Critical Care, Johns Hopkins University, Baltimore, Maryland
  • 5Division of Pulmonary Critical Care, Vanderbilt University Medical Center, Nashville, Tennessee
  • 6Division of Pulmonary Critical Care, Intermountain Medical Center and University of Utah, Salt Lake City
  • 7Department of Surgery, Emory University School of Medicine, Atlanta, Georgia
  • 8Division of Pulmonary Critical Care, Virginia Commonwealth University, Richmond
  • 9Johns Hopkins Critical Illness, Brain Dysfunction, and Survivorship (CIBS) Center, Baltimore, Maryland
  • 10Veteran’s Affairs Tennessee Valley Geriatric Research Education Clinical Center (GRECC), Nashville
  • 11Division of Brain Injury Outcomes, Johns Hopkins University, Baltimore, Maryland
  • 12Department of Emergency Medicine, Thomas Jefferson University Hospital, Philadelphia, Pennsylvania
  • 13Department of Critical Care, Montefiore Medical Center, Bronx, New York
  • 14Department of Emergency Medicine, Emory University, Atlanta, Georgia
  • 15Division of Pulmonary Critical Care, Sentara Healthcare, Norfolk, Virginia
  • 16Division of Pulmonary Critical Care, Oregon Health & Science University, Portland
  • 17Molecular and Clinical Nutrition Section, National Institutes of Health, Bethesda, Maryland
  • 18Department of Emergency Medicine, Harbor-UCLA Medical Center, Torrance, California
  • 19Berry Consultants LLC, Austin, Texas
  • 20Department of Emergency Medicine, David Geffen School of Medicine at UCLA, Los Angeles, California
  • 21Lundquist Institute for Biomedical Innovation, Torrance, California
  • 22Department of Biostatistics, Vanderbilt University Medical Center, Nashville, Tennessee
  • 23Vanderbilt Coordinating Center, Vanderbilt University Medical Center, Nashville, Tennessee
  • 24Department of Pathology and Laboratory Medicine, Emory University School of Medicine, Atlanta, Georgia
  • 25Department of Biostatistics, University of Kentucky, Lexington
JAMA. 2021;325(8):742-750. doi:10.1001/jama.2020.24505
Visual Abstract. Vitamin C, Thiamine, and Hydrocortisone and Ventilator- and Vasopressor-Free Days in Sepsis
Vitamin C, Thiamine, and Hydrocortisone and Ventilator- and Vasopressor-Free Days in Sepsis
Key Points

Question  In adults with sepsis-induced respiratory or cardiovascular dysfunction, does treatment with vitamin C, thiamine, and hydrocortisone result in an increase in the number of days alive and free of mechanical ventilation and vasopressor use?

Findings  In this randomized clinical trial that included 501 patients, treatment with vitamin C, thiamine, and hydrocortisone compared with placebo resulted in a median of 25 vs 26 ventilator- and vasopressor-free days within the 30 days following randomization, a difference that was not statistically significant.

Meaning  Among critically ill patients with sepsis, treatment with vitamin C, thiamine, and hydrocortisone did not significantly improve ventilator- and vasopressor-free days, although, the trial was terminated early for administrative reasons and may have been underpowered to detect a clinically important difference.

Abstract

Importance  Sepsis is a common syndrome with substantial morbidity and mortality. A combination of vitamin C, thiamine, and corticosteroids has been proposed as a potential treatment for patients with sepsis.

Objective  To determine whether a combination of vitamin C, thiamine, and hydrocortisone every 6 hours increases ventilator- and vasopressor-free days compared with placebo in patients with sepsis.

Design, Setting, and Participants  Multicenter, randomized, double-blind, adaptive-sample-size, placebo-controlled trial conducted in adult patients with sepsis-induced respiratory and/or cardiovascular dysfunction. Participants were enrolled in the emergency departments or intensive care units at 43 hospitals in the United States between August 2018 and July 2019. After enrollment of 501 participants, funding was withheld, leading to an administrative termination of the trial. All study-related follow-up was completed by January 2020.

Interventions  Participants were randomized to receive intravenous vitamin C (1.5 g), thiamine (100 mg), and hydrocortisone (50 mg) every 6 hours (n = 252) or matching placebo (n = 249) for 96 hours or until discharge from the intensive care unit or death. Participants could be treated with open-label corticosteroids by the clinical team, with study hydrocortisone or matching placebo withheld if the total daily dose was greater or equal to the equivalent of 200 mg of hydrocortisone.

Main Outcomes and Measures  The primary outcome was the number of consecutive ventilator- and vasopressor-free days in the first 30 days following the day of randomization. The key secondary outcome was 30-day mortality.

Results  Among 501 participants randomized (median age, 62 [interquartile range {IQR}, 50-70] years; 46% female; 30% Black; median Acute Physiology and Chronic Health Evaluation II score, 27 [IQR, 20.8-33.0]; median Sequential Organ Failure Assessment score, 9 [IQR, 7-12]), all completed the trial. Open-label corticosteroids were prescribed to 33% and 32% of the intervention and control groups, respectively. Ventilator- and vasopressor-free days were a median of 25 days (IQR, 0-29 days) in the intervention group and 26 days (IQR, 0-28 days) in the placebo group, with a median difference of −1 day (95% CI, −4 to 2 days; P = .85). Thirty-day mortality was 22% in the intervention group and 24% in the placebo group.

Conclusions and Relevance  Among critically ill patients with sepsis, treatment with vitamin C, thiamine, and hydrocortisone, compared with placebo, did not significantly increase ventilator- and vasopressor-free days within 30 days. However, the trial was terminated early for administrative reasons and may have been underpowered to detect a clinically important difference.

Trial Registration  ClinicalTrials.gov Identifier: NCT03509350

Introduction

Sepsis is a common syndrome caused by infection leading to substantial morbidity and mortality.1 Of the estimated 1 750 000 cases of sepsis in the United States in 2014, approximately 55% required intensive care unit (ICU) admission and 20% to 30% died,2 making sepsis the third leading cause of hospital deaths, with an annual estimated cost of nearly $60 billion.3,4 Patients who survive are at risk of worse physical, emotional, and cognitive outcomes and reduced quality of life.5 Although early antibiotics, fluid resuscitation, hemodynamic support, and source control have been associated with improved patient outcomes,6 other sepsis interventions have not demonstrated consistent benefits.7

Recently, high-dose intravenous vitamin C has been proposed as a therapy for sepsis.8 There is biological plausibility for the value of vitamin C in sepsis,9 but the results of clinical studies have been inconsistent.10-12 The addition of thiamine and hydrocortisone to vitamin C has been suggested to be beneficial in preclinical and observational studies.13,14

The Vitamin C, Thiamine, and Steroids in Sepsis (VICTAS) trial was designed to test the hypothesis that combination therapy with vitamin C, thiamine, and hydrocortisone would improve clinically important outcomes in patients with sepsis-induced respiratory and/or circulatory failure.15

Methods
Trial Design and Oversight

The trial was a randomized, double-blind, adaptive-sample-size, placebo-controlled trial conducted at 43 hospitals in the United States. The protocol (Supplement 1) and statistical analysis plan (Supplement 2) have been previously published.15,16 The central institutional review board at Johns Hopkins University approved the trial protocol and consent documents for all sites. Participants or legally authorized representatives provided written informed consent prior to enrollment and randomization. An independent data and safety monitoring board conducted reviews of safety data and predicted probability of success during 4 predetermined interim analyses.

Trial Population

Quiz Ref IDPatients aged 18 years or older with acute respiratory and/or cardiovascular dysfunction caused by suspected infection with planned ICU admission were eligible for enrollment. Infection was defined by the ordering of blood cultures and administration of at least 1 antimicrobial agent. Acute respiratory dysfunction was defined by an arterial partial pressure of oxygen/fraction of inspired oxygen (Fio2) of 300 or less or a pulse oximetry oxygen saturation/Fio2 of 315 or less and the need for (1) intubation and mechanical ventilation, (2) noninvasive positive pressure ventilation, or (3) high-flow nasal cannula at 40 L/min or higher with an Fio2 of at least 0.40. Cardiovascular dysfunction was defined as the need for any vasopressor for more than 1 hour to maintain a mean arterial pressure of 65 mm Hg or higher despite intravenous crystalloid fluid resuscitation of 1 L or more. There were no exclusions based on race or ethnicity. Race and ethnicity were considered possible effect modifiers and were systematically collected from the medical record using fixed categories. Full inclusion and exclusion criteria are detailed in Supplement 1.

Randomization

Patients were randomized in a 1:1 ratio to either the intervention group or the control group within 24 hours of qualifying organ dysfunction, and organ dysfunction was required to be ongoing at randomization. The randomization sequence was generated using statistical software and used permuted small blocks of size 2, 4, or 6, stratified by site. Study group randomization was operationalized via the use of presorted drug kits shipped to study sites from a centralized pharmacy and Vanderbilt University Medical Center, with participants assigned kits in predefined sequence. Randomization blinding was maintained by labeling study drug kits with numbers only. Investigational drug pharmacists were not blinded but were prohibited by protocol from revealing assignment to anyone.

Interventions

Quiz Ref IDParticipants received intravenous vitamin C (1.5 g), thiamine hydrochloride (100 mg), and hydrocortisone sodium succinate (50 mg) or matching placebos within 4 hours of randomization and then every 6 hours thereafter up to 96 hours, death, or discharge from the ICU, whichever occurred first.Quiz Ref ID Participants could be treated with open-label corticosteroids by the clinical team: for daily doses of at least 200 mg of hydrocortisone (or equivalent), hydrocortisone or matching placebo was withheld by the investigational pharmacy.

All other management of participants was at the discretion of the clinical team. A point-of-care device approved for use in the setting of high plasma concentrations of vitamin C or a central laboratory device not affected by vitamin C were used for glucose measurements through 24 hours following the last dose of study drug.17

Outcomes

The primary outcome measure was ventilator- and vasopressor-free days (VVFDs) in the first 30 days following randomization (day 0). Participants who died were assigned 0 VVFDs. A “last status carried forward” approach was used in the calculation of VVFDs for participants discharged from the hospital before day 30.

The key secondary outcome was mortality within 30 days of randomization. A patient discharged alive was considered to have survived to day 30. Exploratory outcomes to support efficacy included ICU mortality, ICU and hospital lengths of stay, ICU delirium- and coma-free days, kidney replacement therapy–free days at day 30, and change between prerandomization and day 4 Sequential Organ Failure Assessment (SOFA) score (range, 0 [best] to 28 [worst]).18 Safety end points included the prespecified potentially associated adverse events of nephrolithiasis, hemolysis, hypersensitivity reactions, and injection site reactions. At discharge or at 30 days, whichever occurred sooner, participants were invited to participate in a 180-day telephone follow-up. For patients who died between discharge and 180 days, date of death was obtained at this follow-up.

Statistical Analysis

In the absence of phase 2 data at the time of trial design, an adaptive approach was used. The trial was designed to detect a difference in the primary end point of 1.5 VVFDs with recruitment of up to 2000 patients. These calculations were based on a mortality rate of 25% and the proposed end point of 1.5 VVFDs was based on a vote from the executive committee as meaningful (Supplement 3) while allowing early stoppage if a large effect on mortality was observed.15,16 At each planned interim analysis, all monitored and unmonitored data for participants with day 30 observations were used, and bayesian predictive distributions were used to impute the outcomes of participants with incomplete data. Together, these data were used to compute a predictive probability of success on the mortality end point when 200, 300, and 400 participants were enrolled. Predefined stopping rules for efficacy, but not futility, were in place for these early interim analyses. Beyond N = 400, interim analyses were planned at N = 500, N = 1000, and N = 1500, at which both efficacy and futility were to be assessed using the VVFD outcome (Supplement 3).

The success thresholds at each interim were calibrated to control the 1-sided type I error rate at 2.5% (Supplement 2).16 If a possible 20% mortality benefit were real, which is lower than the absolute risk reduction in mortality seen in a prior study,14 the trial was very likely (approximately 97% chance) to stop at or before 400 patients with a power of approximately 99%.15

Categorical variables are reported using frequencies and percentages. Continuous variables are reported as means with standard deviations and medians with interquartile ranges (IQRs). In the primary analysis, patients were analyzed according to their randomization group. Given the approach to calculating VVFDs, the primary outcome was available for all participants. The VVFDs are described using medians and IQRs; bivariable differences between groups are reported with 95% confidence intervals and compared using the Wilcoxon rank sum test. To account for the multiple interim analyses, the critical P value was set to a 1-sided threshold of P = .022. Mortality is described using frequencies and percentages. A gatekeeping strategy was used to decide whether to formally test for a treatment effect on mortality. Under this strategy, the mortality end point would be formally tested using a χ2 test with a 1-sided α = .024 if a difference in the primary end point of VVFDs was observed, but not otherwise. After completion of the study but before conducting the final statistical analysis, a post hoc comparison of survival between groups using a Cox proportional hazards model was added to the analysis plan. The proportionality assumption was tested using the Schoenfeld test. All models were summarized with hazard ratios assessing the risk of death following randomization.

Three sensitivity analyses were conducted to explore possible sources of bias that may inform interpretation of the main analysis. These prespecified analyses used a 2-sided P = .05 and included (1) a per-protocol analysis among participants who received at least 4 doses of the assigned study treatment without any major protocol deviations; (2) an analysis of participants for whom day 30 outcomes were directly observed; and (3) an analysis that excluded participants in the placebo group who received open-label steroids.

In an adjusted analysis, generalized linear mixed models were used to estimate the conditional effects of treatment, with site adjusted as a random-effect variable. Generalized estimating equations were used to estimate the marginal effect of study interventions. A proportional odds model was specified for VVFDs. Mortality was modeled assuming a logit link function. A full model was prespecified to include age, body mass index, Acute Physiology and Chronic Health Evaluation (APACHE) II score, race, sex, need for ventilation or vasopressors at baseline, whether sepsis was the primary reason for admission, whether a patient was admitted via the emergency department, and source of infection. Post hoc, body mass index was replaced with weight. No variable selection techniques were used. Multiple imputation based on predictive mean matching was used to overcome any missingness in covariates; missing VVFDs were estimated based on last value carried forward or, for the sensitivity analysis, observed outcomes only. Restricted cubic splines were used to address nonlinearities in the association between continuous variables and observed outcomes. To determine whether subgroup analyses were warranted, the interaction between treatment group assignment and baseline features of participants were evaluated. This was done by fitting the main covariate-adjusted model including all main effects, then individually testing for the interaction between each covariate and the treatment indicator using P ≤ .20. Interaction terms reaching this threshold were assumed to indicate possible effect modification and, thus, the need to explore for differential treatment effects within subgroups. Also, in a post hoc exploratory analysis, we used a graphical approach to further evaluate the possibility that time to treatment modifies the treatment effect.

Prespecified exploratory end points were compared by treatment group in a similar manner with the primary and secondary end points. All exploratory analyses used a 2-sided critical P = .05 without adjustment for multiplicity. All analyses, including generation of the randomization sequence, used R version 3.4.3 (R Foundation for Statistical Computing).

Early Trial Termination

The interim analyses planned for N = 500 proceeded as expected. After 501 participants were enrolled, additional funding for the trial was withheld due to a change in the funder’s priorities, and no further enrollments occurred, leading to administrative termination of the trial prior to meeting any prespecified stopping criterion. Aside from the knowledge that the trial had not yet stopped at an earlier interim analysis, the funder had no information regarding unblinded trial results at the time funding was terminated, which occurred on October 15, 2019. In a post hoc conditional analysis at the time of the N = 500 analysis, the predictive probability of reaching the protocol-specified criteria to declare efficacy on VVFD at N = 2000 was computed using Markov chain Monte Carlo methods. From the posterior distributions of the bayesian analysis model, we performed multiple imputation of the VVFD outcomes for all participants with unknown outcomes. This imputation resulted in 10 000 imputed data sets of size n = 2000, each of which was assessed for meeting the final analysis success threshold (Wilcoxon P < .022).

Results
Enrollment and Patient Characteristics

Of 3243 patients screened at 43 hospitals between August 2018 and July 2019, 501 were enrolled (Figure 1). The most common reasons for exclusion were need for organ support for a diagnosis other than sepsis (n = 735), home oxygen use (n = 514), patient refusal (n = 406), and limitations in care (n = 346). The median age of enrolled patients was 62 years, 30% were Black, and 46% were female (Table 1). Patients qualified for enrollment based on a need for both ventilatory and vasopressor support (41%), ventilatory support alone (21%), or vasopressor support alone (38%). The median baseline APACHE II and SOFA scores were 27.0 (IQR, 20.8-33.0) and 9.0 (IQR, 7.0-12.0), respectively. Following randomization, 252 were assigned to the intervention group and 249 to the control group. The median time between the onset of qualifying organ dysfunction and first dose of study drug was 14.7 hours (IQR, 7.9-20.9 hours). Thirty-three percent of patients in the intervention group and 32% of control patients received clinician-prescribed corticosteroids at a dose of at least 200 mg of hydrocortisone daily equivalent.

Baseline and clinical characteristics of enrolled patients are provided in Table 1; 67% vs 73% of patients admitted from the emergency department were randomized to the intervention group vs the control group, respectively. Severity of illness, comorbidities, source of infection, and type of infecting organism were similar between groups. Protocol adherence was high, with only 20 participants missing more than 1 dose of vitamin C or matching placebo, 6 missing more than 1 dose of thiamine or matching placebo, and 3 missing more than 1 dose of hydrocortisone or matching placebo.

Primary and Key Secondary Outcomes

Quiz Ref IDThere was no statistically significant difference between the intervention and control groups in VVFDs (25 days [IQR, 0-29 days] vs 26 days [IQR, 0-28 days], respectively). The median difference in VVFDs between the intervention and control groups was −1 day (95% CI, −4 to 2 days; P = .85). In both the intervention and control groups, the majority of patients with 0 VVFDs died (71% and 70%, respectively). Thirty-day all-cause mortality was 22% in the intervention group vs 24% in the control group. The adjusted analyses are given in Section 2.5 of Supplement 4. The proportional odds model for VVFDs did not reveal a statistical difference between the intervention and control groups (P = .16; model R2 = 0.32; model C statistic = 0.73). Requiring ventilatory support at enrollment was significantly associated with higher odds of 30-day mortality compared with not requiring ventilatory support at enrollment (odds ratio, 2.54; 95% CI, 1.33-5.38).

Long-term Outcome

At 180 days, mortality in the intervention and control groups was 40.5% vs 37.8% (difference, 2.7%; 95% CI, −11.3% to 5.8%) (Figure 2). The Cox proportional hazard model yielded an unadjusted hazard ratio for death in the intervention group of 1.08 at 180 days (95% CI, 0.82-1.43).

Differential Treatment Effects and Exploratory Outcomes

Prespecified exploratory end points in Table 2 showed no statistically significant difference between groups in bivariable comparisons of change in SOFA score to day 4, length of ICU stay, length of hospital stay, delirium-/coma-free days, or kidney replacement therapy–free days. Post hoc, weight replaced body mass index as the selected covariate and differential treatment effects by weight were explored. The VVFD model yielded an interaction between treatment and weight (P = .02) (Table 14 in Supplement 4). The nonlinear association between weight and outcomes for the treatment groups is shown in Section 2.5.5 in Supplement 4. In the per-protocol analysis, a similar interaction between treatment and weight was observed (P = .001) (Table 44 and Section 3.6.5 in Supplement 4). Treatment response was not significantly associated with any prespecified variable, including time to treatment (Table 14 in Supplement 4). The distribution of outcomes within quartile of time to treatment stratified by treatment group is shown in Section 4 of Supplement 4. Overall, 64 participants in the intervention group and 69 participants in the control group were treated within 8 hours, with median VVFDs of 28 (IQR, 24-29) and 27 (0-29), and there was no significant interaction between time to treatment and group assignment (P = .27).

Safety and Adverse Event Outcomes

There were no reported serious adverse events in the study. There were 2 adverse events (hemorrhagic shock and worsening kidney function) in the intervention group assessed as potentially related to study participation.

Sensitivity Analysis

Of the 501 enrolled patients, 83 were excluded because of protocol deviations, leaving 417 in the per-protocol analysis set. The median number of VVFDs for patients in the per-protocol analysis was 26 (IQR, 0-29) in the intervention group vs 26 (IQR, 0-28) in the control group, with a median group difference of 0 VVFDs (95% CI, −3.00 to 2.00) and an odds ratio of 1.09 (95% CI, 0.78-1.53) (Table 35 in Supplement 4). At 30 days, 18% of patients in the per-protocol intervention group and 22% in the control group had died. Adjusted analyses, both conditional and marginal, were similar to unadjusted analyses, although the odds of 30-day mortality among participants with urinary tract infections was significantly lower than in patients with other types of infection (odds ratio, 0.31; 95% CI, 0.12-0.81) (Table 17 in Supplement 4). No association between 180-day mortality and study group was observed (intervention, 40%; control, 37%; difference, 3%; 95% CI, −12% to 7%). The unadjusted Cox proportional hazard ratio for 180-day death was 1.07 (95% CI, 0.78-1.46). The main findings were unchanged when restricting analyses to the 458 patients for whom outcomes were fully observed (Section 2.6 in Supplement 4) and when restricting analyses to the 402 patients who did not receive open-label steroids (Section 2.6.14 in Supplement 4).

Conditional Power Analysis Due to Administrative Early Trial Termination

At the time of the N = 500 analysis, the predictive probability that the trial would meet the protocol-specified criteria to declare efficacy on VVFDs at N = 2000 was 30.7%. At that time, there was about a 0.3-day difference in mean VVFDs between groups.

Discussion

Quiz Ref IDIn this randomized clinical trial conducted with an adaptive design and terminated early for administrative reasons, treatment with vitamin C, thiamine, and hydrocortisone did not significantly increase VVFDs over 30 days compared with placebo. This held true for all adjusted and per-protocol analyses for the primary, key secondary, and exploratory outcomes. However, due to the early termination, the study may have been underpowered to detect a clinically important difference in VVFDs.

Patients enrolled in each group were well matched and protocol adherence was high. Importantly, observed mortality in the control group was in the range used to conduct power calculations for the study.

This study’s results are concordant with prior clinical trials of vitamin C in sepsis. In patients with septic shock, the ACTS trial found that treatment with vitamin C, thiamine, and hydrocortisone did not reduce SOFA score or mortality,20 while the VITAMINS trial found that the combination did not increase days alive and free of vasopressors.10 Of note, all patients in the control group in the VITAMINS trial were treated with corticosteroids, compared with 33% of the control group in this trial. Comparison between this trial and the CITRIS-ALI study is less straightforward.18 In CITRIS-ALI, a higher dosage of intravenous vitamin C (50 mg/kg every 6 hours) was given to patients with acute lung injury in the setting of sepsis. Although the CITRIS-ALI study, similar to VICTAS, did not show differences in organ dysfunction scores over time, the CITRIS-ALI trial did show that a decrease in mortality in patients treated with vitamin C remains plausible. It is important to note that mortality in CITRIS-ALI was 1 of 46 secondary outcome measures, and no adjustment was made for multiple comparisons in the analysis18 It is possible that differences in study populations, differences in ancillary therapies such as corticosteroids and thiamine, and the higher dose of vitamin C used in CITRIS-ALI contributed to this different outcome.11,21,22

Limitations

This trial has several limitations. First, it was administratively stopped due to a lack of funding and, thus, may have been underpowered to show clinically meaningful differences in VVFD. Second, a fixed dose of vitamin C was tested; a higher dose or dosing based on plasma vitamin C concentrations might yield different results. Third, clinicians were allowed to administer corticosteroids at their discretion. If corticosteroids have a beneficial effect on VVFD, this would bias the trial results toward the null, although a post hoc sensitivity analysis demonstrated similar findings in patients not receiving open-label corticosteroids. Fourth, enrollment was limited to patients with sepsis-induced cardiovascular or respiratory failure; inclusion of patients with different types of organ dysfunction may have yielded different results. Fifth, the median time to receipt of intervention was 14.7 hours (IQR, 7.9-20.9 hours), and while this timing was earlier than in the VITAMINS trial, it remains unknown if still earlier administration could improve outcomes. In this study, treatment response was not associated with time to treatment, and there was no interaction between time and group assignment.

Conclusions

Among critically ill patients with sepsis, treatment with vitamin C, thiamine, and hydrocortisone, compared with placebo, did not significantly increase ventilator- and vasopressor-free days within 30 days. However, the trial was terminated early for administrative reasons and may have been underpowered to detect a clinically important difference.

Section Editor: Christopher Seymour, MD, Associate Editor, JAMA (christopher.seymour@jamanetwork.org).
Back to top
Article Information

Corresponding Author: Jonathan E. Sevransky, MD, MHS, Emory University Hospital, Room D226, Atlanta, GA 30322 (jsevran@emory.edu).

Accepted for Publication: January 19, 2021.

Author Contributions: Dr Lindsell and Mr Nwosu had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Sevransky, Rothman, Hager, Bernard, Buchman, Busse, Coopersmith, DeWilde, Ely, Fowler, Gaieski, Gong, Hall, Hinson, Hooper, Kelen, Khan, Levine, Lewis, Marlin, McGlothlin, Rice, Rudolph, Sanfilippo, Viele, Martin, Wright.

Acquisition, analysis, or interpretation of data: Sevransky, Rothman, Hager, Bernard, Brown, Busse, Coopersmith, DeWilde, Ely, Eyzaguirre, Fowler, Gaieski, Gong, Hall, Hinson, Hooper, Kelen, Khan, Levine, Lewis, Lindsell, Marlin, McGlothlin, Moore, Nugent, Nwosu, Polito, Rice, Ricketts, Viele, Martin, Wright.

Drafting of the manuscript: Sevransky, Rothman, Hager, Bernard, Busse, DeWilde, Ely, Hall, Hinson, Khan, Rice, Rudolph, Viele, Wright.

Critical revision of the manuscript for important intellectual content: Sevransky, Rothman, Hager, Bernard, Brown, Buchman, Busse, Coopersmith, Ely, Eyzaguirre, Fowler, Gaieski, Gong, Hall, Hinson, Hooper, Kelen, Khan, Levine, Lewis, Lindsell, Marlin, McGlothlin, Moore, Nugent, Nwosu, Polito, Rice, Ricketts, Sanfilippo, Viele, Martin, Wright.

Statistical analysis: Ely, Lewis, Lindsell, McGlothlin, Nwosu, Viele.

Obtained funding: Sevransky, Rothman, Bernard, Buchman, Hall, Kelen, Sanfilippo, Wright.

Administrative, technical, or material support: Sevransky, Rothman, Hager, Bernard, Buchman, DeWilde, Ely, Eyzaguirre, Fowler, Gaieski, Hall, Hinson, Hooper, Kelen, Khan, Levine, Lewis, Marlin, Rice, Ricketts, Rudolph, Sanfilippo, Martin, Wright.

Supervision: Sevransky, Rothman, Hager, Bernard, Brown, DeWilde, Ely, Gaieski, Gong, Hall, Hinson, Hooper, Kelen, Khan, Marlin, Nugent, Sanfilippo, Martin, Wright.

Conflict of Interest Disclosures: Dr Sevransky reported receipt of a stipend from Critical Care Medicine for work as an associate editor and grants from the CDC Foundation. Dr Brown reported receipt of personal fees from Hamilton, Oxford University Press/Brigham Young University, and New York University and grants from Faron, Sedana, Janssen, the National Institutes of Health (NIH), and the Department of Defense. Dr Buchman reported that he is senior advisor and intergovernmental personnel assignment to BARDA DRIVe, detailed to the Solving Sepsis program, that he is editor in chief of Critical Care Medicine and Critical Care Explorations, and that he received grants and salary support as site principal investigator for the Surgical Critical Care Initiative managed by the Uniformed Services University of the Health Sciences. Dr Busse reported receipt of speakers bureau fees from La Jolla Pharmaceutical Company. Dr Ely reported receipt of grants from Bioxcel as a trial principal investigator and honoraria for continuing medical education teaching activities from Pfizer and Orion and serving as an unfunded investigator for Eli Lilly. Dr Fowler reported receipt of grants from the National Heart, Lung, and Blood Institute (NHLBI). Dr Gong reported receipt of grants from the NHLBI and the Agency for Healthcare Research and Quality (AHRQ). Dr Hall reported receipt of grants from the Centers for Disease Control and Prevention (CDC), Johnson and Johnson, the NIH, and Nico Corporation. Dr Hinson reported receipt of grants from the AHRQ and the CDC. Dr Khan reported receipt of grants from United Therapeutics, Janssen, Reata Pharma, and the NIH. Dr Lewis reported that he is the senior medical scientist at Berry Consultants LLC, a statistical consulting firm that specializes in adaptive clinical trial design. Dr Lindsell reported receipt of grants from the NIH, the CDC, and the Department of Defense; receipt of research contracts to his institution from Entegrion Inc, Endpoint Health Research, and bioMérieux; and holding stock options in Bioscape Digital; in addition, Dr Lindsell has a patent for risk stratification in pediatric septic shock issued to Cincinnati Children’s Hospital Medical Center. Dr Rice reported consulting for Cumberland Pharmaceuticals Inc, Avisa Pharma LLC, and Cytovale Inc. Dr Martin reported receipt of research advisory board fees from Beckman Coulter, Regeneron, and Grifols. No other disclosures were reported.

Funding/Support: Funding for the trial was provided by the Marcus Foundation to Emory University, which supervised the distribution of resources to clinical sites. Johns Hopkins served as the clinical coordinating center and Vanderbilt University as the data coordinating center. Nova Biomedical Inc provided support to 6 sites with point-of-care glucometers. Vitamin C, thiamine, and hydrocortisone were purchased from McGuff Pharmaceuticals Inc. SCCM Discovery, a critical care research infrastructure program, facilitated the design and conduct of the study, provided administrative support to conducting the study, and was involved in critical manuscript revision. SCCM Discovery reviewed and critiqued the study, which met criteria for SCCM Discovery endorsement.

Role of the Funder/Sponsor: SCCM Discovery was involved in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation and review of the manuscript. SCCM Discovery was not involved in the approval of the manuscript or the decision to submit the manuscript for publication. The other funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.

Group Information: For a list of the VICTAS Investigators, see Supplement 5.

Data Sharing Statement: See Supplement 6.

Additional Contributions: We thank Roger J. Lewis, MD, PhD, Harbor-UCLA Medical Center, Jason Connor, PhD, Confluence Stats, R. Phillip Dellinger, MD, Cooper Medical School of Rowan University, Steven Opal, MD, Brown University, and Tiffany M. Osborne, MD, MPH, Washington University, for participation in the trial data and safety monitoring board.

References
1.
Singer  M, Deutschman  CS, Seymour  CW,  et al.  The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3).   JAMA. 2016;315(8):801-810. doi:10.1001/jama.2016.0287PubMedGoogle ScholarCrossref
2.
Rhee  C, Dantes  R, Epstein  L,  et al; CDC Prevention Epicenter Program.  Incidence and trends of sepsis in US hospitals using clinical vs claims data, 2009-2014.   JAMA. 2017;318(13):1241-1249. doi:10.1001/jama.2017.13836PubMedGoogle ScholarCrossref
3.
Torio  CM, Moore  BJ. National Inpatient Hospital Costs: The Most Expensive Conditions by Payer, 2013. Agency for Healthcare Research and Quality; 2016. Healthcare Cost and Utilization Project Statistical Brief No. 204.
4.
Buchman  TG, Simpson  SQ, Sciarretta  KL,  et al.  Sepsis among Medicare beneficiaries, 1: the burdens of sepsis, 2012-2018.   Crit Care Med. 2020;48(3):276-288. doi:10.1097/CCM.0000000000004224PubMedGoogle ScholarCrossref
5.
Yende  S, Austin  S, Rhodes  A,  et al.  Long-term quality of life among survivors of severe sepsis: analyses of two international trials.   Crit Care Med. 2016;44(8):1461-1467. doi:10.1097/CCM.0000000000001658PubMedGoogle ScholarCrossref
6.
Rhodes  A, Evans  LE, Alhazzani  W,  et al.  Surviving Sepsis Campaign: international guidelines for management of sepsis and septic shock: 2016.   Crit Care Med. 2017;45(3):486-552. doi:10.1097/CCM.0000000000002255PubMedGoogle ScholarCrossref
7.
Marshall  JC.  Why have clinical trials in sepsis failed?   Trends Mol Med. 2014;20(4):195-203. doi:10.1016/j.molmed.2014.01.007PubMedGoogle ScholarCrossref
8.
Spoelstra-de Man  AME, Elbers  PWG, Oudemans-Van Straaten  HM.  Vitamin C: should we supplement?   Curr Opin Crit Care. 2018;24(4):248-255. doi:10.1097/MCC.0000000000000510PubMedGoogle ScholarCrossref
9.
Wilson  JX.  Evaluation of vitamin C for adjuvant sepsis therapy.   Antioxid Redox Signal. 2013;19(17):2129-2140. doi:10.1089/ars.2013.5401PubMedGoogle ScholarCrossref
10.
Fujii  T, Luethi  N, Young  PJ,  et al; VITAMINS Trial Investigators.  Effect of vitamin C, hydrocortisone, and thiamine vs hydrocortisone alone on time alive and free of vasopressor support among patients with septic shock: the VITAMINS randomized clinical trial.   JAMA. 2020;323(5):423-431. doi:10.1001/jama.2019.22176PubMedGoogle ScholarCrossref
11.
Fowler  AA  III, Truwit  JD, Hite  RD,  et al.  Effect of vitamin C infusion on organ failure and biomarkers of inflammation and vascular injury in patients with sepsis and severe acute respiratory failure: the CITRIS-ALI randomized clinical trial.   JAMA. 2019;322(13):1261-1270. doi:10.1001/jama.2019.11825PubMedGoogle ScholarCrossref
12.
Putzu  A, Daems  AM, Lopez-Delgado  JC, Giordano  VF, Landoni  G.  The effect of vitamin C on clinical outcome in critically ill patients: a systematic review with meta-analysis of randomized controlled trials.   Crit Care Med. 2019;47(6):774-783. doi:10.1097/CCM.0000000000003700PubMedGoogle ScholarCrossref
13.
Barabutis  N, Khangoora  V, Marik  PE, Catravas  JD.  Hydrocortisone and ascorbic acid synergistically prevent and repair lipopolysaccharide-induced pulmonary endothelial barrier dysfunction.   Chest. 2017;152(5):954-962. doi:10.1016/j.chest.2017.07.014PubMedGoogle ScholarCrossref
14.
Marik  PE, Khangoora  V, Rivera  R, Hooper  MH, Catravas  J.  Hydrocortisone, vitamin C, and thiamine for the treatment of severe sepsis and septic shock: a retrospective before-after study.   Chest. 2017;151(6):1229-1238. doi:10.1016/j.chest.2016.11.036PubMedGoogle ScholarCrossref
15.
Hager  DN, Hooper  MH, Bernard  GR,  et al.  The Vitamin C, Thiamine and Steroids in Sepsis (VICTAS) protocol: a prospective, multi-center, double-blind, adaptive sample size, randomized, placebo-controlled, clinical trial.   Trials. 2019;20(1):197. doi:10.1186/s13063-019-3254-2PubMedGoogle ScholarCrossref
16.
Lindsell  CJ, McGlothlin  A, Nwosu  S,  et al.  Update to the Vitamin C, Thiamine and Steroids in Sepsis (VICTAS) protocol: statistical analysis plan for a prospective, multicenter, double-blind, adaptive sample size, randomized, placebo-controlled, clinical trial.   Trials. 2019;20(1):670. doi:10.1186/s13063-019-3775-8PubMedGoogle ScholarCrossref
17.
Hager  DN, Martin  GS, Sevransky  JE, Hooper  MH.  Glucometry when using vitamin C in sepsis: a note of caution.   Chest. 2018;154(1):228-229. doi:10.1016/j.chest.2018.03.018PubMedGoogle ScholarCrossref
18.
Vincent  JL, Moreno  R, Takala  J,  et al; Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine.  The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure.   Intensive Care Med. 1996;22(7):707-710. doi:10.1007/BF01709751PubMedGoogle ScholarCrossref
19.
Ely  EW, Margolin  R, Francis  J,  et al.  Evaluation of delirium in critically ill patients: validation of the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU).   Crit Care Med. 2001;29(7):1370-1379. doi:10.1097/00003246-200107000-00012PubMedGoogle ScholarCrossref
20.
Moskowitz  A, Huang  DT, Hou  PC,  et al; ACTS Clinical Trial Investigators.  Effect of ascorbic acid, corticosteroids, and thiamine on organ injury in septic shock: the ACTS randomized clinical trial.   JAMA. 2020;324(7):642-650. doi:10.1001/jama.2020.11946PubMedGoogle ScholarCrossref
21.
de Grooth  HJ, Elbers  PWG, Vincent  JL.  Vitamin C for sepsis and acute respiratory failure [letter].   JAMA. 2020;323(8):792. doi:10.1001/jama.2019.21981PubMedGoogle ScholarCrossref
22.
Fowler  AA  III, Fisher  BJ, Kashiouris  MG.  Vitamin C for sepsis and acute respiratory failure [letter reply].   JAMA. 2020;323(8):792-793. doi:10.1001/jama.2019.21987PubMedGoogle ScholarCrossref
×