Flag of the European Union EU Clinical Trials Register Help

Clinical trials

The European Union Clinical Trials Register   allows you to search for protocol and results information on:
  • interventional clinical trials that were approved in the European Union (EU)/European Economic Area (EEA) under the Clinical Trials Directive 2001/20/EC
  • clinical trials conducted outside the EU/EEA that are linked to European paediatric-medicine development

  • EU/EEA interventional clinical trials approved under or transitioned to the Clinical Trial Regulation 536/2014 are publicly accessible through the
    Clinical Trials Information System (CTIS).


    The EU Clinical Trials Register currently displays   44334   clinical trials with a EudraCT protocol, of which   7366   are clinical trials conducted with subjects less than 18 years old.   The register also displays information on   18700   older paediatric trials (in scope of Article 45 of the Paediatric Regulation (EC) No 1901/2006).

    Phase 1 trials conducted solely on adults and that are not part of an agreed paediatric investigation plan (PIP) are not publicly available (see Frequently Asked Questions ).  
     
    Examples: Cancer AND drug name. Pneumonia AND sponsor name.
    How to search [pdf]
    Search Tips: Under advanced search you can use filters for Country, Age Group, Gender, Trial Phase, Trial Status, Date Range, Rare Diseases and Orphan Designation. For these items you should use the filters and not add them to your search terms in the text field.
    Advanced Search: Search tools
     

    < Back to search results

    Download PDF

    Clinical Trial Results:
    Early administration of Vitamin C in patients with sepsis or septic shock in emergency departments: a multicentre, double blinded, randomized controlled trial: the C-EASIE trial.

    Summary
    EudraCT number
    2020-001862-12
    Trial protocol
    BE  
    Global end of trial date
    15 Nov 2023

    Results information
    Results version number
    v1(current)
    This version publication date
    30 Nov 2024
    First version publication date
    30 Nov 2024
    Other versions
    Summary report(s)
    Article

    Trial information

    Close Top of page
    Trial identification
    Sponsor protocol code
    s63213
    Additional study identifiers
    ISRCTN number
    -
    US NCT number
    NCT04747795
    WHO universal trial number (UTN)
    -
    Other trial identifiers
    KCE Trials: INV19-1237, CTR Pilot: 389
    Sponsors
    Sponsor organisation name
    UZ Leuven
    Sponsor organisation address
    Herestraat 49, Leuven, Belgium, 3000
    Public contact
    Emergency Department Didier Desruelles C-EASIE Trial Team, UZ Leuven, +32 16343952, ceasie@uzleuven.be
    Scientific contact
    Stefanie Vandervelden ZAS Augustinus, Antwerp, UZ Leuven, +32 498221142, stefanie.vandervelden@zas.be
    Paediatric regulatory details
    Is trial part of an agreed paediatric investigation plan (PIP)
    No
    Does article 45 of REGULATION (EC) No 1901/2006 apply to this trial?
    No
    Does article 46 of REGULATION (EC) No 1901/2006 apply to this trial?
    No
    Results analysis stage
    Analysis stage
    Final
    Date of interim/final analysis
    26 Jan 2024
    Is this the analysis of the primary completion data?
    Yes
    Primary completion date
    15 Nov 2023
    Global end of trial reached?
    Yes
    Global end of trial date
    15 Nov 2023
    Was the trial ended prematurely?
    No
    General information about the trial
    Main objective of the trial
    The objective of the trial was to evaluate the efficacy of early administration of Vitamin C in addition to the standard of care in patients with sepsis or septic shock. This was done by determining the average post-baseline patient SOFA score (on day 1-5 after patient inclusion). The SOFA score is widely employed in daily monitoring of acute morbidity and is an acceptable surrogate marker of efficacy in exploratory trials of novel therapeutic agents in sepsis. We aimed to proof that through early admission of Vitamin C, patients who present at an early stage of the disease course will get less sick, resulting in a lower max SOFA score. However, administration of Vitamin C in our sickest population might not necessarily reflect in a lower max SOFA score, but rather in faster recovery (higher delta SOFA between 2 points in time). To capture both phenomena, we looked at the average post-baseline SOFA score as primary outcome supported by a maximum SOFA score as secondary outcome.
    Protection of trial subjects
    - Peer review was conducted by expert referees to the professional and scientific standards expected for clinical studies. The final protocol was reviewed by 2 independent experts and the principal investigators (PIs) of participating sites. - Exclusion criteria protected participation without (delayed) informed consent, antibiotics as single dose, prophylactic treatment or without accompaning body fluid culture according to the timeframe, 'do no intubate’ or ‘comfort measures only’ status, failure to randomize within 6 hours after ED presentation, weight < 45 kg, pregnant or breastfeeding, known allergy for vitamin C, oxalate nephropathy or hyperoxaluria, glucose-6-phosphate dehydrogenase deficiency, chronic iron overload due to iron storage and other diseases, being already on IV steroids for a reason other than septic shock, proven active COVID-19 infection and participation in an interventional trial with an investigational medicinal product or device. - Due to the illness of the patient, also the Legally Authorized Representative was able to give informed consent. - Participants could voluntarily discontinue trial treatment and/or prematurely end their participation in the trial for any reason at any time. - The Trial Steering Committee included 2 sepsis survivors. This way we combined experts in the field of sepsis, emergency medicine and research as well as patient representatives. - During the trial 2x an Annual Progress Report has been send to the EC/FAMH, the third year the Clinical Study Report. These reports contained all occurred AE's, trial (recruitment) issues and advices from the DSMB / TSC / Audits. - The Data and Safety Monitoring Board reviewed the results after a planned blinded interim analysis for sample size recalculation (after 203 patients had completed follow-up for the primary outcome). The power level was at least 80% with the observed values, resulting in an unaltered sample size. - Monitoring and audit were performed regulary
    Background therapy
    Looking at all previous vitamin C studies reporting a study drug administration time frame, the average time to treatment varied between 3.3 and 18 hours (or <24hrs) after ICU admission. The C-EASIE trail aimed a 1st study drug administration within 6 hours after ED presentation, making it the only trial recruiting patients in the ED and investigating a true early administration effect. The large variety in study set-ups, mixed outcomes, and the inconsistency in the reported results, have made it difficult to estimate the optimal dose and timing of vitamin C administration. As a result of increased metabolism due to enhanced inflammatory response, high doses of vitamin C, up to 3 to 6 gram daily, are needed to normalize the vitamin C plasma level. Based on published clinical data and the pharmacokinetics of vitamin C most studies decided to administer 6g of vitamin C per day divided in four equal doses i.e. every 6 hours. We adopted the same regimen in our protocol. Since this was a pragmatic trial, we did not measure the baseline levels of vitamin C in patients. We deemed this appropriate since a recent prospective pharmacokinetic study on the patients in treatment group of the VITAMINS trial showed that the 6-hourly doses regimen of 1.5g should sufficient to achieve and maintain normal to supranormal vitamin C plasma levels. Treatment with vitamin C was deemed effective when a difference of 1 in average post-baseline patient SOFA score (calculated over day 2-5) between both groups, based on a constrained longitudinal data analysis (cLDA) model with α set at 0.0525 with at least 80% power could be detected. The power was calculated using the approach presented by Stroup. Based on study level data aggregated in a systematic review, Delta fixed-day SOFA scores appeared to be most responsively and consistently associated with mortality.
    Evidence for comparator
    We opted for a high dose IV vitamin C only adjuvant therapy, in order to investigate the benefit of mono therapy. All other aspects of care, including the administration of glucocorticoids were performed at the discretion of the treating teams. During the feasibility visits, we inquired about the treatment regimens used regarding sepsis at the participating sites. It was ensured that these were in accordance with current practice guidelines and that they were similar between the different participating hospitals (e.g., corticosteroid dosing and timing). Next, to standardize as much as possible the management of sepsis patients, a sepsis protocol was made. This protocol includes suggested treatment options in regard to antimicrobial therapy, fluid resuscitation, hemodynamic support and RRT, according to international guidelines. Where applicable recommended therapy regimens were mentioned as they were standardized for the C-EASIE-trial protocol. Unless otherwise mentioned, local policy was not allowed to replace treatment strategy for participating patients. For treatment compliance the IMP administration was documented in the electronic patient file and on a Drug Accountability Log. Together with the counted left over ampoules from the treatment kits it was possible to create transparency about the (in)correct IMP administration to obtain the Per Protocol Set for the statistical analysis. Next to the Health Questionnaires all medical trial details were obtained from the electronic patient files. Next to the procalcitonin samples they were all part of the standard of care. During the data cleaning process questions about the recorded trial data could easily be checked with the corresponding medical file.
    Actual start date of recruitment
    04 Jun 2021
    Long term follow-up planned
    No
    Independent data monitoring committee (IDMC) involvement?
    Yes
    Population of trial subjects
    Number of subjects enrolled per country
    Country: Number of subjects enrolled
    Belgium: 292
    Worldwide total number of subjects
    292
    EEA total number of subjects
    292
    Number of subjects enrolled per age group
    In utero
    0
    Preterm newborn - gestational age < 37 wk
    0
    Newborns (0-27 days)
    0
    Infants and toddlers (28 days-23 months)
    0
    Children (2-11 years)
    0
    Adolescents (12-17 years)
    0
    Adults (18-64 years)
    110
    From 65 to 84 years
    167
    85 years and over
    15

    Subject disposition

    Close Top of page
    Recruitment
    Recruitment details
    Patients ≥18 yr admitted to the ED with suspected infection and NEWS ≥ 5 were eligible. (Suspected infection: combination of antibiotic administration + body fluid cultures <first 6 hours after ED presentation). Exclusion criteria: contraindications to vitamin C, or a ‘do do no intubate’ or ‘comfort measures only’ status.

    Pre-assignment
    Screening details
    0

    Period 1
    Period 1 title
    overall trial (overall period)
    Is this the baseline period?
    Yes
    Allocation method
    Randomised - controlled
    Blinding used
    Double blind
    Roles blinded
    Subject, Investigator, Monitor, Data analyst, Carer, Assessor
    Blinding implementation details
    The digital platform Randomize.net randomly assigned in 1:1 allocation ratio stratified by site with block size 4. A pre-made treatment kit with 52 blinded IV ampoules for the entire treatment period was allocated to the patient. The vitamin C and Normal Saline ampoules had identical sizes, were blinded by a cap and sticker and were both colourless and odourless. Bedside administration was identical (dilute 3 ampoules in 50ml of Normal Saline)

    Arms
    Are arms mutually exclusive
    Yes

    Arm title
    Vitamin C
    Arm description
    Patients in the intervention group received 1,5 g intravenous vitamin C (3 ampoules of vitamin C 500mg/5ml) diluted in 50ml of Normal Saline, administered over 15 minutes, every 6 hours for 4 days (4 times a day in total 16 administrations). A maximum of 8 hours was allowed between 2 doses and only one dose could be missed. Parameters were obtained as part of routine clinical care. On day 1 and 4 a targeted blood sample was collected for procalcitonin determination. Baseline data were obtained as close as possible to the time of randomization. The EQ-5D-5L health questionnaire was obtained on admission, day 5, day 28 and 3 months. ABG analysis was part of routine clinical care in ICU. If the patient was discharged to the ward before day 5, ABG sampling was not considered standard of care. To be able to calculate the SOFA score in this population, we used the SpO2/FiO2 ratio to impute for PaO2/FiO2 ratio in the respiratory component.
    Arm type
    Experimental

    Investigational medicinal product name
    Vitamin C
    Investigational medicinal product code
    Other name
    Pharmaceutical forms
    Solution for injection/infusion
    Routes of administration
    Infusion
    Dosage and administration details
    1,5 g intravenous vitamin C (3 ampoules of vitamin C 500mg/5ml) diluted in 50ml of Normal Saline, administered over 15 minutes, every 6 hours for 4 days (4 times a day in total 16 administrations).

    Arm title
    Placebo
    Arm description
    Patients in the placebo group received a matching placebo with Normal Saline. The administration was identical to the vitamin C treatment: 3 ampoules of Noral Saline 5ml diluted in 50ml of Normal Saline, administered over 15 minutes, every 6 hours for 4 days (4 times a day in total 16 administrations). A maximum of 8 hours was allowed between 2 doses and only one dose could be missed. The obtained parameters, blood samples and health questionnaires and the routine clinical care were identical to the Vitamin C arm. Ardena Gent NV, a Drug Product Development & Manufacturing firm blinded, packaged, relabelled the IMP kits, and provided randomization: 1st batch: number 21C22, expiry date September 2022, 2nd batch: number 22I01, expiry date October 2023.
    Arm type
    Placebo

    Investigational medicinal product name
    Placebo
    Investigational medicinal product code
    Other name
    Pharmaceutical forms
    Solution for injection/infusion
    Routes of administration
    Infusion
    Dosage and administration details
    3 ampoules of 5ml Normal Saline diluted in 50ml of Normal Saline, administered over 15 minutes, every 6 hours for 4 days (4 times a day in total 16 administrations).

    Number of subjects in period 1
    Vitamin C Placebo
    Started
    147
    145
    Completed
    147
    145

    Baseline characteristics

    Close Top of page
    Baseline characteristics reporting groups
    Reporting group title
    overall trial
    Reporting group description
    -

    Reporting group values
    overall trial Total
    Number of subjects
    292 292
    Age categorical
    Units: Subjects
        Adults ≥18 year
    292 292
    Gender categorical
    Units: Subjects
        Female
    103 103
        Male
    189 189
    FAS
    In accordance with the intent-to-treat (ITT) principle, the full analysis set (FAS) contained all randomized patients according to their randomized treatment. However, some randomized patients who did not receive any medication were excluded from the FAS, as decided during the Blind Review Meeting.
    Units: Subjects
        Vitamin C
    147 147
        Placebo
    145 145
    Subject analysis sets

    Subject analysis set title
    Per Protocol Set
    Subject analysis set type
    Per protocol
    Subject analysis set description
    Patients from the FAS with substantial protocol deviations (7 hours or more between hospital admission and first administration of study medication, 2 or more successive doses missed, all planned doses missing on the first day) were excluded from the per protocol set (PPS). Both FAS and PPS were used for the evaluation of all efficacy endpoints. Only the FAS was used for the evaluation the safety endpoints.

    Subject analysis set title
    Subgroup Analysis FAS SOFA <6
    Subject analysis set type
    Sub-group analysis
    Subject analysis set description
    For the baseline SOFA score, subgroups were predefined (SOFA <6 vs >=6). Subgroup analyses were only performed for the FAS.

    Subject analysis set title
    Subgroup Analysis FAS SOFA ≥6
    Subject analysis set type
    Sub-group analysis
    Subject analysis set description
    For the baseline SOFA score, subgroups were predefined (SOFA <6 vs >=6). Subgroup analyses were only performed for the FAS.

    Subject analysis sets values
    Per Protocol Set Subgroup Analysis FAS SOFA <6 Subgroup Analysis FAS SOFA ≥6
    Number of subjects
    268
    188
    104
    Age categorical
    Units: Subjects
        Adults ≥18 year
    268
    188
    104
    Age continuous
    Units:
        
    ( )
    ( )
    ( )
    Gender categorical
    Units: Subjects
        Female
    94
        Male
    174
    FAS
    In accordance with the intent-to-treat (ITT) principle, the full analysis set (FAS) contained all randomized patients according to their randomized treatment. However, some randomized patients who did not receive any medication were excluded from the FAS, as decided during the Blind Review Meeting.
    Units: Subjects
        Vitamin C
    139
    96
    51
        Placebo
    129
    92
    53

    End points

    Close Top of page
    End points reporting groups
    Reporting group title
    Vitamin C
    Reporting group description
    Patients in the intervention group received 1,5 g intravenous vitamin C (3 ampoules of vitamin C 500mg/5ml) diluted in 50ml of Normal Saline, administered over 15 minutes, every 6 hours for 4 days (4 times a day in total 16 administrations). A maximum of 8 hours was allowed between 2 doses and only one dose could be missed. Parameters were obtained as part of routine clinical care. On day 1 and 4 a targeted blood sample was collected for procalcitonin determination. Baseline data were obtained as close as possible to the time of randomization. The EQ-5D-5L health questionnaire was obtained on admission, day 5, day 28 and 3 months. ABG analysis was part of routine clinical care in ICU. If the patient was discharged to the ward before day 5, ABG sampling was not considered standard of care. To be able to calculate the SOFA score in this population, we used the SpO2/FiO2 ratio to impute for PaO2/FiO2 ratio in the respiratory component.

    Reporting group title
    Placebo
    Reporting group description
    Patients in the placebo group received a matching placebo with Normal Saline. The administration was identical to the vitamin C treatment: 3 ampoules of Noral Saline 5ml diluted in 50ml of Normal Saline, administered over 15 minutes, every 6 hours for 4 days (4 times a day in total 16 administrations). A maximum of 8 hours was allowed between 2 doses and only one dose could be missed. The obtained parameters, blood samples and health questionnaires and the routine clinical care were identical to the Vitamin C arm. Ardena Gent NV, a Drug Product Development & Manufacturing firm blinded, packaged, relabelled the IMP kits, and provided randomization: 1st batch: number 21C22, expiry date September 2022, 2nd batch: number 22I01, expiry date October 2023.

    Subject analysis set title
    Per Protocol Set
    Subject analysis set type
    Per protocol
    Subject analysis set description
    Patients from the FAS with substantial protocol deviations (7 hours or more between hospital admission and first administration of study medication, 2 or more successive doses missed, all planned doses missing on the first day) were excluded from the per protocol set (PPS). Both FAS and PPS were used for the evaluation of all efficacy endpoints. Only the FAS was used for the evaluation the safety endpoints.

    Subject analysis set title
    Subgroup Analysis FAS SOFA <6
    Subject analysis set type
    Sub-group analysis
    Subject analysis set description
    For the baseline SOFA score, subgroups were predefined (SOFA <6 vs >=6). Subgroup analyses were only performed for the FAS.

    Subject analysis set title
    Subgroup Analysis FAS SOFA ≥6
    Subject analysis set type
    Sub-group analysis
    Subject analysis set description
    For the baseline SOFA score, subgroups were predefined (SOFA <6 vs >=6). Subgroup analyses were only performed for the FAS.

    Primary: Primary: Average post-baseline SOFA score

    Close Top of page
    End point title
    Primary: Average post-baseline SOFA score
    End point description
    The primary outcome was the average post-baseline SOFA score (on day 2-5 after patient inclusion). Following analyses were performed: Primary Outcome: • Calculated SOFA scores for day 1-5 FAS and PPS • Primary outcome: Average post-baseline SOFA score FAS and PPS Subgroup analysis: • Subgroup analysis SOFA <6 FAS • Subgroup analysis SOFA ≥6 FAS Interaction models: • Interaction model baseline SOFA <6 and ≥6 FAS • Interaction model continuous baseline SOFA FAS and PPS • Interaction model baseline NEWS FAS and PPS Maximal SOFA: descriptives FAS and PPS and their comparison. We used a constrained' longitudinal data analysis (cLDA) model, which constrains means of baseline to be equal between arms. The average SOFA scores (d2-d5) was compared, after correction for baseline SOFA (=primary outcome).
    End point type
    Primary
    End point timeframe
    SOFA scores for day 1-5
    End point values
    Vitamin C Placebo Per Protocol Set Subgroup Analysis FAS SOFA <6 Subgroup Analysis FAS SOFA ≥6
    Number of subjects analysed
    147
    145
    139
    96
    51
    Units: SOFA score
        geometric mean (confidence interval 95%)
    1.982 (1.688 to 2.316)
    2.194 (1.870 to 2.563)
    1.989 (1.685 to 2.336)
    1.176 (0.949 to 1.431)
    4.442 (3.617 to 5.444)
    Attachments
    Screening Failures and Post-Random. Exclusions
    Subject Disposition and Follow-Up FAS
    Subject Disposition and Follow-Up PPS
    Subject Disposition Numbers for Patients with FU
    Subject Flow Number of Days with Follow-Up
    Missed Doses
    Definition of Septic Shock
    Baseline Characteristics FAS
    Baseline Characteristics PPS
    Reasons for Exclusions
    Primary Outcome SOFA Score FAS
    Primary Outcome SOFA Score FAS
    Primary Outcome SOFA Score FAS
    Primary Outcome SOFA Score PPS
    Primary Outcome SOFA Score PPS
    Primary Outcome SOFA Score PPS
    SOFA Score Subgroup Analysis FAS SOFA under 6
    SOFA Score Subgroup Analysis FAS SOFA under 6
    SOFA Score Subgroup Analysis FAS SOFA under 6
    SOFA Score Subgroup Analysis FAS SOFA 6 or above
    SOFA Score Subgroup Analysis FAS SOFA 6 or above
    SOFA Score Subgroup Analysis FAS SOFA 6 or above
    SOFA Score Interaction Model for Baseline SOFA FAS
    Interaction Model for Continuous Baseline SOFA FAS
    Interaction Model for Continuous Baseline SOFA FAS
    Interaction Model for Continuous Baseline SOFA PPS
    Interaction Model for Continuous Baseline SOFA PPS
    Interaction Model for Continuous Baseline NEWS FAS
    Interaction Model for baseline NEWS (FAS)
    Interaction Model for Continuous Baseline NEWS PPS
    Interaction Model for Baseline NEWS (PPS)
    Descriptives Maximal SOFA Score
    Statistical analysis title
    Calculated SOFA score for day 1-5
    Statistical analysis description
    Observed SOFA scores for day 1-5 for the FAS and PPS, also displayed in boxplots. A maximal score of 24 was imputed for deceased patients. See charts 'Primary Endpoint'.
    Comparison groups
    Vitamin C v Placebo
    Number of subjects included in analysis
    292
    Analysis specification
    Pre-specified
    Analysis type
    superiority [1]
    P-value
    = 0.272 [2]
    Method
    t-test, 2-sided
    Confidence interval
    Notes
    [1] - From the cLDA model used for the primary endpoint, mean values at d2, d3, d4 and d4 will be compared with two-sided tests and 95% confidence intervals will be reported.
    [2] - day 2 p=0.272 day 3 p= 0.7431 day 4 p= 0.3215 day 5 p= 0.2258
    Statistical analysis title
    Average post-baseline SOFA score (day 2-5)
    Statistical analysis description
    The average post-baseline score compared between Vitamin C and Placeb with a two-sided test derived from cLDA model using a likelihood approach with α=0.05, Kenward-Roger degrees of freedom and the covariance structure based on the Aikake criterion. Site was added as a fixed factor. The cLDA analysis was valid under the MAR and assumes that the probability that an observation is missing depends only on observed values of the individual, but not on the missing ones.
    Comparison groups
    Vitamin C v Placebo
    Number of subjects included in analysis
    292
    Analysis specification
    Pre-specified
    Analysis type
    superiority [3]
    P-value
    = 0.2955
    Method
    cLDA
    Confidence interval
    Notes
    [3] - An inverse hyperbolic sign transformation was applied to handle the (right-)skewed distribution of the SOFA scores. Results were presented after back-transformation to the original scale. As such, the estimate for the treatment effect referred to a ratio instead of to a difference. Results on untransformed SOFA scores were reported as sensitivity analyses. A maximal score of 24 was imputed for deceased patients. See charts 'Primary Endpoint'.
    Statistical analysis title
    Interaction model for baseline SOFA <6 and ≥6 FAS
    Statistical analysis description
    The interaction between the subgroups and treatment effect were tested to assess whether the treatment effect differs according to subgroup: the SOFA <6 and SOFA ≥6 data compared for the Vitamin C and Placebo group with interaction. See charts 'Primary Endpoint'.
    Comparison groups
    Placebo v Vitamin C
    Number of subjects included in analysis
    292
    Analysis specification
    Pre-specified
    Analysis type
    superiority [4]
    P-value
    = 0.1513 [5]
    Method
    interaction model
    Confidence interval
    Notes
    [4] - Estimates of SOFA (95%CI) obtained from the cLDA model fitted on inverse hyperbolic sign transformed data. Means and 95% confidence intervals are obtained after backtransformation to the original scale. Due to the transformation, the comparison of both groups refers to a ratio. Interactions are allowed with the categorised baseline SOFA score (<6 versus >=6). Focus of interest is the comparison of the intervention effects (at each day separately and the average) between subjects with SOFA<6 and
    [5] - baseline SOFA score <6: 0.8010 baseline SOFA score >=6: 0.0419 P-interaction 0.1513
    Statistical analysis title
    Interaction model for continuous baseline SOFA
    Statistical analysis description
    It was also verified if the treatment effect depends on the baseline variables baseline SOFA, both FAS and PPS. Comparisons of Vitamin C with placebo obtained from a LDA model fitted on inverse hyperbolic sign transformed data, allowing the effect of Vitamin C to depend on baseline SOFA score (an interaction between the effect of Vitamin C and baseline SOFA is allowed). Due to the transformation of the SOFA score (as outcome variable), the comarison of both groups refers to a ratio (and 95%Cl).
    Comparison groups
    Vitamin C v Placebo
    Number of subjects included in analysis
    292
    Analysis specification
    Pre-specified
    Analysis type
    other [6]
    P-value
    = 0.3346 [7]
    Method
    Interaction model
    Confidence interval
    Notes
    [6] - The effect to the effect averaged over d2-5. Results from 2 versions of this interaction model are presented: 1: assuming linearity for the effect of the baseline SOFA score and for the interaction effect 2: allowing non-linearity in the interaction between the baseline SOFA score and for the interaction effect To this purpose, restricted cubic splines with 4 knots were used. See charts 'Primary Endpoint'.
    [7] - 1: P-value for interaction: 0.3346 for FAS and 0.3487 for PPS 2: P-value for interaction: 0.2846 for FAS and 0.3661 for PPS The model fitted on the values from d2-d5, hence not including day 1 as outcome variable (model is not a cLDA, but a LDA)
    Statistical analysis title
    Interaction model for continuous baseline NEWS
    Statistical analysis description
    It was also verified if the treatment effect depends on the baseline variable baseline NEWS score, both for FAS and PPS. Comparisons of Vitamin C with placebo obtained from the cLDA model fitted on inverse hyperbolic sign transformed data, allowing the effect of Vitamin C to depend on baseline NEWS score (i.e. an interaction between the effect of vitamin C and baseline NEWS is allowed). Due to the transformation of the SOFA score, the comparison of both groups refers to a ratio (and 95%Cl).
    Comparison groups
    Vitamin C v Placebo
    Number of subjects included in analysis
    292
    Analysis specification
    Pre-specified
    Analysis type
    other [8]
    P-value
    = 0.3852 [9]
    Method
    interaction model
    Confidence interval
    Notes
    [8] - The effect refers to the effect averaged over d2-5. Results from 2 versions of this interaction model: 1: assuming linearity for the effect of the baseline NEWS score and for the interaction effect 2: allowing non-linearity in the effect of the baseline NEWS score and the interaction with vitamin C. To this purpose, restricted cubix splines with 4 knts were used. See charts 'Primary Endpoint'.
    [9] - 1: P-value for interaction for the FAS: 0.3852, for the PPS: 0.3343 2: P-value for interaction for the FAS: 0.3377, for the PPS: 0.3551
    Statistical analysis title
    Maximal SOFA
    Statistical analysis description
    Descriptives for the maximal SOFA score and estimates of mean maximal SOFA for the FAS and PPS and their comparison. Note that by definition the maximal value over all 20 imputations is at least as high as the observed value. See charts 'Primary Endpoint'.
    Comparison groups
    Vitamin C v Placebo
    Number of subjects included in analysis
    292
    Analysis specification
    Pre-specified
    Analysis type
    superiority [10]
    P-value
    = 0.4421
    Method
    t-test, 2-sided
    Confidence interval
    Notes
    [10] - The maximal SOFA score will be compared using a two-sided unpaired t-test after inverse hyperbolic sign transformation. Estimates of the mean maximal SOFA (95%Cl), obtained from a linear model correcting for study site on transformed (inverse hyperbolic sign transformation) as well as on untransformed data. in the transformed case, the means and 95%Cl) were obtained after backtransformation and the comparison between both groups refers to a ratio.
    Statistical analysis title
    Subgroup Analysis SOFA <6
    Statistical analysis description
    Subgroup Analysis SOFA <6 FAS and PPS, observed information and estimates of SOFA. Estimates of SOFA (95%Cl) obtained from the cLDA model fitted on inverse hyperbolic sign transformed data. Means and 95% confidence intervals are obtained after backtransformation to the original scale. Due to the transformation, the comparison of both groups refers to a ratio. See charts 'Primary Endpoint'.
    Comparison groups
    Placebo v Vitamin C
    Number of subjects included in analysis
    292
    Analysis specification
    Pre-specified
    Analysis type
    superiority [11]
    P-value
    = 0.801
    Method
    cLDA
    Confidence interval
    Notes
    [11] - Estimates of SOFA (95%CI) obtained from the cLDA model fitted on inverse hyperbolic sign transformed data. Means and 95% confidence intervals are obtained after backtransformation to the original scale. Due to the transformation, the comparison of both groups refers to a ratio.
    Statistical analysis title
    Subgroup Analysis SOFA ≥6
    Statistical analysis description
    Subgroup Analysis SOFA ≥6, observed information SOFA score and estimates. Estimates of SOFA (95%Cl) obtained from the cLDA model fitted on inverse hyperbolic sign transformed data. Means and 95% confidence intervals are obtained after backtransformation to the original scale. Due to the transformation, the comparison of both groups refers to a ratio. See charts 'Primary Endpoint'.
    Comparison groups
    Vitamin C v Placebo
    Number of subjects included in analysis
    292
    Analysis specification
    Pre-specified
    Analysis type
    superiority [12]
    P-value
    = 0.0419
    Method
    cLDA
    Confidence interval
    Notes
    [12] - Estimates of SOFA (95%CI) obtained from the cLDA model fitted on inverse hyperbolic sign transformed data. Means and 95% confidence intervals are obtained after backtransformation to the original scale. Due to the transformation, the comparison of both groups refers to a ratio.

    Secondary: Secondary: 28d Mortality

    Close Top of page
    End point title
    Secondary: 28d Mortality
    End point description
    Analysed for both FAS and PPS
    End point type
    Secondary
    End point timeframe
    Death at day 28 after inclusion
    End point values
    Vitamin C Placebo Per Protocol Set
    Number of subjects analysed
    147
    145
    139
    Units: deaths
    17
    11
    17
    Attachments
    28d Mortality Descriptives FAS
    28d Mortality FAS
    28d Mortality Descriptives PPS
    28d Mortality PPS
    Statistical analysis title
    Secondary: 28d Mortality
    Statistical analysis description
    28-day mortality was compared using a stratified χ² test (study site as stratum). 95% confidence intervals were reported for the percentage mortality in each treatment group and for the odds ratio. However, when there were subjects lost to follow-up before 28 days, 28-day estimates were derived from the Kaplan-Meier curve and compared using a Z-test (on the complementary log-log scale).
    Comparison groups
    Vitamin C v Placebo
    Number of subjects included in analysis
    292
    Analysis specification
    Pre-specified
    Analysis type
    superiority [13]
    P-value
    = 0.3516
    Method
    Z-test after log transformation
    Confidence interval
    Notes
    [13] - Event rates were estimated using Kaplan-Meier methodology. Confidence intervals were calculated using a log(-log)-transformation. The comparison at 28 days (secondary outcome) is based on a Z-test after log-log transformation of the survival estimates. The difference (95% Cl) in survival at 28 days equals 4.0% (-3.0%; 11.0%) for FAS and 4.5% (-2.9%; 11.9%) for PPS. An effect size has been added in the table, i.e. the difference in percentage survival at 28 days with a 95% confidence interval.

    Secondary: Secondary: Length of Hospital Stay

    Close Top of page
    End point title
    Secondary: Length of Hospital Stay
    End point description
    Death during hospital stay was treated as a competing risk
    End point type
    Secondary
    End point timeframe
    Between inclusion and discharge
    End point values
    Vitamin C Placebo Per Protocol Set
    Number of subjects analysed
    147
    145
    139
    Units: death during hospital stay
    18
    14
    18
    Attachments
    Length of Hospital Stay Descriptives FAS
    Length of Hospital Stay FAS
    Length of Hospital Stay Descriptives PPS
    Length of Hospital Stay PPS
    Statistical analysis title
    Secondary: Length of Hospital Stay
    Statistical analysis description
    To evaluate differences in length of hospital stay, death during hospital stay was treated as a competing risk, using a stratified Gray’s test for the comparison of both groups. Patients not yet discharged were censored. The same approach is used for the secondary outcome Length of ICU Stay.
    Comparison groups
    Vitamin C v Placebo
    Number of subjects included in analysis
    292
    Analysis specification
    Pre-specified
    Analysis type
    superiority [14]
    P-value
    = 0.7063
    Method
    Stratified Grays test
    Confidence interval
    Notes
    [14] - Event rates were estimated as cumulative incidence funtions (CIF), taking into account the competing ristk of in-hospital death (competing risk). Confidence intervals were calculated using a log(-log)-transformation. The hazard ratio (95%Cl) from a Fine&Gray model equals 0.971 (0.765, 1.233) for the FAS and 0.955 (0.743, 1.228) for the PPS. An effect size has been added in the legend of the table, i.e. hazard ratio from a Fine&Gray model with 95% confidence interval.

    Secondary: Secondary: Length of ICU Stay

    Close Top of page
    End point title
    Secondary: Length of ICU Stay
    End point description
    Analysed for both FAS and PPS.
    End point type
    Secondary
    End point timeframe
    Between admission to the ICU and discharge from ICU.
    End point values
    Vitamin C Placebo Per Protocol Set
    Number of subjects analysed
    147
    145
    139
    Units: days at the ICU
        median (inter-quartile range (Q1-Q3))
    5 (3 to 9)
    4 (3 to 7)
    4 (3 to 9)
    Attachments
    Length of ICU Stay FAS
    Length of ICU Stay FAS
    Length of ICU Stay PPS
    Length of ICU Stay PPS
    Statistical analysis title
    Secondary: Length of ICU Stay
    Statistical analysis description
    To evaluate differences in length of ICU Stay, death during ICU stay was treated as a competing risk, using a stratified Gray’s test for the comparison of both groups. Patients not yet discharged were censored. The analysis was restricted to patients admitted in ICU. Note however that the time starts at the moment of ICU admission and not at the moment of randomisation (this to handle the possibility that patients need to go the ICU at a later moment during the follow-up).
    Comparison groups
    Vitamin C v Placebo
    Number of subjects included in analysis
    292
    Analysis specification
    Pre-specified
    Analysis type
    superiority [15]
    P-value
    = 0.6336
    Method
    Grays test
    Confidence interval
    Notes
    [15] - Event rates at specific timepoints were estimated as cumulative incidence functions (CIF), taking into account the competing risk of in-ICU death (competing risk). Confidence intervals were calculated using a log(-log)-transformation. 3 patients were randomised the day after admission to the ICU. For these patients the ICU admission day is set at day 1. The hazard ratio (95%Cl) from a Fine%Gray model equals 1.041 (0.739, 1.467) for the FAS and 1.045 (0.732, 1.491) for the PPS.

    Secondary: Secondary: Mechanical Ventilation

    Close Top of page
    End point title
    Secondary: Mechanical Ventilation
    End point description
    Mechanical Ventilation needed on day 1-5. Is put on zero when patient is discharged. When a subject is deceased, the value is considered missing from that day on (in case MV information was available on the day of death, the information is considered missing from the day after death on).
    End point type
    Secondary
    End point timeframe
    Day 1-5
    End point values
    Vitamin C Placebo Per Protocol Set
    Number of subjects analysed
    147
    145
    139
    Units: days MV needed
        log mean (confidence interval 95%)
    0.061 (0.037 to 0.100)
    0.070 (0.043 to 0.114)
    0.058 (0.033 to 0.098)
    Attachments
    Mechanical Ventilation FAS
    Mechanical Ventilation PPS
    Statistical analysis title
    Secondary: Mechanical Ventilation
    Statistical analysis description
    The probability of MV was compared using a generalised linear mixed model (i.e. a logistic regression model with a random effect of patient). Odds ratio was reported for the comparison of the ‘mean probability’ over d2-d5, and for day-specific odcomparisons. Note that the model was fit on daily information from d1-d5 and the mean duration of MV (i.e. in the period d2-d5) could derived directly from this model as the sum of the daily probabilities.
    Comparison groups
    Vitamin C v Placebo
    Number of subjects included in analysis
    292
    Analysis specification
    Pre-specified
    Analysis type
    superiority [16]
    P-value
    = 0.6958
    Method
    Regression, Logistic
    Confidence interval
    Notes
    [16] - Observed information on mechanical ventilation and predicted probability. Predicted probability (and 95%Cl) of mechanical ventilation is obtained from a generalised linear mixed model (logistic regression with a random effect of patient) with group and time as fixed effects and correcting for study site (as fixed effect). Note that the estimates refer to the predicted probability and the predicted effect for the average patient.

    Secondary: Secondary: Vasopressor Requirement

    Close Top of page
    End point title
    Secondary: Vasopressor Requirement
    End point description
    Day 1-5 Vasopressor Requirement: Observed Information on vaso-active agents and Comparison of the Probability: the Dose of vasoactive agent, the Probability of receiving vaso-active agents, the Probability of receiving at least Dopamine >5, Epinephrine = 0.1 or Norepinephrine = 0.1 and the Probability of receiving Dopamine >15, Epinephrine >0.1 or Norepinephrine >0.1, for both the FAS and PPS. Note that the option Dopamine ≤5 or Dobutamine (any dose) was never chosen. Vasopressor requirement was an ordinal scale (4 levels starting with ‘no’) and results were reported from separate models using the three possible cut points (for example ‘no’ versus ‘yes’).
    End point type
    Secondary
    End point timeframe
    day 1-5. VR put on zero when patient is discharged. When a subject is deceased, the value is considered missing from that day on (in case information was available on the day of death, the information is considered missing from the day after death on).
    End point values
    Vitamin C Placebo Per Protocol Set
    Number of subjects analysed
    147
    145
    139
    Units: days VR needed
        log mean (confidence interval 95%)
    0.060 (0.036 to 0.097)
    0.047 (0.027 to 0.080)
    0.058 (0.034 to 0.097)
    Attachments
    Vasopressor Requirement FAS
    Vasopressor Requirement PPS
    Statistical analysis title
    Secondary: Vasopressor Requirement
    Statistical analysis description
    The vasopressor requirement was compared using a generalised linear mixed model (i.e. a logistic regression model with a random effect of patient). An odds ratio was reported comparing the use over d2-d5, as well as the timepoint specific use. Study site was added in the model. Note that the model was fit on information from d1-d5. Note that the mean duration of vasopressor requirement (i.e. in the period d2-d5) could be derived directly from this model as the sum of the daily probabilities.
    Comparison groups
    Vitamin C v Placebo
    Number of subjects included in analysis
    292
    Analysis specification
    Pre-specified
    Analysis type
    superiority [17]
    P-value
    = 0.5303
    Method
    Regression, Logistic
    Confidence interval
    Notes
    [17] - Comparison: Note that multiplying the average predicted probability over d2-d5 yields the mean number of days with vasopressor requirement. Estimates (95%Cl) obtained from a generalised lineair mixed model (logistic regression with a random effect of patient) with group and time as fixed effects and correcting for study site (as fixed effect). Note that the estimates refer to the predicted probability and the predicted effect for the average patient.

    Secondary: Secondary: Renal Replacement Therapy

    Close Top of page
    End point title
    Secondary: Renal Replacement Therapy
    End point description
    The probability of RRT was compared using a generalised linear mixed model (i.e. a logistic regression model with a random effect of patient). An odds ratio was reported comparing the ‘mean probability’ over d2-d5. Given the expected low number of events, for the timepoint specific RRT only descriptives were reported. RRT was put on zero when the patient was discharged. When a subject was deceased, the value was considered missing from that day on (in case RRT information was available on the day of death, the information was considered missing from the day after death on). Note that the model was fit on daily information from d1-d5 and that the mean duration of RRT (i.e. in the period d2-d5) could be derived directly from this model as the sum of the daily probabilities.
    End point type
    Secondary
    End point timeframe
    Number of days with RRT need during day 1-5
    End point values
    Vitamin C Placebo Per Protocol Set
    Number of subjects analysed
    147
    145
    139
    Units: days of RRT need
        log mean (confidence interval 95%)
    0.012 (0.006 to 0.027)
    0.022 (0.011 to 0.042)
    0.007 (0.002 to 0.020)
    Attachments
    Renal Replacement Therapy FAS
    Renal Replacement Therapy PPS
    Statistical analysis title
    Secondary: Renal Replacement Therapy
    Statistical analysis description
    Observed Information and Predicted Probability. Predicted probability (and 95%Cl) of renal replacement therapy obtained from a generalised linear mixed model (logistic regression with a random effect of patient) with group and time as fixed effects. Given the low number of events no correction has been made for study site and daily-specific comparisons where not tested. Note that the estimates refer to the predicted probability and the predicted efect for the average patient.
    Comparison groups
    Vitamin C v Placebo
    Number of subjects included in analysis
    292
    Analysis specification
    Pre-specified
    Analysis type
    superiority [18]
    P-value
    = 0.258
    Method
    Regression, Logistic
    Confidence interval
    Notes
    [18] - Renal replacement therapy is put on zero when patient is discharged. When subject is deceased, the value is missing. The probability of renal replacement therapy (RRT) will be compared using a generalized linear mixed model (i.e. a logistic regression model with a random effect of patient). An odds ratio will be reported comparing the ‘mean probability’ over d2-d5

    Secondary: Secondary: Need of Steroids

    Close Top of page
    End point title
    Secondary: Need of Steroids
    End point description
    The daily dose of steroids given was compared between both groups using a multivariate normal model for longitudinal measurements, with the Aikake criterion to select the covariance matrix. The mean dosage over d2-d5 was compared, as well as the timepoint specific dosages. Study site was added in the model. Note that the model fitted on information from all patients in the period d1-d5. For discharged, lost-to-Follow Up and deceased patients, the dose was put on missing. The need of steroids was compared using a generalised linear mixed model (i.e. a logistic regression model with a random effect of patient) with site added as a fixed effect. An odds ratio was reported for the comparison of the ‘mean probability’ over d2-d5, and for the day-specific comparisons.
    End point type
    Secondary
    End point timeframe
    Days of steroids needed during day 1-5
    End point values
    Vitamin C Placebo Per Protocol Set
    Number of subjects analysed
    147
    145
    139
    Units: days of steroids needed
        log mean (confidence interval 95%)
    0.149 (0.094 to 0.229)
    0.114 (0.069 to 0.183)
    0.165 (0.103 to 0.254)
    Attachments
    Need of Steroids FAS
    Need of Steroids PPS
    Statistical analysis title
    Secondary: Need of Steroids
    Statistical analysis description
    Observed Information and Predicted Probability. Predicted probability (and 95%Cl) of steroid use obtained from a generalised linear mixed model (logistic regression with a random effect of patient) with group and time as fixed effects and correcting for study site (as fixed effect). Note that the estimates refer to the predicted probability and the predicted effect for the average patient. Use of steroids is put on zero when patient is discharged. When subject is deceased, the value is missing.
    Comparison groups
    Vitamin C v Placebo
    Number of subjects included in analysis
    292
    Analysis specification
    Pre-specified
    Analysis type
    superiority [19]
    P-value
    = 0.43
    Method
    Mixed models analysis
    Confidence interval
    Notes
    [19] - The need of steroids (MV) will be compared using a generalized linear mixed model (i.e. a logistic regression model with a random effect of patient) with site added as a fixed effect. An odds ratio will be reported for the comparison of the ‘mean probability’ over d2-d5, and for the day-specific comparisons.

    Secondary: Secondary: Dose of Steroids

    Close Top of page
    End point title
    Secondary: Dose of Steroids
    End point description
    The daily dose of steroids given was compared between both groups using a multivariate normal model for longitudinal measurements, with the Aikake criterion to select the covariance matrix. The mean dosage over d2-d5 was compared, as well as the timepoint specific dosages. Study site was added in the model. Note that the model fitted on information from all patients in the period d1-d5. For discharged, lost-to-Follow Up and deceased patients, the dose was put on missing. The need of steroids was compared using a generalised linear mixed model (i.e. a logistic regression model with a random effect of patient) with site added as a fixed effect. An odds ratio was reported for the comparison of the ‘mean probability’ over d2-d5, and for the day-specific comparisons.
    End point type
    Secondary
    End point timeframe
    Dose of Corticosteroids needed during day 1-5
    End point values
    Vitamin C Placebo Per Protocol Set
    Number of subjects analysed
    147
    145
    139
    Units: Dose of steroids
        log mean (confidence interval 95%)
    1.565 (1.058 to 2.222)
    1.214 (0.776 to 1.769)
    1.696 (1.145 to 2.422)
    Attachments
    Dose of Steroids FAS
    Dose of Steroids PPS
    Statistical analysis title
    Secondary: Dose of Steroids
    Statistical analysis description
    Descriptive information including patients who do not take steroids and descriptive information excluding patients who do nt take steroids and estimates of steroid dose. For discharged, lost-to-FU and deceased patients, the dose was put on missing (unless there was information on the day of discharge or the day of death.)
    Comparison groups
    Vitamin C v Placebo
    Number of subjects included in analysis
    292
    Analysis specification
    Pre-specified
    Analysis type
    superiority [20]
    P-value
    = 0.3564
    Method
    multivariate normal model
    Confidence interval
    Notes
    [20] - The daily dose of steroids given will be compared between both groups using a multivariate normal model for longitudinal measurements, with the Aikake criterion to select the covariance matrix Estimates of steroid dose (95%Cl) obtained from a linear model for longitudinal measurements fitted on inverse hyperbolic sign transformed data. Means and 95% confidence intervals are obtained after backtransformation to the original scale. Due to the transformation, the comparison of both groups refers t

    Secondary: Secondary: IV Fluid Balance

    Close Top of page
    End point title
    Secondary: IV Fluid Balance
    End point description
    Within the group of patients admitted in the ICU the daily fluid balance (ml) was compared between both groups using a multivariate normal model for longitudinal measurements, with the Aikake criterion to select the covariance matrix. The mean over d2-d5 was compared, as well as the timepoint specific value. Study site was added in the model. Note that the model fitted on information from d1-d5, but restricted to the days the patient was in the ICU. For patients who died during ICU, the balance was put on missing.
    End point type
    Secondary
    End point timeframe
    IV Fluid balance during day 1-5
    End point values
    Vitamin C Placebo Per Protocol Set
    Number of subjects analysed
    147
    145
    139
    Units: ml fluids balance
        log mean (confidence interval 95%)
    137.6 (-191.1 to 466.2)
    2.6 (-344.8 to 350.0)
    134.9 (-204.5 to 474.4)
    Attachments
    IV Fluid Balance FAS
    IV Fluid Balance PPS
    Statistical analysis title
    Secondary: IV Fluid Balance
    Statistical analysis description
    Descriptive information, only including patients in ICUs. For patients deceased in ICU, the fluid balance was put on missing (unless there was information on the day of death). Estimates of mean fluid balance (95%Cl) obtained from a linear model for longitudinal measurements correcting for study site.
    Comparison groups
    Vitamin C v Placebo
    Number of subjects included in analysis
    292
    Analysis specification
    Pre-specified
    Analysis type
    superiority [21]
    P-value
    = 0.5284
    Method
    multivariate normal model
    Confidence interval
    Notes
    [21] - Within the group of patients admitted in the ICU the daily fluid balance (ml) will be compared between both groups using a multivariate normal model for longitudinal measurements, with the Aikake criterion to select the covariance matrix. The mean over d2-d5 will be compared, as well as the timepoint specific value. Study site will be added in the model

    Secondary: Secondary: Quality of Life

    Close Top of page
    End point title
    Secondary: Quality of Life
    End point description
    The EQ-5D-5L questionnaire consists of two parts. The first part is a descriptive system that contains five dimensions; it can be used to estimate the utility index (EQ index). An index of 1 equals perfect health, 0 for death. The second part is the EuroQol visual analogue scale (EQ-VAS), which is a 0–100 scale where respondents indicate their overall health status.
    End point type
    Secondary
    End point timeframe
    The EQ-5D-5L questionnaires were obtained on day 1, day 5, day 28 and after 3 months.
    End point values
    Vitamin C Placebo Per Protocol Set
    Number of subjects analysed
    147
    145
    139
    Units: utility index
        log mean (confidence interval 95%)
    0 (0 to 0)
    0 (0 to 0)
    0 (0 to 0)
    Attachments
    EQ Index FAS
    EQ-VAS FAS
    EQ Index PPS
    44 EQ-VAS PPS
    Statistical analysis title
    Secondary: Quality of life - EQ VAS
    Statistical analysis description
    The EQ-5D VAS score. For patients deceased, discharged or lost-to follow the value was put on missing. If the patient was discharged before 5 days, the value refers to the day of discharge. Estimates of mean EQ-D VAS (95%Cl) obtained from a linear model for longitudinal measurements correcting for study site.
    Comparison groups
    Vitamin C v Placebo
    Number of subjects included in analysis
    292
    Analysis specification
    Pre-specified
    Analysis type
    superiority [22]
    P-value
    = 0.8255 [23]
    Method
    Mixed models analysis
    Confidence interval
    Notes
    [22] - Estimates of mean EQ-5D VAS (95%CI) obtained from a linear model for longitudinal measurements correcting for study site.
    [23] - Day 5 = 0.8255 Day 28 = 0.7213 Month 3 = 0.9997
    Statistical analysis title
    Secondary: Quality of life - EQ index
    Statistical analysis description
    The total scores were compared between both groups using a multivariate normal model for longitudinal measurements, with the Aikake criterion to select the covariance matrix. Study site was added in the model. Note that the model fitted on information from baseline – day 5 (or day of discharge if discharge before day 5), day 28 and month 3. For lost-to-Follow Up and deceased patients, the score were put on missing. Comparisons referred to the three specific timepoints after baseline.
    Comparison groups
    Vitamin C v Placebo
    Number of subjects included in analysis
    292
    Analysis specification
    Pre-specified
    Analysis type
    superiority [24]
    P-value
    = 0.6547 [25]
    Method
    Mixed models analysis
    Confidence interval
    Notes
    [24] - EQ-5D index calculated using the Belgian EQ-5D-5L value set (Bouckaert et al. PharmacoEconomics - Open (2022) 6:823-836). For patients deceased, discharged or lost-to follow the value was put on missing. If the patient was discharged before 5 days, the value refers to the day of discharge. Estimates of mean EQ-5D indec (95%Cl) obtained from a linear model for longitudinal measurements correcting for study site.
    [25] - Day 5 = 0.6547 Day 28 = 0.3119 Month 3 = 0.9505

    Secondary: Secondary: Return to Work or Daily Activities

    Close Top of page
    End point title
    Secondary: Return to Work or Daily Activities
    End point description
    For patients who were at work before inclusion in the study, the time until return to work was compared. Patients lost-to-Follow Up and patients not returned to work (at three months) were censored. Deceased patients were treated as a competing risk. The comparison was based on a stratified Gray’s test.
    End point type
    Secondary
    End point timeframe
    The EQ-5D-5L questionnaires were obtained on day 1, day 5, day 28 and after 3 months.
    End point values
    Vitamin C Placebo Per Protocol Set
    Number of subjects analysed
    147
    145
    139
    Units: days till return to work or activities
        median (confidence interval 95%)
    37 (25 to 70)
    34 (21 to 65)
    37 (25 to 80)
    Attachments
    Return to Work or Daily Activities FAS
    Return to Work or Daily Activities FAS
    Return to Work or Daily Activities PPS
    Return to Work or Daily Activities PPS
    Statistical analysis title
    Secondary: Return to Work or Daily Activities
    Statistical analysis description
    The EQ-5D-5L questionnaires were obtained on day 1, day 5, day 28 and after 3 months.
    Comparison groups
    Vitamin C v Placebo
    Number of subjects included in analysis
    292
    Analysis specification
    Pre-specified
    Analysis type
    superiority [26]
    P-value
    = 0.7416
    Method
    Grays test
    Confidence interval
    Notes
    [26] - Event rates were estimated as cumulative incidence functions (CIF), taking into account the competing risk of death (competing risk). Confidence intervals were calculated using a log(-log)-transformation. It is assumed that deceased subjects with missing information on return, did not return to normal activities. The hazard ratio (95%Cl) from a Fine&Gray model equals 0.936 (0.623,1.407) for the FAS and 0.931 (0.607,1.426) for the PPS.

    Adverse events

    Close Top of page
    Adverse events information
    Timeframe for reporting adverse events
    From inclusion date till last visit at 3 months. First inclusion June 4th, 2021, last patient last visit November 15th, 2023. Only adverse events (AEs) that were not expected in the disease progression needed to be recorded.
    Adverse event reporting additional description
    In total 45 AEs in 40 patients, 5 patients had 2 events reported. In total 44 deaths of which 31 expected and unrecorded (18 Vit C, 13 Placebo) Unexpected deaths/SAEs: Vit C: 5 deaths and 2 deaths as fatal outcome, Placebo 1 death and 5 deaths as fatal outcome
    Assessment type
    Non-systematic
    Dictionary used for adverse event reporting
    Dictionary name
    MedDRA
    Dictionary version
    0
    Reporting groups
    Reporting group title
    Vitamin C
    Reporting group description
    Patients in the intervention group received 1,5 g intravenous vitamin C (3 ampoules of vitamin C 500mg/5ml) diluted in 50ml of Normal Saline, administered over 15 minutes, every 6 hours for 4 days (4 times a day in total 16 administrations). A maximum of 8 hours was allowed between 2 doses and only one dose could be missed. Parameters were obtained as part of routine clinical care. On day 1+4 a targeted blood sample was collected for procalcitonin determination. Baseline data were obtained as close as possible to the time of randomization. The EQ-5D-5L health questionnaire was obtained on admission, day 5, d28 and 3 months. ABG analysis was part of routine clinical care in ICU. If the patient was discharged to the ward before day 5, ABG sampling was not considered standard of care. To be able to calculate the SOFA score in this population, we used the SpO2/FiO2 ratio to impute for PaO2/FiO2 ratio in the respiratory component.

    Reporting group title
    Placebo
    Reporting group description
    Patients in the placebo group received a matching placebo with Normal Saline. The administration was identical to the vitamin C treatment: 3 ampoules of Noral Saline 5ml diluted in 50ml of Normal Saline, administered over 15 minutes, every 6 hours for 4 days (4 times a day in total 16 administrations). A maximum of 8 hours was allowed between 2 doses and only one dose could be missed. The obtained parameters, blood samples and health questionnaires and the routine clinical care were identical to the Vitamin C arm. Ardena Gent NV, a Drug Product Development & Manufacturing firm blinded, packaged, relabelled the IMP kits, and provided randomization: 1st batch: number 21C22, expiry date September 2022 2nd batch: number 22I01, expiry date October 2023

    Serious adverse events
    Vitamin C Placebo
    Total subjects affected by serious adverse events
         subjects affected / exposed
    19 / 147 (12.93%)
    16 / 145 (11.03%)
         number of deaths (all causes)
    25
    19
         number of deaths resulting from adverse events
    7
    6
    Vascular disorders
    Embolism
    Additional description: Embolic and thrombotic event (SMQ)
         subjects affected / exposed
    1 / 147 (0.68%)
    0 / 145 (0.00%)
         occurrences causally related to treatment / all
    0 / 1
    0 / 0
         deaths causally related to treatment / all
    0 / 0
    0 / 0
    Shock
         subjects affected / exposed
    1 / 147 (0.68%)
    0 / 145 (0.00%)
         occurrences causally related to treatment / all
    0 / 1
    0 / 0
         deaths causally related to treatment / all
    0 / 0
    0 / 0
    Hypovolaemic shock
         subjects affected / exposed
    1 / 147 (0.68%)
    0 / 145 (0.00%)
         occurrences causally related to treatment / all
    0 / 1
    0 / 0
         deaths causally related to treatment / all
    0 / 0
    0 / 0
    Cardiac disorders
    Cardio-respiratory arrest
         subjects affected / exposed
    0 / 147 (0.00%)
    1 / 145 (0.69%)
         occurrences causally related to treatment / all
    0 / 0
    0 / 1
         deaths causally related to treatment / all
    0 / 0
    0 / 1
    Hypertensive cardiomyopathy
         subjects affected / exposed
    1 / 147 (0.68%)
    0 / 145 (0.00%)
         occurrences causally related to treatment / all
    0 / 1
    0 / 0
         deaths causally related to treatment / all
    0 / 1
    0 / 0
    Stress cardiomyopathy
         subjects affected / exposed
    1 / 147 (0.68%)
    0 / 145 (0.00%)
         occurrences causally related to treatment / all
    0 / 1
    0 / 0
         deaths causally related to treatment / all
    0 / 0
    0 / 0
    Ventricular tachyarrhythmia
         subjects affected / exposed
    0 / 147 (0.00%)
    1 / 145 (0.69%)
         occurrences causally related to treatment / all
    0 / 0
    0 / 1
         deaths causally related to treatment / all
    0 / 0
    0 / 0
    Cardiac arrest
         subjects affected / exposed
    0 / 147 (0.00%)
    1 / 145 (0.69%)
         occurrences causally related to treatment / all
    0 / 0
    0 / 1
         deaths causally related to treatment / all
    0 / 0
    0 / 1
    Cardiac failure
         subjects affected / exposed
    0 / 147 (0.00%)
    1 / 145 (0.69%)
         occurrences causally related to treatment / all
    0 / 0
    0 / 1
         deaths causally related to treatment / all
    0 / 0
    0 / 0
    Surgical and medical procedures
    Resuscitation
         subjects affected / exposed
    1 / 147 (0.68%)
    0 / 145 (0.00%)
         occurrences causally related to treatment / all
    0 / 1
    0 / 0
         deaths causally related to treatment / all
    0 / 1
    0 / 0
    Nervous system disorders
    Coma
         subjects affected / exposed
    0 / 147 (0.00%)
    1 / 145 (0.69%)
         occurrences causally related to treatment / all
    0 / 0
    0 / 1
         deaths causally related to treatment / all
    0 / 0
    0 / 0
    General disorders and administration site conditions
    Death
         subjects affected / exposed
    5 / 147 (3.40%)
    1 / 145 (0.69%)
         occurrences causally related to treatment / all
    0 / 5
    0 / 1
         deaths causally related to treatment / all
    0 / 5
    0 / 1
    Gastrointestinal disorders
    Small intestinal obstruction
         subjects affected / exposed
    0 / 147 (0.00%)
    1 / 145 (0.69%)
         occurrences causally related to treatment / all
    0 / 0
    0 / 1
         deaths causally related to treatment / all
    0 / 0
    0 / 0
    Neutropenic colitis
         subjects affected / exposed
    0 / 147 (0.00%)
    1 / 145 (0.69%)
         occurrences causally related to treatment / all
    0 / 0
    0 / 1
         deaths causally related to treatment / all
    0 / 0
    0 / 1
    Melaena
         subjects affected / exposed
    0 / 147 (0.00%)
    1 / 145 (0.69%)
         occurrences causally related to treatment / all
    0 / 0
    0 / 1
         deaths causally related to treatment / all
    0 / 0
    0 / 0
    Respiratory, thoracic and mediastinal disorders
    Dyspnoea
         subjects affected / exposed
    1 / 147 (0.68%)
    0 / 145 (0.00%)
         occurrences causally related to treatment / all
    0 / 1
    0 / 0
         deaths causally related to treatment / all
    0 / 0
    0 / 0
    Hydrothorax
         subjects affected / exposed
    0 / 147 (0.00%)
    1 / 145 (0.69%)
         occurrences causally related to treatment / all
    0 / 0
    0 / 1
         deaths causally related to treatment / all
    0 / 0
    0 / 1
    Hypercapnia
         subjects affected / exposed
    0 / 147 (0.00%)
    1 / 145 (0.69%)
         occurrences causally related to treatment / all
    0 / 0
    0 / 1
         deaths causally related to treatment / all
    0 / 0
    0 / 0
    Pneumonia aspiration
         subjects affected / exposed
    1 / 147 (0.68%)
    0 / 145 (0.00%)
         occurrences causally related to treatment / all
    0 / 1
    0 / 0
         deaths causally related to treatment / all
    0 / 0
    0 / 0
    Pulmonary oedema
         subjects affected / exposed
    1 / 147 (0.68%)
    0 / 145 (0.00%)
         occurrences causally related to treatment / all
    0 / 1
    0 / 0
         deaths causally related to treatment / all
    0 / 0
    0 / 0
    Respiratory disorder
         subjects affected / exposed
    0 / 147 (0.00%)
    1 / 145 (0.69%)
         occurrences causally related to treatment / all
    0 / 0
    0 / 1
         deaths causally related to treatment / all
    0 / 0
    0 / 1
    Infections and infestations
    Pneumonia
         subjects affected / exposed
    2 / 147 (1.36%)
    1 / 145 (0.69%)
         occurrences causally related to treatment / all
    0 / 2
    0 / 1
         deaths causally related to treatment / all
    0 / 0
    0 / 0
    Erysipelas
         subjects affected / exposed
    0 / 147 (0.00%)
    1 / 145 (0.69%)
         occurrences causally related to treatment / all
    0 / 0
    0 / 1
         deaths causally related to treatment / all
    0 / 0
    0 / 0
    Sepsis
         subjects affected / exposed
    2 / 147 (1.36%)
    0 / 145 (0.00%)
         occurrences causally related to treatment / all
    0 / 2
    0 / 0
         deaths causally related to treatment / all
    0 / 0
    0 / 0
    Urinary tract infection
         subjects affected / exposed
    1 / 147 (0.68%)
    1 / 145 (0.69%)
         occurrences causally related to treatment / all
    0 / 1
    0 / 1
         deaths causally related to treatment / all
    0 / 0
    0 / 0
    Metabolism and nutrition disorders
    Dehydration
         subjects affected / exposed
    0 / 147 (0.00%)
    1 / 145 (0.69%)
         occurrences causally related to treatment / all
    0 / 0
    0 / 1
         deaths causally related to treatment / all
    0 / 0
    0 / 0
    Frequency threshold for reporting non-serious adverse events: 0%
    Non-serious adverse events
    Vitamin C Placebo
    Total subjects affected by non serious adverse events
         subjects affected / exposed
    4 / 147 (2.72%)
    6 / 145 (4.14%)
    Vascular disorders
    Deep vein thrombosis
         subjects affected / exposed
    0 / 147 (0.00%)
    1 / 145 (0.69%)
         occurrences all number
    0
    1
    Hypotension
         subjects affected / exposed
    0 / 147 (0.00%)
    1 / 145 (0.69%)
         occurrences all number
    0
    1
    Cardiac disorders
    Atrial fibrillation
         subjects affected / exposed
    1 / 147 (0.68%)
    1 / 145 (0.69%)
         occurrences all number
    1
    1
    Surgical and medical procedures
    Hospitalisation
         subjects affected / exposed
    1 / 147 (0.68%)
    0 / 145 (0.00%)
         occurrences all number
    1
    0
    Intensive care
         subjects affected / exposed
    0 / 147 (0.00%)
    1 / 145 (0.69%)
         occurrences all number
    0
    1
    Gastrointestinal disorders
    Stomatitis
         subjects affected / exposed
    0 / 147 (0.00%)
    1 / 145 (0.69%)
         occurrences all number
    0
    1
    Infections and infestations
    Cholangitis infective
         subjects affected / exposed
    1 / 147 (0.68%)
    0 / 145 (0.00%)
         occurrences all number
    1
    0
    Erythema
         subjects affected / exposed
    0 / 147 (0.00%)
    1 / 145 (0.69%)
         occurrences all number
    0
    1
    Staphylococcal sepsis
         subjects affected / exposed
    1 / 147 (0.68%)
    0 / 145 (0.00%)
         occurrences all number
    1
    0

    More information

    Close Top of page

    Substantial protocol amendments (globally)

    Were there any global substantial amendments to the protocol? No

    Interruptions (globally)

    Were there any global interruptions to the trial? No

    Limitations and caveats

    Limitations of the trial such as small numbers of subjects analysed or technical problems leading to unreliable data.
    - optimal dose and timing of vitamin C administration not known - no measurement of baseline vitamin C levels - one hospital included about 40% of the patients
    For support, Contact us.
    The status and protocol content of GB trials is no longer updated since 1 January 2021. For the UK, as of 31 January 2021, EU Law applies only to the territory of Northern Ireland (NI) to the extent foreseen in the Protocol on Ireland/NI. Legal notice
    As of 31 January 2023, all EU/EEA initial clinical trial applications must be submitted through CTIS . Updated EudraCT trials information and information on PIP/Art 46 trials conducted exclusively in third countries continues to be submitted through EudraCT and published on this website.

    European Medicines Agency © 1995-Fri May 02 16:56:49 CEST 2025 | Domenico Scarlattilaan 6, 1083 HS Amsterdam, The Netherlands
    EMA HMA