12. Vitamin C, Thiamine & Hydrocortisone in Sepsis (The Marik Protocol Controversy)

Quick Recap

Cross-cutting protocol — companion to Corticosteroids in Septic Shock & Critical Illness protocol. Addresses one of the more publicly visible evidence reversals in recent critical care history: a single-center, retrospective before-after study with dramatic, "biologically implausible" results generated widespread media attention and rapid clinical adoption, before a series of increasingly rigorous multicenter RCTs progressively failed to replicate benefit — culminating in a large trial (LOVIT) that found a statistically significant signal of harm.

1. Definition

The "Marik protocol" / HAT therapy (Hydrocortisone, Ascorbic acid [Vitamin C], Thiamine): a combination adjunctive therapy for sepsis/septic shock, originally proposed as intravenous vitamin C 1.5g every 6 hours, hydrocortisone 50mg every 6 hours, and thiamine 200mg every 12 hours — based on a proposed mechanism of reduced oxidative stress/inflammation, improved endothelial and mitochondrial function, and preserved catecholamine synthesis.

High-dose intravenous vitamin C: subsequently studied both as a component of combination HAT therapy and as a standalone intervention, using either the original fixed-dose regimen (1.5g every 6 hours) or weight-based dosing (50mg/kg every 6 hours) in later, larger trials — the specific dosing strategy and trial design have varied considerably across this literature, contributing to some of the interpretive complexity discussed below.

2. Pathophysiology

The proposed mechanistic rationale for vitamin C in sepsis includes: reduction of oxidative stress and reactive oxygen species generated during the inflammatory response, stabilization of the microvascular endothelium, support of mitochondrial function, and — mechanistically linked to the rationale for combining it with hydrocortisone — preservation of endogenous catecholamine synthesis (vitamin C is a required cofactor for the enzyme dopamine beta-hydroxylase in catecholamine biosynthesis) and thiamine's role as a cofactor supporting aerobic mitochondrial metabolism, particularly relevant in patients with thiamine deficiency (a recognized, correctable co-morbidity in some critically ill patients).

Why the mechanistic rationale, however plausible, did not straightforwardly translate into clinical benefit: this is a recurring pattern across several other topics in this library (early aggressive nutrition, tight glycemic control, early neuromuscular blockade) — a biologically coherent mechanism does not reliably predict net clinical benefit once tested rigorously at scale, and this protocol's evidence arc is one of the more dramatic illustrations of that general principle, given how large the initially reported effect size was and how starkly it failed to replicate.

3. Immediate Stabilization (ABCDE) — Not Applicable

This protocol addresses an adjunctive pharmacotherapy decision rather than acute stabilization; cross-reference Septic Shock (revised) and Corticosteroids in Septic Shock & Critical Illness protocols for the underlying acute management framework.

4. Focused History

• Illness severity (baseline SOFA score) — relevant given a specific, plausible severity-dependent effect modification observed across the trial literature (Section 11)
• Timing since sepsis/septic shock onset — relevant given proposed (though not definitively established) timing-dependent effects
• Known or suspected thiamine deficiency (a distinct, separately valid indication for thiamine repletion regardless of the broader HAT therapy controversy — cross-reference Alcohol & Sedative Withdrawal Management in the ICU protocol, where thiamine repletion is addressed for Wernicke's encephalopathy prevention specifically)
• Concurrent corticosteroid use for vasopressor-refractory shock (cross-reference Corticosteroids in Septic Shock & Critical Illness protocol — the hydrocortisone component of HAT therapy overlaps with, but was tested in a different context than, the corticosteroid evidence in that protocol)

5. Comprehensive System-wise Examination — Not Applicable

Not a primary examination-driven protocol.

6. Syndrome Identification — Not Applicable

This protocol addresses a specific adjunctive therapy decision rather than a syndrome classification framework; cross-reference Septic Shock (revised) protocol for syndrome identification.

7. Differential Diagnosis — Not Applicable

8. Severity/Risk Assessment

A specific, recurring observation across multiple trials: several studies (including a comparison discussed in one RCT's own discussion section) have noted that trials showing less benefit tended to enroll patients with higher baseline severity of illness (e.g., mean SOFA score 10 in one negative RCT vs. mean SOFA 8 in Marik's original cohort), and one trial (ViCTOR) specifically found mortality benefit only in a subset of patients with lower SOFA scores — this represents a genuinely plausible, though not definitively confirmed, effect-modification pattern: any real benefit from this therapy, if it exists, may be concentrated in less severely ill patients, rather than the more severely ill, vasopressor-dependent populations that later, larger, more rigorous trials (including LOVIT) specifically enrolled.

9. Investigations — Not Applicable

10. Point-of-Care Ultrasound — Not Applicable

11. Evidence-Based Management — The Full Evidence Arc

The Origin — Marik's Single-Center Study and the Media Response

• Marik et al., Chest 2017: a retrospective, before-after study (not a randomized trial) from a single center, reporting a dramatic reduction in mortality with the HAT protocol in patients with severe sepsis/septic shock — the reported effect size was large enough that news outlets described the treatment as a potential "miracle" cure, generating substantial public and clinical interest well before any confirmatory randomized evidence existed
• A pointed, methodologically-focused editorial response (Møller, Laake, Myburgh, Alhazzani, Perner — explicitly titled "The Magic Bullet in Sepsis or the Inflation of Chance Findings?", Chest 2017) directly criticized the study for "spurious findings, biased results, and overstated conclusions," specifically noting that a reported absolute mortality risk reduction exceeding 30% is biologically implausible for an adjunctive metabolic therapy of this kind — a genuinely important, prescient methodological critique that anticipated the subsequent trial evidence's failure to replicate anything close to this magnitude of effect
• This is a valuable, explicit teaching example for evidence appraisal: a retrospective, single-center, before-after study design (lacking randomization, blinding, and contemporaneous controls) is inherently vulnerable to confounding, secular trends in general sepsis care improving over the study period, and selection effects — the appropriate response to such a striking finding is exactly what happened here: rigorous, adequately powered RCT confirmation before clinical adoption, not immediate widespread practice change based on the initial report alone

CITRIS-ALI (2019) — First Major RCT, Mixed Signal

• CITRIS-ALI trial (Fowler et al., JAMA 2019): vitamin C infusion (as monotherapy, not combined HAT therapy) in patients with sepsis and severe acute respiratory failure — primary outcomes (organ failure scores, biomarkers of inflammation/vascular injury) showed no significant improvement; however, a secondary outcome of reduced mortality was observed
• This secondary mortality finding was specifically and immediately criticized on methodological grounds: a wide confidence interval, the study not being powered to detect a mortality difference, and the large number of secondary outcomes examined (increasing the likelihood of a chance positive finding through multiplicity) — a critique structurally analogous to the primary-vs-secondary-outcome caution emphasized in the HFNC vs. NIV protocol's discussion of FLORALI, and worth recognizing as the same general evidence-interpretation principle recurring here
• A subsequent reanalysis (Hemilä and Chalker, Front Med 2020) argued that "important findings" regarding the mortality signal were "dismissed in the trial report" — illustrating that even the interpretation of CITRIS-ALI's own secondary finding has remained a point of genuine, ongoing disagreement among investigators, not a fully settled matter

VITAMINS (2020) — Combination Therapy, No Benefit

• VITAMINS trial (Fujii et al., JAMA 2020): vitamin C + hydrocortisone + thiamine vs. hydrocortisone alone in septic shock — no significant difference in the primary outcome (time alive and free of vasopressor support) between groups; no difference in 90-day mortality
• One secondary outcome (change in SOFA score) was significantly lower in the combination-therapy group, but the trial's overall, authors' conclusion was that combination therapy did not lead to faster resolution of septic shock compared to hydrocortisone alone
• The trial's open-label design and limited capture of certain ICU care details (antibiotic duration, adverse events) were noted as limitations

VICTAS — Stopped Early, Essentially Negative

• VICTAS trial (Sevransky et al., JAMA): a multicenter, double-blind, adaptive-sample-size, placebo-controlled RCT specifically designed to rigorously test the combination therapy (contrasting with Marik's uncontrolled original design) — stopped early due to a lack of funding rather than for efficacy or futility, limiting its statistical power to definitively answer the question, though the trial's results as reported did not demonstrate the dramatic benefit originally proposed

LOVIT (2022) — The Definitive, Rigorous Trial Finding Harm

• LOVIT trial: tested high-dose intravenous vitamin C alone (50mg/kg every 6 hours for 96 hours, a weight-based regimen distinct from the original fixed-dose HAT protocol) vs. placebo in ICU adults with sepsis receiving vasopressors — not stopped early, achieving high protocol adherence with minimal open-label contamination, representing a methodologically stronger, more definitive test than several of the earlier trials
• Primary outcome (death or persistent organ dysfunction at day 28): significantly higher with vitamin C — RR 1.21 (95% CI 1.04–1.40, p=0.01), driven mainly by mortality — a genuine, statistically significant harm signal, not merely an absence of benefit
• No meaningful improvement in organ support-free days or 6-month survival was demonstrated
• Practical significance: LOVIT is widely regarded as a landmark trial specifically because it rigorously tested a widely discussed, inexpensive, seemingly low-risk adjunctive therapy and identified a signal for harm rather than mere lack of benefit — this result substantially shifted the evidentiary center of gravity away from empiric high-dose vitamin C in vasopressor-dependent sepsis, and catalyzed subsequent meta-analyses and guideline updates recommending against routine administration
• A genuinely important, still-debated post-hoc critique of LOVIT's own interpretation: a secondary analysis specifically examining the timing of adverse events found a striking 2.28-fold increase in mortality specifically during the 3 days immediately after vitamin C administration was terminated (at the end of the 96-hour treatment protocol), with no significant evidence of harm during the treatment period itself or afterward — this analysis argues that LOVIT's own conclusion ("receiving intravenous vitamin C" caused harm) may be misleading, and that the more precise causal story may be that abrupt discontinuation of vitamin C, rather than its administration per se, drove the observed mortality signal — a genuinely important, unresolved interpretive nuance that complicates a simple "vitamin C causes harm" reading of LOVIT, without at all invalidating the trial's core finding that the tested strategy, as actually implemented, was associated with worse outcomes

Subgroup and Later-Trial Signals — Genuine, Unresolved Complexity

• C-EASIE trial: a multicenter trial of early vitamin C in sepsis/septic shock — did not identify a clinically important benefit, reinforcing the lack of efficacy across a range of trial designs and timing strategies rather than suggesting earlier administration resolves the discordance
• A LOVIT subgroup analysis found a trend toward benefit in patients with COVID-19 specifically, though this trend was not clinically significant — a genuinely intriguing, unconfirmed signal in a specific viral-sepsis subpopulation that should not be over-interpreted given its non-significance, but is worth flagging as an area warranting further, dedicated investigation rather than dismissal
• ACTS trial: found more shock-free days with vitamin C in one analysis — a positive process-outcome signal amid an otherwise largely null-to-harmful overall literature, illustrating genuine heterogeneity of findings across this body of trials that has not been fully reconciled
• A recent (2024) opinion/reevaluation piece specifically raises the possibility of a potential benefit concentrated in severe cases, and proposes that dosing strategy (fixed-dose vs. weight-based), timing of intervention, and patient severity may all be under-characterized effect modifiers requiring further, more targeted study — reflecting that the field has not fully closed the door on a more narrowly-defined beneficial population, even as the broad, unselected-population evidence has turned decisively unfavorable

Current Guideline Position

• A rapid recommendation guideline (Lamontagne et al.) explicitly concluded that intravenous vitamin C should not be used routinely in adult sepsis or septic shock — a recommendation against routine use, reflecting very low certainty of benefit and emerging harm concerns
• Japanese guideline updates similarly moved away from routine vitamin C use in sepsis, reflecting international convergence on this more cautious position

Practical Synthesis

Current evidence does not support routine use of intravenous vitamin C (as monotherapy or as part of combination HAT therapy) in adult sepsis/septic shock, and the most rigorous, definitive trial (LOVIT) found a statistically significant harm signal specifically with high-dose vitamin C in vasopressor-dependent patients. This represents one of critical care's clearer, more complete evidence reversals — from single-center, retrospective enthusiasm and substantial media attention, through a series of progressively more rigorous RCTs, to a definitive, guideline-changing finding of harm. Genuinely unresolved nuances remain (the abrupt-discontinuation reinterpretation of LOVIT's mortality signal, the possible severity- and timing-dependent effect modification suggested across several smaller trials, and the unconfirmed COVID-19 subgroup signal) that argue against treating the question as fully, mechanistically closed — but these nuances should not be used to justify continued routine clinical use outside of a dedicated research protocol, given the current, guideline-endorsed recommendation against routine administration. Thiamine repletion remains separately, validly indicated in specific clinical contexts (e.g., Wernicke's encephalopathy prevention, cross-reference Alcohol & Sedative Withdrawal Management protocol) independent of this broader HAT-therapy controversy, and should not be conflated with or discontinued based on the vitamin C-specific evidence discussed here.

12. Organ Support — Not Applicable

Cross-reference Corticosteroids in Septic Shock & Critical Illness protocol for the hydrocortisone-specific evidence base, which is separate from and should not be conflated with this protocol's vitamin C-specific findings.

13. Disease-Specific Therapy

• Routine intravenous vitamin C (fixed-dose or weight-based) is not currently recommended for adult sepsis/septic shock outside of a research protocol, per current guideline consensus
• Thiamine repletion remains appropriately indicated in specific contexts (documented or suspected deficiency, alcohol use disorder-related Wernicke's encephalopathy prevention) independent of the broader HAT therapy controversy
• Hydrocortisone: governed by its own separate evidence base for vasopressor-refractory septic shock (cross-reference Corticosteroids in Septic Shock & Critical Illness protocol), not by the vitamin C-specific findings in this protocol

14. Consultation Matrix

15. Monitoring Framework — Not Applicable

Not applicable as a standalone monitoring protocol; this content is informational/decision-support regarding a specific adjunctive therapy's evidence base.

16. ICU Bundle Checklist

17. Complications

Of high-dose intravenous vitamin C specifically, per LOVIT: increased risk of death or persistent organ dysfunction at day 28 (RR 1.21, p=0.01) — a genuine, statistically significant harm signal in vasopressor-dependent sepsis patients, the population and dosing strategy LOVIT specifically tested.

Of abrupt discontinuation (per the reinterpretive secondary analysis): a specific, striking mortality increase in the 3 days immediately following treatment cessation — an unresolved, genuinely important nuance suggesting the harm mechanism may be more specifically tied to discontinuation timing than to vitamin C administration itself, though this has not displaced LOVIT's own primary conclusion.

Prevention: avoiding routine use outside a research context, per current guideline recommendation.

18. Escalation & De-escalation — Not Applicable

19. ICU Discharge Criteria — Not Applicable

20. Documentation & Medicolegal Checklist

• If vitamin C is used in any non-standard context, rationale and awareness of current guideline recommendation against routine use documented
• Thiamine repletion, where indicated, documented separately from any broader HAT-therapy discussion

21. Key Guidelines

• Rapid recommendation guideline (Lamontagne et al.): recommendation against routine intravenous vitamin C use in adult sepsis/septic shock, reflecting very low certainty of benefit and emerging harm concerns
• Japanese sepsis guideline updates: similarly moved away from routine vitamin C use, reflecting international convergence

22. Landmark Trials

23. Controversies

• This is one of the clearest examples in critical care of the appropriate evidence hierarchy functioning as intended: a striking, single-center, uncontrolled finding generated enthusiasm and media attention, but was appropriately met with methodological skepticism and followed by a rigorous confirmatory trial sequence, culminating in a large, definitive RCT reversing the initial claim entirely (finding harm, not merely absence of benefit) — this protocol treats the full arc, including the initial editorial skepticism, as itself an important teaching example in evidence appraisal, not merely a historical footnote to the final answer.
• The abrupt-discontinuation reinterpretation of LOVIT remains genuinely unresolved and worth taking seriously: if the mortality signal is substantially attributable to discontinuation timing rather than vitamin C administration itself, this would meaningfully change the practical implication (suggesting a tapered discontinuation strategy might avoid the harm, rather than avoiding the drug entirely) — however, this reinterpretation has not displaced the trial's own primary conclusion or the guideline recommendations built upon it, and should be presented as an important, active area of ongoing discussion rather than an accepted alternative explanation.
• Possible severity- and timing-dependent effect modification remains a genuinely open, unconfirmed question: several smaller trials and subgroup analyses (ViCTOR's lower-SOFA-score subgroup, the COVID-19 subgroup trend in LOVIT, ACTS's shock-free-days finding) suggest the possibility of a more narrowly beneficial population or context, but none of these have been independently, prospectively confirmed as a validated treatment-selection criterion — this represents a legitimate area for continued, more targeted research rather than either a basis for current clinical practice or a fully closed question.
• Dosing strategy heterogeneity across the trial literature (fixed-dose 1.5g q6h vs. weight-based 50mg/kg q6h) complicates direct cross-trial comparison and may itself be a genuine, under-examined source of the discordant findings across this body of evidence, separate from the population- and timing-related considerations above.

24. References

1. 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.
2. Møller MH, Laake JH, Myburgh JA, Alhazzani W, Perner A. The magic bullet in sepsis or the inflation of chance findings? [editorial]. Chest. 2017;152(1):222-223.
3. Fowler AA 3rd, 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.
4. Hemilä H, Chalker E. Reanalysis of the effect of vitamin C on mortality in the CITRIS-ALI trial: important findings dismissed in the trial report. Front Med. 2020;7:590853.
5. Fujii T, Luethi N, Young PJ, et al. 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.
6. Sevransky JE, Rothman RE, Hager DN, et al. Effect of vitamin C, thiamine, and hydrocortisone on ventilator- and vasopressor-free days in patients with sepsis: the VICTAS randomized clinical trial. JAMA. 2021.
7. Lamontagne F, et al. Intravenous vitamin C in adults with sepsis: a rapid practice guideline. Intensive Care Med. 2022.
8. Abrupt termination of vitamin C from ICU patients may increase mortality: secondary analysis of the LOVIT trial. Eur J Clin Nutr. 2022.
9. Mohamed ZU, et al. Vitamin C therapy for routine care in septic shock (ViCTOR) trial. Indian J Crit Care Med. 2020;24(8):653-661.
10. Reevaluating vitamin C in sepsis and septic shock: a potential benefit in severe cases? Front Med. 2024.
11. Intravenous vitamin C in severe sepsis: a systematic review of evidence from 2020 to 2025. 2025.
12. The Washington Manual of Critical Care, 4th ed. 2025 — sepsis adjunctive therapy chapter.
13. ICU Protocols: A Step-wise Approach, 2nd ed. — relevant sepsis content.