A combination therapy that many emergency and critical care physicians had hoped could shield the injured brain after cardiac arrest has failed its most rigorous test to date. In a multicentre, placebo-controlled phase 2 trial known as STAR, researchers in South Korea found that early intravenous treatment with hydrocortisone, thiamine, and ascorbic acid did not reduce neuronal injury in comatose survivors of out-of-hospital cardiac arrest, and it did not improve survival or neurological recovery. The results, published in eClinicalMedicine, deliver a sobering verdict on one of the most debated ‘metabolic resuscitation’ strategies in critical care.
Out-of-hospital cardiac arrest affects more than 3.7 million people worldwide each year, yet barely 10 percent of patients survive with favourable neurological function. For those who regain a pulse, the leading cause of death and long-term disability is hypoxic-ischaemic brain injury, produced by the initial global loss of blood flow during the arrest and amplified by reperfusion injury once circulation returns. This post-cardiac arrest syndrome involves a dysregulated cascade of systemic inflammation, oxidative stress, endothelial dysfunction, mitochondrial impairment, and neuroinflammation that continues to damage neurons even after resuscitation appears successful.
The STAR trial was designed around the idea that attacking several of these pathways simultaneously might succeed where single-agent approaches had failed. Hydrocortisone was chosen to dampen glucocorticoid-responsive inflammatory signalling and cytokine production; ascorbic acid to counter oxidative stress and support endothelial function; and thiamine to sustain pyruvate dehydrogenase-dependent mitochondrial energy metabolism. The combination had shown mixed results in sepsis, but no multicentre randomised trial had ever evaluated its early administration in comatose cardiac arrest survivors. The research team, led by Youn-Jung Kim and Won Young Kim, enrolled 160 adults aged 19 to 79 at four tertiary academic hospitals in South Korea between December 2021 and December 2024, with final follow-up completed in June 2025.
Participants, all comatose after resuscitation from presumed cardiac-cause arrest and all undergoing targeted temperature management, were randomly assigned in a one-to-one ratio to receive the drug combination or matching saline placebo. Treatment began quickly, at a median of 3.9 hours after cardiac arrest, and consisted of six intravenous infusions given every 12 hours over three days: hydrocortisone 100 mg, thiamine 200 mg, and ascorbic acid dosed at 50 mg/kg up to a maximum of 3 g per dose. The primary endpoint was the peak serum concentration of neuron-specific enolase, a well-validated biomarker of hypoxic-ischaemic brain injury, measured at 48 and 72 hours after return of spontaneous circulation. Secondary endpoints included survival at multiple timepoints, functional neurological outcomes on the Cerebral Performance Category and modified Rankin scales, organ failure trajectory, and length of intensive care and hospital stay.
The verdict on the primary endpoint was unambiguous. In the modified intention-to-treat population of 147 patients, peak neuron-specific enolase did not differ significantly between the treatment and placebo groups, with a median between-group difference of minus 5.9 ng/mL and a confidence interval spanning from minus 34.5 to plus 40.8, and a p value of 0.54. A sensitivity analysis in the full intention-to-treat population using multiple imputation for missing values produced the same conclusion. Serial biomarker modelling across the first 72 hours likewise showed no divergence in the trajectory of neuronal injury between the groups.
Clinical outcomes told the same story. At 180 days, 63.5 percent of treated patients in the primary analysis population were alive compared with 69.9 percent of placebo patients, a difference that was not statistically significant, and good neurological outcome was achieved by 54.1 percent versus 56.2 percent. There were no meaningful differences in time to awakening, organ failure scores, vasopressor requirements, or intensive care and hospital length of stay. Safety signals were balanced as well: roughly half of patients in each group experienced at least one serious adverse event, and none of these events were judged probably or definitely related to the study medications.
Intriguingly, the intervention was pharmacologically active. Serum vitamin C and thiamine concentrations rose sharply and significantly in the treated group by 72 hours, confirming that patients received and absorbed the drugs as intended. Yet this clear biochemical exposure failed to translate into biological protection of the brain. The authors suggest several explanations. All patients were already receiving targeted temperature management, which itself modulates cerebral metabolism and inflammatory responses, potentially leaving little incremental benefit for the drug cocktail. More fundamentally, the trial measured drug levels but not engagement of the intended molecular targets, so it cannot distinguish between failure to modulate the relevant pathways and failure of that modulation to prevent neuronal injury.
The trial also revealed something important about who was being treated: no participant had thiamine deficiency at baseline, and only about 31 percent had vitamin C deficiency. Universal supplementation of patients who are not deficient, the findings imply, is unlikely to help, and subgroup analyses provided no clear evidence that vitamin C status modified the treatment effect, though the trial was not powered to detect such interactions. This points the field toward phenotype-directed therapy, in which objectively confirmed micronutrient deficiency, rather than diagnosis alone, determines who receives supplementation.
The STAR results also align with a broader and increasingly cautionary literature. The VITAMINS trial found no benefit of the same triple combination in septic shock. The LOVIT trial reported that intravenous vitamin C increased the risk of death or persistent organ dysfunction in adults with sepsis, and the VITaCCA trial in cardiac arrest found worse organ-function, myocardial, renal, and neurological outcomes with a supraphysiological 10 g/day vitamin C regimen. Thiamine trials in both in-hospital and out-of-hospital cardiac arrest survivors showed no improvement in lactate clearance and were stopped early after mortality signals raised safety concerns. Together, these studies challenge the assumption that pharmacological doses of vitamins are inherently benign or beneficial in critically ill patients.
The trial’s investigators acknowledge limitations, including the exclusion of patients who died before 48 hours from the primary analysis, the awareness of treatment allocation by bedside clinicians, some open-label co-interventions in the placebo group, and restriction to four South Korean centres treating a predominantly cardiac-cause, temperature-managed population. Nevertheless, the central conclusion stands: early empiric administration of hydrocortisone, thiamine, and ascorbic acid cannot be recommended as a neuroprotective strategy for unselected comatose survivors of out-of-hospital cardiac arrest. Future research, the authors argue, should move beyond one-size-fits-all metabolic cocktails toward rigorously validated biological targets, confirmed deficiency states, and demonstrated target engagement before any new combination reaches the bedside.
Subject of Research: A randomised phase 2 trial of hydrocortisone, thiamine, and ascorbic acid for neuroprotection in comatose out-of-hospital cardiac arrest survivors
Article Title: Steroid, thiamine, and ascorbic acid for comatose out-of-hospital cardiac arrest survivors (STAR): a randomised, placebo-controlled, phase 2 trial
Article References: Kim, Y.-J., Ko, B. S., Roh, Y.-I., Kwon, H. J., Kim, S., Cha, K.-C., Kim, Y. H., & Kim, W. Y. (2026). Steroid, thiamine, and ascorbic acid for comatose out-of-hospital cardiac arrest survivors (STAR): a randomised, placebo-controlled, phase 2 trial. eClinicalMedicine, 100, Article 104191. https://doi.org/10.1016/j.eclinm.2026.104191
Image Credits: AI Generated
DOI: 10.1016/j.eclinm.2026.104191
Keywords: cardiac arrest, neuroprotection, hydrocortisone, thiamine, vitamin C, neuron-specific enolase, post-cardiac arrest syndrome, targeted temperature management, randomised controlled trial, ischaemia-reperfusion injury, critical care, metabolic resuscitation
Cite Scienmag News
APA MLA Chicago
Cassandra Pierce. (September 12, 2026). Vitamin cocktail fails to protect the brain after cardiac arrest in landmark trial. Scienmag. https://scienmag.com/vitamin-cocktail-fails-to-protect-the-brain-after-cardiac-arrest-in-landmark-trial/
Cassandra Pierce. “Vitamin cocktail fails to protect the brain after cardiac arrest in landmark trial.” Scienmag, 12 September 2026, https://scienmag.com/vitamin-cocktail-fails-to-protect-the-brain-after-cardiac-arrest-in-landmark-trial/. Accessed 12 September 2026.
Cassandra Pierce. “Vitamin cocktail fails to protect the brain after cardiac arrest in landmark trial.” Scienmag. September 12, 2026. https://scienmag.com/vitamin-cocktail-fails-to-protect-the-brain-after-cardiac-arrest-in-landmark-trial/
Copy citation Download RIS
Tags: cardiac arrestCardiac arrest brain injurycritical carecritical care treatment for cardiac arrest survivorsefficacy of combination therapy in neuroprotectionhydrocortisonehypoxic-ischaemic brain injuryischaemia-reperfusion injurylong-term neurological outcomes post-cardiac arrestmetabolic resuscitationmetabolic resuscitation therapyneuron-specific enolaseneuronal injury prevention strategiesNeuroprotectionneuroprotection after out-of-hospital cardiac arrestphase 2 placebo-controlled clinical trialpost-cardiac arrest syndromerandomised controlled trialreperfusion injury mechanismsSTAR trial on hydrocortisone thiamine ascorbic acidsystemic inflammation and oxidative stress in cardiac arresttargeted temperature managementthiaminevitamin C
