Thiamine for Septic Shock: Take Your Vitamins?

Sepsis remains a leading cause of death in hospitals (1). In the presence of timely antibiotics, infection source control, and performance improvement projects, most ancillary therapies for sepsis have not demonstrated benefits in clinical trials (2). Thiamine, also known as vitamin B1, has been investigated for the treatment of sepsis and septic shock, as thiamine deficiency has been reported in patients with sepsis (3). Many clinicians are familiar with the use of thiamine to correct low concentrations in patients with alcohol use disorder, and it has previously been studied in sepsis in conjunction with vitamin C and hydrocortisone (4, 5).

Thiamine is an essential nutrient that plays a key role in energy generation and regulation of gene transcription, and its deficiency can lead to impaired cellular metabolism and organ dysfunction (6). Thiamine pyrophosphate (TPP), the active form of thiamine, enters mitochondria to play as a cofactor in the Krebs cycle and facilitate transformation of pyruvate to acetyl-coenzyme A. Therefore, thiamine deficiency can raise lactate and lower ATP production. TPP also aids the pentose phosphate pathway and antioxidant synthesis. Thiamine deficiency can increase oxidative stress and might affect inflammation responses (6).

Thiamine replacement can reduce lactate concentrations in critically ill patients (7) and appeared to benefit certain patients with septic shock at increased risk for thiamine deficiency; however, the potential benefits and risks have yet to be fully understood (8–10).

In this issue of the Journal, articles by Moskowitz and colleagues (pp. 570–578) and Pereira and colleagues (pp. 616–618) report the results of modest-sized randomized clinical trials on thiamine for septic shock (11, 12). Moskowitz and colleagues conducted an exploratory, multicenter, randomized trial (TRPSS trial) investigating the effects of thiamine versus placebo on kidney function (measured by serum creatinine changes at 72 h from enrollment) in adult patients with septic shock (11). The trial included 95 adult patients with septic shock and serum creatinine >1.0 mg/dl at three ICUs in the United States; 88 patients received one or more doses of the drug. The thiamine group received 200 mg of intravenous thiamine every 12 hours for 3 days, and the placebo group received a matching volume of 0.9% NaCl. The trial was quadruple blinded, with participants, investigators, clinical teams, and outcomes assessors blinded to the allocation.

There were no significant differences between the thiamine and placebo groups concerning the primary outcome (change in creatinine concentrations). In terms of secondary outcomes, the thiamine group demonstrated a significantly higher number of ICU-free days than the placebo group, which was one of the 10 predefined secondary outcomes. Of note, the point estimates for many outcome measures, including the primary endpoint, favored the thiamine group, but the trial was underpowered to show potentially clinically important differences. Twenty-three patients (27%) had a plasma or serum thiamine <8 mmol/L and were labeled thiamine deficient. There was no statistically significant interaction between baseline thiamine deficiency status and the effect of thiamine.

Pereira and colleagues reported a randomized, double-blind pilot trial conducted at a single ICU in Brazil to compare the effects of thiamine and placebo for septic shock (12). One hundred twenty-two patients were randomized in the first 24 hours of ICU admission to receive 200 mg of intravenous thiamine or placebo every 12 hours for 5 days. There were no significant differences between the groups at baseline regarding most variables; however, patients allocated to the thiamine group had lower hemoglobin concentrations and higher plasma thiamine concentrations and creatinine concentrations at baseline than the placebo group. The trial measured plasma thiamine: 8% of patients were thiamine deficient, defined as <16 ng/ml, with no difference between groups at baseline. No statistically significant difference was observed in the 28-day mortality between the thiamine group and the placebo group: 32 (56.1%) versus 35 (60.3%) (P = 0.789). The length of ICU and hospital stays, the vasoactive drug-free days, and mechanical ventilation–free days also showed no significant difference between the two groups.

Both modest-sized trials explored the potential benefits of the administration of thiamine intravenously in patients with septic shock in a randomized trial. The investigators conducted the trials before and during the coronavirus disease (COVID-19) pandemic (for which we commend the authors), which led to the usual enrollment challenges.

We wish to highlight some limitations of these studies. Both trials measured plasma or serum thiamine to assess the thiamine deficiency. However, the physiological activity of thiamine for metabolic reactions is exhibited primarily in the form of TPP, and TPP constitutes the majority of total thiamine content in whole blood predominantly within erythrocytes (13). The biological rationale for measuring plasma or serum thiamine was not provided in these trials. Furthermore, the cut-off values to identify a population with thiamine deficiency were different between the two trials. These uncertainties and inconsistencies suggest that basic biological and physiological research on thiamine physiology is needed to inform future clinical trials, including which thiamine to measure and what concentration merits treatment.

Another limitation would be the high control event rate in the two trials. In-hospital mortality or 28-day mortality in the control groups was around 50%, which might be higher than anticipated from a severity score or the patients’ baseline description when compared with other recent clinical research on septic shock (14, 15), which may limit the generalizability of the trial results.

It is important that thiamine replacement is based on a biologically and physiologically valid rationale and is also safe. A recent study demonstrated that 200 mg of intravenous thiamine twice daily achieved and sustained supraphysiological concentrations of thiamine pyrophosphate in whole blood (16). Finally, both of the trials follow the traditional mode of taking a biologically plausible intervention in sepsis and testing it in frequentist clinical trials. To date, this approach has not led to successful ancillary treatments for sepsis; potential alternate approaches (which have yet to be proven superior to this traditional methodology) include better identification of “treatable traits” that may be amenable to interventions and the use of platform-adaptive randomized controlled trials (17).

The two trials published here have added to the evolving research investigating the potential promise of essential vitamins in managing septic shock. We would suggest that additional work is essential to define the appropriate thiamine measurement methods and explore the target population that may benefit from thiamine administration, possibly those with thiamine deficiency measured at the bedside. For now, thiamine administration for the treatment of septic shock remains in the realm of research, rather than clinical care.