Sepsis is a life-threatening condition arising from a dysregulated host response to infection, leading to systemic inflammation and organ dysfunction, and posing a significant global public health challenge (1). Recent data from 2025 indicate that the overall mortality rate among sepsis patients is 35%, which increases to 66% in high-risk populations (2–4). Survivors of severe sepsis or septic shock frequently suffer from long-term disability and poor outcomes, largely attributable to tissue and organ injury caused by thrombosis, hypoperfusion, and inflammatory imbalance (5). In addition, sepsis is associated with profound metabolic dysregulation, including mitochondrial dysfunction, lactate accumulation-induced acidosis, and impaired adenosine triphosphate (ATP) synthesis (6). These disturbances compromise cellular energy metabolism and organ perfusion, thereby directly contributing to multi-organ failure (7, 8). This vicious cycle highlights the crucial role of metabolic dysfunction in driving sepsis-related inflammatory injury.

Vitamin B1 (VB1) is an essential coenzyme required for cellular energy metabolism. It serves as a key cofactor for the pyruvate dehydrogenase (PDH) and α-ketoglutarate dehydrogenase (KGDH) complexes, central enzymes in aerobic glucose metabolism and the tricarboxylic acid (TCA) cycle (9). Previous research suggests that the beneficial effects of VB1 in sepsis are mediated through the correction of metabolic abnormalities, improved energy generation (10), reduction of lactate levels (10), and attenuation of oxidative stress (11, 12). Moreover, growing evidence indicates that VB1 may also play a role in modulating the inflammatory response in sepsis. For instance, it has been shown to lower C-reactive protein (CRP) levels in patients, suggesting an anti-inflammatory effect (13).

While existing evidence indicates that VB1 helps overcome sepsis-induced metabolic disturbances by providing essential metabolic support, thereby facilitating the restoration of homeostasis (14), its role in the inflammatory response remains underexplored. This study therefore aims to evaluate the association between VB1 levels and inflammatory indicators in sepsis, as well as its potential value for clinical.

Sepsis, as a systemic inflammatory response, is often accompanied by metabolic disturbances and increased energy expenditure (19). The pathophysiological process of sepsis is highly complex, primarily involving enhanced glycolysis, mitochondrial dysfunction, excessive ROS production (20), and the resulting oxidative stress and tissue damage (21, 22). Early metabolic alterations in sepsis include the accumulation of pyruvate and its conversion to lactate, leading to increased tissue levels of lactate and glycerol, as well as enhanced anaerobic oxidation (21). Concurrently, sepsis involves mitochondrial dysfunction and increased ROS levels, which readily induce heightened oxidative stress and insufficient ATP supply, further contributing to patient shock and multiple organ failure (21, 22). Sepsis is a disease caused by severe infectious diseases, with pneumonia being one of the most common infectious etiologies of sepsis (23). A study by Khalid et al. found that VB1, as an adjuvant therapy, significantly reduced the in-hospital mortality rate in pneumonia patients (24). Therefore, VB1, as a coenzyme in the body’s metabolism, still holds potential clinical value in anti-inflammatory effects and improving sepsis.

In the study by Xu et al. (10), it was confirmed that VB1 primarily alleviates the inflammatory response in sepsis treatment by modulating metabolic processes and subsequently regulating inflammatory cytokine levels. This approach has been applied in clinical practice. However, systematic reviews of its efficacy reveal heterogeneity in the existing evidence: some studies reported improvements in patient outcome scores (e.g., SOFA score) or critical illness indicators (e.g., CRP), while others failed to demonstrate significant benefits regarding ICU length of stay, need for mechanical ventilation, or mortality (25). Furthermore, the study by Khan et al. (26) found that VB1 can be used as an adjunct in managing sepsis-related complications (such as Wernicke’s encephalopathy). When combined with other vitamins and trace elements, it may improve metabolic disturbances, nutritional deficiencies, and neurological complications induced by sepsis, thereby reducing sepsis-induced parenchymal organ damage. In our study, we also found that VB1 levels were significantly negatively correlated with patient creatinine and cTnI levels. Therefore, sepsis patients may develop VB1 deficiency due to inadequate intake, increased metabolic demands, or accelerated renal clearance, potentially exacerbating lactic acidosis and organ damage (e.g., myocardial and renal injury) (13, 27). Although the present study confirmed an association between low VB1 levels and sepsis severity, it did not incorporate indicators such as lactate into the analysis. Furthermore, while multivariable logistic regression showed a significant independent association between VB1 deficiency and sepsis compared to patients with sufficient VB1 levels, the confidence interval for the VB1 deficientcy was extremely wide (spanning an order of magnitude), indicating unstable estimates likely resulting from low event counts in this group. Thus, the pathophysiological mechanisms linking VB1 levels to sepsis and its related complications—including lactic acidosis and organ injury—as well as the therapeutic efficacy of VB1 monotherapy in improving outcomes in sepsis and its complications, require further validation through high-quality studies. Nonetheless, these findings further highlight the potential clinical relevance of VB1 in sepsis and its associated complications.

Currently, the relationship between VB1 and sepsis remains an area of limited and inconsistent evidence. The study by Moskowitz and Donnino (27) highlighted that VB1 deficiency is relatively common in sepsis patients, particularly among high-risk groups such as those with malnutrition, and suggested that supplementation may improve lactic acidosis and organ injury. However, that study also acknowledged that the effect of VB1 may be limited by significant inter-individual variability. In contrast, Mishra et al. (13) directly compared high-dose VB1 with vitamin C (VC) in sepsis patients. While VC is a recognized antioxidant with known anti-inflammatory and organ-protective properties (28), VB1 performed comparably to VC in reducing CRP levels, indicating a potential role in modulating the sepsis-induced inflammatory response. Nevertheless, in subsequent analyses, VB1 was inferior to VC in improving SOFA scores or reducing vasopressor duration, with no significant improvement observed in the VB1 group.

These observations lead to two pertinent considerations regarding the inconsistent findings across studies. First, the present study identified a significant correlation between VB1 levels and SOFA score but not with other clinical indicators, whereas other reports have shown divergent associations (25, 28). This discrepancy may be partly explained by the nature of the SOFA score as a composite measure of overall disease severity—higher scores reflect more critical illness. In line with this, our data demonstrated a significant negative correlation between VB1 levels and SOFA score. Moreover, the sepsis group exhibited significantly higher SOFA scores than the control group (patients with ordinary pneumonia), and lower VB1 levels were consistently observed in patients with higher SOFA scores, a pattern further confirmed in the VB1-deficient subgroup. Thus, lower VB1 levels in patients with higher SOFA scores may reflect or contribute to worsened outcomes through impaired cellular energy metabolism (29, 30). Second, VB1 primarily functions as a coenzyme in energy metabolism rather than as a direct anti-inflammatory or organ-protective agent. Its measurable effects in sepsis may therefore be modulated by various patient-specific factors, including age, sex, lifestyle, and BMI (31), which could account for the variability across different study populations and designs.

It must be emphasized that, as a prospective single-center study, our research cannot establish a causal relationship between VB1 levels and outcomes in sepsis. The observed association may be interpreted in two ways: pre-existing VB1 deficiency could impair mitochondrial function and increase susceptibility to sepsis and poorer outcomes (32); alternatively, the hypermetabolic state and inflammatory response during sepsis may rapidly deplete VB1 stores, making low VB1 a marker of severity rather than a causative factor (14). To clarify this relationship, further investigations are necessary—such as dynamic monitoring of VB1 and its metabolites in early sepsis or high-risk cohorts, or interventional randomized controlled trials assessing whether VB1 supplementation can improve clinical outcomes.

Sepsis patients commonly present with coagulation dysfunction and disseminated intravascular coagulation (DIC) (33, 34). Previous studies generally considered VB1 to be related to metabolism, with no direct link to coagulation parameters. Interestingly, our study found that D-dimer levels were significantly different between the sepsis and control groups, and VB1 levels were negatively correlated with D-dimer levels. In contrast, FIB showed no statistically significant correlation. VB1 deficiency can disrupt energy metabolism, including glucose metabolism, thereby impairing hepatic function (35, 36). As the primary site of coagulation factor synthesis, liver dysfunction may disturb coagulation homeostasis and lead to coagulation abnormalities (36). Moreover, metabolic disorders are often accompanied by inflammation or oxidative stress, which can activate coagulation pathways and alter hemodynamics, ultimately elevating D-dimer levels (27). As mentioned in the study by Fujii et al. (37), vitamin supplementation in sepsis might improve vascular function by reducing oxidative stress, but this does not directly equate to changes in coagulation parameters. Whether VB1 directly affects coagulation parameters requires further investigation.

In this study, we conducted a first-of-its-kind comparison between patients with sepsis and non-septic patients with ordinary pneumonia infection to assess the association between VB1 levels and sepsis. Our data demonstrated a significant intergroup difference in VB1 levels between the control and sepsis groups. Further analysis revealed that sepsis patients had lower VB1 levels compared to the control group. Additional analyses showed negative correlations between VB1 levels and D-dimer, PCT, SOFA score, creatinine, and troponin I—indicators related to coagulation, inflammation, and organ injury. VB1 level was positively correlated with ALB. Moreover, multivariable logistic regression analysis revealed that VB1 levels were independently associated with sepsis in the adult study population.

This study has several limitations. First, this single-center study had a limited sample size and primarily focused on an adult patient with sepsis. Future research with larger-scale, more age-diverse samples is necessary to obtain a more precise assessment of the relationship between VB1 and sepsis. Second, while selecting patients with ordinary pneumonia as the control group aimed to compare against individuals with less severe infection than sepsis—thereby minimizing the influence of general acute-phase response following infection, but this choice itself may introduce confounding factors. The inflammatory milieu of pneumonia may also affect VB1 levels, potentially attenuating the observed between-group differences. Finally, as a single-center study, despite accounting for multiple variables such as anthropometric measures (e.g., BMI) and biochemical markers (e.g., albumin), we lacked other essential data including detailed dietary intake history, lifestyle, socioeconomic status, as well as follow-up information on subsequent treatments and long-term outcomes. It must be emphasized that as VB1 is a coenzyme for pyruvate dehydrogenase, lactate serves as a direct clinical biomarker of thiamine-related metabolic dysfunction. Unfortunately, due to incomplete lactate data in our prospective dataset, we were unable to include lactate in the correlation or regression analyses, which limits the strength of the metabolic argument. Future studies should prioritize the inclusion of lactate measurements. Consequently, we were unable to clearly distinguish between nutritional deficiency and the acute metabolic effects of critical illness. This limitation also hindered our ability to evaluate long-term prognosis or establish causality. These points underscore the need for more comprehensive nutritional data collection in future studies.

In summary, this prospective single-center study demonstrates that circulating VB1 levels at admission are significantly lower in patients with sepsis compared to those with ordinary pneumonia. Furthermore, correlation analysis revealed that low VB1 levels are significantly associated with inflammatory markers and disease severity (SOFA score). Concurrently, low VB1 levels showed a significant correlation with a higher prevalence of sepsis. These findings suggest that maintaining sufficient VB1 levels may have potential clinical implications for preventing severe infection in sepsis, though whether supplementation to achieve this can prevent severe infection or improve outcomes in sepsis patients awaits confirmation from future intervention trials.