• Meta-analysis evaluates thiamine’s efficacy for sepsis/septic shock. • Meta-analysis includes RCTs on intravenous thiamine for sepsis/septic shock. • Thiamine reduces the need for Renal Replacement Therapy (RRT). • Thiamine increases 24 h SOFA score changes and improves organ dysfunction. • Overall short-term mortality unchanged; subgroup has lower mortality. To perform a meta-analysis evaluating thiamine’s efficacy for sepsis and septic shock to support its clinical use. The meta-analysis of randomized controlled studies on patients with sepsis and septic shock evaluated the efficacy of intravenous thiamine. Articles were retrieved from online databases. Qualities of included studies were assessed. Nine studies enrolling 520 patients were included. Regarding bias risk assessment of studies, two articles were classified as low risk, three were categorized as raising some concerns, and four were identified as high risk. Patients in the thiamine group were significantly associated with decreased proportion of Renal Replacement Therapy (RRT, Odds Ratio OR = 0.26, p = 0.0007), decreased 24 h lactate level (Mean Difference MD = -0.35, p = 0.004), and larger 24 h SOFA score change (MD = 1.15, p < 0.00001). No statistical significance was detected in short-time mortality (OR = 0.78, p = 0.18) and 24 h lactate change (MD = 0.09, p = 0.36) between the two groups. In the thiamine-deficient subgroup (2 studies, 51 patients), exploratory analysis indicated a potential short-term mortality reduction with thiamine (OR = 0.18, p = 0.01), though limited by the small sample size. This meta-analysis preliminarily suggested that thiamine may exert a potential role in improving organ dysfunction among patients with sepsis and septic shock. No significant benefit of thiamine in reducing short-term mortality was observed. It is important to note that the total number of patients included in this meta-analysis is relatively small, which limits the statistical power and external validity of the results.
Xu et al. (Thu,) studied this question.