Objective:To investigate the relationship between the lowest hemoglobin value during hospitalization and the prognosis in patients with extensive burns, in order to explore the hemoglobin warning threshold for blood transfusion in patients with extensive burns.Methods:The research was a retrospective observational study. From October 2012 to October 2022, 288 patients with extensive burns who met the inclusion criteria were admitted to the First Affiliated Hospital of Army Medical University (the Third Military Medical University), including 243 males and 45 females, aged 18 to 65 years. These patients were assigned to the death group ( n=54) and the survival group ( n=234) based on their final prognosis. The clinical data including gender, age, body mass index, total burn area, full-thickness burn area, time of first operation after injury, preoperative prothrombin time (PT) and activated partial thromboplastin time (APTT) and hemoglobin level of the first surgery, complication of inhalation injury, number of surgeries, total surgical area, total surgical time, total length of hospital stay, and highest procalcitonin value, lowest platelet count and hemoglobin values, and occurrence of sepsis during hospitalization were compared between the two groups of patients. According to the lowest hemoglobin value during hospitalization, the patients were assigned to <65 g/L group, ≥65 g/L and <75 g/L group, ≥75 g/L and <85 g/L group, and ≥85 g/L group. The total length of hospital stay, mortality and incidence of sepsis during hospitalization, and mortality within 90 days after injury were compared among the four groups of patients. The relationship between the lowest hemoglobin value during hospitalization and the mortality risk of patients with extensive burns was analyzed using a restricted cubic spline model before and after adjusting covariates. A logistic regression model was adopted to analyze the relationship between the lowest hemoglobin value during hospitalization and the mortality risk of patients with extensive burns after adjusting covariates, with the lowest hemoglobin value during hospitalization as a continuous variable and a categorical variable, separately. Results:Compared with those in survival group, the total burn area, full-thickness burn area, and total surgical area of patients in death group were significantly increased, the preoperative APTT of the first surgery was significantly prolonged, the number of surgeries was significantly reduced, the total length of hospital stay was significantly shortened, the highest procalcitonin value during hospitalization was significantly increased, the lowest platelet count and hemoglobin values during hospitalization were significantly decreased, and the incidence proportion of sepsis during hospitalization was significantly increased (with Z values of -6.72, -5.40, -2.15, -2.99, -2.21, -7.84, -6.23, -7.03, and -3.43, respectively, χ2=161.95, P values all <0.05). There were no statistically significant differences in the other clinical data of patients between the two groups ( P>0.05). There were statistically significant differences in mortality and incidence of sepsis during hospitalization, and mortality within 90 days after injury of patients among the four groups divided according to the lowest hemoglobin value during hospitalization (with χ2 values of 12.12, 15.93, and 10.62, respectively, P<0.05). There was no statistically significant difference in the total length of hospital stay of patients among the four groups ( P>0.05). The restricted cubic spline model analysis revealed an approximately linear relationship between the lowest hemoglobin value during hospitalization and the mortality risk of patients with extensive burns before and after adjusting covariates (with χ2 values of 0.81 and 0.75, respectively, P>0.05). After adjusting covariates, the logistic regression model analysis showed that the mortality risk of patients with extensive burns increased with decreasing hemoglobin when the lowest hemoglobin value during hospitalization was analyzed as a continuous variable (with odds ratio of 0.96, with 95% confidence interval of 0.92 to 0.99, P<0.05). When using the median value of 75.5 g/L as the cut-off value for categorizing the lowest hemoglobin value during hospitalization, there was no statistically significant difference in the mortality risk between patients with hemoglobin <75.5 g/L and those with hemoglobin ≥75.5 g/L ( P>0.05). When the patients were divided into four groups based on the lowest hemoglobin value during hospitalization as above, using ≥85 g/L group as a reference, only patients in <65 g/L group had a significantly increased mortality risk (with odds ratio of 5.37, with 95% confidence interval of 1.57 to 18.29, P<0.05). Conclusions:There is an approximately linear correlation between the lowest hemoglobin value during hospitalization and the mortality risk of patients with extensive burns. When the hemoglobin level drops to 65 g/L or lower, the mortality risk of patients increases significantly, suggesting that a hemoglobin level of 65 g/L could serve as a warning threshold for blood transfusion in patients with extensive burns.