BACKGROUND Septic shock is a life-threatening subset of sepsis, and blood urea nitrogen (BUN) and serum albumin are inexpensive and easily available biomarkers. BUN to serum albumin ratio (BAR) has been identified as a valuable prognostic marker in various diseases. Therefore, we conducted a retrospective study to explore the relationship between BAR and mortality risk in septic shock.
METHODS From 2008 to 2019, we enrolled 3, 220 patients diagnosed with septic shock from the MIMIC-IV database. Restricted cubic spline (RCS) was used to visualize the relationship between BAR and mortality. Kaplan-Meier survival curves were generated to evaluate differences in survival rates among three groups: low-BAR (<9.6), medium-BAR (≥9.6, <20.0), and high-BAR (≥20.0). Box plot analysis was performed to compare the distribution of BAR between survivors and non-survivors. Cox regression analysis was used to estimate the independent association between BAR and all-cause mortality. The receiver operating characteristic (ROC) curve analysis was conducted to assess the performance of BAR for predicting mortality.
RESULTS The mean age was 66.1 years. The 28-day mortality risk increased non-linearly with BAR values as shown by RCS. Compared to the low-BAR or medium-BAR group, the high-BAR group had significantly higher mortality according to Kaplan-Meier curves for in-hospital, 28-day, 90-day, and 1-year mortalities. The boxplot demonstrated that patients who survived had lower BAR compare to non-survived in terms of in-hospital, 28-day, 90-day, and 1-year. The results of both the univariate and multivariate Cox regression analysis showed that BAR was an independent risk factor for predicting in-hospital, 28-day, 90-day, and 1-year mortalities, with higher BAR values associated with increasing mortality.
For sensitivity analysis, the Cox regression analysis also showed that compared with the low-BAR group, the high-BAR group had a higher risk of in-hospital, 28-day, 90-day, and 1-year mortality rates in both the unadjusted and adjusted models.
Conclusions The study showed that BAR was an independent risk factor for predicting both short-term and 1-year mortality rates in septic shock patients.