MIMIC-IV database: should serum osmolality be a critical factor in assessing prognosis for adult patients following hemorrhagic stroke surgery?

Osmotherapy is commonly applied to postoperative patients with hemorrhagic stroke (HS) in the neuro-intensive care unit, necessitating precise monitoring of internal environment parameters such as serum osmolality. Although serum osmolality imbalances are associated with poor outcomes, definitive evidence regarding its prognostic significance in postoperative HS patients remains limited. This study analyzes the correlation between serum osmolality and in-hospital all-cause mortality in postoperative HS patients based on a large sample of postoperative HS patients from the Medical Information Mart for Intensive Care-IV database. A retrospective cohort of 342 HS patients from Medical Information Mart for Intensive Care-IV was analyzed. Consensus clustering was applied to group patients into 6 clusters, with COX proportional hazards regression models used to assess the relationship between serum osmolality and in-hospital mortality. Restricted cubic spline curves were drawn to evaluate the non-linear associations between variables and outcomes. Survival analysis was performed using the Kaplan-Meier method to estimate patient outcomes across groups, and intergroup differences were statistically compared with the log-rank test. Subgroup analyses were conducted to assess differences across various populations. The study included 342 patients, with consensus clustering identifying 6 groups. A significant difference in survival was found between Cluster 4 and the combined group (Cluster 1 + 2 + 3 + 5 + 6), with Cluster 4 showing a lower mortality rate (Log-rank P = .042). COX regression models adjusting for age, gender, comorbidities, and common clinical scores consistently showed that Cluster 4 had a lower risk of mortality. Furthermore, serum osmolality was not found to significantly influence mortality risk after adjusting for potential non-linear effects (P = .742). Serum osmolality does not appear to be a strong predictor of mortality in postoperative patients with surgically treated HS in the neuro-intensive care unit. Age (>65 years), gender, absence of hypertension, lack of hypertonic therapy, no invasive mechanical ventilation, and no diuretic use were significant factors influencing mortality risk.