Association between serum osmolarity and mortality in patients who are critically ill: a retrospective cohort study

OBJECTIVES: This research aims to explore the association between serum osmolarity and mortality in patients who are critically ill with specific categories of disease. DESIGN: A retrospective cohort study. SETTING AND PARTICIPANTS: Data were extracted from an online database named ‘Multiparameter Intelligent Monitoring in Intensive Care II’. 16 598 patients were included. METHODS: Patients were divided into six disease subgroups based on the diagnosis at admission: cardiac, cerebral, vascular, gastrointestinal, respiratory and non-respiratory. The association between maximum osmolarity (osmolarity(max)) and hospital mortality in each subgroup was evaluated using osmolarity(max) as a design variable (six levels). RESULTS: Analysis of the 16 598 patients revealed a ‘U’-shaped relationship between osmolarity and mortality with a threshold of 300 mmoL/L. For patients with non-respiratory disease, both hypo-osmolarity and hyperosmolarity(max) were associated with increased mortality, with the OR increasing from osmolarity(max) level 3 (OR: 1.98, 95% CI 1.69 to 2.33, p<0.001) to level 6 (OR: 4.45, 95% CI 3.58 to 5.53, p<0.001), using level 2 (290–309 mmoL/L) as the reference group. For patients with respiratory disease, however, neither hypo-osmolarity nor hyperosmolarity(max) was significantly associated with mortality (levels 1 to 5) except for extreme hyperosmolarity(max) (≥340 mmoL/L, OR: 2.03, 95% CI 1.20 to 3.42, p=0.007). ORs of mortality in the other four subgroups (cardiac, cerebral, vascular, gastrointestinal) were similar, with OR progressively increasing from level 3 to 6. In all six subgroups, vasopressin use was consistently associated with increased mortality. CONCLUSIONS: Hyperosmolarity is associated with increased mortality in patients who are critically ill with cardiac, cerebral, vascular and gastrointestinal admission diagnoses, with thresholds at 300 mmoL/L. For patients with respiratory disease, however, no significant association was detected.