Electrolyte disorders in the critically ill: a retrospective analysis.

Several electrolyte disorders have been associated with adverse outcomes, but data on the coincidence of multiple disorders, interactions between different electrolytes and the impact of substitution on electrolyte levels are scarce. This study aimed to describe electrolyte disorders' prevalence, incidence, and interaction in a consecutive cohort of ICU patients. Single-center retrospective study, including 2,056 consecutive adult ICU patients. Patients without laboratory data, those declining participation, and those receiving renal replacement therapy were excluded. Electrolyte levels and intravenous electrolyte administration were analyzed during the first 96 h after ICU admission using descriptive statistics, regression models and AUC-ROC analysis. Of 2392 admitted patients, 2056 were included. On admission, 312 patients (15.2%) had no electrolytes measured, 316 (15.4%) did not have any electrolyte disorder, 643 (31.3%) had one disorder, and 785 (38.2%) had multiple disorders. The most common electrolyte disorders on admission were hyperchloremia (56.0%, 977/1,744), hyperkalemia (18.7%, 326/1,744), hypophosphatemia (16.8%, 293/1,744), and hyperphosphatemia (16.4%, 286/1,744). Most patients (77.4%, 1,592/2,056) developed at least one new disorder during their ICU stay. A considerable number of patients (19.7%, 344/1,744) experienced electrolyte disorders in both directions ('hypo' and 'hyper') during their ICU stay. Linear regression models revealed that changes in one electrolyte level are often associated with alterations in other electrolyte levels. Most patients (99.2% 2,039/2,056) received intravenous electrolytes. Electrolyte overcorrection occurred in 89/365 patients (24.4%) for hypokalemia and 50/575 patients (8.7%) for hypophosphatemia. AUC-ROC analysis revealed cut-off points with the highest sensitivity and specificity at 30 mmol of potassium within 6 h and 45 mmol of phosphate within 15 h. Electrolyte disorders are highly prevalent in ICU patients, with significant co-incidence and interplay between these disorders. Knowledge regarding overcorrection risks may inform safe electrolyte administration protocols.