Hidden sources of fluids, sodium and potassium in stabilised Swedish ICU patients: A multicentre retrospective observational study.

BACKGROUND Fluid overload in ICU patients is associated with increased morbidity and mortality. Although studies report on optimisation of resuscitation fluids given to ICU patients, increasing evidence suggests that maintenance fluids and fluids used to administer drugs are important sources of fluid overload. OBJECTIVES We aimed to evaluate the volume of maintenance fluids and electrolytes on overall fluid balance and their relation to mortality in stabilised ICU patients. DESIGN Multicentre retrospective observational study. SETTING Six mixed surgical and medical ICUs in Sweden. PATIENTS A total of 241 adult patients who spent at least 7 days in the ICU during 2018. MAIN OUTCOME MEASURES The primary endpoint was the volume of maintenance, resuscitation and drug diluent fluids administered on days 3 to 7 in the ICU. Secondary endpoints were to compare dispensed amounts of maintenance fluids and electrolytes with predicted requirements. We also investigated the effects of administered fluids and electrolytes on patient outcomes. RESULTS During ICU days 3 to 7, 56.4% of the total fluids given were maintenance fluids, nutritional fluids or both, 25.4% were drug fluids and 18.1% were resuscitation fluids. Patients received fluids 1.29 (95% confidence interval 1.07 to 1.56) times their estimated fluid needs. Despite this, 93% of the cohort was treated with diuretics or renal replacement therapy. Patients were given 2.17 (1.57 to 2.96) times their theoretical sodium needs and 1.22 (0.75 to 1.77) times their potassium needs. The median [IQR] volume of fluid loss during the 5-day study period was 3742 [3156 to 4479] ml day, with urine output the main source of fluid loss. Death at 90 days was not associated with fluid or electrolyte balance in this cohort. CONCLUSION Maintenance and drug fluids far exceeded resuscitative fluids in ICU patients beyond the resuscitative phase. This excess fluid intake, in conjunction with high urinary output and treatment for fluid offload in almost all patients, suggests that a large volume of the maintenance fluids given was unnecessary. TRIAL REGISTRATION ClinicalTrials.org (identifier: NCT03972475).