April 6, 2016


In patients admitted to the intensive care unit (ICU) with acute kidney injury (AKI) the need for renal replacement therapy (RRT) is associated with greater ICU mortality and greater ICU length of stay. This association persists following correction for disease severity [1]. Whilst it is probable that this association is – at least in part – due to unmeasured confounding exposures, RRT has numerous potential adverse effects including increased risk of bleeding, haemodynamic instability, inadequate antibiotic dosing and loss of micronutrients. In addition it is easy to conceive of benefits of being able to predict when a patient with AKI will not require further RRT (including early removal of central-venous dialysis catheters and early discharge from critical care).

In their retrospective single-centre cohort study [2] Aniort et al. aimed to identify urinary biomarkers that could predict which patients would require no further sessions of intermittent haemodialysis (IHD) for AKI. Over a six-year period (January 2006 to December 2011) 406 patients received RRT for AKI, of whom 144 (35.5%) received IHD for ≥7 days and ≥4 sessions. After excluding patients who remained oliguric (<100ml/day urine output), had a decision to withdraw life-sustaining therapies or had a renal transplant, 67 patients were analysed. Patients were considered ‘weaned’ from IHD if they remained free of RRT for 7 days (n=37) or ‘unweaned’ if IHD continued to be provided after discharge from ICU (n=30). No weaned patient required RRT following ICU discharge, whilst only 4 (13.3%) of unweaned patients had adequate renal recovery to subsequently discontinue RRT.

The investigators were restricted to studying data that had been collected as part of standard patient care during the admission. Urine output was considered in 24 hour blocks (07:00 to 07:00). Blood and urine were sampled at 07:00 daily for concentrations of electrolytes, creatinine and urea. For convenience daily urinary urea excretion was defined as the product of urine output and urinary urea concentration.

For weaned patients results measured 2 days following the last dialysis session were analysed. These were compared with the last results prior to ICU discharge in unweaned patients. In multivariate analysis 24-hour urine output and urinary urea concentration were independent predictors of successful weaning. Daily urinary urea excretion had the greatest ability to predict weaning (AUROC 0.96, 95%CI 0.93-0.99) and was significantly superior to either urine output (AUROC 0.86) or urinary urea concentration (AUROC 0.83) alone. With a daily urinary urea excretion threshold of >1.35mmol/kg/day, sensitivity and specificity were 89.2% and 96.7% respectively.

This study corroborates previous studies which have suggested urine output can predict renal recovery [3,4] and suggests that accuracy can be improved by including urinary urea concentration in the predictive model. The study also provides a threshold that can be tested in a prospective cohort to see if using daily urinary urea excretion for identification of patients with renal recovery is feasible, and if it can translate to improved clinical outcomes and cost savings.

Article review was submitted by ESICM Journal Review Club member Richard Fisher on behalf of the AKI section.


1.    Elseviers MM et al. Renal replacement therapy is an independent risk factor for mortality in critically ill patients with acute kidney injury. Critical Care. 2010;14:R221.
2.    Aniort et al. Daily urinary urea excretion to guide intermittent hemodialysis weaning in critically ill patients. Critical Care. 2016; 20:43.
3.    Uchino S et al. Discontinuation of continuous renal replacement therapy: A post hoc analysis of a prospective multicentre observational study. Critical Care Medicine. 2009;37:2576-82.
4.    Wu VC et al. Risk factors of early redialysis after weaning from postoperative acute renal replacement therapy. Intensive Care Medicine. 2008;34:101-8.


Comment on this news