Good Morning
  This week we review an interesting article related to the diagnosis of AKI using an easily measured parameter. Identifying AKI at KDIGO > 1 stage is difficult. However, early identification helps to minimize the morbidity associated with Kidney Injury and RRT. Several biomarkers are available and several more are under investigation. Burns et al used K excretion and compared it to  creatinine clearance and attempted to correlate this with a diagnosis of AKI.
Urinary potassium excretion and its association with acute kidney injury in the intensive care unit.
Burns et al
Journal of Critical Care 46 (2018) 58–62

This is a prospective cohort study which attempted to measure 2 hr K excretion and correlate it with conventionally measured creatinine clearance. The data obtained was used to predict AKI over the subsequent 7 days in ICU. The basic hypothesis was that K excretion is reduced in AKI and could be used to discriminate those who developed AKI over the next 7 days from those who did not. Urine sample collected between 0400-0600 hrs was analysed for K Na and Creatinine concentrations. Around 60 patients were enrolled in the study. 62% were male. Post operative and trauma patients made up close to 50% of the cohort. A quarter of the cohort needed vasopressors. The mean APACHE II score of the cohort was 15. The total amount of urinary potassium excreted in 2-h was calculated by multiplying the 2-hour urine sample potassium concentration (mmol/l) by the urine volume (litres in the 2 h collection period) to give m mol of potassium excreted in 2 h. All patients were followed up to hospital discharge capturing data on the peak plasma creatinine concentration within 7 days of enrolment and during the whole hospital stay, KDIGO AKI grading after enrolment, RRT, and mortality. In patients who did not receive frusemide the urinary potassium excretion correlated linearly with the simultaneously calculated creatinine clearance. Frusemide seemed to be decrease this correlation. This correlation was found to be better that Na excretion or Renal SOFA for the prediction on AKI.
           Based on the AUROC curve of urinary potassium excretion, using a cut-point of urinary potassium excretion ≤3.8mmol in 2 h would have a specificity of 85% and sensitivity of 77%  in predicting subsequent AKI (KDIGO stage ≥1) within 7 days of testing. This is different from the relationship between urinary sodium excretion and the calculated Cr Cl or risk of subsequent
AKI.

What it means: This study gives an easily measurable parameter to predict the onset of AKI in a cohort of patients at risk of AKI. However, it remains to be seen whether it can be applied to patients with higher APACHE scores. The effect of augmented renal clearance on this parameter also needs to be evaluated