Clinician assessment of kidney function from plasma creatinine values during critical illness: A scenario-based...

During critical illness interpretation of serum creatinine is affected by non-steady state conditions, reduced creatinine generation, and altered distribution. We evaluated healthcare professionals' ability to adjudicate underlying kidney function, based on simulated creatinine values.

We developed an online survey, incorporating 12 scenarios with simulated trajectories of creatinine based on profiles of muscle mass, GFR and fluid balance using bespoke kinetic modelling. Participants predicted true underlying GFR (<5, 5-14, 15-29, 30-44, 45-59, 60-90, >90 ml.min^-1.1.73 m^-2) and AKI stage (stages 1-3, defined as 33 %, 50 %, 66 % decrease in GFR from baseline) during the first 7-days and at ICU discharge.

100 of 103 respondents from 16 countries, 94 completed 1 or more scenarios. 43(43 %) were senior physicians, 74(74 %) critical care and 31(31 %) nephrology physicians. Over the first 7-days, true GFR was correctly estimated 43 % of the time and underlying AKI stage in 57 % of patient days. At ICU discharge GFR was predicted 35 % of the time. At all timepoints, over and under-estimation of GFR was observed.

Participants displayed marked variation in estimation of kidney function, suggesting difficulty in accounting for multiple confounders. There is need for alternative, unbiased measures of kidney function in critical illness to avoid misclassifying kidney disease.