Urinary metabolites predict mortality or need for renal replacement therapy after combat injury.

BACKGROUND

Traditionally, patient risk scoring is done by evaluating vital signs and clinical severity scores with clinical intuition. Urinary biomarkers can add objectivity to these models to make risk prediction more accurate. We used metabolomics to identify prognostic urinary biomarkers of mortality or need for renal replacement therapy (RRT). Additionally, we assessed acute kidney injury (AKI) diagnosis, injury severity score (ISS), and AKI stage.

METHODS

Urine samples (n = 82) from a previous study of combat casualties were evaluated using proton nuclear magnetic resonance (

RESULTS

Eighty-four (84) metabolites were identified and quantified in each urine sample. Of these, 11 were identified as drugs or drug metabolites and excluded. The PLS-DA models for ISS and AKI diagnosis did not have acceptable model statistics. Therefore, only mortality/RRT and AKI stage were analyzed further. Of 73 analyzed metabolites, 9 were significantly associated with mortality/RRT (p < 0.05) and 11 were significantly associated with AKI stage (p < 0.05). 1-Methylnicotinamide was the only metabolite to be significantly associated (p < 0.05) with all outcomes and was significantly higher (p < 0.05) in patients with adverse outcomes. Elevated lactate and 1-methylnicotinamide levels were associated with higher AKI stage and mortality and RRT, whereas elevated glycine levels were associated with patients who survived and did not require RRT, or had less severe AKI. ROC curves for each of these metabolites and the combined panel had good predictive value (lactate AUC = 0.901, 1-methylnicotinamide AUC = 0.864, glycine AUC = 0.735, panel AUC = 0.858).

CONCLUSIONS

We identified urinary metabolites associated with AKI stage and the primary outcome of mortality or need for RRT. Lactate, 1-methylnicotinamide, and glycine may be used as a panel of predictive biomarkers for mortality and RRT. 1-Methylnicotinamide is a novel biomarker associated with adverse outcomes. Additional studies are necessary to determine how these metabolites can be utilized in clinically-relevant risk prediction models.