The effect of glucocorticoids on serum cystatin C in identifying acute kidney injury: a propensity-matched cohort study.

ABSTRACT

BACKGROUND: Glucocorticoids may impact the accuracy of serum cystatin C (sCysC) in reflecting renal function. We aimed to assess the effect of glucocorticoids on the performance of sCysC in detecting acute kidney injury (AKI) in critically ill patients. METHODS: A prospective observational cohort study was performed in a general intensive care unit (ICU). Using propensity score matching, we successfully matched 240 glucocorticoid users with 960 non-users among 2716 patients. Serum creatinine (SCr) and sCysC were measured for all patients at ICU admission. Patients were divided into four groups based on cumulative doses of glucocorticoids within 5 days before ICU admission (Group I non-users; Group II 0 mg < prednisone ≤50 mg; Group III 50 mg < prednisone ≤150 mg; Group IV prednisone > 150 mg). We compared the performance of sCysC for diagnosing and predicting AKI in different groups using the area under the receiver operator characteristic curve (AUC). RESULTS: A total of 240 patients received glucocorticoid medication within 5 days before ICU admission. Before and after matching, the differences of sCysC levels between glucocorticoid users and non-users were both significant (P <  0.001). The multiple linear regression analysis revealed that glucocorticoids were independently associated with sCysC (P <  0.001). After matching, the group I had significantly lower sCysC levels than the group III and group IV (P <  0.05), but there were no significant differences in sCysC levels within different glucocorticoids recipient groups (P > 0.05). Simultaneously, we did not find significant differences in the AUC between any two groups in the matched cohort (P > 0.05). CONCLUSIONS: Glucocorticoids did not impact the performance of sCysC in identifying AKI in critically ill patients.