Background:

Falsely elevated plasma creatinine due to glucose and ketone interference is well documented in routine Jaffe assay (1,2). Although Jaffe assays have been recently standardised since the introduction of eGFR, the effect of this interference is unknown since the alignment process.  We assessed plasma creatinine in DKA patients using 4 Jaffe assays against an enzymatic method as the gold standard.

Methods:

Samples were collected from DKA patients presenting to the Emergency Department and ten outpatients with normal glucose and bicarbonate. Aliquots were frozen, transported on dry ice, and analysed on 4 Jaffe assays (Architect, Roche, Beckman Coulter and Siemens) and the Vitros Enzymatic assay. Assay bias (Jaffe creatinine – enzymatic creatinine)/ enzymatic creatinine x 100%) was correlated with pH, glucose, ketone and baseline creatinine.

Results:

Of the 22 episodes of DKA from 20 patients (16 M, 4 F, age 32.4 ± 13.4 years), admission venous pH was 7.14 ± 0.15, bicarbonate was 9.4 ± 5.5 mmol/L, glucose was 35.9 ± 14.9 mmol/L and capillary ketone was 5.9 ± 1.1 mmol/L. The average enzymatic creatinine value was 111.5 ± 64.1 µmol/L. All Jaffe creatinine results were significantly higher (p <0.001), in order of magnitude: Abbott (154.6 ± 68.2 µmol/L), Beckman Coulter (147.1 ± 70.5 µmol/L), Roche (128.0 ± 56.4 µmol/L) and Siemens (127.0 ± 59.4 µmol/L). β-hydroxybutyrate and bicarbonate did not correlate with creatinine bias, whereas baseline enzymatic creatinine correlated with all Jaffe methods (p < 0.001) and glucose was negatively correlated (p< 0.05). No significant creatinine bias was detected in control subjects.

Discussion:

The routine Jaffe methods led to falsely elevated plasma creatinine of up to 39% in DKA patients. This interference was prominent in lean female patients whereby the baseline creatinine was low and the bias was up to 58%. A falsely high creatinine can potentially influence fluid and potassium replacement.