Evaluation of a handheld point-of-care analyser for measurement of creatinine in cats. Issue 2 (February 2017)

• Record Type: Journal Article
• Authors: Reeve, Jenny; Warman, Sheena; Lewis, Daniel; Watson, Natalie; Papasouliotis, Kostas
• Is Part Of: Journal of feline medicine and surgery. Volume 19:Issue 2(2017)
• Page Start: 207
• Page End: 215
• Publication Date: 2017-02
• Subjects: Cats -- Diseases -- Periodicals; Cats -- Surgery -- Periodicals; 636.8089
• DOI: 10.1177/1098612X15622676
• Languages: English
• ISSNs: 1098-612X

Abstract

Objectives: The aim of the study was to evaluate whether a handheld creatinine analyser (StatSensor Xpress; SSXp), available for human patients, can be used to measure creatinine reliably in cats. Methods: Analytical performance was evaluated by determining within- and between-run coefficient of variation (CV, %), total error observed (TEobs, %) and sigma metrics. Fifty client-owned cats presenting for investigation of clinical disease had creatinine measured simultaneously, using SSXp (whole blood and plasma) and a reference instrument (Konelab, serum); 48 paired samples were included in the study. Creatinine correlation between methodologies (SSXp vs Konelab) and sample types (SSXpwhole blood vs SSXpplasma ) was assessed by Spearman's correlation coefficient and agreement was determined using Bland–Altman difference plots. Each creatinine value was assigned an IRIS stage (1–4); correlation and agreement between Konelab and SSXp IRIS stages were evaluated. Results: Within-run CV (4.23–8.85%), between-run CV (8.95–11.72%), TEobs (22.15–34.92%) and sigma metrics (⩽3) did not meet desired analytical requirements. Correlation between sample types was high (SSXpwhole blood vs SSXpplasma ; r = 0.89), and between instruments was high (SSXpwhole blood vs Konelabserum ; r = 0.85) to very high (SSXpplasma vs Konelabserum ; r = 0.91). Konelab and SSXpwhole blood IRIS scores exhibited high correlation ( r = 0.76). Packed cell volume did not significantly affect SSXp determination of creatinine. Bland–Altman difference plots identified a positive bias for the SSXp (7.13 μmol/l SSXpwhole blood ; 20.23 μmol/l SSXpplasma ) compared with the Konelab. Outliers (1/48 whole blood; 2/48 plasma) occurred exclusively at very high creatinine concentrations. The SSXp failed to identify 2/21 azotaemic cats. Conclusions and relevance: Analytical performance of the SSXp in feline patients is not considered acceptable. The SSXp exhibited a high to very high correlation compared with the reference methodology but the two instruments cannot be used interchangeably. Improvements in the SSXp analytical performance are needed before its use can be recommended in feline clinical practice.