Hemolysis detection with GEM Premier 7000 in whole-blood specimens from patients undergoing hemodialysis: an observational study

Abstract Objectives To evaluate the prevalence and degree of hemolysis in whole-blood specimens collected for blood gas analysis after routine hemodialysis and to investigate potential demographic, clinical, and procedural determinants of the hemolysis index (HI). Methods This single-center observational study included consecutive patients undergoing maintenance hemodialysis at a tertiary-care hospital between December 2025 and January 2026. Whole-blood specimens were collected after completion of the dialysis session and analyzed using the GEM Premier 7000 blood gas analyzer, which incorporates an automated hemolysis detection module. Associations between HI values and patient- and procedure-related variables were assessed using a linear mixed-effects model accounting for repeated measurements within individuals. Results A total of 438 specimens were collected from 116 patients, of which 435 were available for HI analysis. The median HI was 5 mg/dL (IQR, 2–7 mg/dL), and no specimen exceeded the clinically interference threshold of 116 mg/dL, defined as the manufacturer-specified cutoff above which hemolysis may cause clinically relevant interference with potassium measurement. Higher hemoglobin concentration, and arteriovenous fistula compared with tunneled dialysis catheter, were independently associated with higher HI values. No significant associations were observed for other demographic, clinical, or procedural variables. Conclusions Whole-blood specimens collected after routine maintenance hemodialysis displayed no clinically significant hemolysis. These findings support the analytical reliability of contemporary post-dialysis blood gas sampling and suggest that automated hemolysis detection may represent a useful quality assurance tool in this setting.