A-133 Trusted Test Goes Awry: Investigating Dilution Aberrations for Total Bilirubin

Abstract Background A patient with acute decompensated cirrhosis was hospitalized due to lightheadedness and jaundice, undergoing evaluation for a liver transplant. The attending physician brought to the laboratory’s attention a perplexing pattern in the patient's total bilirubin results. Initially, the patient displayed a consistent trend of elevated total bilirubin levels above the upper limit of linearity, ULOL ( 30 mg/dL), which then plummeted to under the limit of quantification ( 0.1 mg/dL) before surging back to 30 mg/dL. Methods The total bilirubin in heparinized plasma was measured on the AU5800 by Beckman Coulter. Samples ULOL are diluted automatically 1:3 utilizing the instrument’s on-board dilution feature. The final test results are reported automatically to an electronic medical record. Following the reported incident, an investigation and troubleshooting plan was developed, encompassing the following steps: Repeat patient testing with total bilirubin concentration 30 mg/dL and 0.1 mg/dL; Rule out the previously reported potential interferences, including Immunoglobin M and acetaminophen; Conduct a 90-day quality control (QC) review; Perform split test analysis on two instruments within the laboratory and at a reference laboratory; Conduct patient chart reviews to identify discrepancies based on previously reported values; Complete a root cause analysis and develop corrective/preventive action plan. Results Patient sample 0.1 mg/dL was tested on an alternative instrument, resulting 30 mg/dL. Patient with normal IgM of 161 mg/dL didn’t reveal any other potential drug interferences in chart review. The 3-month QC review found consistent with peer performance. Split comparison of 6 patients revealed discrepancies for samples 30 mg/dL on one of the two instrument units. When total bilirubin 30 mg/dL, the malfunctioning unit erroneously reported values 0.1 mg/dL after dilution. Review of 34 patient charts showed the impact dates were spanned for 39 days. Since the on-board dilution feature is utilized by 10 other analytes, 892 diluted samples were analyzed for discrepancies. 36 discrepant results were identified among 4 affected assays including total bilirubin, direct bilirubin, creatine kinase, and alanine aminotransferase. These results were corrected, and providers were notified. Conclusions The root cause was traced to a malfunction in one instrument unit, despite no malfunction flags. The laboratory QC concentrations fell within the analytical measurement range, missing the dilution issue during the daily QCs. During an extensive repair period, the on-board dilution was disabled on the affected unit and samples were re-routed to alternative instrument for testing. Policy changes and operator training were initiated for a new result verification process, impacting the laboratory productivity. Autoverification of diluted sample results was disabled for all chemistry analyzers. The vendor identified the issue as sample transfer unit crane misalignment in the up/down direction. It is worth noting that proficiency testing or QCs often do not extend beyond the analytical measurement ranges, leading to undetected incidents. Additional quality measures like periodic patient chart reviews, patient sample comparisons, and other quality assurance tools can aid in on-going dilution performance verification. Open communication between the physicians and laboratories is crucial for prompt issue identification and resolution.