BACKGROUND: Although false elevation of automated leukocyte (white blood cell [WBC]) counts has been described in the setting of hip corrosion, the more general correlation between manual and automated cell counts among synovial fluid aspirates from hip and knee arthroplasties has not been studied. METHODS: This retrospective review at one laboratory identified 8607 consecutive synovial fluid samples from arthroplasties and 812 from native knees, each with an automated WBC count > 3000 cells/μL and a corresponding paired reflex manual count. The correlation between automated and manual counts was evaluated, as was the rate of false-positive automated WBC counts. RESULTS: The correlation between automated WBC counts for native knees, total knee arthroplasties, and total hip arthroplasties was near-perfect, strong, and moderate, respectively. The false-positive rates for automated counts were 4.4%, 10.1%, and 34.3%, respectively (P < .0001). International Consensus Meeting scores and culture positivity demonstrated that manual counts, not automated counts, were correct. CONCLUSION: The presence of a hip or knee arthroplasty appears to substantially increase the risk of a false-positive automated synovial fluid WBC count. Clinicians evaluating an arthroplasty should exercise caution when interpreting positive automated WBC counts, and consider requesting a reflex manual count, to verify the accuracy of automated cell counting.
BACKGROUND: Although false elevation of automated leukocyte (white blood cell [WBC]) counts has been described in the setting of hip corrosion, the more general correlation between manual and automated cell counts among synovial fluid aspirates from hip and knee arthroplasties has not been studied. METHODS: This retrospective review at one laboratory identified 8607 consecutive synovial fluid samples from arthroplasties and 812 from native knees, each with an automated WBC count > 3000 cells/μL and a corresponding paired reflex manual count. The correlation between automated and manual counts was evaluated, as was the rate of false-positive automated WBC counts. RESULTS: The correlation between automated WBC counts for native knees, total knee arthroplasties, and total hip arthroplasties was near-perfect, strong, and moderate, respectively. The false-positive rates for automated counts were 4.4%, 10.1%, and 34.3%, respectively (P < .0001). International Consensus Meeting scores and culture positivity demonstrated that manual counts, not automated counts, were correct. CONCLUSION: The presence of a hip or knee arthroplasty appears to substantially increase the risk of a false-positive automated synovial fluid WBC count. Clinicians evaluating an arthroplasty should exercise caution when interpreting positive automated WBC counts, and consider requesting a reflex manual count, to verify the accuracy of automated cell counting.