OBJECTIVE: To measure free phenytoin (FP) concentrations in filtered specimens using the Abbott ARCHITECT iPhenytoin assay and to compare results from this method with results from the Abbott TDx/FLx assays. METHODS: We verified accuracy, analytic measurement range, and precision for FP measurements. For correlation and therapeutic interval studies, we used filtered calibrators, controls, proficiency-testing materials, and surplus clinical samples. After implementation, we determined proficiency testing results. RESULTS: The analytic measurement range was 2.0 to 25.0 micromol/L. Quality control materials (6.1, 12.6, and 20.1 micromol/L) provided mean (SD) recoveries of 96.1 (5.0%), 99.2 (5.0%), and 99.3 (5.7%), respectively, and coefficients of variation of 5.2%, 5.0%, and 5.8%, respectively. Clinical specimens produced mean (SD) FP recovery levels of 103.7 (10.6%) (bias, 0.1 [0.3] micromol/L). Altering the FP therapeutic range (4.0-8.0 micromol/L) was unnecessary. Proficiency testing yielded consistently acceptable results. CONCLUSION: Our accuracy, precision, and correlation results were similar for the TDx/FLx and ARCHITECT assays, which demonstrates that the ARCHITECT iPhenytoin assay is acceptable for clinical FP measurements.
OBJECTIVE: To measure free phenytoin (FP) concentrations in filtered specimens using the Abbott ARCHITECT iPhenytoin assay and to compare results from this method with results from the Abbott TDx/FLx assays. METHODS: We verified accuracy, analytic measurement range, and precision for FP measurements. For correlation and therapeutic interval studies, we used filtered calibrators, controls, proficiency-testing materials, and surplus clinical samples. After implementation, we determined proficiency testing results. RESULTS: The analytic measurement range was 2.0 to 25.0 micromol/L. Quality control materials (6.1, 12.6, and 20.1 micromol/L) provided mean (SD) recoveries of 96.1 (5.0%), 99.2 (5.0%), and 99.3 (5.7%), respectively, and coefficients of variation of 5.2%, 5.0%, and 5.8%, respectively. Clinical specimens produced mean (SD) FP recovery levels of 103.7 (10.6%) (bias, 0.1 [0.3] micromol/L). Altering the FP therapeutic range (4.0-8.0 micromol/L) was unnecessary. Proficiency testing yielded consistently acceptable results. CONCLUSION: Our accuracy, precision, and correlation results were similar for the TDx/FLx and ARCHITECT assays, which demonstrates that the ARCHITECT iPhenytoin assay is acceptable for clinical FP measurements.