Assessment of lipemia interference on phenobarbital quantification using particle-enhanced turbidimetric inhibition immunoassay with Dimension clinical chemistry system and comparison to high-performance liquid chromatography.

Phenobarbital, an anti-seizure medication with a narrow therapeutic window, requires therapeutic drug monitoring (TDM) for safe and effective use. Lipemia is a well-known interferent in turbidimetric immunoassays, yet its impact on phenobarbital quantification compared with high-performance liquid chromatography (HPLC) has not been fully characterized. We evaluated the effect of lipemia on phenobarbital measurement using a particle-enhanced turbidimetric inhibition immunoassay (PETINIA; Dimension Xpand Plus clinical chemistry system, Siemens) versus HPLC. Simulated lipemic plasma was prepared by spiking normal canine plasma with lipid emulsion to achieve predefined lipemia indices. Naturally lipemic TDM samples (index ≥ 4; n = 40) and non-lipemic controls (index ≤ 3, n = 40) were also analyzed. Phenobarbital concentrations were compared before and after lipid reduction by double microcentrifugation using both methods. HPLC performance was stable across all lipemia indices, with consistent retention times (5.0 min), accuracy and precision within acceptance limits, and high linearity (R² > 0.999). In contrast, PETINIA had a significant positive bias in lipemic samples, with median bias increasing from 9.1% at index 1 to 100% at index 6. Bias was significantly higher in naturally lipemic samples (median: 74.5%) compared with non‑lipemic controls (1.2%; p < 0.001). PETINIA and HPLC had strong agreement in non-lipemic samples (Passing-Bablok regression, r = 0.99), but poor agreement and proportional positive bias in lipemic samples (r = 0.16; slope CI [1.01, 1.89]). PETINIA overestimated phenobarbital concentrations in lipemic plasma. Microcentrifugation effectively mitigated lipemic interference and improved agreement with HPLC, supporting the use of microcentrifugation in clinical TDM workflows.