(T1130-11-63) Prediction of Rifampin Exposure Using a Single Concentration Point in Patients with Tuberculosis

Purpose: The plasma concentrations of rifampin (Cp) are highly variable when measured in patients with tuberculosis (TB) after fixed doses, which can result in unpredictable treatment outcome and adverse drug reactions. Recent studies have suggested performing the therapeutic drug monitoring (TDM) to optimize the dosage regimen of rifampin based on the area under the concentration-time curve of rifampin (AUC). However, the AUC prediction requires extensive serial blood sampling, which is inconvenient and reluctant for both patients and clinicians. This study investigated a single time point of Cp that would be the best to predict the AUC in TB patients.

Methods: The rifampin concentration-time data were obtained from the TB patients who visited the outpatient clinic of respiratory disease. The study patients had been receiving a fixed daily dose of rifampin 450mg (for bodyweight < 50 kg) or 600 mg (for bodyweight ≥ 50 kg). Serial blood samples were collected from the patients’ arm vein to measure the Cp at pre-dose, and 0.5, 1, 2, 4, 6, 9, 12 and 24 hours after the last dose of rifampin. The plasma was separated from the blood and frozen at -70℃ until concentration measurement. The Cp was measured using a high-performance liquid chromatographic method with tandem mass spectrometric detection. The Cp at 3 hours post dose (C₃) that had not been collected was estimated by taking the arithmetic mean of Cp at 2 hours (C₂) and 4 hours (C₄). The pharmacokinetic parameter values of rifampin were determined using a non-compartmental method. The AUC_obs was defined as the calculated AUC of rifampin from time zero to infinity. The values of AUC were predicted (AUC_pred) using the formulas obtained from a linear regression based on the Cp at each time point of blood sampling (e.g., AUC_pred,2h, AUC_pred,3h, etc). Passing-Bablock regression was used to determine the conformity between AUC_obs and AUC_pred. Bland-Altman analysis was used to assess the accuracy and precision of AUC_pred by computing the mean bias and 95% limits of agreement (LoA) of the difference between AUC_obs and AUC_pred.

Results: A total of twenty-six patients were included in the analyses among thirty-three patients who were enrolled in the study. Among the twenty-six patients, twenty-two had been receiving rifampin 600 mg and four 450 mg. Excluded were six patients whose terminal phase of the Cp-time curve could not be determined (e.g., below the lower limit of quantification) to calculate AUC_obs and one with unanalyzable Cp-time curve. The median values (interquartile range, IQR) of maximum concentrations (C_max), time to C_max (T_max), AUC_obs, and t_1/2 were 11.3 (7.01 - 20.8) μg/mL, 2 (1 - 2) hours, 50.4 (30.4 - 76.8) μg*h/mL, and 1.83 (1.45 - 2.09) hours, respectively. In linear regression, C₃ predicted the AUC_obs of rifampin best (r² = 0.95, p < 0.0001) followed by C₄ (r² = 0.91, p < 0.0001) and C­₂ (r² = 0.84, p < 0.0001). The formulas were AUC_pred = 5.23*C₃ + 6.19 (p < 0.0001 and = 0.045 for slope and intercept, respectively), 7.04*C₄ + 5.61 (p < 0.0001 and = 0.18) and 3.59*C₂ + 13.6 (p < 0.0001 and = 0.01) for the C₃, C₄ and C₂, respectively. In Passing-Bablock regression, C₃ demonstrated the best conformity between AUC_obs and AUC_pred with the slope and intercept (95% confidence interval, 95% CI) of 1.04 (0.90 to 1.18) and -1.36 (-9.5 to 3.9), respectively. The slope and intercept (95% CI) for AUC_pred,4h and AUC_pred,2h were 1.09 (0.91 to 1.26) and -2.21 (-9.97 to 3.7), and 1.21 (0.89 to 1.47) and -10.9 (-22.4 to 2.70), respectively. Bland-Altman plot demonstrated the random scattering of the points of AUC_obs-AUC_pred. The accuracy (i.e., mean bias) of AUC_pred,3h, AUC_pred,4h and AUC_pred,2h was 0.041, 0.012 and 0.042 μg*h/mL, respectively. The precision (i.e., 95% LoA) -15.6 to 15.7, -21.1 to 21.2 and -27.6 to 27.7 μg*h/mL, respectively.

Conclusion: Based on the results, the concentration at three hours post dose predicted the AUC of rifampin best with the highest coefficient of determination, conformity and precision. The predicted exposure (i.e., AUC) of rifampin using the concentration at three hours post dose is applicable to the TDM of rifampin for improving its effectiveness and minimizing adverse drug reactions in patients with tuberculosis.

Acknowledgements: We thank all study participants and staff at clinic of respiratory disease, Bundang Hospital of Seoul National University, Republic of Korea.