Cyclosporine A area-under-time-concentration-curve in rheumatologic patients after the first dose. Which of the sparse sampling strategies will predict the best?

Objective: The aim of the present study was to validate the limited sampling strategies (LSS:s) for prediction of AUC of cyclosporine A (CsA) after the first dose in rheumatologic patients. Methods: 22 patients suffering from rheumathoid arthritis, systemic lupus erythematodus, ankylosing spondylitis dermato(poly)myositis or seronegative spondylarthritis were treated with Neoral(R) (female/male: 11/3, mean +/- SD: age 49 +/- 14 y, body weight 75 +/- 12 kg, height 166 +/- 7 cm, dose 71 +/- 25 mg, dose per kg 1.0 +/- 0.3 mg/kg), or Consupren(R) (7/1, 78 +/- 36, 175 +/- 8, 82 +/- 22, 1.1 +/- 0.3). Two patients whose C-12h were missing were excluded from the AUC(0-12) calculation. Whole blood levels of CsA were analyzed with HPLC. Blood samples were collected at 0, 0.5, 1, 1.5, 2, 3, 4, 6, 8, and 12 hours after taking the first dose. Altogether 115 LSS:s obtained from the literature were validated. A linear trapezoidal rule was used as a reference method. Mean percentage prediction error (%PE) < +/- 15% and maximal one value of absolute %PE > 30% were considered to be acceptable. The root mean squared error (RMSE) was evaluated for equations that passed the criteria. Results: The best performance with all values of the absolute %PE < 30% was found in three LSS:s for AUC(0-12) and two for AUC(0-8): AUC(0-12) = 123.792 + 1.165 x C-1h + 3.021 x C-3h + 7.33 x C-8h; 97.6 + 1.27 x C-1h + 3.14 x C-3h + 4.06 x C-6h; or 124.3 + 1.34 x C-1h - 0.16 x C-2h 3.27 x C-3h + 3.96 x C-6h; AUC(0-8) = -19.8 + 1.99 x C-2h + 2.38 x C-4h + 3.15 x C-6h or -22.4 + 2.51 x C-2h + 5.49 x C-6h. Validation criteria were further fulfilled in AUC(0-12) = 24 + 3.66 x C-0h + 2.11 x C-1.5h + 4.54 x C-4h or 0.2 + 2 x C-2h + 10.2 x C-6h; AUC(0-8) = 55.37 + 2.89 x C-0h + 1.08 x Cl + 0.9 x C-2h 2.23 x C-3h; and AUC(0-4) = -41 + 1.17 x C-1h + 1.85 x C-2h. Only one equation proposed for AUC(0-6) did not pass the validation criteria. Conclusions: Equations validated for prediction of AUC(0-12), AUC(0-8) and AUC(0-4) might be used for LSS:s of CsA independently of the length of treatment, indication, dosage or galenic: formulation.