Population pharmacodynamic modeling of zolmitriptan following administration of conventional tablets

Purpose To build a population pharmacodynamic (PD) model of zolmitriptan (ZOL) following administration of ZOL conventional tablets. Methods ZOL concentration and headache scores (HS) were obtained in 235 migraine patients from AstraZeneca studies 8311CIL/006, 311CIL/007 and 311CIL/017. NONMEM likelihood estimation method was used to build a population PD model. The individual probabilities of headache scores (HS)≤ 1 (no pain), 2 (mild pain), 3 (moderate pain), or 4 (severe pain) at various times following administration of the drug were modeled using a logistic model. The placebo (time) effect was modeled as a first‐order increase with an upper bound. A sigmoidal Emax‐type equation with a high sigmoidicity factor was chosen to mimic the step function of pain relief. Results The built population PD model yielded 78% perfect match, 10% under prediction and 12% over prediction of the observed HS, respectively. Baseline HS is a significant covariate for pain relief. Conclusions A logistic model adequately described the time courses of HS following the administration of ZOL conventional tablets. Patients with moderate baseline HS had a higher probability of improvement of pain. Demographic factors do not impact the PD relationship between HS and ZOL concentration. Clinical Pharmacology & Therapeutics (2005) 77 , P90–P90; doi: 10.1016/j.clpt.2004.12.237