Comparative Predictive Accuracy for CYP3A-Mediated Drug-Drug Interactions
A 2025 study developed a mechanistic static pharmacokinetic model to predict CYP3A-mediated drug-drug interactions using the urinary 6β-OHF/cortisol ratio. When applied to 29 different perpetrator drugs, the model achieved a 73.1% success rate in predicting the area under the curve ratio (AUCR) within a 2-fold error of observed values . Notably, the prediction success rate for CYP3A inhibitors was approximately 80%, significantly outperforming the 61% success rate for inducers . In comparison, the same study explored a model using 6β-hydroxylation clearance (CLm[6β]) and achieved a 90.3% success rate across 5 perpetrators, though the limited sample size precludes a robust comparative claim .
| Evidence Dimension | Predictive accuracy of AUCR within 2-fold error for CYP3A-mediated DDIs |
|---|---|
| Target Compound Data | 73.1% (overall, n=29 perpetrators); ~80% (inhibitors); 61% (inducers) |
| Comparator Or Baseline | CLm(6β) model: 90.3% (n=5 perpetrators) |
| Quantified Difference | CLm(6β) model shows a higher success rate (90.3% vs 73.1%) but with a much smaller and potentially biased sample size (n=5 vs n=29). |
| Conditions | Mechanistic static pharmacokinetic model using literature-derived fold-change values for biomarkers and AUCR from 29 clinical DDI studies. |
Why This Matters
This quantitative evidence allows researchers to select the urinary 6β-OHF/cortisol ratio over less validated alternatives or the more invasive CLm(6β) method for broad-spectrum CYP3A inhibitor screening in early clinical development.
- [1] Takubo H, et al. Prediction of CYP3A-mediated drug-drug interactions using endogenous biomarkers 6β-hydroxycortisol and cortisol. Drug Metab Dispos. 2025;53(7):100109. View Source
