146: Clinical PK Sampling Strategies, Precise Accurate Estimates of Population Clearance. Is Trough Sampling Enough? Practical NOT Optimal Sampling

Statement of Purpose: To inform a planned phase 3 study design, stochastic simulation and estimation (SSE) exercises were performed. Different PK sampling strategies that would provide a reliable estimate of the apparent population clearance (CL/F) were evaluated. A final proposal was made to support an adequate sampling strategy based on the true uncertainty to provide a reliable estimate of clearance that was clinically advantageous.

Description of Methods & Materials: A Population PK model had been developed using PK data collected from phase 1 studies (2 compartment PK with the first order absorption and elimination). The simulated dosing scheme included gradual dose titrations until the final dose and steady-state was achieved. Evaluated PK sampling strategies included S1, two sets of one trough sample followed by 2-hour post dose sample, one week after the new dose during the dose titration period (i.e., 2 trough plus 2 post dose samples over 2 PK collection days); S2, 2-hour post dose sample on Day 1, followed by three trough samples during the dose titration (one post dose sample plus three trough samples, over 4 sample collection days); and S3 (Trough only): four trough samples prior to each dose level during the titration (4 trough samples over 4 PK collection days). Each scheme consisted of four PK samples per subject, 60 subjects in total. Using SSE output, relative root mean square error (RMSE) of estimation error (ee), (where ee = estimated-true)/true), as well as relative bias, the mean of ee, were compared as an assessment for precision and accuracy for CL/F estimates, respectively. Perl-speaks-NONMEM (v5.2.6) in conjunction with NONMEM (v7.4.3) was used for SSE tool, and 1,000 samples were repeated.

Data & Results: Both relative RMSE (%) and the magnitude of relative bias (%) decreased with the addition of at least one post-dose sample, compared to trough only sampling (Table 1). The boxplot of ee (Figure 1) also supported the improvement in precision and accuracy.

Interpretation, Conclusion or Significance: Compared to trough only sampling, precision and accuracy improved with an addition of post-dose sample. S1 was more clinical advantageous as it requires only 2 visits for PK data collection than S2. Consideration for clinical feasibility in concert with precision and bias in parameter estimates, were also used to make the final clinical recommendation for PK sampling to be used in the Phase 3 study.

Disclosures: