No abstract

Background Patients with severe and moderate haemophilia A are treated prophylactically with factor VIII (FVIII) concentrate. Individualization of prophylaxis can be achieved by pharmacokinetic (PK)-guided dosing. Aim In this study, the performance of three PK tools (myPKFiT, Web-Accessible Population Pharmacokinetic Service-Hemophilia [WAPPS] and NONMEM) is compared. Methods In 39 patients, with severe or moderate haemophilia A, blood samples were collected 4, 24 and 48 hours after administration of 50 IU kg−1 of recombinant FVIII (Advate [n = 30] or Kogenate [n = 9]). FVIII dose, FVIII activity and patient characteristics were entered into the three PK tools. Obtained PK parameters and dosing advises were compared. Results myPKFiT provided PK parameters for 24 of 30 patients receiving Advate, whereas WAPPS and NONMEM provided estimates for all patients. Half-life was different among the three methods: medians were 12.6 hours (n = 24), 11.2 hours (n = 30) and 13.0 hours (n = 30) for myPKFiT, WAPPS and NONMEM (p < 0.001), respectively. To maintain a FVIII trough level of 0.01 IU mL−1 after 48 hours, doses for myPKFiT and NONMEM were 15.1 and 11.0 IU kg−1 (p < 0.01, n = 11) and for WAPPS and NONMEM were 9.0 and 8.0 IU kg−1 (p < 0.01, n = 23), respectively. In nine patients receiving Kogenate, WAPPS and NONMEM produced different PK-parameter estimates; half-life was 15.0 and 12.3 hours and time to 0.05 IU mL−1 was 69.2 and 60.8 hours, respectively (p < 0.01, n = 9). However, recommended doses to obtain these levels were not different. Conclusion The three evaluated PK tools produced different PK parameters and doses for recombinant FVIII. Haematologists should be aware that recommended doses may be influenced by the choice of PK tool.

Background: Osimertinib is the cornerstone in the treatment of epidermal growth factor receptor-mutated non-small cell lung cancer (NSCLC). Nonetheless, ±25% of patients experience severe treatment-related toxicities. Currently, it is impossible to identify patients at risk of severe toxicity beforehand. Therefore, we aimed to study the relationship between osimertinib exposure and severe toxicity and to identify a safe toxic limit for a preventive dose reduction. Methods: In this real-life prospective cohort study, patients with NSCLC treated with osimertinib were followed for severe toxicity (grade ⩾3 toxicity, dose reduction or discontinuation, hospital admission, or treatment termination). Blood for pharmacokinetic analyses was withdrawn during every out-patient visit. Primary endpoint was the correlation between osimertinib clearance (exposure) and severe toxicity. Secondary endpoint was the exposure–efficacy relationship, defined as progression-free survival (PFS) and overall survival (OS). Results: In total, 819 samples from 159 patients were included in the analysis. Multivariate competing risk analysis showed osimertinib clearance ( c.q. exposure) to be significantly correlated with severe toxicity (hazard ratio 0.93, 95% CI: 0.88–0.99). An relative operating characteristic curve showed the optimal toxic limit to be 259 ng/mL osimertinib. A 50% dose reduction in the high-exposure group, that is 25.8% of the total cohort, would reduce the risk of severe toxicity by 53%. Osimertinib exposure was not associated with PFS nor OS. Conclusion: Osimertinib exposure is highly correlated with the occurrence of severe toxicity. To optimize tolerability, patients above the toxic limit concentration of 259 ng/mL could benefit from a preventive dose reduction, without fear for diminished effectiveness.

Nonlinear mixed effect (NLME) models are the gold standard for the analysis of patient response following drug exposure. However, these types of models are complex and time‐consuming to develop. There is great interest in the adoption of machine‐learning methods, but most implementations cannot be reliably extrapolated to treatment strategies outside of the training data. In order to solve this problem, we propose the deep compartment model (DCM), a combination of neural networks and ordinary differential equations. Using simulated datasets of different sizes, we show that our model remains accurate when training on small data sets. Furthermore, using a real‐world data set of patients with hemophilia A receiving factor VIII concentrate while undergoing surgery, we show that our model more accurately predicts a priori drug concentrations compared to a previous NLME model. In addition, we show that our model correctly describes the changing drug concentration over time. By adopting pharmacokinetic principles, the DCM allows for simulation of different treatment strategies and enables therapeutic drug monitoring.

Background Patients with severe hemophilia B regularly administer prophylactic intravenous doses of clotting factor IX concentrate to maintain a trough level of at least 0.01 IU mL in order to prevent joint bleeds. Assessment of individual pharmacokinetic (PK) parameters allows individualization of the recombinant factor IX (rFIX) dose. Aim To evaluate the predictive performance of limited sampling strategies (LSSs) with one to three samples to estimate individual PK parameters of rFIX. Methods Monte Carlo simulations were performed to obtain 5000 concentration-time profiles by the use of population PK parameters for rFIX from literature. Eleven LSSs were developed with one, two or three samples taken within an 80-h interval following administration of 100 IU kg rFIX. Clearance (CL), half-life (t ), time to 1% and steady-state distribution volume (V ) were estimated for each simulated individual by the use of Bayesian analysis. Results For each LSS, average bias was small for CL (range - 1.5% to 1.4%), t (range - 4.5% to - 0.7%), time to 1% (range - 2.9% to 0%), and V (range - 3.7% to 0.3%). Imprecision for these parameters ranged from 6.4% to 11.9%, from 10.3% to 15.6%, from 7.3% to 10.9%, and from 9% to 20.1%, respectively. The best predictive performance was achieved with one sample taken between 10 min and 3 h and two samples taken between 48 h and 56 h after administration of rFIX. Conclusions This study demonstrates that limited sampling strategies, used for individualized dosing of rFIX in hemophilia B patients, can be developed and evaluated by in silico simulation.