A personalized limited sampling approach to better estimate terminal half-life of FVIII concentrates

INTRODUCTION

Hemophilia A is a bleeding disorder characterized by a deficiency of a coagulation factor VIII and optimally treated using pharmacokinetics (PK)-guided prophylactic replacement therapy. To decrease patient burden, PK can be estimated from sparse sampling leveraging population PK modeling. However, recommendations for sampling times meant for patients with hemophilia A as a group may not be optimal at the individual level.

OBJECTIVE

To evaluate a personalized limited sampling approach (Personalized LSA) that suggests a next sampling time point that would provide a more accurate estimation of terminal half-life of FVIII concentrates when using a population PK approach.

METHODS

331 PK studies with rich sampling were extracted from the WAPPS-Hemo database. Two sampling approaches were evaluated and compared: 974 PK studies consisting of two samples were built from the rich sampling data including one sample selected using the personalized LSA prediction; 974 PK studies consisting of two samples were built from the rich sampling data including one sample selected randomly. Half-life values were estimated on the sparse data and compared within patients to the estimates obtained on the rich data for assessing the error on half-life values.

RESULTS

Relative errors between estimates from sparse sampling data using personalized LSA and from rich sampling data were always lower than 20% and significantly lower than the comparative approach that used random sampling (median-95th percentile were 3.8%-13.1% vs. 7.0%-23.5%, respectively, p-value < 10-10 ). Moreover, less than 4% of the samples suggested by the personalized LSA were below the limit of quantification.

CONCLUSIONS

Identifying the most informative sampling points for PK assessment using a Personalized LSA approach that accounts for individual differences in PK improves the precision of FVIII terminal half-life estimates in sparse sampling.

Population pharmacokinetics of plasma-derived factor IX: procedures for dose individualization

BACKGROUND

Population pharmacokinetic (POPPK) models describing factor IX (FIX) activity levels in plasma, in combination with individual FIX measurements, may be used to individualize dosing in the treatment of hemophilia B.

OBJECTIVES

The aim was to reevaluate a previously developed POPPK model for FIX activity and to explore the number and timing of FIX samples required in pharmacokinetic (PK) dose individualization.

METHODS

The POPPK model was reevaluated using an extended data set. Several sampling schedules, varying with respect to the timing and number of samples, were evaluated in a simulation study with relative dose errors compared between schedules. The performance of individually calculated doses was compared with commonly prescribed FIX doses with respect to the number of patients with a trough FIX activity > 0.01 U mL(-1) .

RESULTS AND CONCLUSIONS

A three-compartment PK model best described the FIX activity levels. The number and timing of samples greatly influenced imprecision in dose prediction. Schedules with single samples taken on both day 2 and day 3 were identified as being convenient schedules with an acceptable performance level. Individually calculated doses performed better with respect to patient target attainment than a fixed 40 U kg(-1) dose regardless of how many samples were available to calculate individual doses. The results of this study suggest that PK dose tailoring with limited sampling may be applicable for plasma-derived FIX products.

Dynamic prediction modeling approaches for cardiac surgery

BACKGROUND

The calibration of several cardiac clinical prediction models has deteriorated over time. We compare different model fitting approaches for in-hospital mortality after cardiac surgery that adjust for cross-sectional case mix in a heterogeneous patient population.

METHODS AND RESULTS

Data from >300 000 consecutive cardiac surgery procedures performed at all National Health Service and some private hospitals in England and Wales between April 2001 and March 2011 were extracted from the National Institute for Cardiovascular Outcomes Research clinical registry. The study outcome was in-hospital mortality. Model approaches included not updating, periodic refitting, rolling window, and dynamic logistic regression. Covariate adjustment was made in each model using variables included in the logistic European System for Cardiac Operative Risk Evaluation model. The association between in-hospital mortality and some variables changed with time. Notably, the intercept coefficient has been steadily decreasing during the study period, consistent with decreasing observed mortality. Some risk factors, such as operative urgency and postinfarct ventricular septal defect, have been relatively stable over time, whereas other risk factors, such as left ventricular function and surgery on the thoracic aorta, have been associated with lower risk relative to the static model.

CONCLUSIONS

Dynamic models or periodic model refitting is necessary to counteract calibration drift. A dynamic modeling framework that uses contemporary and available historic data can provide a continuously smooth update mechanism that also allows for inferences to be made on individual risk factors. Better models that withstand the effects of time give advantages for governance, quality improvement, and patient-level decision making.