Interview with Irina Bondareva, Research Institute of Physical–Chemical Medicine, Moscow, Russia

This month’s interview comes from Moscow from Irina Bondareva, a colleague who is a mathematician, but who none the less influences decisions for patients based on Therapeutic Drug Monitoring combined with modelling. I first heard about Irina’s modelling work with early anti-epileptic drugs at the Stuttgart congress, and recently took the opportunity to get in touch with a query. It’s been a pleasure to find out about Irina’s 30-year experience in Pharmacometrics and PK/PD modelling, which is used for personalising treatment in complex cases, primarily patients with epilepsy.


Irina Bondareva
Research Institute of Physical–Chemical Medicine
Moscow, Russia


How did you become interested in your area of expertise?

With a background in Applied Mathematics, I have been working in medical research at the Research Institute of Physical-Chemical Medicine in Moscow for almost 30 years. I started mathematical modelling in nuclear medicine, and in 1990 my PhD thesis was “A Non-linear Mathematical Model for in-vivo Evaluation of the Reticuloendothelial System Phagocytic Function”. Later in the 90’s I switched to PK/PD modelling. The mathematical models in these two fields are quite similar in general, and a background in nuclear medicine was an advantage. During those years it was an honour and a great experience for me to work in collaboration with Prof. Roger Jelliffe and his colleagues from the USC Laboratory of Applied Pharmacokinetics (LAPK) of the Keck School of Medicine at the University of Southern California. That was when I became enthusiastic about optimizing drug therapy for individual patients based on population models and TDM data. Since 1995, I’ve been using USCPACK software from LAPK for processing TDM data. One of the fruits of that collaboration was a two day workshop on “Principles of Pharmacokinetics – Population PK and PD Modelling: Applications to Therapeutic Drug Monitoring and to Optimal Individualization of Drug Therapy” in Moscow and Saint-Petersburg (Russia), which Professor Jelliffe and I coordinated annually from 2000 till 2008. Thanks to that intensive program, physicians and clinical pharmacologists from Russia and CIS countries had an opportunity to improve their knowledge of PK/PD concepts and find out about recent advances in TDM and optimizing drug therapy. This workshop also contributed to the popularization of TDM among medical professionals in Russia.

My research in TDM relates to the population modelling and dosage individualizing for antiepileptic drugs, mostly the so-called old AEDs, still widely used in Russia. Epilepsy is a chronic disease requiring long-term treatment. It is very important that therapy be individualized with respect to correct drug choice and appropriate drug dosage. Antiepileptic drugs have widely varying pharmacokinetic and pharmacodynamic patient responses, and TDM based only on serum drug concentrations without software for adaptive control of dosage is frequently ineffective, even for AEDs with well-known therapeutic ranges.

Through experience we know that even well established general therapeutic ranges cannot guarantee freedom of seizures for all patients. Selection of a specific therapeutic goal based on an individual patient’s clinical need for the drug and acceptable risk of toxicity, and calculation of optimal dosage regimens are important for effective therapy. Many investigators now support the individualization of anticonvulsant dosage regimens for better seizure control. There is a significant movement away from rigid therapeutic ranges toward “individual therapeutic concentration” among physicians concerned with epilepsy management.

Patient data of anticonvulsant monitoring have been routinely collected in several PK services in Russia since the 90’s. Usually adult and paediatric epileptic outpatients from different epilepsy clinics attended these consultations to evaluate potential reasons for lack or loss of efficacy and/or for toxicity of their antiepileptic therapy, as well as to establish or to check their “baseline” effective concentrations during a period of remission. These relatively rich TDM data (peak-trough sampling strategy for all patients, 1 – 3 additional samples in other dosing intervals for some patients, on multiple occasions in some patients) of older AEDs were used to develop the population models that we use for individualization of regimens.

In 2001 I obtained a Doctor of Bioscience Degree in Mathematical Modelling and Clinical Pharmacology for this research. My modelling experience includes: 1) traditional one-compartment linear PK models to describe VPA and CBZ behaviour in the post-induction period; 2) nonparametric population PK analysis and simultaneous fitting, performed using the Pmetrics software (LAPK), based on a compartmental model with first-order absorption and linear elimination kinetics for both CBZ and its primary metabolite CBZE; 3) parameters of a developed nonlinear model for CBZ time-dependent pharmacokinetic behaviour during its auto-induction period, estimated from TDM data of drug-naive adult epileptic patients started on CBZ monotherapy; 4) a nonlinear model for PHN from steady-state and/or non-steady-state routinely collected TDM data of adult epileptic patients on monotherapy.

While monotherapy is considered “gold standard”, patients who don’t respond are prescribed AED polytherapy. The older AEDs are involved in many interactions due to their pharmacokinetic properties. Among others, potential PK drug-drug interactions are indications for TDM of AEDs. The objective of my recent modelling work is to develop nonlinear models describing PK drug-drug interaction during AED polytherapy, for example, CBZ hetero-induction when another AED is added to CBZ monotherapy. The majority of these population models are applicable for PK guided dosage individualization in clinical settings. Recent improvements in Pmetrics make it possible to develop more complex PK models and to use them to calculate optimal dosage regimen to achieve a desired target for a particular patient.

What sort of research do you have on the horizon?

For the physician and for the patient, TDM is an added burden. To have a wider spread of individualizing AED dosing based on TDM for management of epilepsy, physicians need clear evidence of the benefits, a clear explanation of the approach, knowledge of pharmacokinetics and pharmacodynamics, help for interpreting the results if required.

An important step is to evaluate how well a method “works”, and how helpful this method is in real clinical settings. The aim of my recent work is to test external validity and to evaluate the predictability of a Bayesian procedure for AED individualized dosage regimens based on TDM data and the developed population PK models. The repeated measurements in some patients after their dosage regimen corrections and adjustments make such validation of the procedure and the proposed population models possible, and in real clinical settings.

These results help explain the advantages of TDM and individualizing AED dosage regimens to sceptical physicians. I believe that in many epileptic patients seizure control can be improved when the physician utilizes measured PK data in conjunction with clinical data (seizure control, toxicity, patient compliance, concomitant medications, concomitant diseases, etc). A modelling approach for planning, monitoring and adjusting dosage appropriately, based on a population PK/PD model as the prior, and using Bayes’ theorem to fit the serum concentration data, can help develop an individualized model that best describes the behaviour of the drug in a particular individual patient. I plan to continue development and validation of nonlinear models for PK drug-drug interaction during AED polytherapy,

Can you tell us a little bit about your respective roles? What is a typical day like for you?

My “typical” day is a mixture of different calculations for PK modelling, research projects and clinical trials as well as consulting for clinicians and colleagues who need help in interpretation or processing of medical data. I take part in Russian and international scientific conferences, work as a lecturer in the fields of pharmacometrics and mathematical statistics in medical research.

What do you consider is the future for TDM and CT? What are you excited about? What are the challenges we face?

TDM is an important part of personalized medicine. There are drugs and patients, for which the balance between therapy efficacy and toxicity requires TDM. Improved assay methods, understanding of pharmacokinetic–pharmacodynamic relationships for many drugs result in therapeutic drug monitoring becoming a useful adjunct to administration of wide range of drugs. Opportunity to measure drug levels on-site or at home, reliable tools for real-time monitoring combined with software allowing dose optimization should have positive effect on therapeutic outcomes and convenience for clinicians and patients.

For TDM of AEDs, future developments include AEDs measurement in saliva and other less invasive and better accepted sampling methods. Portable near-patient testing devices, minimally-invasive, simple, and low-cost methods for online therapeutic drug monitoring enable wider and more effective implementation of TDM for epileptic patients. Non- or minimally-invasive intensive and continuous sampling strategies provide tracking of the drug concentrations, hence individual parameters of more complex PK models for drugs or drug combinations may be identified from such “rich” TDM data and used for dosage individualization. The typical methodology is to consider the clinical effect of AED therapy as an improvement of seizure control with minimal toxicity. To date, no PK measure has been established to evaluate the short-term or immediate pharmacodynamic effect of an AED in humans. More intensive TDM procedures offer a possibility for investigation and analysis of relationship between pharmacokinetics and pharmacodynamic measures and PK/PD modelling of AEDs.


The content of the IATDMCT Blog does not necessarily have the endorsement of the Association.
Irina Bondareva