HIV model parameter estimates from interruption trial data including drug efficacy and reservoir dynamics

Area of Science:

• Mathematical modeling
• Virology
• Immunology

Background:

• Ordinary differential equation (ODE) models are crucial for understanding HIV dynamics and treatment.
• Parameter identifiability from clinical data is essential for predictive modeling.
• Previous HIV model studies often used limited, sparsely sampled decay-phase data.

Purpose of the Study:

• To identify model parameters using frequently sampled viral load data from HIV patients.
• To enable more accurate predictions of HIV disease dynamics and treatment outcomes.
• To estimate drug efficacy and reservoir contribution rates, which are difficult to identify from decay-phase data alone.

Main Methods:

• Utilized frequently sampled viral load data (69-114 measurements per patient) from ten patients in the AutoVac HAART interruption study.
• Employed a Markov-Chain Monte-Carlo (MCMC) method for parameter estimation, with initial estimates from nonlinear least-squares.
• Analyzed data from two experimental conditions to allow direct estimation of drug efficacy and reservoir contribution.

Main Results:

• Successfully identified model parameters from a larger, more frequent dataset than previously used.
• Estimated drug efficacy and reservoir contribution rates, providing insights not possible with decay-phase data alone.
• Reported and compared posterior distributions of parameter estimates across all patients.

Conclusions:

• Frequent viral load sampling significantly enhances the identifiability of parameters in HIV models.
• The findings contribute to more robust mathematical models for predicting HIV disease progression and optimizing antiretroviral therapy.
• This methodology provides a foundation for personalized treatment strategies based on individual patient data.