Mathematical and statistical modelling of CD4 T-cell reconstitution in HIV-infected children starting antiretroviral therapy

• Main Author: Lewis, J. E. A
• Format: Dissertation
• Language: English
• Published: University College London (University of London) 2013

Children's immune systems develop over the first 20 years of life and are fundamentally different to adults' in terms of lymphocyte numbers, turnover rates and population structure. Children perinatally infected with the human immunodeficiency virus (HIV) can expect their infection and any antiretroviral therapy (ART) administered to affect the developing population dynamics of the immune system and long-term immunological health in adulthood. Describing, understanding and predicting these effects is important for guiding individualised treatment regimes, public health policy and future research. The work described here develops an empirical model for CD4 T-cell reconstitution in HIV-infected children starting ART based on a monophasic, asymptotic recovery function and applied to data in a nonlinear mixed-effects framework. The model quantifies effects of age on both pre-ART and long-term CD4 count, and an association between poor pre-ART and poor long-term immune status. It allows prediction, by extrapolation, of CD4 counts in adulthood given ART initiation at different ages and CD4 levels. These predictions suggest that current ART initiation guidelines preserve potential for good long-term CD4 numbers in young children, but that this potential is progressively damaged as age increases. The model also identifies groups of children with qualitatively different CD4 trajectories and illustrates fundamental differences between recovery in children and adults probably due to children’s more active thymuses. Next, a mechanistic model of T-cell homeostasis is described, aiming to develop a biological understanding of some of the effects identified using the empirical model. A model of homeostasis in HIV proves a valid, useful tool for understanding pathogenesis and the reasons for incomplete recovery on ART, and illustrates the power of mixed-effects modelling for gleaning information about differences between children’s T-cell dynamics. In the future, more complex and complete models building on this foundation will help understanding of the interactions between development, HIV and ART.