Determinants of CD4+ counts on babies from HIV + mothers: A semi-parametric modelling approach

Abstract

Background: The variability in CD4+ cell counts within mothers and babies at birth, in human immunodeficiency virus (HIV) positive mothers, has not been explained adequately. As one of the predictors of HIV, the CD4 + cell count plots the course of HIV progression. The important implication has to be understood, in relation to HIV-related opportunistic infections. Therefore, this study proposes to evaluate factors that determine CD4+ cell count in infants born to women infected with HIV and to determine the relationship between CD4+ cell counts in maternal viraemia and CD4+ cell counts of babies at birth. Methods: A semi-parametric approach was used in modeling CD4+ cell counts in the newly born babies to HIV-positive women. Gestational age of the mothers was including in the model using a smooth function to cater for non-linearity. Analysis was restricted to antiretroviral therapy (ART) - naive, HIV-infected pregnant women. At enrollment, maternal median age was 26 years (interquartile range (IQR): 5) and median CD4+ counts was 459 cells/mm 3 (IQR: 566). The mean (SD) CD4+ cell counts of HIV-seropositive women were 691 cells/mm 3 (803 cells/mm 3 ). Results: Maternal CD4+ counts and gestational age were found to be the significant determinants of CD4+ counts in the new born babies. Maternal CD4+ was positively associated with CD4+ counts in babies (0.239; p=0.004). We noted a non-linear trend in CD 4+ cell activity with gestational age, the optimum value at occurs at 37 weeks of pregnancy. Conclusions: The model provides a flexible way to estimate the changing distribution of CD4+ cell counts in babies with weeks of pregnancy. Results obtained from the semi-parametric modeling could substantially improve the planning of health services, including the need for antiretroviral therapy.