An anti-VZV booster response was experimentally

defined as a >4-fold increase in anti-VZV IgG levels between two consecutive samples or a >2-fold increase resulting in an absolute increase of ≥1000 IU/L (not shown). Antibody avidity increases during the maturation process of memory B cells, such that re-exposure to endogenous or exogenous antigen results in antibodies of higher avidity. Accordingly, antibody avidity is an indirect marker for the reactivation of memory responses [15]. The avidity of anti-VZV antibodies was determined by adding various dilutions (0–3 M) of sodium thiocyanate to serum-containing antigen-coated wells, as previously described [16–18]. Results are expressed as the avidity index (AI), defined as the thiocyanate concentration at which 50% of the VZV-specific antibodies were eluted. As AI may fail to identify differences attributable to a small pool of high- or low-avidity antibodies, analyses were completed by calculating the percentage IDH inhibitor of antibodies dissociated at each thiocyanate concentration (AVISCAN) [19,20]. The Aviscan gives information about the distribution of different avidities within an antibody population of heterogenous avidities. All P-values were two-tailed. P-values <0.05 were considered statistically significant. Continuous variables were assessed using parametric or nonparametric tests when appropriate, whereas categorical selleck products data were assessed

using the χ2 or Fisher’s exact test. Linear regression was used to analyse potential risk factors for low anti-VZV IgG levels and AI, Amoxicillin whereas conditional logistic

regression was used to identify potential risk factors for a complete loss of VZV antibodies. All variables were examined at the univariate level. Thereafter, only variables with a P-value <0.25 by univariate analysis were included in the multivariate model [21]. Change in anti-VZV IgG levels over time in HIV-infected children and adults were analysed using mixed linear models. This statistical model takes into account the repeated measurement of each individual across time. We included as predictors the group of patients (HIV-infected children or adults), the time of measure (linear trend) and the time of measure squared (quadratic trend) to account for a downward trend that could be faster for high VZV levels and slower for low levels. Finally, we adjusted for age, CD4 T-cell count and VZV serological reactivation. Statistical analyses were performed using spss (v15.0; SPSS Inc., Chicago, IL), with the exception of longitudinal analyses, which were performed using the lme statistical package of the R software, v 2.9.2 [22]. Ninety-seven vertically HIV-infected children (541 samples) and 78 HIV-infected adults (440 samples) met the study inclusion criteria (Table 1). In 2008, the CD4 T-cell count and percentage (P<0.001 for both) and the HIV RNA level (P=0.007) were higher in HIV-infected children than adults.