Structural-equation-modelling of the tropism impact on achieving viral suppression within six months in naïve HIV patients

INTRODUCTION: Aim of the study was to evaluate the relevance of baseline (BL) plasma tropism of HIV on the achievement of a viral suppression within six months of antiviral therapy (ARV) in naïve patients by a structural-equation-modelling. MATERIALS AND METHODS: Two-hundred and twenty-seven patients were enrolled; viral tropism on plasma was determined at baseline (BL) by sequencing and interpretation by genotopheno algorithm. Booster atazanavir or lopinavir , or efavirenz or nevirapine were used, in combination with either abacavir/lamivudine or tenofovir-emtricitabine. RESULTS: X4-tropism correlate negatively with CD4 cell count at BL and follow-up (FU), and CD4 correlate negatively with BL-plasma viremia (PLV). BL-PLV correlate positively with FU-PLV. We have developed the hypothesis that the variables BL-CD4 and BL-PLV represent a mediators chain among X4-tropism and outcome of plasma viraemia at six months. This model, after structural-equation-modelling (SEM, Stata13), is shown in Figure 1. The indirect effect of X4-tropism on Fup-PLV is significant (p<0.01) but about 10 fold lower than the direct effect by BL-PLV. X4-tropism also has a direct negative effect on BL-CD4 (p<0.001) and an indirect positive effect on BL-PLV (p<0.001), irrespective of the drug regimen. Path model explaining direct and mediated effects of “tro (tropism),” “gender,” “age,” “cd0 (BL-CD4)” and “lrna0 (BL-PLV)” on the final outcome (“lrna1-Fup-PLV),” where “tro,” “gender,” and “age” are exogenous, cd0 and lrna0 are endogenous (mediators). Numbers on the arrows indicate direct effects. Circles indicate residuals related to endogenous/dependent variables; numbers near to circles are the corresponding variances. CONCLUSIONS: This model shows the relevance of BL-tropism on the outcome of plasma viraemia in naïve patients after six months of therapy, irrespective of drug regimen used.