It is debated whether HIV cerebrospinal fluid (CSF) escape (HIVE) may be linked to patients’ features or to ineffective antiretroviral (ARV) CSF penetration. Aim was to assess CSF penetration and activity of ARVs and correlation with blood brain barrier (BBB) and to estimate risk and predictors of HIVE.
Retrospective study on CSF/plasma paired samples from HIV+ patients taking cART undergoing lumbar puncture (LP) because neurologically symptomatic or for CNS staging. HIVE was: a) detectable CSF HIV-RNA with concurrent plasma <50 cp/mL; b) CSF HIV-RNA >1.0 log10 higher than in concomitant plasma. ARVs concentrations in CSF and plasma were measured by mass spectrometry methods. CSF/plasma concentration ratio was calculated and 95% CSF inhibitory quotients (IQ95) derived, and considered optimal if >1. Impaired BBB was defined by CSF/plasma albumin quotient x 103 (Qalb), according to age-normalized reference ranges (Reiber, Clin Chem, 1995). Adjusted ORs of HIVE were estimated by fitting a logistic multivariable regression model.
133 CSF/plasma pairs from 120 patients (2000-2014). Male 79.7%; median age 47y (IQR 40-51); CD4 count <200 cells/mm3 47.4%. At LP, 58.7% was on TDF/FTC, 3% ABC/3TC, 10.5% ZDV/3TC, 36.8% EFV, 27.1% LPV/r, 13.5% ATV/r, 11.3% DRV/r. Proportion of samples with IQ95 >1 was lower for LPV/r and ATV/r (p=0.014). After stratification by Qalb, the proportion of CSF with IQ95 >1 in the two strata was still similar for EFV, ATV/r and DRV/r, but not for LPV/r (p=0.05 at interaction test). A positive correlation between Qalb and CSF/plasma ARV concentration ratio (R=0.686, p<0.001) and ARV IQ95 (R=0.658, p<0.001) was found. HIVE was detected in 19/133 samples (14.4%). Only in a subset of 73 asymptomatic pts, duration of current cART [AOR 1.98 per year more (95%CI 1.27-3.12)] and CSF log10 HIV-RNA [AOR 17.91 (95%CI 1.04-307.55)] was associated to an increased HIVE risk by multivariable logistic regression. IQ95 of third drug was not associated with HIVE [AOR 0.84 95%CI 0.20-3.50], whereas Qalb was associated to increased HIVE risk [AOR 5.75; 95%CI 1.10-30.1].
BBB dysfunction may influence neuropenetration (CSF/plasma drug concentration ratio) and efficacy (IQ95) of ARVs and play a role on asymptomatic HIVE mechanism. Despite the fact that action of ARVs in the CNS, except for DRV/r, was far below, the role of drug exposure and activity in CSF on HIVE pathogenesis remains controversial.