Open Access
Prevalence and risk factors for HIV CSF Viral Escape: Results from the CHARTER and HNRP cohorts
Author(s) -
Perez Valero I,
Letendre S,
Ellis R,
Deutsch R,
Franklin D,
Clifford D,
Heaton R,
McCutchan J,
McArthur J,
Morgello S,
Gelman B,
Collier A,
Grant I
Publication year - 2012
Publication title -
journal of the international aids society
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.724
H-Index - 62
ISSN - 1758-2652
DOI - 10.7448/ias.15.6.18189
Subject(s) - medicine , viral load , odds ratio , atazanavir , gastroenterology , immunology , human immunodeficiency virus (hiv) , antiretroviral therapy
Background During HAART, HIV RNA can be detectable (>50 cop/mL) in CSF when it is undetectable in plasma, a condition termed CSF viral escape (CVE). The aim of the current analysis was to determine the prevalence and risk factors for CVE in two large US cohorts. Methods 1,264 volunteers enrolled in CHARTER or HNRP at their most recent visit between 2003 and 2011 were included in this cross‐sectional analysis if their HIV RNA level in plasma was undetectable while on stable HAART (>6 months) and if they had CSF collected. Potential risk factors were identified using univariable and multivariable analysis. Odds ratios for detected risk factors were calculated. Results Mean age was 46 years, 82% were men, 70% had AIDS, 22% were HCV+, 49% were Caucasians, median CD4 nadir was 129, and 38% were cognitively impaired. CVE was present in 55 (4.35%) with a median HIV RNA in CSF of 155 (IQR 80‐283). The table summarizes the main analysis results. CVE was associated with longer durations of HIV disease, higher platelet count, higher total serum protein, and higher CSF white blood cells (WBCs). CVE was also associated with treatment‐associated factors, including use of boosted PIs and unboosted atazanavir.Variable Measure CSF viral load >50 (n=55) CSF viral load <50 (n=1209) Univariable analysis Multivariable analysis Odds ratio Odds ratio infoAge (years) Mean±SD 45.3±7.48 46±9.50 p=0.663Gender (male) n (%) 43 (78.18) 978 (81.98) p=0.486Ethnicity (white) n (%) 25 (45.45) 590 (49.50) p=0.692Years since first HIV+ Median [IQR] 16.02 [11.55–19.9] 12.72 [6.79–18.24] p=0.018 p=0.016 1.367 [1.058–1.785] (each 5 years) HCV (positive) n (%) 10 (22.22) 219 (22.42) p=0.976CD4 nadir (cells/mL) Median [IQR] 71 [8.75–188.5] 133 [27.5–240] p=0.025AIDS (CDC) n (%) 41 (83.67) 775 (69.07) p=0.021Cognitively impair (Yes) n (%) 21 (38.18) 447 (37.69) p=0.914Global Deficit Score Median [IQR] 0.4 [0.11–0.87] 0.33 [0.11–0.72] p=0.943CD4 in blood (cells/mL) Median [IQR] 506 [268–711] 508.5 [340.25–710] p=0.678Hb in blood (mg/dL) Median [IQR] 14 [12.9–15.1] 14.4 [13.4–15.3] p=0.104Platelets in blood (x10 3 /mL) Median [IQR] 232 [202–283] 230 [189–273] p=0.056 p=0.022 1.314 [1.041–1.645] (each 50,000) Protein in serum (g/dL) Median [IQR] 7.7 [6.90–8.30] 7.4 [7–7.9] p=0.013 p=0.035 1.649 [1.036–2.614] (each 1 g/dL) WBC in CSF (cells/mL) Median [IQR] 4 [2.5–16.5] 2 [1–3] p≤0.001 p≤0.001 3.416 [2.204–5.582] (each 10 cells) Protein in CSF (mg/dL) Median [IQR] 47 [31.5–55.5] 38 [30–48] p=0.002Glucose in CSF (mg/dL) Median [IQR] 62 [57.5–68.5] 63 [58–68] p=0.740Previous blips (Y/N) n (%) 6 (10.91) 111 (9.2) p=0.677Months VL <50 in plasma Median [IQR 13.07 [6.04–37.66] 18.63 [7.51–40.8] p=0.543Months on current HAART Median [IQR] 13.07 [6.05–29.78] 18.32 [6.7–35.89] p=0.18Months on HAART ever Median [IQR] 77.55 [39.65–124.15] 72.87 [34.94–117.07] p=0.419CPE score Median [IQR] 7 [6–8] 7 [7–9] p=0.625NNRTI+NRTIs n (%) 8 (14.55) 438 (36.23) p<0.001PI/r + NRTIs n (%) 31 (56.36) 501 (41.47) p=0.03 p=0.006 2.749 [1.340–5.976] (Yes vs No) ATV + NRTIs n (%) 4 (7.27) 27 (2.23) p=0.052 p=0.024 6.006 [1.302–21.57] (Yes vs No) Other HAART regimens n (%) 12 (21.82) 243 (20.07) p=0.747Conclusions In this large, cross‐sectional analysis, CVE was uncommon in subjects on effective HAART. A combination of disease and treatment factors were associated with CVE. The associations with higher levels of CSF WBCs, blood platelets, and serum total protein may reflect greater immune activation. Treatment with PI‐based HAART was particularly associated with CVE, especially if unboosted atazanavir was part of the regimen. CVE was not associated with neurocognitive impairment. Prospective analyses are needed for better characterization of CVE.