Cerebrospinal fluid abacavir concentrations in HIV‐positive patients following once‐daily administration

Abstract

Abacavir is a widely used nucleotide reverse transcriptase inhibitor, for which cerebrospinal fluid (CSF) exposure has been previously assessed in twice‐daily recipients. We studied abacavir CSF concentrations in 61 and nine HIV‐positive patients taking abacavir once daily and twice daily, respectively. Patients on once‐daily abacavir had higher plasma and CSF concentrations (96 vs. 22 ng ml⁻¹, P = 0.038 and 123 vs. 49 ng ml⁻¹, P = 0.038) but similar CSF‐to‐plasma ratios (0.8 vs. 0.5, P = 0.500). CSF abacavir concentrations were adequate in patients receiving once‐daily treatment.

Introduction

Abacavir (ABC) is a widely used, once‐daily HIV nucleoside reverse transcriptase inhibitor included in first‐line regimens by international guidelines 1. Even though the clinical relevance is debated, higher penetration of antiretroviral agents into the cerebrospinal fluid (CSF) has been associated with lower concentrations of CSF HIV RNA and, in some studies, with better cognitive function 2, 3. The concentration penetration–effectiveness score (CPE) ranks ABC at 3 (out of 4), so the compound has been included within the medium–high score stratus. However, most of the data were generated from using the 300 mg twice‐daily dose: its CSF to plasma ratio (CPR) was reported to be 36% (95% confidence interval 28–46%), although with significant variability 4. The aim of the present study was to compare CSF concentrations (and CPRs) in patients administered once‐ vs. twice‐daily ABC.

Methods

ABC‐administered patients undergoing lumbar punctures for clinical reasons or included in longitudinal studies were included in either the Turin (Ospedale Amedeo di Savoia, ASL “Città di Torino”) or Rome (Istituto Nazionale per le Malattie Infettive “Lazzaro Spallanzani”) centre. Signed informed consent was given by all subjects, and ethics approval was obtained by each Institution (Comitato Etico Interaziandale di Orbassano, n. 103/2015). Plasma and CSF samples were obtained concomitantly (less than 30 min apart). ABC concentrations were measured at steady state in the same externally validated laboratory (University of Torino), through a validated high‐performance liquid chromatography–tandem mass spectrometry method (with a limit of detection of 4.88 ng ml⁻¹) 5. Trough concentrations were those measured 21–27 h (once daily) and 10–14 h (twice daily) after drug intake.

The CSF to serum albumin ratio [CSAR; calculated as CSF albumin (mg l⁻¹)/serum albumin (g l⁻¹)] was used to evaluate blood–brain barrier (BBB) function. The definition of BBB damage was derived from age‐adjusted reibergrams (normal if below 6.5 in patients aged <40 years and below 8 in patients >40 years) 6.

Data are expressed as medians [interquartile range (IQR)] and compared through nonparametric tests (Spearman's test for linear correlation and Mann–Whitney or Kruskal–Wallis for when a categorical variable was compared with a linear one). All concentrations were used for identifying predictors of CSF penetration, whereas only trough values were used for comparing once‐ and twice‐daily administration. Coefficients of variation (CVs) were calculated as the ratio between the standard deviation and the mean.

Results

Seventy patients (85.7% male, 95% Caucasians) were included. Their ages and body mass indexes were 50.2 (46.2–58.7) years and 23.4 (20.9–25.6) kg m⁻², respectively. In 60 patients, chronic hepatitis status was known: 21 (35%) were hepatitis C virus positive (HCV+), three (5%) were hepatitis B virus positive (HBV+) and six (10%) had a previous diagnosis of liver cirrhosis. Forty‐four subjects (62.8%) were asymptomatic and received lumbar punctures for non‐Hodgkin's lymphoma staging (33; 47.1%) or in the context of longitudinal studies (11; 15.7%). The remaining 26 patients had HIV encephalopathy (n = 8; 11.4%), HIV‐associated neurocognitive impairment (n = 6; 8.6%), opportunistic infections of the central nervous system (n = 6; 8.6%) and various neurological presentations (n = 6, including two cases of seizures of new onset). Baseline features are listed in Table 1.

Table 1.

Baseline characteristics of the included patients stratified according to once‐daily or twice‐daily abacavir intake

	n or median (% or IQR)
	Once daily	Twice daily
n	61	9
Age, years	51 (46.4–58.1)	49.3 (43.7–60.1)
Male gender, n (%)	51 (83.6%)	9 (100%)
BMI, kg m−2	23.4 (20.9–25.6)	25.2 (21.4–25.6)
CD4, Cells μl−1	332 (137–524)	155 (130–275)
CD4 nadir, Cells μl−1	132 (60–288)	94 (16–543)
Plasma HIV RNA, Log10 copies ml−1	1.59 (<1.27–1.75)	1.65 (<1.27–2.17)
CSF HIV RNA: Log10 copies ml−1	1.59 (1.59–2)	1.59 (<1.27–2.03)
CSF cells, n mm−3	1 (0–3)	2 (1–6)
CSF proteins, mg dl−1	42 (33–56)	39 (25–96)
CSAR	5.1 (4.0–6.0)	7 (6.8–7.6)
Time after abacavir intake, h	13 (11.5–24)	12 (11.5–12.5)
ARV classes: NNRTI	20 (32.7%)	0
NRTIs only	2 (3.3%)	2 (22.2%)
PI	35 (57.4%)	7 (77.8%)
INSTI	4 (11.4%)	0
Indication for LP: Asymptomatic	7 (11.5%)	4 (44.4%)
HAND	6 (9.8%)	0
Neurological symptoms	4 (6.6%)	2 (22.2%)
HIV encephalopathy	7 (11.5%)	1 (11.1%)
CNS opportunistic infections	6 (9.8%)	0
NHL	31 (50.8%)	2 (22.2%)

ARV, antiretroviral; BMI, body mass index; CNS, central nervous system; CSF, cerebrospinal fluid; CSAR, CSF to serum albumin ratio; HAND, HIV‐associated neurocognitive disorders; INSTI, integrase strand transfer inhibitor; IQR, interquartile range; LP, lumbar puncture; NHL, non‐Hodgkin's lymphomas (screening for meningeal involvement); NNRTI, non‐nucleoside reverse transcriptase inhibitors; NRTI, nucleos(t)ide reverse transcriptase inhibitors; PI, protease inhibitor

The most frequently coadministered drugs were protease inhibitors (PIs) (67.1%) followed by non‐nucleoside reverse transcriptase inhibitors (28.6%), nucleos(t)ide reverse transcriptase inhibitors only (5.7%) and integrase strand transfer inhibitors (5.7%). The CD4+ T lymphocyte cell count was 302 (137–462) μl^–1. After excluding patients with plasma HIV RNA >1000 copies ml⁻¹ (n = 3), 62 subjects had available plasma and CSF HIV RNA – below 50 copies ml⁻¹ in 47 (74.6%) and 43 (68.3%) subjects. CSF escape (defined as detectable HIV RNA in the CSF, while undetectable in the plasma or with CSF HIV RNA ≥1 log₁₀ higher than the plasma level) was observed in 13 individuals (20.6%); no association was observed between ABC levels, detectability or the schedule of administration, and the prevalence of CSF escape.

Samples were withdrawn over the interval range but after a median of 12 (12–20) h after drug intake; 18 were trough samples. Median ABC plasma and CSF concentrations were 58 ng ml⁻¹ (<1–815; CV 140%) and 106 ng ml⁻¹ (21–304; CV 149%), respectively. ABC CPRs ranged from 0 to 4.92 (median 0.80; IQR 0.17–1.87; CV 99%). Among those with undetectable plasma concentrations (n =17; 25%), nine were nondetectable in the CSF, while eight showed low but detectable CSF concentrations [35 (11–67) ng l⁻¹]. A direct correlation was observed between ABC plasma concentrations and the time from drug intake (rho = 0.406; P = 0.005); this was not observed for CSF concentrations (rho = 0.212; P = 0.148) and CPRs (rho = −0.092, P = 0.582).

The drug was administered once‐daily in 61 (87.1%) patients. CSF ABC concentrations over time are shown in Figure 1 (stratified by frequency of intake). The baseline characteristics of the two groups were comparable, with the exception of a nonsignificantly higher CD4 cell count (332 μl^–1 vs. 155 μl^–1; P = 0.062) in the once‐daily group. Patients on once‐daily ABC had higher plasma and CSF concentrations (96 ng ml⁻¹ vs. 22 ng ml⁻¹; P = 0.038 and 123 ng ml⁻¹ vs. 49 ng ml⁻¹; P = 0.038), but CPRs were similar (0.8 vs. 0.5; P = 0.500). Trough values showed the same trend: higher plasma concentrations (1377 ng ml⁻¹ vs. <4.88 ng ml⁻¹; P = 0.02) but similar CSF values (152 ng ml⁻¹ vs. 24 ng ml⁻¹; P = 0.08) and CPRs (0.2 vs. 0l P = 0.08). In once‐daily recipients, CSF ABC concentrations were associated with ABC plasma levels (rho = 0.865; P < 0.001) and, weakly, with age (rho = 0.275; P = 0.046). Patients receiving PIs had higher CSF [220 (64–529) ng ml⁻¹ vs. 70 (17–134) ng ml⁻¹; P = 0.013] and plasma [500 (35–1531) ng ml⁻¹ vs. 42 (<4.88–100) ng ml⁻¹; P = 0.002] concentrations. HCV+ once‐daily recipients showed higher CPR compared with HCV‐negative subjects [1.57 (0.80–2.35) vs. 0.36 (0.12–1.28); P = 0.007].

Figure 1.

Abacavir cerebrospinal fluid to plasma ratios, according to time after dose and stratified for once‐ or twice‐daily schedule

CSARs were available in 28 patients; the median value was 5.35 (4.12–6.87), with four individuals (14.3%) presenting with altered BBB permeability. No linear correlation was observed between CSAR and ABC CSF concentrations or CPRs, and impairment in the BBB was found to have no effect.

No difference was observed in ABC CPRs according to neurological confounding conditions and/or CSF HIV‐1 escape (Kruskal–Wallis and Mann–Whitney tests; P‐values >0.05).

Discussion

In the present small study, we observed that patients on once‐daily ABC had CSF ABC concentrations and CPRs higher than previously reported (in a median of 80% of patients). Our control group of twice‐daily recipients was small (n = 9) but showed higher CPRs than those reported by Capparelli et al. 4 (36%) and Antinori et al. (4%) 7. This may be explained by the significant interpatient variability, the concomitant use of PIs (not reported in the previous studies) and potentially by the older age of our patients. PIs may influence drug penetration by inhibiting drug efflux transporters, such as P‐glycoprotein and breast cancer resistance protein, as shown in vitro by Marzolini et al. 8. Age may be a contributing factor in increasing drug concentration in the CSF, as increased BBB permeability, lower CSF production and reduced transporter function have been suggested to occur in older adults 9, 10. The only available, but as yet unpublished, study on this topic reported a steep increase in CSF tenofovir and efavirenz concentration with increasing age (Croteau et al., manuscript in preparation). In addition, a high prevalence of CSF escape was observed, and this may have influenced our results (in terms of confounding conditions or pharmacokinetic features following ongoing inflammation). Nevertheless, when ABC (coadministered with lamivudine and dolutegravir) was administered to antiretroviral therapy‐naïve patients, there was excellent antiviral activity in the CSF (both 2 weeks and 16 weeks after treatment initiation), confirming the compartmental efficacy of the drug when used in combination with other active compounds 11.

Limitations of current analysis included the small sample size, the inclusion of single time points and use of non‐trough concentrations. ABC CSF concentrations and CPRs were, however, more stable over the time interval, compared with plasma levels, and as previously reported by Cusini et al. 12.

In conclusion in once‐daily ABC‐treated HIV+ patients, high CSF ABC, and CPRs were observed, suggesting that the current CPE ranking may be appropriate, even with the commonly used once‐daily schedule.

Competing Interests

The authors declare no competing interests.

This study was presented at the 9 ^th Italian conference on AIDS and Antiviral Research, Siena, 2017. The study was supported by internal funding.

Calcagno, A. , Pinnetti, C. , De Nicolò, A. , Scarvaglieri, E. , Gisslen, M. , Tempestilli, M. , D'Avolio, A. , Fedele, V. , Di Perri, G. , Antinori, A. , and Bonora, S. (2018) Cerebrospinal fluid abacavir concentrations in HIV‐positive patients following once‐daily administration. Br J Clin Pharmacol, 84: 1380–1383. doi: 10.1111/bcp.13552.