Detectable cerebrospinal fluid HIV RNA with associated neurological deficits, despite suppression of HIV replication in the plasma compartment

Highly-active antiretroviral therapy (HAART) rapidly suppresses HIV replication in plasma; however, concerns exist over the ability of antiretrovirals (ARVs) to penetrate and suppress viral replication in biological compartments such as the central nervous system (CNS) and genital tract [1,2]. Ongoing cerebrospinal fluid (CSF) HIV replication in the presence of virological suppression in the plasma compartment by HAART is not well described [3]. We describe the case of an individual treated with HAART, successfully maintaining plasma HIV RNA below 50 copies/ml, who had ongoing CSF HIV replication with neurological symptoms. A 33-year-old white man presented with left leg weakness affecting his gait. An HIV test 6 months earlier was negative. Neurological examination demonstrated increased tone, sustained ankle clonus, pyramidal distribution weakness (Medical Research Council grade 4/5), brisk reflexes and an extensor plantar response affecting the left leg, with normal findings elsewhere. Cerebral MRI revealed a nonenhancing right precentral gyrus lesion (Fig. 1). CSF contained 13 mononuclear cells/μl, protein 0.4 g/l, glucose 3.6 mmol/l (serum 5.2 mmol/l), and viral PCRs (John Cunningham virus, herpes simplex virus, varicella-zoster virus, Epstein–Barr virus, enterovirus and cytomegalovirus) were negative, as were CSF tests for cryptococcal antigen, syphilis and culture for bacteria, fungi and tuberculosis. Investigations found no obvious cause for a cerebrovascular event. Serum was positive for HIV-1 antibody, CD4 cell count was 190 cells/μl (27%), plasma HIV viral load was 361 837 copies/ml and human leukocyte antigen-B5701 allele test was negative. Genotypic analysis of plasma revealed wild-type virus (clade B) [4].Fig. 1: T2-weighted flair image demonstrating enhancing lesion in the right precentral gyrus.A clinical diagnosis of recent HIV seroconversion with neurological involvement was made. The patient was commenced on HAART [abacavir, lamivudine (Kivexa) and saquinavir/ritonavir], and plasma HIV RNA of less than 50 copies/ml was achieved by week 12. Follow-up cerebral MRI was unchanged 3 and 6 months later. Twelve months later, the patient experienced worsening of left leg weakness. HAART was unchanged, plasma HIV RNA was consistently less than 50 copies/ml and CD4 cell count had risen to 450 cells/μl (47%). MRI revealed additional diffuse-enhancing lesions throughout both cerebral hemispheres. CSF white cell count was four per microliter, and the CSF HIV RNA level was 718 copies/ml. Two months later, plasma HIV RNA remained less than 50 copies/ml, and CSF HIV RNA was 676 copies/ml. CSF sequencing of reverse transcriptase/polymerase confirmed clade B HIV-1 virus, and M184V and D67E were observed. Furthermore, the CSF on this occasion contained 44 mononuclear cells/μl, protein 0.53 g/l and glucose 3.2 mmol/l (serum 5.2 mmol/l). Tuberculosis culture and all other CSF tests were negative. HAART was switched to abacavir, zidovudine, lamivudine (Trizivir) and nevirapine [200 mg twice daily (b.i.d)] with the aim of improving CNS penetration of ARV agents. MRI performed 3 months later showed that the prefrontal gyrus lesion was unchanged; however, the small enhancing lesions had resolved. The CSF pleocytosis had fully resolved, and CSF HIV viral load was now less than 50 copies/ml. The case demonstrates the ability of HIV to continue replicating within the CNS compartment, despite the plasma viral load being suppressed by HAART to below 50 copies/ml. This phenomenon has not previously been well described. Along with other investigations, CSF HIV RNA quantification and genotype should be considered early in any individual with neurological symptoms, irrespective of plasma HIV RNA level. Further research will establish optimal ARV combinations for such individuals and demonstrate whether longitudinal monitoring of CSF RNA and genotype (with HAART adjustment accordingly) can benefit clinical outcome. The work conducted by Letendre et al.[5] demonstrated the use of a CNS penetration efficacy ranking score (based on available data of ARV chemical properties, concentration in the CSF and/or results in clinical studies) that is associated with a lower level of CSF HIV RNA. Long-term clinical outcome data for this strategy, however, do not yet exist. Of interest, our patient developed a late CSF pleocytosis without evidence of additional CNS infections, and after changing HAART, this resolved. It is possible that CNS inflammation developed in response to persistent HIV replication and resolved with ‘CNS-optimized’ HAART. This mechanism maybe similar to the brief CSF pleocytosis demonstrated in approximately half of the patients on stopping therapy [6]. The shortcomings of this report include that ultrasensitive techniques were not used for HIV RNA quantification, and minority level resistance mutations may have been present in plasma at baseline. Furthermore, CSF HIV RNA and genotype were not performed at baseline and, therefore, it cannot be confirmed that the CSF HIV resistance mutations that developed were due to the therapy administered. Clinicians managing individuals with neurological symptoms and an undetectable plasma HIV RNA should be aware of the possibility of CSF virological replication and the selection of viral variants associated with drug resistance.