Human Pegivirus Encephalitis With Brain Detection and Response to Sofosbuvir Ledipasvir Authors: Antoine Moulignier, Françoise Heran, Francis Lallemand, Pierre Bourdillon Published: 2026 ABSTRACT ABSTRACTHuman pegivirus (HPgV‐1) has been associated with severe encephalomyelitis in immunocompromised patients. Its neurological spectrum remains poorly defined. We report a slowly progressive encephalitis in a person living with well‐controlled HIV, characterized by white matter abnormalities and inflammatory cerebrospinal fluid (CSF). HPgV RNA was detected in CSF and brain tissue by metagenomic sequencing, with no alternative pathogen identified. Following off‐label treatment with sofosbuvir/ledipasvir, the patient showed sustained clinical improvement, normalization of CSF findings, and disappearance of detectable HPgV RNA. This observation expands the clinical context of HPgV‐1 detection and supports further investigation of its role in central nervous system disease.Keywords: encephalitis, HIV infection, HPgV, human pegivirus Introduction Central nervous system (CNS) infections remain a major diagnostic challenge in neurology, as approximately 50% of encephalitis cases remain of undetermined etiology despite extensive investigations, with mortality rates of 7%–18% and severe long‐term disability in up to 56% of survivors [1]. Metagenomic next‐generation sequencing (mNGS) has demonstrated pooled sensitivities of 70%–77% and specificities of 93%–96% in CNS infections, enabling the identification of previously unrecognized infectious causes of encephalomyelitis and facilitating more targeted diagnostic approaches [2]. Human pegivirus (HPgV‐1), a blood‐borne RNA virus of the Pegivirus genus within the Flaviviridae family, has recently emerged as a potential cause of CNS infection, particularly in immunocompromised individuals [3, 4, 5, 6, 7, 8, 9, 10, 11, 12]. HPgV‐1 is phylogenetically related to hepatitis C virus (HCV), raising the hypothesis that direct‐acting antivirals developed against HCV could exert off‐target antiviral activity against HPgV‐1 [5]. Although long considered non‐pathogenic because of its high worldwide seroprevalence (≈3% in blood donors and up to 30%–40% in immunocompromised populations) [13], detection of HPgV RNA in blood or cerebrospinal fluid (CSF) alone may reflect incidental viremia or passive trafficking across a disrupted blood–brain barrier rather than true CNS infection. The neurotropic potential of HPgV‐1 therefore remains debated [4, 7, 11, 12]. Recent neuropathological observations suggest that HPgV‐1 may access the CNS via infected lymphocytes and, in some cases, exhibit features consistent with viral persistence or neuroadaptation, although these mechanisms remain to be confirmed [3, 14]. Overinterpretation of CSF‐based detection in the absence of tissue confirmation has consequently been cautioned [15]. Several independent neurological teams have reported a relatively homogeneous and often severe encephalomyelitis in immunocompromised patients characterized by combined optic nerve, spinal cord, and brain involvement, with clinico‐radiological features reminiscent of neuromyelitis optica spectrum disorder (NMOSD) [3, 4, 5, 10]. In some of these cases, HPgV RNA was identified directly within brain tissue in the absence of alternative pathogens, providing stronger evidence for a possible causal role. This NMOSD‐like phenotype may represent the most severe end of the currently described disease spectrum [3, 4, 10]. More heterogeneous or less fulminant neurological presentations have recently been described [6, 7, 9, 12], suggesting that HPgV‐associated CNS disease may be broader and potentially underrecognized in neurological practice. We report a slowly progressive inflammatory encephalitis in a person living with HIV with sustained virological control in whom HPgV RNA was detected in both cerebrospinal fluid (CSF) and brain tissue by metagenomic sequencing, followed by sustained clinical improvement after treatment with sofosbuvir/ledipasvir. This observation highlights the diagnostic and interpretative challenges posed by HPgV‐1 detection in unexplained inflammatory encephalitis and may contribute to refining the clinical spectrum and diagnostic significance of HPgV‐1 in neurological disease. Case Report A 60‐year‐old surgeon living with long‐standing HIV infection and sustained virological suppression for more than 5 years (plasma HIV RNA < 50 copies/mL) while receiving abacavir and ritonavir‐boosted lopinavir presented with a 2‐month history of slowly progressive upper‐limb clumsiness and executive dysfunction interfering with surgical practice. HIV infection had been diagnosed in the early 1980s, with a nadir CD4+ T‐cell count of 80 cells/μL before initiation of antiretroviral therapy in the mid‐1990s. At neurological presentation, the CD4+ T‐cell count was 852 cells/μL with a CD4/CD8 ratio of 1.24. Neurological examination revealed mild pyramidal signs and proprioceptive impairment without major cognitive decline. Ophthalmological assessment, including Humphrey visual field testing and optical coherence tomography, was unremarkable. Brain magnetic resonance imaging demonstrated bilateral supratentorial and pontine T2/FLAIR hyperintensities without gadolinium enhancement, diffusion restriction, or mass effect, suggesting a low‐grade inflammatory process (Figure 1). Brain MRI did not demonstrate selective involvement of the pyramidal tracts or visual pathways. Spinal imaging was normal. Brain MRI: FLAIR sequences (A, B), T1‐weighted spin‐echo with contrast enhancement (C), and diffusion‐weighted imaging (D). Hyperintense, non‐enhancing white matter lesions involving the pons, external capsule, and periventricular regions, without diffusion restriction. Initial CSF analysis revealed lymphocytic pleocytosis (21 cells/μL), elevated protein (0.89 g/L), and normal glucose. Extensive investigations for infectious, autoimmune, metabolic, and neoplastic etiologies were unrevealing. Plasma and CSF HIV RNA levels remained below the assay detection thresholds (< 20 copies/mL and < 100 copies/mL, respectively) at the time of neurological deterioration and throughout follow‐up. Repeat CSF analysis performed 6 months after neurological onset showed a decrease in lymphocytosis to 8 cells/μL and a reduction in CSF protein concentration to 0.52 g/L. By the time stereotactic brain biopsy was performed, CSF inflammatory abnormalities had already markedly improved. Because neurological deterioration persisted over 6 months, targeted virological testing was performed and detected human pegivirus (HPgV‐1) RNA in plasma and CSF. Semi‐quantitative RT‐PCR yielded cycle threshold (Ct) values of 24 in plasma and 30 in CSF, while HPgV‐1 RNA was not detected in urine. A stereotactic brain biopsy was performed 7 months after onset in a right frontal white matter lesion using optical neuronavigation guidance. Histopathological examination showed perivascular lymphocytic and plasmocytic inflammation with astrocytic gliosis, without evidence of demyelination, vasculitis, granuloma, or malignancy. Metagenomic next‐generation sequencing of brain tissue (79.9 million reads) identified eight HPgV‐1‐specific sequences mapping to non‐overlapping genomic regions and covering approximately 9% of the reference genome. No alternative viral, bacterial, fungal, or parasitic pathogens were detected, and no HPgV reads were identified in negative controls processed in parallel. Attempts to detect HPgV‐1 RNA using in situ hybridization and immunohistochemical analyses for HPgV‐1 remained negative. Brain biopsy was performed 7 months after neurological onset, when CSF inflammatory abnormalities had already markedly improved. In the absence of an alternative diagnosis and in light of previously reported severe HPgV‐associated encephalomyelitis, off‐label treatment with sofosbuvir/ledipasvir was initiated in September 2021 using the standard HCV regimen after informed consent. Neurological symptoms gradually improved over the following weeks, allowing resumption of professional activity. Follow‐up CSF analysis demonstrated normalization of the cell count (3 lymphocytes/μL) and protein concentration (0.42 g/L). HPgV‐1 RNA was no longer detectable in subsequent CSF testing. CD4+ T‐cell counts remained stable between 800 and 900 cells/μL throughout follow‐up, including after initiation of sofosbuvir/ledipasvir. After more than 5 years of follow‐up, the patient remains clinically stable without recurrence. Discussion HPgV‐1 has recently been reported in association with severe encephalomyelitis in immunocompromised individuals, predominantly with rapidly progressive and frequently fatal NMOSD‐like phenotypes. These initial cases shared a homogeneous NMOSD‐like phenotype involving optic nerves, spinal cord, and brain. The recurrence of this distinctive triad across independent reports, together with the absence of alternative pathogens, makes a coincidental association less likely. A review of previously reported cases highlights several recurrent features (Table S1). Most published observations have involved immunocompromised patients and have presented with severe NMOSD‐like phenotypes, often associated with poor outcomes. Importantly, direct detection of HPgV‐1 in brain tissue has been documented only in a limited number of cases. In contrast, the present case illustrates a slowly progressive inflammatory encephalitis with subtle white matter involvement and sustained clinical improvement following treatment with sofosbuvir/ledipasvir. Notably, this presentation occurred in a patient with preserved CD4+ T‐cell count (852 cells/μL, CD4/CD8 ratio 1.24). However, HIV infection dated back to the early 1980s, with a nadir CD4 count of 80 cells/μL and delayed initiation of antiretroviral therapy (~15 years), a context associated with incomplete immune restoration despite apparent immune recovery. Indeed, approximately 20%–30% of pre–highly effective ARV‐era PLWHIVs failed to fully restore their immunity [16]. This observation suggests that HPgV‐associated CNS involvement may occur outside conditions of overt immunosuppression, while still reflecting persistent immunological vulnerability. This observation broadens the clinical context in which HPgV‐1 RNA may be encountered in neurological practice, without implying a distinct clinicopathological entity. Reports of reversible meningitic or encephalitic presentations associated with HPgV‐1 detection, including isolated meningoencephalitis and opsoclonus–myoclonus syndrome, further support the existence of less aggressive forms. Such presentations may escape recognition or be misattributed to HIV‐associated leukoencephalopathy. The NMOSD‐like phenotype may therefore represent only the most severe manifestation of HPgV‐associated CNS disease. Establishing causality remains challenging. Detection of HPgV‐1 RNA in blood or CSF alone is insufficient given HPgV's relatively high prevalence and lymphotropic nature. In this context, the detection of HPgV‐1 RNA directly within brain tissue by metagenomic sequencing represents a key element supporting CNS involvement, particularly in the absence of alternative pathogens, although it does not by itself establish causality. The distribution of HPgV‐specific reads across non‐overlapping genomic regions further reduces the likelihood of sequencing artifact. However, no immunohistochemical evidence of viral protein expression was obtained, and attempts to detect HPgV‐1 RNA using in situ hybridization remained negative. Brain biopsy was performed at a relatively late stage of the disease course, when CSF inflammatory abnormalities had already markedly improved, which may have reduced the sensitivity of tissue‐based detection methods. Detection in brain tissue may therefore reflect trafficking of infected lymphocytes rather than direct infection of resident CNS cells. Nevertheless, prior neuropathological observations have demonstrated HPgV‐1 RNA and viral antigens within brain‐infiltrating lymphocytes and glial cells, supporting the possibility of viral persistence and replication within the CNS. In the present case, semi‐quantitative RT‐PCR showed a higher viral burden in plasma (Ct 24) than in CSF (Ct 30), consistent with the predominantly lymphotropic nature of HPgV‐1 and a limited viral signal within the CNS compartment. This interpretation is further supported by the progressive decrease in CSF pleocytosis over time (from 21 to 8 cells/μL, followed by normalization), suggesting lower inflammatory activity at the time of brain biopsy. Taken together, the concordance of tissue‐based detection, inflammatory CSF findings, compatible clinico‐radiological features, and exclusion of competing diagnoses supports, but does not establish, a potential pathogenic role for HPgV‐1 in this presentation. The differential diagnosis included HIV‐associated encephalitis; however, HIV replication remained suppressed, neuropathological examination did not reveal HIV encephalopathy features, and deterioration persisted despite antiretroviral intensification. The absence of alternative infectious, autoimmune, or neoplastic causes despite extensive investigation further supports the potential relevance of the virological finding, while acknowledging that incidental viral presence cannot be fully excluded. Clinical improvement was observed after initiation of sofosbuvir/ledipasvir. The subsequent normalization of inflammatory CSF parameters and disappearance of detectable HPgV‐1 RNA in CSF provide biological coherence to the temporal association. While spontaneous fluctuation cannot be excluded and causality cannot be established from a single observation, the prolonged neurological deterioration before treatment initiation followed by sustained improvement shortly after sofosbuvir/ledipasvir initiation supports a temporal association. Direct‐acting antivirals targeting the NS5B polymerase of HVC virus may exert off‐target activity against phylogenetically related flaviviruses, although specific in vitro or in vivo data for HPgV‐1 remain limited. Recent clinical virological studies in HCV/HPgV‐coinfected patients further support a potential antiviral effect of sofosbuvir‐containing regimens against HPgV‐1. In particular, sofosbuvir/ledipasvir was associated with sustained suppression of HPgV‐1 viremia, whereas viral rebound was observed after treatment discontinuation in some patients, supporting a partial but biologically plausible antiviral activity against HPgV‐1 [17]. This provides a biologically plausible, albeit unproven, framework for the observed clinical response. Previously reported antiviral treatment failures largely occurred in advanced disease, suggesting that timing of intervention may be critical. A single‐case observation cannot establish causality and requires confirmation in independent cohorts. Nonetheless, from a clinical standpoint, three practical points emerge: (1) HPgV‐1 should be considered in unexplained inflammatory encephalitis in immunocompromised patients; (2) while HPgV‐1 detection in CSF alone is insufficient, HPgV‐1 RNA detection in brain tissue strengthens diagnostic confidence; (3) early‐stage or mild presentations may exist beyond the NMOSD‐like phenotype. Additional well‐documented cases will be required to determine whether HPgV‐1 represents a rare neurotropic pathogen, a trigger of immune‐mediated injury, or an incidental bystander in selected inflammatory contexts, and to explore whether antiviral therapy could influence the course of selected inflammatory encephalitides. Author Contributions A.M. designed the study; A.M., F.H., F.L., P.B. analyzed the results; A.M. wrote the paper. All authors read and approved the final version of the manuscript. A.M. is the guarantor. Funding The authors have nothing to report. Conflicts of Interest The authors declare no conflicts of interest. Supporting information Table S1: Selected reported cases of human pegivirus–associated central nervous system disease with clinical, radiological, and virological characteristics. Data Availability Statement The data supporting the study findings are available upon reasonable request of qualified researchers. References Associated Data Supplementary Materials Table S1: Selected reported cases of human pegivirus–associated central nervous system disease with clinical, radiological, and virological characteristics. Data Availability Statement The data supporting the study findings are available upon reasonable request of qualified researchers.