Emergence of invasive cerebral aspergillosis in an HIV-positive patient on voriconazole therapy

Abstract

Invasive aspergillosis has an incidence of less than 10% in patients with HIV infection [1, 2]. Although neutropenia and corticosteroids are risk factors, invasive disease usually occurs in the context of severe immunodeficiency (CD4 count<20 cells/μL) [3]. Mortality rates remain high while invasive cerebral disease is universally fatal [4]. A 40-year-old male was diagnosed with HIV infection in January 2003 with a CD4 count of 40 cells/μL and a viral load of >100 000 HIV-1 RNA copies/mL. The patient declined to start prophylaxis and highly active antiretroviral therapy (HAART). Two months later (March) the patient was hospitalized for presumptive Pneumocystis jiroveci pneumonia, which responded to intravenous trimethoprim-sulphamethoxazole therapy and coadministration of prednisone. Lamivudine, stavudine and lopinavir/ritonavir therapy was commenced 6 weeks later (May) with a good virological response after 1 month of therapy (viral load 380 copies/mL; CD4 count 80 cells/μL). However, during this period he had an episode of disseminated cutaneous shingles which responded to valaciclovir. Six weeks following the introduction of HAART (May), there was a recurrence of his fevers with nodules and a halo sign on chest computed tomography (CT) scan, consistent with invasive fungal infection. As the patient declined bronchoscopy, oral voriconazole was commenced for probable invasive aspergillosis. There was complete clinical and radiological resolution on repeat imaging 3 weeks later (June). One month later (July), the patient presented with a witnessed seizure. Magnetic resonance imaging (MRI) revealed multiple bilateral ring enhancing lesions. Although he was thought to be compliant with cotrimoxazole prophylaxis, he was known to be immunoglobulin G (IgG) positive for Toxoplasma gondii and treatment for this was commenced. Seizures continued despite therapeutic levels of anticonvulsants. Dexamethasone was added after the patient presented with global dysphasia 1 month later (August) and a repeat MRI revealed that the lesions had increased in size with marked white matter oedema. A brain biopsy was performed which revealed invasive apergillosis. Aspergillus fumigatus was isolated from culture with a minimum inhibitory concentration (MIC) to fluconazole of >256 mg/L, voriconazole of 0.25 mg/L and amphotericin of 0.25 mg/L. As the patient had been on voriconazole for 3 months, combination therapy with intravenous capsofungin and liposomal amphotericin was initiated with adjunct cytokine therapy, granulocyte-macrophage colony-stimulating factor (GM-CSF). Despite combination therapy, the patient continued to deteriorate and died in October, with no recurrence of his pulmonary lesions. His HIV viral load was undetectable for the last 5 months of his life. Invasive cerebral aspergillosis occurs in 10–20% of all cases of invasive disease [5]. It is an uncommon complication in HIV infection, with 33 reported cases. The majority of cases had an additional site involved, most commonly the respiratory tract, while eight cases had isolated cerebral disease [3]. A brain biopsy is required in order to differentiate invasive aspergillosis from the more common diseases (toxoplasmosis and lymphoma) as presentations are nonspecific [6]. Treatment options for invasive aspergillosis have increased with the introduction of voriconazole and caspofungin [7, 8]. Combination therapy with caspofungin and liposomal amphotericin has been shown to have synergy in vitro [9]. Cytokine therapy, as an adjunct to antifungals, has been used previously [10]. GM-CSF and interferon-γ increase phagocytosis in vitro [11]. However, whether these effects and combination antifungal therapy translate into clinically improved outcomes is still to be determined. Although nonspecific, nodules and a halo sign on CT scan are suggestive of a fungal infection with a positive predictive value of 80% [12]. Furthermore, the patient's pulmonary disease responded appropriately to antifungal therapy. The emergence of his cerebral aspergillosis 3 months later is of concern. Voriconazole has superior (at least in animals) CSF and brain tissue penetration compared to other antifungals [13]. Fungal cells die when azoles reach adequate drug concentration to reduce ergosterol synthesis to zero. Thus it is possible that voriconazole pharmacokinetic and pharmacodynamic parameters were suboptimal despite a fully sensitive A. fumigatus isolate. There have been no reported cases of immune restoration disease (IRD) to aspergillosis. However, the time sequence (4 weeks after commencement of HAART) and presentation are consistent with an IRD [14]. The delay in the neurological presentation suggests site-specific immune responses, with the brain a known sanctuary site. This is supported by the MRI scan which revealed ongoing inflammation (white matter oedema). Thus this case may represent the first case of IRD to invasive aspergillosis. In conclusion, we wish to make the following recommendations. All immunocompromised patients with invasive fungal pulmonary disease should undergo appropriate imaging for disseminated disease. Although voriconazole has good oral bioavailability, intravenous induction with oral maintenance should be considered for cerebral disease. As the prognosis of cerebral disease is poor, clinicians should be aware of combination antifungal and adjunct cytokine therapy.