Relapsing meningitis caused by persistent cryptococcal antigens and immune reconstitution after the initiation of highly active antiretroviral therapy

We report a peculiar case of cryptococcal meningitis and subsequent immune reconstitution inflammatory syndrome (IRIS) occurring shortly after the initiation of highly active antiretroviral therapy (HAART). A 49-year-old homosexual man with headache, meningeal signs and abducens paralysis was found to be HIV-1 positive, with 16 CD4 lymphocytes/mm3 and 110 000 HIV-1-RNA copies/ml. Lumbar puncture revealed an elevated opening pressure (850 mm H2O). The cerebrospinal fluid (CSF) contained 20 white blood cells (mononuclear)/mm3, glucose 20 mg/dl, protein 79 mg/dl; coccal yeasts were seen, and antigen detection and culture were positive for Cryptococcus neoformans. Serum cryptococcal antigen detection was positive. Brain contrast-enhanced magnetic resonance imaging (MRI) was normal (Fig. 1A). The cryptococcal meningitis was successfully treated for one month with amphotericin B (1 mg/kg a day) and flucytosine (100 mg/kg a day), and weekly lumbar punctures were performed to decrease intracranial pressure. Therapy was then switched to oral fluconazole (400 mg/day) and HAART was initiated (zidovudine, lamivudine and ritonavir-boosted lopinavir). Eight days later, the patient developed fever, vomiting and headache with normal neurological findings and fundoscopy. Three weeks later, the CD4 lymphocyte count was 38 cells/mm3 and plasma HIV-1 RNA was 166 copies/ml. Lumbar puncture contained 35 white blood cells/mm3, glucose 37 mg/dl and protein 120 mg/dl. CSF India ink preparation and culture were negative, whereas cryptococcal antigen remained positive. Brain MRI showed contrast enhancement not of the leptomeninges as previously reported [1,2], but strongly of the choroid plexus (Fig. 1B, lower part) as well as of the bilateral linear structures in the frontal sulcus (Fig. 1B, upper part), suggestive of inflammatory vasculitis. HAART-related IRIS towards cryptococcal antigen(s) was suspected. Fluconazole was continued, HAART was withdrawn one month after its initiation, and methylprednisolone (48 mg/day) was started, together with hydroxychloroquine (600 mg/day). The fever and neurological symptoms disappeared. After 8 weeks, the methylprednisolone dose was tapered to 16 mg/day. On brain MRI 9 weeks after HAART withdrawal and anti-inflammatory treatment, the previously observed contrast enhancement was strongly diminished (Fig. 1C). Two weeks later, the initial HAART combination was reinstituted. At latest follow-up one month after HAART reinstitution, no relapse has occurred and anti-inflammatory treatment is being tapered further.Fig. 1.: Brain magnetic resonance images at the times of initial meningitis (A, left panel), of immune reconstitution inflammatory syndrome-related meningitis relapse (B, middle panel) and after treatment of the latter (C, right panel), respectively.This case illustrates IRIS presenting as a relapse of meningitis, which occurred after the successful treatment of initial C. neoformans meningitis and soon after HAART initiation. Restricting the literature analysis to sufficiently detailed case reports, six patients have been reported with cryptococcal IRIS presenting as relapsing meningitis [1,2] (case no. 5), [3] (case no. 3), [4] (case nos. 4 and 5), [5]. Relapse occurred a median of 3.5 weeks after HAART initiation (range 10 days to 10 months). Whereas the initial meningitis was CSF culture-positive for C. neoformans, the IRIS-related meningitis relapse was, in six out of six cases plus ours, CSF cryptococcal antigen-positive but CSF yeast culture-negative. Contrast-enhanced brain MRI, reported in only two out of six cases, showed diffuse meningeal enhancement [1,2]. Our case allows the comparison of sequential brain MRI at the time of the initial meningitis, IRIS meningitis relapse, and after treatment of the latter, respectively. Of particular interest is the MRI at the time of IRIS (Fig. 1B), showing strong enhancement of the choroid plexus and of linear structures within the sulci, corresponding to choroid plexitis and (peri)-vasculitis. These findings were initially absent (Fig. 1A), and faded during IRIS treat ment (Fig. 1C). They differ from tm the spontaneously hyperdense, non-enhancing cryptococcal loads in the Virchow–Robin spaces, reporported in AIDS-related cryptococcal meningitis [6]. They probably reflect intense IRIS towards cryptococcal antigen(s) rather than cryptococcal infectious relapse. These MRI findings may be helpful in considering cryptococcus-related IRIS and starting appropriate treatment. The therapy of meningitis relapse caused by IRIS towards cryptococcal antigen(s) has not been studied prospectively. In the six literature cases, HAART interruption is not mentioned, and corticosteroids were administered to only two out of six patients [1–5]. As HAART-associated immune reconstitution causes the inflammatory syndrome, presenting here as meningitis relapse, both the interruption of HAART and the institution of anti-inflammatory treatment are pathophysiologically sound and were applied to our patient. Hydroxychloroquine was given for two reasons: for its anti-inflammatory activity that may facilitate tapering of the corticosteroid dosage, and for the modest anti-HIV activity it shares with chloroquine [7]. (Hydroxy)chloroquine was not used for its anticryptococcal activity [8], as IRIS-associated relapsing meningitis is CSF yeast culture-negative. Our case does not allow a conclusion as to which treatment modality (HAART withdrawal, anti-inflammatory treatment or both) was most important. A trial should be initiated to determine the optimal timing to start HAART after cryptococcal meningitis, and the best therapeutic approach in post-HAART cryptococcal IRIS. Acknowledgements The authors would like to thank Dr Olivier Lortholary, Institut Pasteur, Paris, for helpful discussions.