Cerebral thrombophlebitis as a paradoxical reaction in HIV-patients

Abstract

The immune reconstitution inflammatory syndrome (IRIS) occurs frequently in patients infected with HIV and an opportunistic pathogen, usually a few weeks after the initiation of a combination antiretroviral therapy (cART) [1]. We report here two cases of HIV-infected patients who developed cerebral thrombophlebitis after the diagnosis and treatment of cryptococcal meningitidis, suggesting unusual presentation of IRIS. Case 1 A 46-year-old HIV-1-infected African man was admitted on 13 May 2005 for right hemiparesia, diplopia, left VII and bilateral VI palsies. CD4 cell counts were 10 cells/μl and plasma viral load (pVL) was 1911 copies/ml despite cART with zidovudine, lamivudine and abacavir. His cerebrospinal fluid (CSF) showed 15 lymphocytes/μl, protein and glucose levels of 0.43 g/l and 2.4 mM/l, respectively. CSF direct examination and culture found Cryptococcus neoformans. Latex agglutination test for cryptococcal polysaccharide antigen was 68 enzyme immunoassay (EIA) in serum and 597 EIA in CSF. Cerebral magnetic resonance imaging (MRI) showed left frontal hemorrhagic lacuna and ischemic arteriolar lesions in lenticular nuclei. He was treated with intravenous amphotericin B and flucytosine, and high dose steroids. On 17 June 2005, while on fluconazole, steroids were interrupted and antiretroviral therapy was switched to emtricitabine, tenofovir and boosted atazanavir. On 26 July, he was hospitalized for headache, fever and visual acuity loss. CD4 cell count was 28 cells/μl and pVL was less than 50 copies/ml. CSF showed 14 cells/μl and cryptococcal yeasts, but culture was negative. The C. neoformans antigen was 45 EIA in blood and 12 EIA in CSF. Brain MRI showed thrombophlebitis of the superior sagittal and left transverse sinuses. Efficient anticoagulation and prednisone were initiated, with improvement of brain MRI 2 months later. Case 2 On 19 November 2005, a 26-year-old African woman was hospitalized for photophobia, neck stiffness, right VI palsy revealing HIV infection and cryptococcal meningitis [20 cells/μl, 78% lymphocytes, positive cryptococcal antigen (1000 EIA) and C. neoformans in CSF]. CD4 cell count was 3 cells/μl and pVL was 200 000 copies/ml. She was treated with intravenous amphotericin B and flucytosine, then fluconazole. Eight days later, she presented a visual acuity loss. Brain MRI showed bilateral and symetric dilatation of the optical nerve sheathes and intracranial hypertension. Neurological signs disappeared after introduction of prednisone. On 17 January 2006, she started cART with emtricitabine, tenofovir and boosted atazanavir. On 6 March, she presented with blindness, confusion, neck stiffness, left VI palsy. CSF showed an opening pressure of 45 cmH2O, no cellularity, normoglycorachia, positive India ink coloration but negative culture for C. neoformans. MRI showed disappearance of venal flow in superior sagittal sinus and both transverse sinuses and intracranial hypertension. CD4 cell count was 9 cells/μl and pVL was 52 copies/ml. She received intravenous methylprednisolone, acetazolamide and curative anticoagulation. A ventriculo-peritoneal derivation was performed without subsequent visual improvement. Discussion IRIS may develop in 30% of HIV-infected patients with cryptococcal infection [2]. The most common neurological presentations are a recurrent meningitis with negative fungal culture, multiple cerebral cryptococcoma, diffuse microabscesses, marked edema, or abnormal contrast media uptake in meninges [1,3]. Cerebral venous occlusion has already been described in association with viral, bacterial or fungal infections and cancer, in HIV-infected patients [4–5] and in non-HIV patients [6–11], but never in association with cryptococcosis. In one retrospective study, risk factors of cryptococcosis-associated IRIS in HIV-infected patients were previously unknown HIV infection, CD4 cell count lower than 7 cells/μl, fungemia and cART initiation within 2 months of cryptococcosis diagnosis [1]. Median CD4 cell counts at the time of IRIS were 215 cells/μl [1]. Our two patients maintained low CD4 cell counts below 50 cells/μl at the time of IRIS, and none of them had significant increase in their CD4 count. Nevertheless, cART might have beneficial immune effects mediated by mechanisms other than CD4 cell increment, as previously suggested [12,13], and low CD4 count does not rule out diagnosis of IRIS. In our patients, the time from initial diagnosis of cryptococcosis to effective cART was 1 and 2 months and the time from initiation of an efficient cART to IRIS was 1 and 1.5 months, which were shorter than that reported in other cases [1,14]. Both had cART introduced shortly after the diagnosis of cryptococcosis (one of them switched to a new cART due to bad compliance and virological failure of his previous treatment). Both also received high dosages of corticosteroids, but thrombophlebitis developed subsequently when therapy was decreased. Optimal treatment of patients at high risk of IRIS is still under discussion, with early or delayed introduction of cART and preventive long high dosage corticosteroids with potential adverse effects. Additional studies are thus needed.