Management of Cryptococcal Meningitis with Extremely High Intracranial Pressure in HIV-Infected Patients

SIR—Cryptococcal meningitis develops in antiretroviral-naive patients with AIDS at an estimated incidence of 5%–10% [1] and is a condition that is often associated with elevated intracranial pressure (ICP) of ⩾200 mm H2O in >50% of cases [2]. Elevated ICP may result in increased mortality and morbidity. In a study by Van der Horst et al. [3], almost all of the early deaths (13 of 14) and 40% of the deaths during weeks 3–10 were associated with elevated ICP. Visual loss was also observed as a consequence of elevated ICP [4]. Frequent lumbar punctures have been advocated for the management of increased ICP [2, 5]. However, the risk of brain herniation as a result of frequent lumbar punctures for elevated ICP has been a major concern among clinicians [6, 7]. Here, we present our experience with frequent lumbar punctures and large-volume drainage of CSF in the management of extremely high ICP (>600 mm H2O) in HIV-infected patients who had cryptococcal meningitis. Between June 1994 and August 2003, a total of 35 patients were given a diagnosis of cryptococcal meningitis on the basis of a CSF culture positive for Cryptococcus neoformans or a positive result of a CSF cryptococcal antigen test or CSF India ink smear. During the study period, lumbar puncture was performed according to a standardized protocol after obtaining written informed consent from patients or their family members. CT of the brain was performed to rule out mass lesions before performing lumbar puncture for patients suspected of having cryptococcal meningitis. The frequency of the lumbar puncture was determined by previous pressure levels and the symptoms of the patients. If the patient's initial opening pressure was ⩾350 mm H2O, lumbar puncture was performed daily to drain 20–30 mL of CSF until the pressure was <350 mm H2O. Later, we performed lumbar puncture every 2–3 days until the pressure measured <200 mm H2O, followed by 1 lumbar puncture per week until the patient's discharge from the hospital. For each patient, a standardized case-record form was used to document the initial opening and closing pressures of lumbar punctures and the results of CSF analysis, which included fungal culture, India ink smear, WBC count, and tests for levels of glucose, protein, and cryptococcal antigen. Age, sex, risk factors, CD4/CD8 cell count, peripheral virus load when cryptococcal meningitis was diagnosed, and antiretroviral therapy received were also recorded. Amphotericin B at a daily dosage of 0.8–1 mg/kg was administered as induction therapy for 2–3 weeks, followed by maintenance therapy with oral fluconazole. The patients were followed up until death, loss to follow-up, or the end date of the study (31 December 2003). There were 9 patients who had ever had extremely high ICP; 5 patients had extremely high ICP at presentation, and 4 developed extremely high ICP during hospitalization. They were all male patients, with a median age of 32 years (range, 23–57 years). Seven patients developed concurrent cryptococcemia, and 2 had cryptococcal pneumonia. Three patients developed diplopia, and 1 had blurred vision. Five patients had altered consciousness at presentation, and 2 other patients developed this complication during the hospital stay. The median initial WBC count in the CSF was 0 cells/μL (range, 0–5 cells/μL), and the median initial CSF cryptococcal antigen titer was 1 : 1024 (range, 1 : 512 to 1 : 65536). The median CD4 count at the time of diagnosis of cryptococcal meningitis was 25 cells/μL (range, 9–53 cells/μL). The patients underwent multiple lumbar punctures and received glycerol or mannitol for increased ICP. The median number of lumbar punctures performed was 6 (range, 1–29). A lumbar drain was placed in one patient, and another patient underwent ventriculoperitoneal (VP) shunting for refractory elevated ICP. The median duration of follow-up was 360 days (range, 1–1009 days). The condition of the patient who developed visual loss because of extremely high ICP improved with external CSF drainage. The median duration from hospital admission to the initiation of treatment with amphotericin B was 1 day (range, 1–3 days). The median durations of treatment with amphotericin B and fluconazole were 22 days (range, 1–44 days) and 312 days (range, 0–900 days), respectively. Death directly attributed to cryptococcal meningitis occurred in 2 patients. One patient presented with nausea and vomiting followed by altered consciousness and seizures, which occurred before lumbar puncture was performed. Signs of increased ICP (headache, nausea, and vomiting) resolved gradually, and the altered consciousness returned to normal after another lumbar puncture and concurrent administration of amphotericin B and fluconazole. However, later, altered consciousness and profound shock developed, and a follow-up lumbar puncture revealed an opening CSF pressure of ⩾600 mm H2O. The second patient, who had a 1-month history of cavitary pneumonia, presented with altered consciousness and septic shock due to cryptococcal fungemia, meningitis, and pneumonia before a lumbar puncture was performed. The patient died on the third day of hospitalization, despite the initiation of treatment with amphotericin B soon after the lumbar puncture. Except for the 2 patients who died and the patient with VP shunt placement, the patients had an ICP of <400 mm H2O after a median of 3.5 lumbar punctures (range, 1–27). Only 1 patient had a relapse episode, which occurred after treatment with amphotericin B for 13 days and oral fluconazole for 15 days. Of the other 26 patients without extremely high ICP, 11 patients remained alive and 14 patients died, with 3 deaths directly caused by cryptococcal meningitis. Elevated ICP due to cryptococcal meningitis was postulated to be the result of failed CSF reabsorption in the channels of subarachnoid villi and lymphatic vessels due to elevated CSF viscosity, the accumulation of fungal polysaccharides that form microscopic plugs, and the fungal cells themselves [4, 8]. External lumbar drainage by daily lumbar puncture, insertion of a lumbar drain, or implantation of a VP shunt is recommended if neither focal neurologic signs nor obtunded consciousness with space-occupying lesions visible by radiographic imaging were observed [2, 9]. However, frequent lumbar punctures and large-volume drainage in the presence of elevated ICP have raised concerns among clinicians about the risk of brain herniation [6]. Antinori and colleagues [6] reported 2 cases of cryptococcal meningitis in patients who experienced loss of consciousness and died soon after lumbar puncture. At autopsy, brain herniation was considered by Antorini et al. [6] to be the cause of death; however, inadequate CSF drainage and the resultant brain herniation—not the lumbar puncture—was suspected by Graybill and Sobel [10] to be the cause of death. In our patients with cryptococcal meningitis who had extremely high ICP and poor prognostic predictors of death during initial therapy (i.e., abnormal mental status, a CSF antigen titer of >1:1024, and a CSF WBC count of <20/μL [11]), we found that, with frequent, multiple lumbar punctures combined with antifungal therapy consisting of amphotericin B and fluconazole, 7 patients survived the disease. The condition of the patient who developed visual loss because of extremely high ICP improved with our aggressive external CSF drainage. The 2 deaths were most likely caused by delayed presentation for appropriate HIV care. In 1 of the 2 patients, signs of increased ICP resolved gradually, and altered consciousness returned to normal after multiple procedures. However, extremely high ICP recurred and resulted in death. This case should alert clinicians to the possibility of recurrent high ICP despite antifungal therapy, and careful monitoring of the ICP is indicated. In conclusion, the findings of our study support the recommendation of aggressive use of lumbar puncture for the management of elevated ICP to reduce the risk of early mortality and late morbidity, even for patients with extremely high ICP. Multiple lumbar punctures and large-volume CSF drainage in HIV-infected patients with elevated ICP due to cryptococcal meningitis is safe after the presence of space-occupying lesions is excluded. Conflict of interest. All authors: No conflict.