Predictors and long-term outcome of seizures in human immuno-deficiency virus (HIV)-negative cryptococcal meningitis

The medical records of patients with HIV-negative cryptococcal meningitis admitted to Kaohsiung Chang Gung Memorial Hospital between 1986 and 2007 were reviewed for blood cultures, microbiologic records, and neuro-imaging findings using pre-existing standardized evaluation forms. Kaohsiung Chang Gung Memorial Hospital is a 2482-bed acute-care teaching hospital that provides both primary and tertiary referral care.

Patients with HIV-negative cryptococcal meningitis were included if they had: (1) positive isolation of Cryptococcus neoformans (C. neoformans) in one or more cerebro-spinal fluid (CSF) cultures, positive CSF cryptococcal antigen titer, or positive CSF India ink and clinical features of meningitis; or (2) isolation of C. neoformans in blood culture, with clinical presentations of meningitis and typical CSF features [1].

Patients were excluded if they had a history of seizures; pre-existing neurologic conditions with various neurologic deficits (e.g. stroke, head trauma, and hypoxic encephalopathy); or regularly took AEDs for epilepsy or other clinical indications (e.g. trigeminal neuralgia or neuropathic pain). Thus, only 180 of 185 patients were enrolled for analysis.

Seizures were classified according to the recommendations of the International League against Epilepsy [8]. Status epilepticus (SE) was defined as a continuous behavioral seizure activity or repetitive seizures without full recovery of neurologic function between seizures, lasting longer than 30 minutes [9]. Seizure was defined according to those used in previous studies [10]-[12]. Seizures occurring after HIV-negative cryptococcal meningitis were causally related to the HIV-negative cryptococcal meningitis itself. A provoked (acute symptomatic) seizure was one that occurred in close temporal relation with HIV-negative cryptococcal meningitis, which was the presumed etiology [10]-[12]. In contrast, an unprovoked seizure was a seizure occurring in the absence of one or more precipitating factors, including events in patients with previously stable (non-progressing) HIV-negative cryptococcal meningitis [10]-[12].

Epilepsy was the occurrence of repeated unprovoked seizures [5],[10]-[12]. Based on seizures onset in relation to the clinical ictus of HIV-negative cryptococcal meningitis, patients with seizures were divided into two sub-types. Early seizures were those occurring within two weeks of the infection, whereas late seizures were those occurring after two weeks.

All of the patients underwent brain computed tomography (CT) scan at the emergency room. Follow-up brain CT scans and/or magnetic resonance imaging (MRI) were performed if there was clinical deterioration, including acute onset focal neurologic deficits, seizures or status epilepticus, progressively disturbed consciousness, and for post-neurosurgical procedure. Hydrocephalus was judged retrospectively by dilated temporal horn of the ventricle without obvious brain atrophy and/or an Evan’s ratio >0.3 on initial CT scan. The Evan’s ratio was the ratio of the ventricular width of the bilateral frontal horn to the maximum biparietal diameter [13],[14].

The hospital’s standard protocol for HIV-negative cryptococcal meningitis was to administer AEDs only to those with acute symptomatic seizures. Prophylactic AED therapy was not given to asymptomatic patients in the acute stage. The AEDs were administered to patients with HIV-negative cryptococcal meningitis during hospitalization and were discontinued if there were no unprovoked seizures on follow-up.

The follow-up period was terminated by death or by the end of the study (December 2007). Most patients were followed-up at the out-patient department after discharge, while others were interviewed by telephone to identify neurologic outcomes. The frequency of seizures was determined by a Seizure Frequency Scoring System, which was slightly modified from Engel et al. [15] and a previous study [16]. Good outcome was defined as survival while poor outcome was defined as fatality during the follow-up period. The hospital’s Institutional Review Committee on Human Research approved the study protocol.

In the study hospital, anti-fungal therapy in patients with HIV-negative cryptococcal meningitis was started according to the preference of the attending physician [1],[17]-[18]. The patients were divided into three major groups: Group I received Amphotericin B intravenously, with or without 5-flucytosine, for 6–10 weeks, for a total intravenous dose of 2–3 g; Group II received Amphotericin B with fluconazole 300–400 mg intravenously in the first 2–3 weeks, before treatment was switched to oral fluconazole alone (300–400 mg daily for 10 weeks); and Group III, received Fluconazole 400–800 mg intravenously daily during the first 2 weeks and then switched to an oral dose of 300–400 mg daily for the duration of primary therapy.

Two separate statistical analyses were performed. First, the demographic data between the good and poor outcome groups were compared. Categorical variables were compared using the Chi-square test or Fischer’s exact test, as appropriate. Continuous variables within two groups were compared using the independent t-test for parametric data and the Mann–Whitney U test for non-parametric data. Second, to determine risk factors independently associated with seizures or fatality, the clinical and laboratory parameters with a p value <0.05 in univariate analysis were entered in a logistic regression model using a stepwise entry system. All statistical analyses were conducted using the SAS software package, version 13.0 (2002, SAS Statistical Institute, Cary, North Carolina).

The 180 patients (age, 18–88 years) with HIV-negative cryptococcal meningitis included 110 males and 70 females (mean age, 50.2 and 56.5 years, respectively). Their characteristics in terms of seizures (Table 1) revealed that 28 had seizures.

The neuro-imaging findings of the HIV-negative cryptococcal meningitis patients with or without seizures were listed in Table 2. Hydrocephalus, gyral enhancement, basal cistern effacement, and cerebral infarction were the four most common findings in the seizure group, while hydrocephalus, dilated Virchow-Robin space, and gyral enhancement were the three most common findings in the non-seizure group.

Regarding seizure onset in relation to the clinical ictus of HIV-negative cryptococcal meningitis, 13 had early seizures and 15 had late seizures (Table 1). The mean time interval from HIV-negative cryptococcal meningitis to first seizure in the early and late seizure groups were 1.5 and 51.4 days, respectively (Figure 1). Nine of the 28 cases (32%) occurred within 24 hours of presentation.

As regards seizure sub-types among the 28 cases, 18 were generalized tonic-clonic (64.3%, 18/28), 5 were focal seizures with secondary generalized seizures (17.8%, 5/28), three were myoclonic seizures (10.7%, 3/28), one was simple partial seizure (3.6%, 1/28), and the remaining one was complex partial seizure (3.6%, 1/28). Three (10.7%, 3/28) of the 28 cases progressed to status epilepticus.

In terms of treatment, 22 were treated with phenytoin alone, three with valproic acid alone, and the remaining three with both valproic acid and phenytoin. The mean duration of AED therapy for those who were discharged on AEDs was 12.9 ± 7.3 months. None of the 28 patients had drug-related complications.

The mean duration of hospitalization for the seizure and non-seizure groups was 77.54 ± 56.14 and 56.9 ± 5.29 days, respectively (p = 0.05). Fifteen (53.6%, 15/28) died in the acute phase of HIV-negative cryptococcal meningitis, while 13 (46.4%, 13/28) survived. In terms of seizure control after discharge among the 13 survivors, 5 (38.4%, 5/13) were seizure-free without AED therapy, 6 (46.2%, 6/13) were seizure-free under AED therapy, one (7.7%, 1/13) had 4–11 seizures per year, and the remaining one (7.7%, 1/13) had 1–3 seizures per month.

Twenty-three (12.8%, 23/180) cases were lost to follow-up after discharge during the study period. In total, 123 of 180 (68.3%, 123/180) patients survived. The fatality rate in the seizure and non-seizure groups was 54% (15/28) and 12% (42/152), respectively. The mean age at onset, median days of hospitalization, and median GCS on admission between the two groups were listed in Table 3.

Comparisons of clinical features and neuro-imaging findings between the seizure and non-seizure groups were listed in Table 3. Statistical analysis between the two patient groups revealed that only thrombocytopenia (p = 0.028) and neuro-imaging findings at presentation showing gyral enhancement (P = 0.042) were significant variables. After stepwise logistic regression analysis, both thrombocytopenia (p = 0.035, 95% CI: 1.09-9.76) and neuro-imaging findings at presentation showing gyral enhancement (p = 0.048, 95% CI 1.01-8.873) were independently associated with seizures.

Comparisons of clinical features and neuro-imaging findings between survivors and non-survivors after a minimum of 5 years of follow-up were listed in Table 4. Statistical analysis between the two groups revealed that mean GCS on admission (p < 0.001), presence of underlying diseases as liver cirrhosis (p = 0.004) and iatrogenic Cushing syndrome (p = 0.016), and the presence of seizures (p = 0.007) and CSF cryptococcal antigen titer ≥1:1024 and lactate level (p = 0.015) were all significant variables. Significant univariate factors used in the stepwise logistic regression were mean GCS on admission, presence of underlying diseases (liver cirrhosis and iatrogenic Cushing syndrome), seizures, and CSF cryptococcal antigen titer (≥1:1024 or not) and lactate level.

After analysis, only mean GCS on admission (OR: 0.835, 95% CI, 0.737-0.946; p = 0.005), presence of acute seizures (OR: 3.683, 95% CI: 1.332-10.18; p = 0.012), presence of underlying diseases such as liver cirrhosis (OR: 6.580, 95% CI: 1.348-32.12; p = 0.02) and iatrogenic Cushing syndrome (OR: 4.574, 95% CI: 1.185-17.66; p = 0.027) and CSF cryptococcal antigen titer ≥1:1024 (OR: 3.614, 95% CI: 1.093-6.247; p = 0.031) were independently associated with fatality. Any reduction of one point of GCS increased the fatality rate by 17%.

The present study examined the predictive factors and outcomes of seizures in patients with HIV-negative cryptococcal meningitis and produced two major findings. First, patients with seizure had longer mean length of hospitalization. Lower mean GCS on admission and the presence of seizures were independently associated with fatality, and any reduction of one point of GCS increased the fatality rate by 17%. Second, seizures in patients with HIV-negative cryptococcal meningitis might have delayed manifestations although most seizures occur during the acute phase of the meningitis. None of the patients had a first seizure occurring more than one year after the diagnosis of HIV-negative cryptococcal meningitis. The occurrence of seizures was commonly observed within 24 hours of presentation in one study [3], as well as in the present study, and accounted for 100% (8/8) and 32% (9/28) of cases, respectively.

The use of prophylactic AED therapy for preventing seizures in patients with HIV-negative cryptococcal meningitis remains unclear. A possible benefit is the reduction in morbidity after seizures. At present, no formal recommendations can be made regarding the use of prophylactic AEDs for patients with HIV-negative cryptococcal meningitis based on currently available well-designed randomized controlled trials. Phenytoin is the most widely used AED in this study and there are no observed adverse effects related to its use.

In this study, lower mean GCS on admission, presence of acute seizures, presence of underlying diseases like liver cirrhosis and iatrogenic Cushing syndrome, and CSF cryptococcal antigen titer ≥1:1024 are independently associated with fatality. The CSF cryptococcal-antigen titer >1:1024 may imply severe infection. It is also a significant prognostic factor in both HIV-negative and HIV-positive cryptococcal meningitis [1],[19]. There is a striking correlation between initial consciousness level and therapeutic outcome, and this is consistent with findings of other reports [20],[21]. The presence of underlying diseases such as liver cirrhosis and iatrogenic Cushing syndrome is also a well-known predisposing factor of poor outcome [22],[23].

Seizures complicating HIV-negative Cryptococcal meningitis have worse outcomes and longer hospitalization. There may be several reasons for the increased mortality and morbidity in patients with seizures. For instance, patients are commonly associated with underlying medical conditions. Moreover, later neurologic morbidity is attributable to hypoxia, lactic acidosis, increased intra-cranial pressure (IICP), and autonomic dysfunction.

First, this was a retrospective analysis and therefore subject to bias of unmeasured factors (e.g. possible reporting bias due to patient selection presented to the hospital). It was also not possible to assess the effects of prophylactic AEDs after the acute stage of HIV-negative cryptococcal meningitis. Second, patients who took AEDs for epilepsy or other clinical indications (e.g. trigeminal neuralgia or neuropathic pain) or those with pre-existing neurologic deficits were excluded. Thus, there was continued uncertainty in assessing the incidence of unprovoked seizures after HIV-negative cryptococcal meningitis in non-selected patients. Third, the impact of surgical intervention for hydrocephalus on the frequency of seizures was not clear. Surgical interventions might cause potential brain insults and lead to unprovoked seizures during follow-up. Lastly, most patients in this study were treated with AEDs after their first acute symptomatic seizure, in accordance with the study protocol. Thus, the findings might underestimate the “true” frequency of seizures associated with the “natural history” of untreated unprovoked seizures.

The study was approved by Chang Gung Memorial Hospital’s Institutional Review Committee on Human Research.