Cerebral infarction in HIV-negative patients with cryptococcal meningitis: its predictors and impact on outcomes

Cryptococcal meningitis (CM) is one of the most common fungal infections in the central nervous system (CNS), characterized by a long disease course with difficult management. Despite the advances in its diagnosis and treatment, CM is still associated with a high mortality rate (20–70%), and about 50% of the survivors have moderate to severe neurological deficits, leading to a great health care burden [1, 2]. More than 80% of CM cases occur in human immunodeficiency virus (HIV)-infected patients, but it can also be found in immunocompetent individuals [3]. In China, most of the CM cases were reported in patients without HIV infection [4].

Previous studies have shown that 13–21.5% of the patients with CM experience cerebral infarction during the course of the illness [5, 6], but for HIV-negative patients with CM, the patterns of vascular involvement, predictors of cerebral infarction and the patients’ outcomes remain poorly documented [5]. In this context, we conducted this retrospective study to evaluate the predictive factors of acute/subacute cerebral infarction (ASCI) in HIV-negative patients with CM and assess the impact of cerebral infarction on the outcomes of the patients.

ASCI can occur in patients with CM, but its presentations are often masked by the manifestations of CM [5, 6, 8]. Mishra et al. [5] reported that the incidence of cerebral infarction was 13% in CM patients, including both HIV-positive and HIV-negative patients, but only 30% of the patients in their study underwent brain MRI, and the routine CT examination has a relatively low detection rate for small acute infarcts. In the study conducted by Chen et al. [8] 37 out of 54 adult HIV-negative patients with CM had brain MRI at admission, and the detection rate of ASCI was 18.9%. Wan-Chen et al. [6] reported an ASCI incidence of 21.5% in adult HIV-negative patients with CM based on brain MRI findings using a 1.5T scanner. In this study, we recorded a higher ASCI incidence of 26.4% among the 53 patients, possibly because all the patients underwent brain MRI studies using a 3.0T scanner early after admission, and the high field MRI with a high resolution is more sensitive for detecting small infarcts.

Most of our patients with ASCI had infarctions of the lacunar type (13/14, 92.9%); multiple cerebral infarctions occurred in the majority of the patients with ASCI (10/14, 71.4%), and over half (8/14, 57.1%) of the patients had bilateral infarcts. These findings suggest that in CM patients, cerebral infarction occurred most likely in the brain region supplied by a perforator-type artery, which may explain the relatively high frequency of multiple lacunar infarctions in these patients. We noted a discrepancy concerning the infarct type from a previous report, where all the infarcts were of the lacunar type and 70% of them were unilateral [5]. This discrepancy may partly be attributed to a high rate (65%) of HIV infection in the enrolled cases in the reported study, as neurovascular complications may occur in CM patients with HIV infection through different mechanisms from those without HIV infection [5]. We found multiple large cerebral infarction only in one patient (7.1%), a rate much lower than those in previous studies [5, 8, 9]. Although lacunar infarction is the most common type of ASCI in HIV-negative patients with CM, cases of large infarctions due to large vessel involvement had been reported, as a result of direct vascular invasion or secondary to associated inflammations [10, 11].

Our finding suggests that in CM patients, the most common site of infarction was the basal ganglia-corona radiate (57.2%, 8/14), followed by the frontal lobe, brainstem, cerebellum, thalamus, parietal lobe, occipital lobe, and the temporal regions. We observed a relatively high incidence of brainstem-cerebellum involvement in CM patients with ASCI (57.2%, 8/14), and this posterior circulation infarction used to be thought uncommon in CM [12]. A previous study suggests that the brain regions supplied by the middle cerebral artery are the most frequent regions of infarct occurrence, and the vertebrobasilar system is less frequently affected by bathing in the basal exudates [5]; but our finding indicates otherwise: infarction can occur in both the anterior and posterior cerebrovascular territories in HIV-negative patients with CM. This difference may be due to our high detection rates (10/14, 71.4%) of posterior fossa exudates, which increase the risk of infarction at the corresponding position. Kumar et al. [13] suggested that caudate head and cerebellar infarctions were also common in CM, which was consistent with our finding.

Posterior fossa exudates was one of the main cranial MRI findings in our patients and a significant factor contributing to the development of ASCI. Wan-Chen Tsai et al. [6] concluded that basal meningeal enhancement was associated with the occurrence of ASCI, possibly due to the intense inflammatory reactions in the adjacent small vessels elicited by meningeal exudates, especially in the basal and cerebellar meninges. Zimelewicz et al. [14] reported a case of central nervous system vasculitis in an immunocompetent CM patient, who showed exudate in the posterior cranial fossa on MRI and bilateral atypical vasculitis in the posterior cranial fossa by digital angiography. In this study, the high detection rate of posterior fossa exudates among the patients with ASCI provides further evidence to support the presumption that posterior fossa exudates increases the risk of ASCI in HIV-negative CM patients.

We found that hydrocephalus was independently associated with the occurrence of ASCI in HIV-negative CM patients. According to the study of Chen et al. [9], hydrocephalus might occur in CM patients (due to the inflammatory exudates, fungal polysaccharides and high protein levels) to cause CSF accumulation in the arachnoid villi or subarachnoid space to block CSF outflow and absorption. Hydrocephalus may stretch the local compromised arteries and potentially causes cerebral infarction [9]. Our finding provides further evidence that hydrocephalus increases the risk of ASCI in HIV-negative CM patients.

Here we show that the CM patients with ASCI had significantly lower GCS scores than those without ASCI (P = 0.015). As GCS scores can not fully evaluate consciousness status of the patients before coma onset, we used the parameter consciousness disturbance for multivariate logistic analysis. Table 3 shows that consciousness disturbance was a significant risk factor for ASCI in the CM patients (P = 0.02). Previous studies also demonstrated that CM patients with ASCI had a higher frequency of consciousness disturbance and a higher rate of positive CSF culture for cryptococcus [2, 8]. Kumar et al. [13] proposed that altered consciousness on admission was associated with an increased risk of infarction. We presume that consciousness disturbance increases the disease burden and causes more intense vascular inflammation to contribute to cerebral infarction in CM patients.

Currently the correlation between CM and cerebral vasculopathy remains unclear. Several mechanisms have been proposed to explain the association between CM and cerebral infarctions. First of all, inflammatory changes in the vascular endothelium, resulting directly from the toxic effect of the Cryptococcus and from vasculitis due to antigen–antibody complex deposition, are most likely to cause vascular complications in CM. Exudative meningitis in the posterior fossa results in strangulation, vasospasm, constriction, periarteritis and even necrotizing panarteritis of the blood vessels, leading eventually to thrombosis and occlusion [14‐16]. Secondly, the extensive inflammatory fibrosis of the subarachnoid space caused by Cryptococcus may mechanically compress the small veins, and thus increases blood flow resistance [17]. Thirdly, the inflamed vessels are stretched due to dilation of the ventricles [18], as shown by a higher incidence of hydrocephalus in the patients with ASCI than those without ASCI (50% vs 20.5%) in this study. In addition, cryptococcal infections are also thought to precipitate the vulnerability of atherosclerotic plaques and hence lead to infarction [19]. Finally, multiple acute infarctions are very likely secondary to septic embolism or mechanical obstruction [20, 21]. These assumptions provide possible explanations for the wide distribution of ASCI across both the anterior and posterior cerebrovascular territories in our CM patients, but the exact mechanisms of ASCI in these patients still remain to be further investigated.

In this study, 9 (64.3%) of the 14 patients with ASCI had poor outcomes. Cerebral infarction, along with other factors such as cerebral hernia, consciousness disorder, hydrocephalus, visual impairment, and increased intracranial pressure, all independently determines the poor outcomes of CM [21‐23]. Chen et al. [8] reported poor outcomes of 7 CM patients with ASCI, among whom 3 died, 2 were bedridden, and 2 were wheelchair bound. The high rate of poor outcomes in that report may result largely from the failure of early detection of asymptomatic ASCI because not all the patients enrolled had MRI examination. We also found that the CM patients with ASCI had a significantly higher likelihood of having poor outcomes than those without ASCI (64.3% vs 12.8%), which is consistent with the findings of previous studies [21‐24].

In summary, cerebral infarction in HIV-negative patients with CM often presents as multiple lacunar infarctions involving all the cerebrovascular territories. Conscious disturbance, hydrocephalus, and posterior fossa exudates are all significant predictors of ASCI in CM patients, who often have unfavorable outcomes. But in view of the retrospective nature and the relative small sample size of this study, the reliability of our conclusions remains to be further tested by future studies in more balanced cohorts and larger sample sizes.