The Diagnostic Value of Metagenomic Next-Generation Sequencing in Angiostrongylus cantonensis Encephalitis/Meningitis

Objective: Angiostrongylus cantonensis (A. cantonensis) is an important parasite, which causes neurological infection in children, and is generally difficult to diagnose. We analyzed the significance of metagenomic next-generation sequencing in diagnosing A. cantonensis encephalitis/meningitis in children to provide a reference for clinical diagnosis. Methods: The case collection of A. cantonensis encephalitis/meningitis was from July 2018 to August 2020. The patients’ clinical characteristics and pathogen were described, and diagnostic sensitivity methods for A. cantonensis encephalitis/meningitis were compared and analyzed, including parasite antibody detection and metagenomic next-generation sequencing (mNGS) detection in different samples. Results: Eleven cases were diagnosed with A. cantonensis encephalitis/meningitis, including six males (54.5%) and five females (45.5%), and the age ranged from 1 to 13 years with a median of 21 months (IQR: 15.6, 96). All patients were undiagnosed upon admission, of which ten cases had neurological symptoms or signs, six cases (54.5%) had a CSF antibody-positive rate was 27.2%, and mNGS-positive rate of CSF was 81.8%, and the positive diagnostic rate of the latter was significantly higher than the former two (p = 0.035 < 0.05 and p = 0.030 < 0.05, respectively). Conclusion: Metagenomic next-generation sequencing has a higher positive rate for diagnosing infection and plays an important role in clarifying diagnosis and reducing misdiagnosis.


Introduction
Angiostrongylus cantonensis is parasites prevalent in Southeast Asia and the Pacific Ocean. Humans can be infected by ingesting uncooked intermediate hosts, such as snails, crabs, freshwater shrimp, frogs, and certain fishes, or through contact with soil or water sources contaminated by parasites [1]. It can develop into stage IV and stage V larvae in the human host but not into adults. Its larvae are neurotrophic and migrate to the nervous tissue after infecting the human body, which can cause acidophilic meningitis, but can also cause encephalitis, encephalomyelitis, and even eye involvement [2].
The diagnosis of A. cantonensis encephalitis/meningitis is mainly based on epidemiological history, clinical manifestations, cerebrospinal fluid acidophilia, and immune examination. The gold standard for diagnosis is the discovery of A. cantonensis in CSF. However, it is usually difficult. Although antibody detection of serum and CSF are the most commonly used clinical diagnostic methods, they both lack sensitivity [3]. Alternatively, the clinical metagenomic next-generation sequencing (mNGS) has developed rapidly in recent years and has played an important role in the accurate diagnosis of infectious diseases. The mNGs technique is advantageous because there is no need to preset, cultivate, and no preferences; results are obtained by extracting the sum of all microbial genetic material (DNA and RNA) from clinical specimens, constructing a library, and analyzing by highthroughput sequencing and suitable bio-informatics algorithms; and it is fast and has comprehensive identification of all types of microorganisms (bacteria, fungi, viruses, parasites, etc.) detection technology [4]. It is also extremely competitive in the pathogenic diagnosis of the most difficult-to-diagnose cases or immunocompromised patients [4].
Although the infection of A. cantonensis is uncommon, it can usually cause severe clinical manifestations and neurological sequelae. Due to the low positive rate of pathogenic diagnosis, the clinical treatment of the disease could be delayed or even missed, which increases the risk of deterioration and sequelae. Here, we study children with A. cantonensis encephalitis/meningitis in the past three years and compared the sensitivity and specificity of different diagnostic methods to provide a clinical diagnostic reference.

Patients and Methods
Clinical data This is a retrospective study of patients diagnosed with A. cantonensis encephalitis/meningitis in Guangzhou Women and Children's Medical Center from July 2018 to August 2020, who were undiagnosed upon admission. The diagnosis meets the following criteria [5]: 1) a clear epidemiological history; 2) clear infection, mainly symptoms of CNS infection; 3) eosinophil elevation in cerebrospinal fluid by more than 10%; 4) pathogen examination: larvae, A. cantonensis antibody positive in serum or cerebrospinal fluid. Comprehensive assessment of the appeal of four criteria, while excluding infection from other causes. All patients were treated with albendazole combined with low-dose glucocorticoids [6].

Methods
The clinical data for all included cases were recorded, including demographic features, clinical manifestations, laboratory and imaging results, treatments, and outcomes. A. cantonensis antibodies in peripheral blood and CSF were tested by Guangzhou Golden Field Medical Laboratory Co., Ltd. Cerebrospinal fluid mNGS were tested by Shenzhen MGI Technology Co., Ltd.
Categorical variables are described as n (%). Continuous variables with normal distribution are expressed as mean ± standard deviation (SD). Continuous variables with non-normal distribution are expressed as the median (interquartile range, IQR). Use the c 2 test to compare differences between groups. The bilateral p-value < 0.05 is considered statistically significant.

Characteristics of cases
A total of 11 patients were diagnosed with A. cantonensis encephalitis/meningitis in Guangzhou Women and Children's Medical Center from July 2018 to August 2020, whose clinical data were collected. All of them were undiagnosed upon admission, including six males (54.5%) and five females (45.5%), and the age ranged from 1 to 13 years with a median of 21 months (IQR: 15.6, 96). Among them, one case (9.1%) had a clear history of raw snail-eating, three cases (27.3%) had edible shrimp and oysters, two cases (18.2%) had contact with field water sources, and five cases (45.5%) had no obvious exposure history of the parasite.
Among them, nine cases (81.8%) had fever and vomiting, five cases (45.5%) had a headache, one case (9.1%) had convulsions, and five cases (45.5%) had mental changes. The interval between onset and definite diagnosis was 5 -102 d, with a median of 26 d (IQR: 20, 38). The total length of hospital stay was 17 -92 d, with Journal of Behavioral and Brain Science a median of 27 d (IQR: 18, 39) (see Table 1).

Imaging and neurophysiology
A chest X-ray in all patients, found four cases (36.4%) were normal, 7 cases (63.6%) had a lung infection, and one (9.1%) had lung consolidation. All patients underwent head magnetic resonance imaging (MRI) examination, and eight positive  cases (72.7%) were found, including meningitis changes and abnormal reinforcement of brain parenchymal (see Table 3, Figure 1, Figure 2). Except for seven cases (63.6%) with normal electroencephalogram (EEG), one case (9.1%) showed slow background activity and a large amount of 3.5 -5 hz slow-wave burst in each lead, and three cases (27.3%) showed boundary EEG changes with increased slow background activity.

Diagnostic test
The results of A. cantonensis in peripheral blood and cerebrospinal fluid (CSF) were compared with the metagenomic next-generation sequencing (mNGS) of CSF. The antibody-positive rate of peripheral blood was 54.5% (6/11), CSF antibody-positive rate was 27.2% (3/11), and mNGS-positive rate of CSF was 81.8% (9/11), and the positive diagnostic rate of the latter was significantly higher than the former two (p = 0.035 < 0.05 and p = 0.030 < 0.05, respectively) (see Table   4), and after the treatment, it was found that the mNGS reads in the CSF decreased, indicating the anti-A. cantonensis treatment was effective (see Figure   3). Treatment and prognosis The patients were given empirical antibiotic treatment and completed pathogenic sampling and testing after admission. After the diagnosis was confirmed, albendazole (20 mg/kg•d * 10 d, 2 courses) combined with low-dose methylprednisolone (0.2 mg/kg•d * 5 -10 d) treatment were conducted. Except for one     case lost to follow-up, the remaining 10 cases were followed up for 3 months to half a year after discharge, and no sequelae were left.

Discussion
A. cantonensis is a zoonotic pathogen. In the past 20 years, there have been local outbreaks or occasional cases [8] [9] [10]. The basis for the diagnosis of finding the larvae from the cerebrospinal fluid is difficult. Simultaneously, neither clinical manifestations nor cerebrospinal fluid with increased eosinophils are specific for diagnosis. Although advances in immunological technology have improved the diagnostic rate of parasitic infections, this technology is still insufficient for diagnosis A. cantonensis encephalitis [11]. It usually takes a long time to identify A. cantonensis infection, with a median of 26 d in our study. This not only aggravates illness, but also increases the family economic burden. Therefore, it is necessary to find a simple sampling detection method with high diagnostic specificity, and our research found that CSF mNGS is a good alternative.
Since children often like to play with water and soil, once contaminated by A. cantonensis, they can accidentally be infected. Among the study cases in this group, four cases had a history of eating snails, shrimp, or oysters, and two cases had water exposure in the wild, and the remaining five cases failed to trace a suspicious epidemiological contact history and could not provide help in diagnosis. A. cantonensis infection with neurological involvement is more common [2]. In this study, except for one case with fever of unknown origin, ten patients had different neurological symptoms or signs, the most common was headache (45.5%), complained by older children mainly, followed by muscle weakness (21%) and convulsions (9%). Although vomiting occurred in nine cases (81.8%), it might be a nonspecific neurological symptom. Because children cannot clearly describe personal feelings, such as headaches, and activity trajectories, the feasibility of assisting diagnosis through epidemiological history and neurological symptoms is reduced.
A. cantonensis is one of the important pathogen causing eosinophilic meningitis [12], and the increase in eosinophils in CSF is also an important clue for A. cantonensis encephalitis/meningitis. Previous studies have shown that 71% of patients have elevated eosinophils in the cerebrospinal fluid [13], while eight cases (72.7%) increased eosinophils in CSF in our study. However, it has also been reported that only 4% of cases had elevated eosinophils in CSF in the early course of the disease [13], which is not a unique indicator of A. cantonensis encephalitis/meningitis needing ruled out other causes such as other pathogens infections and brain tumors.
Head magnetic resonance imaging (MRI) can find abnormal signals, including multiple nodular enhancement lesions and linear enhancement of the pia mater, which are common in the brain [14] [15]. In our study, five cases (45.5%) showed multinodular enhancement of brain lesions, and three cases (27.3%) showed leptomeningeal enhancement. Although neuroangiostrongyliasis cannot be diagnosed by brain MRI, but may have suggestive, and contribute to other causes of eosinophilic meningitis excluded.
Finding pathogens in human tissues, secretions, and excrement is the basis for diagnosis parasitic infection. However, it is more challenging to detect A. cantonensis directly from the central nervous system (CNS) to confirm the CNS A. cantonensis infection. In contrast, immunological diagnosis methods are relatively feasible, which were used in our study and found that the positive rates of A. cantonensis antibodies in serum and CSF were 54.5% and 27.2%, respectively.
In another aspect of the diagnostic technique, the CSF mNGS-positive rate exceeds 80% in our study, which is significantly higher than immunization me-thods. The patients have a good prognosis by targeted treatment. Metagenomic next-generation sequencing (mNGS) has greatly increased the pathogen detection rate of central nervous system infections [16][17][18]. Additionally, pathogens of mixed infections in CNS can be found more comprehensively by mNGS, especially in patients with a long course of the disease and low immunity, simultaneously, the therapeutic effect can be evaluated by monitoring the reads [18].
Miao et al. suggest that mNGS could yield higher sensitivity for early identification of fastidious and time-consuming microbes (e.g., Mycobacterium tuberculosis, virus, anaerobe, and fungus) [19]. Our study confirmed that mNGS also has a high sensitivity in the diagnosis of parasitic infection. Presently, the mNGS technology is not widely used in clinical practice due to its high cost of testing, however, a clear early diagnosis, followed by targeted treatment in exchange for lower hospital stay, costs and sequelae rate, and comprehensive cost performance might be more beneficial to patients.

Conclusion
Compared with epidemiological history, elevated eosinophils, antibody immune detection, and brain MRI examination, CSF mNGS technology can obtain a higher positive diagnosis rate in children with A. cantonensis encephalitis/meningitis, which not only helps find the CNS pathogens earlier and comprehensively, but also evaluate the efficacy of treatment, and its clinical value is worthy of promotion.

Limitations
This study still has some limitations. First of all, this is a retrospective singlecenter study with a limited number of children. Additionally, due to the limitation of testing costs, it was impossible to conduct a re-examination of CSF mNGS in all patients. Its clinical promotion requires further technical improvement and reduction in testing costs.