INTRODUCTION
Angiostrongyliasis, a disease caused by Angiostrongylus cantonensis, has spread from the tropical endemic area to various regions throughout the world because of extensive international travel and eating habits. 1,2 The route of infection is through ingestion of raw freshwater snails, shrimp, or monitor lizards. A. cantonensis is a neurotropic parasite, which presents as three main clinical manifestations: eosinophilic meningitis, eosinophilic encephalitis, and ocular angiostrongyliasis.3
Angiostrongyliasis is diagnosed on the basis of clinical manifestations,4–8 including the presence of cerebrospinal fluid (CSF) eosinophils and history of larval exposure. The serologic tests vary in sensitivity, and availability is limited. Therefore, their clinical uses are restricted.
The presenting symptom of meningitic angiostrongyliasis, the most common form, is acute severe headache. 6,8 In contrast, encephalitic angiostrongyliasis is rare, but fatal. Similar to viral encephalitis, encephalitic angiostrongyliasis presents with acute deterioration of consciousness to a coma.7 Seizure attacks have never been reported in encephalitic angiostrongyliasis. Limited information is currently available on the risk factors for encephalitic angiostrongyliasis. Here, we perform a hospital-based, comparison study to identify the clinical factors predictive of encephalitis caused by A. cantonensis.
MATERIALS AND METHODS
Study population.
We recruited adult patients hospitalized for angiostrongyliasis at Srinagarind Hospital, Khon Kaen University, Khon Kaen, Thailand. The clinical diagnostic criteria4–8 for angiostrongyliasis were as follows: 1) CSF with a white blood cell count of > 10 cells/mm3, 2) CSF eosinophils constituting > 10% of the total CSF white blood cell count, 3) negative tests for CSF Gram, acid-fast, and India ink staining, cryptococcal antigen testing, and cultures, and 4) history of ingesting raw freshwater snails or other paratenic hosts, such as shrimp and monitor lizards.
Exclusion criteria aim to eliminate other possible causes of CSF eosinophils included history of raw fish consumption, history of migratory swelling, clinical diagnosis of subarachnoid hemorrhage or myeloencephalitis, positive serologic test for gnathostomiasis or cysticercosis, abnormal brain computed tomography or magnetic resonance findings, symptomatic or serology-positive HIV infection, and active or previous history of tuberculosis or malignancy.
The mentioned clinical criteria were applied to both encephalitis and meningitis groups. As previously reported, both conditions were differentiated by a complaining symptom, in that encephalitic angiostrongyliasis presented with acute coma. 6,7 In addition, brain imaging must be normal in the encephalitis group. Clinical factors between both groups were studied, and the predictors were determined for encephalitic angiostrongyliasis.
Sample size.
From previous studies, the proportions of potential parameters such as numbers of patients with fever or neck stiffness in the meningitis and encephalitis group were 10% and 40%, respectively. 6,7,9,10 Using a two-sided significance level of 0.05, power of 80%, and the meningitis/encephalitis sample size ratio of 6:1, the approximate numbers of the encephalitis and meningitis groups were 14 and 86 subjects, respectively. 11 The study protocol was reviewed and approved by the institutional review board and the ethics committee of Khon Kaen University.
Data collection.
We recorded the baseline characteristics, symptoms, physical signs, and laboratory results of all participants. Baseline characteristics included sex, age, season of admission defined by the Thailand meteorologic classification system (winter, summer, or rainy), incubation period (number of days after the last exposure to snails or paratenic hosts to the first day of developing symptoms), duration of headache (days), history of paresthesia, and history of vomiting.
Physical signs included fever (oral temperature of > 38°C), cranial nerve abnormalities, papilledema, and stiff neck. Laboratory examinations comprised complete blood count (CBC), serologic test for A. cantonensis,12 and CSF analyses. The serologic test was done by immunoblotting analysis using IgG antibodies to the 29-kd antigenic polypeptide of A. cantonensis. The specificity of the 29-kd antigen for human angiostrongyliasis is 99.4%. All data were obtained at the initial presentation before administration of any treatment.
Data analysis.
Baseline and clinical characteristics of both groups were compared using descriptive statistics. Wilcoxon rank-sum and Fisher exact tests were applied to compare the differences in medians and proportions between the two groups, respectively.
Univariate logistic regression analyses were applied to calculate the crude odds ratios (ORs) of individual variables for the development of encephalitis. All variables with P < 0.20 in univariate analysis were included in subsequent multivariate logistic regression analyses. All variables with P > 0.15 in the multivariate model were excluded with the stepwise approach, whereas those with P < 0.15 were retained in the final model. To account for possible interaction, interaction terms were forced into the final model and considered significant at P < 0.10. Analytical results were presented as crude ORs, adjusted ORs, and 95% confidence intervals (CIs).
The goodness-of-fit of the final model was evaluated using Hosmer-Lemeshow statistics. 13 Nagelkerke R2 was calculated to estimate the proportion of explained variance in the model. To evaluate the discriminatory power or accuracy of the model, c statistics or area under the receiver operating characteristic curve was examined. 14 All data analyses were performed with SAS software version 8.2.
RESULTS
We enrolled 14 patients diagnosed with encephalitic angiostrongyliasis and 86 unmatched patients with meningitic angiostrongyliasis randomly selected from the hospital registration database (1996–2007). Six patients in meningitis group were excluded because of incomplete clinical information. The mortality rate in the encephalitic group was 79% (11 of 14 cases). On the other hand, no deaths were recorded in the meningitis group.
The baseline characteristics, physical signs, and laboratory findings of the both groups are presented in Tables 1 and 2. Approximately three fourths of the subjects in each group were men. The encephalitis and meningitis groups showed distinct clinical features in terms of age, season of presentation, duration of headache, history of vomiting, presence of fever, seventh cranial nerve palsy, papilledema and stiff neck, percentage of blood eosinophils, CSF white blood cell count, CSF eosinophil count, CSF protein level, and CSF/plasma glucose ratios. The sensitivity of the serologic test was 50% and 62% in the encephalitis and meningitis groups, respectively.
Univariate analyses showed that factors significantly associated with encephalitis were older age, summer season, longer duration of headache, fever, papilledema, neck stiffness, low percentage of eosinophils on CBC, and low CSF/plasma glucose, as specified in Table 3.
Table 4 shows stepwise logistic multivariate analysis data on factors remaining in the final model predictive of encephalitis, which include older age (adjusted OR, 1.22; 95% CI, 1.05–1.42), prolonged duration of headache (adjusted OR, 1.26; 95% CI, 1.03–1.55), and fever (adjusted OR, 37.05, 95% CI, 1.59–862.35). Interaction terms for combinations of headache duration, age, and fever were forced into the model but were not statistically significant. For the final model, the Hosmer-Lemeshow value, Nagelkerke R2, and the c value were 6.30 (P = 0.50), 0.78, and 0.97, respectively.
DISCUSSION
Our results showed that older age, prolonged headache duration, and fever at presentation are predictive factors for encephalitis. After adjustment for other factors, the risk of encephalitis in infected patients increased by 22% for every additional year of age and by 26% for every additional day of headache, whereas fever at presentation was associated with a 37-fold higher risk of encephalitis.
Older age has not been identified as a risk factor for encephalitis angiostrongyliasis until now but is a known risk factor for encephalitis induced by other agents, such as West Nile virus. 15 This may relate to compromised immunity in older individuals.
The A. cantonensis larvae usually localized in the subarachnoid spaces and meninges in humans, leading to nonfatal headaches or meningitic forms. However, if a headache is disregarded by patients or diagnosis is missed or delayed, the larvae may attempt to migrate to the pulmonary arteries, as observed in rats, their definitive hosts. 16 During migration, larvae may damage brain tissues, resulting in severe inflammatory processes, production of pyrogenic cytokines, and development of the encephalitic form. The presence of CSF eosinophils in the encephalitic group supports this theory.
Unlike in bacterial meningitis, only 10% of meningitic angiostrongyliasis experienced fever, compared with 71% in the encephalitic group. We propose that direct invasion of brain tissue by larvae may activate inflammatory processes and produce pyrogenic cytokines. 17 Generation of pyrogenic cytokines may additionally contribute to fever, although the underlying mechanism of action in encephalitic angiostrongyliasis remains unknown.
Overall, the clinical manifestations of encephalitis and meningitis angiostrongyliasis are relatively dissimilar. Clear differences are evident in the duration of headache between the groups, with a median of 7 days for the meningitis group and 18.5 days for the encephalitis group (P = 0.002). Thus, in clinical practice, if meningitic angiostrongyliasis patients have the triad of elderly, febrile, and headache > 7 days, they should be considered at high risk of encephalitis. Such patients may benefit from combination therapy, such as corticosteroids and albendazole. 3,18
The primary limitation of this study was the small number of patients in the encephalitis group. This led to wide confidence limits of ORs for the fever variable. Because of the retrospective design of the study, some information may be missing or unavailable. Finally, despite strict clinical criteria, a minor proportion of cases may be attributed to other causes such as gnathostomiasis. Clinically, the serologic test for angiostrongyliasis is not routinely used, and its sensitivity is only ~50–60%. 12 However, the neurologic manifestations of gnathostomiasis are rare, unique, and distinct from those of angiostrongyliasis.4 History of ingestion of uncooked fish, migratory swelling, sub-arachnoid hemorrhage, myeloencephalitis, and unusual intracerebral hemorrhage are suggestive of gnathostomiasis.
Goodness-of-fit statistics indicated a good fit to the model. The Nagelkerke R2 value suggested that 78% of data were accounted for in the final model. The c statistic value or area under the ROC curve value of 0.97 signified effective discrimination of a random pair of patients with and without encephalitis. However, prospective interventional studies are needed because of the small sample size.
In summary, elderly patients with angiostrongyliasis experiencing fever and prolonged headaches are at the highest risk of developing encephalitis. Awareness, prompt diagnosis, and aggressive treatment are important factors in preventing the development of encephalitic angiostrongyliasis.
Baseline characteristics and physical findings of encephalitis and meningitis group
Laboratory findings of encephalitis and meningitis group
Results of univariate regression analyses showing independent variables predictive for encephalitis and their crude odds ratios
Results of multiple logistic regression analysis showing independent variables for encephalitis and their adjusted ORs
Address correspondence to Kittisak Sawanyawisuth, Department of Medicine, Faculty of Medicine, Khon Kaen University, 123 Mitraparp Friendship Road, Khon Kaen 40002, Thailand. E-mail: kittisak@kku.ac.th
Authors’ addresses: Kittisak Sawanyawisuth, Vichai Senthong, Panita Limpawattana, Somsak Tiamkao, Suthipun Jitpimolmard, and Verajit Chotmongkol, Department of Medicine, Faculty of Medicine, Srina-garind Hospital, Khon Kaen University, 123 Mitraparp Friendship Road, Khon Kaen, 40002, Thailand. Ken Takahashi, Tsutomu Hoshuyama, and Donald Wilson, Department of Environmental Epidemiology, Institute of Industrial Ecological Sciences, University of Occupational and Environmental Health, Iseigaoka 1-1, Yahatanishiku, Kitakyushu City, 807-8555, Japan. Kanlayanee Sawanyawisuth, Department of Biochemistry, Faculty of Medicine, Khon Kaen University, 123 Mitraparp Friendship Road, Khon Kaen, 40002, Thailand. Pewpan M. Intapan, Department of Parasitology, Faculty of Medicine, Khon Kaen University, 123 Mitraparp Friendship Road, Khon Kaen, 40002, Thailand.
Financial support: The Japan Society for Promotion of Sciences (JSPS) provided financial support to the first and second authors on international travel expenses to pursue this study on a collaborative basis.
REFERENCES
- 1↑
Ali AB, Van den Enden E, Van Gompel A, Van Esbroeck M, 2008. Eosinophilic meningitis due to Angiostrongylus cantonensis in a Belgian traveller. Travel Med Infect Dis 6 :41–44.
- 2↑
Malvy D, Ezzedine K, Receveur MC, Pistone T, Crevon L, Lemardeley P, Josse R, 2008. Cluster of eosinophilic meningitis attributable to Angiostrongylus cantonensis infection in French policemen troop returning from the Pacific Islands. Travel Med Infect Dis 6 :301–304.
- 4↑
Ramirez-Avila L, Slome S, Schuster FL, Gavali S, Schantz PM, Sejvar J, Glaser CA, 2009. Eosinophilic meningitis due to Angiostrongylus and Gnathostoma species. Clin Infect Dis 48 :322–327.
- 5
Eamsobhana P, Yong HS, 2009. Immunological diagnosis of human angiostrongyliasis due to Angiostrongylus cantonensis (Nematoda:Angiostrongylidae). Int J Infect Dis 13 :425–431.
- 6↑
Chotmongkol V, Sawanyawisuth K, Thavornpitak Y, 2000. Corticosteroid treatment of eosinophilic meningitis. Clin Infect Dis 31 :660–662.
- 7↑
Chotmongkol V, Sawanyawisuth K, 2002. Clinical manifestations and outcome of patients with severe eosinophilic meningoencephalitis presumably caused by Angiostrongylus cantonensis. Southeast Asian J Trop Med Public Health 33 :231–234.
- 8↑
Punyagupta S, Bunnag T, Juttijudata P, Rosen L, 1970. Eosinophilic meningitis in Thailand. Epidemiologic studies of 484 typical cases and the etiologic role of Angiostrongylus cantonensis. Am J Trop Med Hyg 19 :950–958.
- 9↑
Chotmongkol V, Sawadpanitch K, Sawanyawisuth K, Louhawilai S, Limpawattana P, 2006. Treatment of eosinophilic meningitis with a combination of prednisolone and mebendazole. Am J Trop Med Hyg 74 :1122–1124.
- 10↑
Chotmongkol V, Wongjitrat C, Sawadpanit K, Sawanyawisuth K, 2004. Treatment of eosinophilic meningitis with a combination of albendazole and corticosteroid. Southeast Asian J Trop Med Public Health 35 :172–174.
- 11↑
Pezzullo JC. Proportion Difference Power/Sample Size Calculation. Available at: http://statpages.org/proppowr.html. Accessed May 31, 2009.
- 12↑
Maleewong W, Sombatsawat P, Intapan PM, Wongkham C, Chotmongkol V, 2001. Immunoblot evaluation of the specificity of the 29-kDa antigen from young adult female worms Angiostrongylus cantonensis for immunodiagnosis of human angiostrongyliasis. Asian Pac J Allergy Immunol 19 :267–273.
- 13↑
Hosmer DW, Hosmer T, Le Cessie S, Lemeshow S, 1997. A comparison of goodness-of-fit tests for the logistic regression model. Stat Med 16 :965–980.
- 14↑
Hanley JA, McNeil BJ, 1982. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 143 :29–36.
- 15↑
Petersen LR, Roehrig JT, Hughes JM, 2002. West Nile virus encephalitis. N Engl J Med 347 :1225–1226.
- 16↑
Hochberg NS, Park SY, Blackburn BG, Sejvar JJ, Gaynor K, Chung H, Leniek K, Herwaldt BL, Effler PV, 2007. Distribution of eosinophilic meningitis cases attributable to Angiostrongylus cantonensis, Hawaii. Emerg Infect Dis 13 :1675–1680.
- 17↑
Dinarello CA, Gelfand JA, 2005. Fever and hyperthermia. Braunwald E, Fauci AS, Kaspers DS, eds. Harrison’s Principles of Internal Medicine, 16th Edition. New York: McGraw-Hill Publishers, 104–108.
- 18↑
Tu WC, Lai SC, 2006. Angiostrongylus cantonensis: efficacy of albendazole-dexamethasone co-therapy against infection-induced plasminogen activators and eosinophilic meningitis. Exp Parasitol 113 :8–15.