Management of Central Nervous System Infections, Vientiane, Laos, 2003–2011

During 2003–2011, we recruited 1,065 patients of all ages admitted to Mahosot Hospital (Vientiane, Laos) with suspected central nervous system (CNS) infection. Etiologies were laboratory confirmed for 42.3% of patients, who mostly had infections with emerging pathogens: viruses in 16.2% (mainly Japanese encephalitis virus [8.8%]); bacteria in 16.4% (including Orientia tsutsugamushi [2.9%], Leptospira spp. [2.3%], and Rickettsia spp. [2.3%]); and Cryptococcus spp. fungi in 6.6%. We observed no significant differences in distribution of clinical encephalitis and meningitis by bacterial or viral etiology. However, patients with bacterial CNS infection were more likely to have a history of diabetes than others. Death (26.3%) was associated with low Glasgow Coma Scale score, and the mortality rate was higher for patients with bacterial than viral infections. No clinical or laboratory variables could guide antibiotic selection. We conclude that high-dependency units and first-line treatment with ceftriaxone and doxycycline for suspected CNS infections could improve patient survival in Laos.

S. pneumonia isolates, when available, were sent to Murdoch Children Research Institute.
Serotyping was performed by latex agglutination using a combination of commercial and inhouse typing reagents (19), and results were confirmed using the Quellung reaction.  (17,21). Modification has been made as improvement of CtrA System and the use of freeze-dried Taqman Quadruplex assay (22). The quadruplex contains primers against N. meningitides capsule transporter (ctrA), serogroup B sialyltransferase (siaDB), S. pneumoniae pneumolysin (ply), and an internal control (Cucurbita cv. Kurokawa amakuri hydroxypyruvate reductase). The assay was prepared by Applied Biosystems (Thermo Fisher Scientific) in a lyophilized format, with primer and probe sequences provided by the MRU; the components of the master mix have not been disclosed by the company. Lyophilized reagents were rehydrated with 20 μL of molecular-grade water, and then, 5 μL of DNA was added. Amplification and detection was done on TaqMan 7500 (Applied Biosystems, Thermo Fisher Scientific) using fast cycling conditions (2 min at 95°C, 45 cycles of 95°C for 3 sec and 60°C for 30 sec).

Orientia tsutsugamushi PCR
This hydrolysis probe qPCR was based on that described by Jiang et al. (23) targeting the 47-kD gene. In total, 1 µL of DNA extract from EDTA buffy coat and 5 µL for CSF was used in a 25-µL reaction with the Platinum Quantitative PCR SuperMix-UDG (Invitrogen) kit and 100 nmol/L of each primer and 200 nmol/L of probe. The thermal cycling program was 50°C for 2 min, 95°C for 2 min, and 45 cycles of 95°C for 15 sec and 60°C for 60 sec. All positive qPCRs were confirmed by sequencing (Macrogen Inc) or conventional PCR targeting 56 kDa as previously described (24).

Rickettsia genus and Rickettsia typhi PCR
This assay is a hydrolysis probe qPCR targeting the 17-kDa gene of Rickettsia spp. (23,25) using 1 µL of DNA extracted from the EDTA buffy coat and 5 µL for CSF, in a 25-µL reaction volume. The Platinum Quantitative PCR SuperMix-UDG kit (Invitrogen, Thermo Fisher Scientific) was used in a final volume of 25 μL, with 400 nmol/L of each primer and probe. The thermal cycling program was 50°C for 2 min, 95°C for 2 min, and 45 cycles of 95°C for 15 sec and 60°C for 30 sec.
The HSV1/2 system permits to detect HSV1 and HSV2 viruses. Samples positive by HSV1/2 PCR were submitted to 2 specific hydrolysis probe qPCRs for the detection of HSV1 and HSV2 (39). Detection of Dengue virus was done using a pan-dengue hydrolysis probe qPCR system designed to detect the 4 dengue serotypes. Positive samples were then submitted to the 4 Page 11 of 50 hydrolysis probe qPCRs specific for the 4 serotypes. The hydrolysis probe qPCR used for the detection of EV is a pan-EV system that permits detection of all enteroviruses. Typing of EVpositive samples was performed following techniques developed by Nix et al. (40), see below, directly on patient sample extract or after inoculation on cell culture, when possible.
The primers and probe for detection of Henipavirus were designed using alignment of all Hendra and Nipah virus sequences available in GenBank.
PCR conditions were adapted to a standard 2-step protocol using TaqMan Reverse Transcription Reagents kit (Roche) for RNA viruses, followed by hydrolysis probe qPCR using  (50°C for 2 min, 95°C for 10 min, and 45 cycles of 95°C for 15 sec and 60°C for 1 min). WNV and TBEV primers and probes were used in a duplex qPCR following the same protocol. Any samples positive with a cycle quantification (Cq) <40 were repeated for confirmation.
Internal controls (MS2 and T4, RNA and DNA bacteriophages, respectively) were added to all specimens and systematically tested by hydrolysis probe qPCR (14). T4 and MS2 qPCRs were performed on 3 µL of DNA (RT product for MS2, nucleic acid extract for T4) in a final volume of 15 µL. In case of no detection of internal control, a new sample was extracted. In case of inhibition of the PCR, the extract sample was diluted 1:10 in AVE buffer (QIAGEN), and all qPCR reactions were repeated from this dilution.
Duplex hydrolysis probe qPCR was performed for the detection of influenza viruses A and B until September 2009 following the protocol above. Influenzavirus A qPCR was shown to not detect pandemic H1N1/09 (41), so primers alone were used in a SYBR Green RT-qPCR. PCRs underwent a heminested PCR using 3 µL of the primary PCR product, the same reverse primer, the forward 2 primer, and the same amplification protocol as in the primary PCR.
Amplicons (162 bp) were sent for sequencing to Macrogen Inc. Then, the sequences were BLASTed (blastn, NCBI website) for identification.

Enterovirus typing
The typing of EV was performed using the protocol from the French reference center for Enterovirus based on Nix et al. (40). When available, clinical samples EV-positive by RT-qPCR were inoculated on MRC5, BGM, and MA104 cells in 12-well plates. In cases of cytopathic effect, cell supernatant was collected, extracted using EZ1 Virus Mini Kit v2.0 (QIAGEN), and submitted to pan-EV hydrolysis probe RT-qPCR using SuperScript III Platinum One-Step qRT-PCR kit (Invitrogen) with 200 nmol/L of each primer, 100 nmol/L of probe, and 5 µL of extract in 25 µL final volume. The thermal cycling program was 50°C for 15 min, 95°C for 2 min, and 45 cycles of 95°C for 15 sec and 60°C for 45 sec. The extracts from EV-positive cultures were submitted for RT-PCR using the Access RT-PCR system (Promega) with 1 μmol/L of each forward primer and reverse primer 1 and 5 µL of extract in a final volume of 50 µL, following the manufacturer's instructions, with 42°C as the annealing temperature. The thermal cycling Page 13 of 50   program was 45°C for 45 min; 94°C for 2 min; and 40 cycles of 94°C for 30 sec, 42°C for 1 min,   and 68°C for 2 min, ending with 68°C for 7 min. For patients, whose EV could not be isolated by cell culture, 5 µL of extract underwent RT using 100 U of SuperScript III Reverse Trancriptase (Invitrogen), 10 mmol/L dithiothreitol, 0.1 mmol/L dNTP, 200 nmol/L of each RT primer (1)(2)(3)(4), and 20 U RNaseOUT Recombinant Ribonuclease (Invitrogen) in a 10-µL final volume. The RT thermal cycling program was 22°C for 10 min, 50°C for 50 min, and 95°C for 5 min. Primary PCR was performed on RT products using 2.5 U of AmpliTaq DNA Polymerase (Applied Biosystems), 1 µmol/L of forward primers and reverse primer 1, and 0.2 mmol/L of dNTP in a 50-µL final volume. The thermal cycling program used was 95°C for 5 min and 40 cycles of 95°C for 30 sec, 42°C for 50 sec, and 60°C for 50 sec.
The primary PCR with primers 1 produce a 990-bp amplicon. In case of negative primary PCR, a nested PCR was performed with 5 µL of the primary PCR product using 2.5 U of FastStart Taq DNA Polymerase (Roche), 800 nmol/L of each primer 2, and 0.2 mmol/L dNTP in a 50-µL final volume. The thermal cycling program was 95°C for 5 min and 40 cycles of 95°C for 30 sec, 60°C for 50 sec, and 72°C for 30 sec.
Nested PCR with primers 2 produce a 375-bp amplicon. Amplicons from primary or nested PCR were sent for sequencing to Macrogen Inc. Then, the sequences were BLASTed (blastn, NCBI website) for identification.

qPCR Interpretation
For quality control, positive and nontemplate controls were included in each run. A PCR was classified as positive if an amplification curve with a Cq value <40 was observed from the same sample in 2 separate PCR runs.
*In September 2016 we reviewed articles published in English in the Medline database in the past 15 years using the terms "encephalitis," "meningitis," "CNS syndrome" "CNS infection" "central nervous system syndrome" "central nervous system infection," with adding the terms "asia," or "south-east asia. Red cell count, n = 886, cells/mm 3 , median (IQR)  §Data collected for children (<15 years old) were excluded for analysis. ⁋Considered as not reliable, the data were excluded from analysis for children <3 y old. #Of these patients, 7 had hemiplegia, 11 had limb weakness, and 1 had paraplegia; 13 patients had admission or discharge diagnoses of Guillain-Barre syndrome. Retrospective evaluation of the likelihood of this diagnosis by using the Brighton system suggested that 4 patients met level 3 criteria for Guillain-Barre syndrome diagnostic certainty (Sejvar et al. 2011). **Including confused and disoriented. † †WHO clinical CNS infection = fever with either GCS score <15, neck stiffness (history or examination), or history of seizure, patients with missing data for one of those criteria were not counted. WHO encephalitis = fever with either GCS score <15 or history of seizure. WHO meningitis = fever with GCS score <15 and/or neck stiffness. WHO meningoencephalitis = meeting both WHO encephalitis and WHO meningitis criteria. ‡ ‡Elevated and low parameters = above or below normal ranges (Appendix Table 3), anemia: hematocrit below normal range. In elevated CSF white cells count, were not taken into account the cases that could not be counted because of high turbidity. Eosinophilia = CSF eosinophils >10%. § §Elevated serum sodium: higher than 150 mmol/L, low serum sodium: lower than <130 mmol/L. Five patients (0.6%) had serum sodium <115 mmol/L. ⁋ ⁋Hyperglycemia = blood glucose higher than 7.7 mmol/L, severe hyperglycemia: blood glucose higher than 11.1 mmol/L. ##Mortality includes patients who died at hospital and the ones who were taken to die at home = moribund.
Confirmed etiology was determined according to positive results by the tests presented in Table 3, consisting in direct detection of the pathogen in CSF or blood, IgM detection in CSF, antibody seroconversion or 4-fold rise in antibody titter between admission and follow-up serum. When >1 pathogens were detected in a same patient, the confirmed etiology was determined by giving the priority to direct detection over indirect detection then to CSF over blood. Confirmed co-infection was defined when > one pathogens were detected by the same kind of test in the same matrix. List of confirmed coinfections in supplemental data (Appendix Table 4). The other etiologies confirmed in <20 patients were cytomegalovirus in 12 patients, herpes simplex virus in 15, Enterovirus in 10, varicella zoster virus in 6, mumps virus in 5, Plasmodium falciparum in 4, and other bacteria in 48 patients (the list of bacteria is provided in Appendix Table 5). Among 35 patients with CSF eosinophils >10%, 4 were found positive for Angiostrongylus cantonensis by PCR (55). Among 662 patients tested for syphilis by the SD. Bioline RDT (Cat No. 06FK10) on serum then confirmed by VDRL and TPHA on serum and CSF, 2 patients could be classified as possible neurosyphilis, as per the UK and European guidelines (TPHA positive in CSF). Other bacterial antibiotic susceptibility data are given in Appendix Table 6. Typing information for Cryptococcus spp. is presented in Appendix Table 5  ¶Data collected for children (<15 y old) were excluded for analysis. #Considered as not reliable, the data were excluded from analysis for children <3 years old. **Including confused and disoriented. † †WHO clinical CNS infection= fever with either GCS score <15, neck stiffness (history or examination), or history of seizure, patients with missing data for 1 of those criteria were not counted. WHO encephalitis = fever with GCS score <15 or history of seizure or both. WHO meningitis = fever with GCS score <15 or neck stiffness or both. WHO meningoencephalitis = meeting both WHO encephalitis and WHO meningitis criteria. ‡ ‡Elevated and decreased parameters = above or below normal ranges (Appendix Table 3), anemia: hematocrit below normal range. In elevated CSF white cell count, were not taken into account the cases that could not be counted because of high turbidity.
§ §Hyperglycemia: blood glucose higher than 7.7 mmol/L, severe hyperglycemia: blood glucose higher than 11.1 mmol/L. ¶ ¶CSF eosinophils >10%.  . History or physical examination were taken into account for rash, confusion, neck stiffness, fever (history of fever or >37.5°C during physical examination). ALP, alkaline phosphatase; ALT, alanine transaminase; AST, aspartate transaminase; CRP, C-reactive protein; CNS, central nervous system; CSF, cerebrospinal fluid; GCS, Glasgow coma scale; IQR, interquartile range; LP, lumbar puncture; TB, Mycobacterium tuberculosis; WHO, World Health Organization. †Population density of the village of residence: Population densities per village were from population census 2005, recovered from Lao DECIDE info Web site (platform of Government of Lao PDR, www.decide.la). ‡Occupation: work indoors = teacher, government official, business, factory worker, accountant; work outdoors = driver, building worker, merchant, carpenter, soldier, mechanic; other: housewife, no job, monk, retired, singer, health worker. §Data collected for children (<15 years old) were excluded for analysis. ¶Considered as not reliable, the data were excluded from analysis for children <3 years old. #Including confused and disoriented. **WHO clinical CNS infection = fever with either GCS score <15, neck stiffness (history or examination), or history of seizure, patients with missing data for 1 of those criteria were not counted. WHO encephalitis = fever with GCS score <15 or history of seizure or both. WHO meningitis = fever with GCS score <15 or neck stiffness or both. WHO meningoencephalitis = meeting both WHO encephalitis and WHO meningitis criteria. † †Elevated and low parameters = above or below normal ranges (Appendix Table 3), anemia: hematocrit below normal range. In elevated CSF white cell count, were not taken into account the cases that could not be counted because of high turbidity.

Appendix
‡ ‡Hyperglycemia: blood glucose higher than 7.7 mmol/L, severe hyperglycemia: blood glucose higher than 11.1 mmol/L. § §Eosinophilia: CSF eosinophils >10%. *The factors that showed p<0.01 in univariate analysis were submitted to multivariate analysis. Some factors were excluded (e.g., HIV seropositivity), since the choice for patient testing was biased. Clinical meningitis was correlated with neck stiffness, neutrophils was correlated with white cell count, and hyperglycemia was correlated with diabetes (a model was run replacing diabetes with hyperglycemia or blood glucose, which turned out to be not significant). aOR, adjusted odds ratio; CSF, cerebrospinal fluid; MICE, multiple imputation by chained equation. †Complete case analysis was repeated with only significant factors (p<0.05) identified by stepwise approach (n = 607). ‡Final model with imputed values with only significant variables included (n = 1,043). §Variables with imputed values. Other variables included in the imputation model: bacterial infection (outcome), sex, age, fever, and neck stiffness. ¶The aOR for a 10-U increase in C-reactive protein or CSF lactate. §Data collected for children (<15 years old) were excluded for analysis. ¶Considered as not reliable, the data were excluded from analysis for children <3 y old. #Including confused and disoriented. **WHO clinical CNS infection: fever with either GCS score<15, neck stiffness (history or examination), or history of seizure, patients with missing data for 1 of those criteria were not counted. WHO encephalitis = fever with GCS score <15 or history of seizure or both. WHO meningitis = fever with GCS score <15 or neck stiffness or both. WHO meningoencephalitis = meeting both WHO encephalitis and WHO meningitis criteria. † †Elevated and low parameters = above or below normal ranges (Appendix Table 3), anemia: hematocrit below normal range. In elevated CSF white cell count, were not taken into account the cases that could not be counted because of high turbidity.

Appendix
‡ ‡Hyperglycemia: blood glucose higher than 7.7 mmol/L, severe hyperglycemia: blood glucose higher than 11.1 mmol/L. § §Eosinophilia: CSF eosinophils >10%. §Data collected for children (<15 years old) were excluded for analysis. ¶Considered as not reliable, the data were excluded from analysis for children <3 y old. #Including confused and disoriented. **WHO clinical CNS infection: fever with either GCS score <15, neck stiffness (history or examination), or history of seizure, patients with missing data for 1 of those criteria were not counted. WHO encephalitis = fever with GCS score<15 or history of seizure or both. WHO meningitis = fever with GCS score <15 or neck stiffness or both. WHO meningoencephalitis = meeting both WHO encephalitis and WHO meningitis criteria. † †Elevated and low parameters = above or below normal ranges (Appendix Table 3), anemia: hematocrit below normal range. In elevated CSF white cell count, were not taken into account the cases that could not be counted because of high turbidity.