Utility of FilmArray Meningitis/Encephalitis Panel during Outbreak of Brainstem Encephalitis Caused by Enterovirus in Catalonia in 2016

LETTER

We read with interest the article by Leber et al. (1). The authors reported the performance of the FilmArray meningitis/encephalitis (FA-M/E) panel in a multicenter evaluation that included a large number of infants. Enterovirus (EV) was the main detection, and they described very high positive and negative percentages of agreement with comparator PCR assays (95.7% and 99.5%, respectively) for EV detections.

In April 2016, the first cases of an outbreak of brainstem encephalitis were observed in Catalonia, Spain. They were related to an increase of EV detections in nasopharyngeal aspirates and/or stools but not in normally sterile samples (blood or cerebrospinal fluid [CSF]) (2). EV is rarely detected by culture or molecular methods in CSF samples of children presenting with brainstem encephalitis, due to the low viral loads there (3). Despite the fact that clinical diagnosis was compatible with the EV neurological disease in other outbreaks, physicians were faced with multiple diagnostic dilemmas in an effort to ensure that bacterial or other treatable infections were not missed.

We report the performance and utility of the FA-M/E panel in this setting.

Commencing in April 2016, the first 20 children fulfilling clinical case definitions of brainstem encephalitis or encephalomyelitis (4) who were admitted in a referral pediatric hospital were included. The protocol for assessment included blood and CSF bacterial cultures, a real-time PCR for detection of herpes simplex virus 1 and 2 in CSF samples, and an in-house pan-EV real-time PCR assay (5) in plasma and CSF samples. All the negative CSF samples in the previous assays underwent additional testing by the FA-M/E panel. In these cases, the pan-EV real-time PCR was also performed in nasopharyngeal aspirate and fecal samples.

The median patient age was 27 months (interquartile range [IQR], 27.7 to 77.6). Pan-EV real-time PCR was positive in noninvasive samples in all patients (13/20 in respiratory sample and 13/18 in fecal samples) (Table 1). In CSF samples, the pan-EV real-time PCR was negative in all patients, whereas EV RNA was detected in 4/20 using the FA-M/E panel. These four patients had a lower median duration of symptoms before sampling than the patients who were FA-M/E negative (30 h [IQR, 8 to 45] versus 48 h [IQR, 21 to 72], respectively), but this difference was not statistically significant. There were no differences in clinical severity between patients with positive and negative detections in CSF samples. EV genotyping (6) was performed in 19 positive samples at the National Centre for Microbiology, Madrid, and all strains were identified as EV-A71.

TABLE 1.

Microbiological results of 20 cases with clinical diagnosis of brainstem encephalitis or encephalomyelitis

No. of cases	Result for indicated sample type and assaya
	Cerebrospinal fluid				Plasma/blood		Respiratory specimens	Feces
	Real-time PCR-HSV 1 and 2	Pan-EV real-time PCR	Gram stain and bacterial culture	FA-M/Eb	Pan-EV real-time PCR	Bacterial culture	Pan-EV real-time PCRc	Pan-EV real-time PCRd
7	Negative	Negative	Negative	No detection	Negative	Negative	Negative	Positive
4	Negative	Negative	Negative	No detection	Negative	Negative	Positive	Negative/NT
2	Negative	Negative	Negative	EV	Negative	Negative	Positive	Positive
2	Negative	Negative	Negative	EV	Negative	Negative	Positive	Negative/NT
2	Negative	Negative	Negative	HHV-6	Negative	Negative	Positive	Positive
1	Negative	Negative	Negative	No detection	Positive	Negative	Positive	Positive
1	Negative	Negative	Negative	No detection	Negative	Negative	Positive	Positive
1	Negative	Negative	Negative	HHV-6	Negative	Negative	Positive	Negative

^aHSV, herpes-simplex virus; EV, enterovirus; HHV-6, human herpesvirus 6; FA-M/E, FilmArray meningitis/encephalitis panel; NT, not tested.

^bThe median lag time (interquartile range) (h) between onset of fever and sampling was 30 (8 to 45) for samples with EV-positive results and 48 (21 to 72) for samples with EV-negative results.

^cThe median lag time (interquartile range) (h) between onset of fever and sampling was 48 (27 to 90) for samples with EV-positive results and 48 (18 to 72) for samples with EV-negative results.

^dThe median lag time (interquartile range) (h) between onset of fever and sampling was 48 (24 to 66) for samples with EV-positive results and 96 (30 to 96) for samples with EV-negative results.

With regard to other pathogens in CSF samples, human herpesvirus 6 (HHV-6) was detected in 3 patients who were negative for EV. Typical HHV-6 symptoms (such as exanthema or seizures) were not observed in them. On the other hand, HHV-6 was not found in either nasopharyngeal aspirate or blood samples, suggesting a viral latency in CSF samples. The neurological clinical features and the clinical course were similar to those in the rest of the patients.

To conclude, the FA-M/E panel was prospectively used to test samples from 20 children with brainstem symptoms with/without paresis in whom EV was detected in respiratory samples and/or fecal samples but not in CSF samples using a pan-EV real-time PCR. The FA-M/E detected EV in the CSF samples of 4 cases. These detections were confirmatory of the ongoing outbreak due to an EV neuroinfection, which was subsequently identified as EV-A71. We hypothesize that, like other nested multiplex PCRs (7), the FA-M/E panel could have a higher sensitivity to detect low viral loads than single-step real-time reverse transcription (RT)-PCR procedures.