Secondary Logo

Journal Logo

Original Studies

Impact of Cerebrospinal Fluid Multiplex Assay on Diagnosis and Outcomes of Central Nervous System Infections in Children

A Before and After Cohort Study

O’Brien, Matthew P. MBBS, BMedSci*; Francis, Joshua R. MBBS, FRACP†,‡; Marr, Ian M. MBBS§; Baird, Robert W. MBBS, FRACP, FRCPA

Author Information
The Pediatric Infectious Disease Journal: September 2018 - Volume 37 - Issue 9 - p 868-871
doi: 10.1097/INF.0000000000001936
  • Free

Abstract

Pediatric central nervous system (CNS) infections are associated with significant mortality and morbidity, including long-term sequela1 with a considerable economic burden to healthcare systems.2 Empiric broad-spectrum antibiotic coverage is recommended for the initial management pending organism identification.3 Delayed diagnostic completion times, low positive pathogen identification and difficulties differentiating a partially treated bacterial meningitis and viral meningitis may all result in unnecessarily prolonged ` antibiotic courses and increased length of hospitalization.4

BioFire Diagnostics (Salt Lake City, UT) developed a qualitative multiplex polymerase chain reaction (PCR) Meningitis and Encephalitis Panel to be used on the FilmArray system (multiplex assay). This diagnostic tool is performed on cerebrospinal fluid (CSF) and is able to identify up to 14 potential pathogens including 6 bacterial [Escherichia coli K1, Haemophilus influenzae (HI), Listeria monocytogenes, Neisseria meningitidis, Streptococcus agalactiae, Streptococcus pneumoniae (SP)], 7 viral [cytomegalovirus (CMV), Enterovirus (EV), Herpes simplex virus (HSV) 1, 2, Human Herpes Virus 6, Human parechovirus (HPeV), Varicella-zoster (VZ) virus] and 1 fungal pathogen, Cryptococcus gattii/neoformans (GG/CN). The most common causes of pediatric viral central nervous system infections include EV and HPeV5,6 and previous studies have demonstrated that rapid diagnosis of EV leads to decreased antimicrobial use, length of hospitalization and reduced hospital costs.7,8

This multiplex assay has been evaluated in several studies including a large multicenter US retrospective evaluation which demonstrated a sensitivity of ≥95% and specificity ≥98.5% for the key target pathogens (EV, HPeV and HSV) and an overall agreement with 2 validated PCR comparator assays of 99.8%. A further retrospective study compared the multiplex assay to conventional assay on 138 stored pediatric CSF specimens and determined a mean 13.3-hour pathogen identification time with conventional methods compared with 3 hours using the multiplex assay (P < 0.0001).9 A further study analyzed 62 neonatal CSF samples and positively identified 12 pathogens, 7 more than by culture alone and all in babies who had received antibiotics before lumbar puncture.10

The aim of this study was to analyze the diagnostic utility and impact on management and clinical outcomes of the multiplex assay for pediatric patients undergoing diagnostic lumbar puncture for suspected CNS infection in a geographically isolated health service.

MATERIALS AND METHODS

The study was undertaken at the Royal Darwin Hospital (RDH), which is located in the Northern Territory, Australia, and services a 1 million square kilometer and 200,000 population catchment. The study duration was 30 months, which included a 15-month prospective cohort from February 2016 to May 2017 and a retrospective cohort November 2014 to February 2016 as this coincided with the introduction of the multiplex assay. The inclusion criteria were pediatric patients 0–16 years of age who underwent lumbar puncture for investigation of possible CNS infection. A diagnosis of definite CNS infection was made on the basis of compatible clinical symptoms (any 2 of fever, vomiting, lethargy, meningism, irritability, headache, photophobia or seizures) plus an organism identified on CSF (Gram stain, culture or nucleic acid testing) and a discharge diagnosis of meningitis or encephalitis. A diagnosis of suspected CNS infection was made on the basis of a compatible clinical syndrome (as above) plus a CSF leukocyte count >5 × 106/L (or >20 × 106/L for infants less than 1 month of age) and a discharge diagnosis of meningitis or encephalitis.

The multiplex assay is fully automated utilizing an enclosed and disposable pouch. A hands-on preparation time of 2 minutes and processing time of an hour per specimen was required.

Data analyzed included the results of biochemistry, microscopy, culture, viral PCR studies, the time to organism identification, duration of antibiotic therapy and length of hospital admission. For determination of sensitivity and specificity, the comparison standard for bacterial/fungal infections was standard culture (performed at the RDH laboratory) and for viral infections the comparator was validated PCR confirmation. HSV and VZV PCR assays were performed in house at the RDH laboratory on demand during weekday work hours (the testing platforms used included the Roche MagNA Pure System–Total Nucleic Acid kit for extraction with a Roche Assay using the LightCycler capillary system for HSV 1/2 and an in-house assay using the Roche LC4800 plate system for VZ in batched daily testing). HPeV and EV PCR were performed interstate at the Victorian Infectious Diseases Reference Laboratory in Melbourne, Victoria (utilizing an Applied Biosystems 7500 fast real-time PCR with a validated in-house designed assay in batched twice daily PCR runs). The start point for all time intervals was the time of presentation to the RDH emergency department. The end-point for organism identification was following organism culture or positive PCR or multiplex result. The end-point for antibiotic duration was the point of antibiotic cessation and the end-point for length of hospitalization was the day of discharge. All time intervals were compared using the Mann–Whitney and a P value of <0.05 was considered significant.

Ethics approval was obtained from the Northern Territory and Menzies School of Health Research Human Research Ethics Committee (HREC Reference Number: 2016–2577).

RESULTS

Over the study period, 251 children underwent lumbar puncture for investigation of CNS infection and 65 (26%) were found to have CNS infection (58 definite, 7 suspected) as illustrated in Figure 1.

FIGURE 1.
FIGURE 1.:
Study inclusion pathway and definitions. Definite CNS infection: compatible clinical symptoms (any two of fever, vomiting, lethargy, meningism, irritability, headache, photophobia or seizures) plus an organism identified on CSF (Gram stain, culture or nucleic acid testing) and a discharge diagnosis of meningitis or encephalitis. Suspected CNS infection: Compatible clinical symptoms (as above) plus a CSF leukocyte count >5 × 106/L (or >20 × 106/L for infants less than 1 month of age) and a discharge diagnosis of meningitis or encephalitis.

Young infants were most commonly affected with 80% (52/65) of cases occurring in children 3 months or younger. The most commonly identified pathogens were EV in 38/65 (57%) children followed by HPeV in 8/65 (12%) children and there were 7/65 (11%) confirmed bacterial CNS infections. Of the 65 patients who underwent lumbar puncture, 64 (98%) had CSF microscopy performed and 21 (33%) of these were found to have normal CSF leukocyte counts, all of whom were <3 months of age. This included 15/38 (39%) of the infections caused by EV and 5/7 (71%) of the HPeV infections. There was 1 further case of HPeV; however, the specimen clotted precluding an accurate cell count. Of the 65 children with CNS infections, 58 (89%) also had CSF biochemistry performed and 48 of these (83%) were found to have an elevated protein count (>0.45 g/L).

In the retrospective cohort, 128 children underwent lumbar puncture and 36 (28%) were determined to have CNS infection (32 confirmed, 4 suspected). Twenty-eight of thirty-six (78%) were <3 months of age (Table, Supplemental Digital Content 1, https://links.lww.com/INF/C981). The most commonly identified organism was EV 22/36 (61%) followed by HPeV 3/36 (8%). One organism was identified as a flavivirus (Murray Valley), 1 as VZ and 1 as CMV. The diagnosis of suspected CMV meningitis was based on a high serum CMV viral load, elevated CSF leukocyte count and raised CSF protein concentration as there was insufficient CSF for CMV PCR assay. There were 4 confirmed cases of bacterial meningitis (2 cases of HI, 1 SP and a 1 Acinetobacter species) and 1 fungal cultured confirmed case of CG. There were no organisms identified in 3/36 children (8.3%) who were also classified as suspected CNS infection based on the specified criteria above.

In the prospective arm, 123 children underwent lumbar puncture and 29 (24%) were determined to have CNS infection (26 confirmed, 3 suspected), of whom 24 (83%) were <3 months of age. Multiplex assay was performed on 55 (45%) children including 26 children with confirmed CNS infection and on 29 patients who were ultimately diagnosed with a non-CNS infection (Table 1).

TABLE 1.
TABLE 1.:
All CSF Samples Taken in Prospective Study Period (February 2016–May 2017)

The multiplex assay identified a potential pathogen in 20 children and the most frequently detected pathogen was EV in 14 children (70%), followed by HPeV in 4 children (20%), HHV-6 in 1 infant and CG in 1 adolescent (Table, Supplemental Digital Content 2, https://links.lww.com/INF/C982). No bacterial pathogens were identified by multiplex assay; however, there was 1 case of bacterial meningitis with an organism in the Enterobacteriaceae family identified on 16s ribosomal RNA PCR sequencing that was unable to be further speciated. Of the 3 children who did not have a multiplex assay performed, 1 cultured positive in the CSF for salmonella species and the other 2 children had insufficient CSF for multiplex though subsequently returned positive for EV on reference laboratory testing.

Of the 20 positive multiplex results for viral pathogens, 16 were referred for interstate reference laboratory PCR and the remaining 4 specimens were of insufficient residual volume for further testing. Fifteen of sixteen specimens sent to the reference laboratory returned positive. Of the 33 negative multiplex results, 10 were on-sent to the reference laboratory and 9 of these returned negative on testing. There were 2 discordant results, the first was a 16-day-old infant with irritability, fevers and vomiting whose CSF was positive for EV on the multiplex assay though was negative for EV and HPeV on reference laboratory testing. The multiplex assay result was treated as a true positive because of strong clinical suspicion of CNS infection. The second discordant result was for a 10-week old with fevers and irritability whose CSF tested negative on the multiplex assay though positive for HPeV on testing at the reference laboratory. This was treated as a false negative result on the basis of a strong clinical suspicion of CNS infection. The multiplex assay therefore was deemed to have 94% sensitivity (15/16) and 90% specificity (9/10) for detection of viral pathogens in the multiplex panel when compared with reference laboratory PCR in this cohort.

As outlined in Table 2, the median time to confirmatory diagnosis reduced from 6.0 to 2.0 days (P value <0.001) in the prospective cohort compared with the retrospective cohort. The turn-around time for reference laboratory testing for EV and HPeV including transportation and processing was between 4 and 14 days (median 7 days) and there was no statistical difference in between the retrospective and prospective cohorts. The median duration of antibiotic treatment also decreased from 3.0 to 2.0 days (P value <0.001) and median hospitalization reduced from 5.0 to 3.0 days (P value 0.016).

TABLE 2.
TABLE 2.:
Comparison of Patient Outcomes Before and After Introduction of Onsite Multiplex Assay

DISCUSSION

This study has been able to characterize 65 pediatric CNS infections in our population over a 30-month time interval. The most frequently identified organisms were EV (58%) and HPeV (12%) and the majority (80%) of the infections occurred in infants 3 months or younger. In the neonatal age range, there was commonly an absence of CSF pleocytosis and/or normal CSF protein concentrations which is consistent with published literature in regards to neonatal viral CNS infections.11,12 A small number of bacterial meningitis cases were detected in our cohort consistent with the low rates of bacterial pathogen identification in the Leber et al8 study, which identified a case rate of 1.4% by multiplex assay and 0.4% by routine CSF culture. These findings reflect the global decrease in cases of pediatric bacterial meningitis following the introduction of the HI, SP and NM immunization programs.13

The multiplex assay performed well against the comparator of reference laboratory testing demonstrating 94% sensitivity and 90% specificity for detection of viral pathogens in the multiplex panel when compared with reference laboratory PCR assay. In addition, this facilitated a statistically significant reduction of the median time to organism identification (reduction of 4.0 days), median duration of antibiotic treatment (reduction of 1.0 days) and median length of hospitalization (reduction of 2.0 days) in children with pediatric CNS infections during the prospective study period. The benefits resulting from this are numerous including a reduced duration of broad-spectrum antibiotic use thereby reducing the risk of short- and long-term adverse drugs effects. In particular, antibiotics may disrupt the neonatal gut microbiome leading to an increased risk of necrotizing enterocolitis and potential negative effects on long-term metabolic and immune development.14 This is particularly relevant for our study population as the majority of the children were younger than 3 months. A reduction in the total duration of antibiotic therapy also reduces the period of time a child requires an IV cannula and therefore decreases the risk of cannula related skin or bloodstream infections. In addition, there are fewer requirements for IV cannula replacement, which is often a distressing procedure for children and their families. From a health economic perspective, there are also benefits in reducing the length of hospitalization for infections. A standard public hospital bed is estimated to cost AUD$584/day based on Independent Hospital Pricing Authority report in 201615 and the multiplex reduced hospitalization by a median of 2.0 days.

The multiplex assay is likely to be of greatest benefit to hospitals and health care centers that do have access to validated in-house PCR assays for the common viral pathogens causing CNS infections in children, such as EV and HPeV. The introduction of a reliable, sensitive and easy to use multiplex assay facilitated same day cessation of antibiotic therapy or in several cases, the withholding of any antibiotic treatment at all. The multiplex can be easily incorporated into similar hospitals that do also not have access to in-house PCR as the multiplex can be easily performed after staff training has occurred.

There are several limitations with the study to acknowledge. This is a single center pre-post evaluation study and therefore the overall sample size was small which may have influenced the statistical certainty of the assay performance calculations. The study also did not have sufficient case numbers to make comment regarding the accuracy or utility of the multiplex assay in the diagnosis of bacterial pediatric CNS infections. Only 30% of negative multiplex assay isolates were on-sent for confirmatory PCR assay which may have impacted on specificity and negative predictive value calculations; however, we note that the majority of these cases had an alternative clinical diagnosis made. Further studies of specificity and sensitivity of the multiplex assay will also need to take into account differing seasonal pathogen occurrences, which our study was not powered to evaluate. Finally, there were also some operational limitations which resulted in the multiplex assay not being able to be performed immediately or shortly after lumbar puncture in many of the cases, reflected in the median time for confirmatory diagnosis of 2.0 days after admission. Had the test had been able to be performed immediately after lumbar puncture in these cases, the potential overall improvement in diagnostic time, duration of antibiotic therapy and length of hospitalization may have been even more significant.

In conclusion, this study supports the use of the multiplex assay as a fast and accurate adjunct diagnostic tool for the most common causes of neonatal and pediatric viral CNS infections. A major benefit of the multiplex is that the assay can be run in a routine laboratory without specialized molecular equipment and may facilitated faster organism identification, decreased duration of antimicrobial therapy and decreased duration of hospital inpatient stay.

ACKNOWLEDGMENTS

The authors acknowledge the pediatric staff and the laboratory staff at the Royal Darwin Hospital for their assistance throughout the study.

REFERENCES

1. Chandran A, Herbert H, Misurski D, et al. Long-term sequelae of childhood bacterial meningitis: an underappreciated problem. Pediatr Infect Dis J. 2011;30:36.
2. Vora NM, Holman RC, Mehal JM, et al. Burden of encephalitis-associated hospitalizations in the United States, 1998-2010. Neurology. 2014;82:443451.
3. Thigpen C, Whitney C, Messonnier N, et al. Bacterial meningitis in the United States – 1998 – 2007. N Engl J Med. 2011; 364.
4. Khumalo J, Nicol M, Hardie D, et al. Diagnostic accuracy of two multiplex real-time polymerase chain reaction assays for the diagnosis of meningitis in children in a resource-limited setting. PLoS One. 2017;12:e0173948.
5. Robart HA. Viral meningitis. Semin Neurol. 2000;20:277.
6. Britton P, Dale RC, Blyth C, et al. The causes and clinical features of childhood encephalitis in Australia (ACE): a multicentre, prospective cohort study. 2016.ESCMID poster presentation.
7. King RL, Lorch SA, Cohen DM, et al. Routine cerebrospinal fluid polymerase chain reaction assaying reduces hospitalization and antibiotic use for infants 90 days or younger. Pediatrics. 2007;120:489496.
8. Leber A, Everhart K, Balada-Llasat JM, et al. Multi-center evaluation of a multiple meningitis/encephalitis PCR Panel for simultaneous detection of bacteria, yeast and viruses in cerebrospinal fluid specimens. J Clin Microbiol. 2016;54:22512261.
9. Messacar K, Breazeale G, Robinson CC, et al. Potential clinical impact of the film array meningitis encephalitis panel in children with suspected central nervous system infections. Diagn Microbiol Infect Dis. 2016;86:118120.
10. Arora HS, Asmar BI, Salimnia H, et al. Enhanced identification of group B Streptococcus and Escherichia Coli in young infants with meningitis using the biofire filmarray meningitis/encephalitis panel. Pediatr Infect Dis J. 2017;36:685687.
11. Mulford WS, Buller RS, Arens MQ, et al. Correlation of cerebrospinal fluid (CSF) cell counts and elevated CSF protein levels with enterovirus reverse transcription-PCR results in pediatric and adult patients. J Clin Microbiol. 2004;42:41994203.
12. Renaud C, Harrison CJ. Human Parechovirus 3: the most common viral cause of meningoencephalitis in young infants. Infect Dis Clin North Am. 2015;29:415428.
13. McIntyre PB, O’Brien KL, Greenwood B, et al. Effect of vaccines on bacterial meningitis worldwide. Lancet. 2012;380:17031711.
14. Gibson MK, Crofts TS, Dantas G. Antibiotics and the developing infant gut microbiota and resistome. Curr Opin Microbiol. 2015;27:5156.
15. Independent Hospital Pricing Authority (IHPA). National Hospital Cost Data Collection: Australian Public Hospital Cost Report 2013–2014 Round 18. 2016.Sydney: IHPA.
Keywords:

CNS infections; pediatrics; multiplex PCR; antibiotic duration; hospitalization

Supplemental Digital Content

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.