Virological Diagnosis of Central Nervous System Infections by Use of PCR Coupled with Mass Spectrometry Analysis of Cerebrospinal Fluid Samples

Abstract

Viruses are the leading cause of central nervous system (CNS) infections, ahead of bacteria, parasites, and fungal agents. A rapid and comprehensive virologic diagnostic testing method is needed to improve the therapeutic management of hospitalized pediatric or adult patients. In this study, we assessed the clinical performance of PCR amplification coupled with electrospray ionization-time of flight mass spectrometry analysis (PCR-MS) for the diagnosis of viral CNS infections. Three hundred twenty-seven cerebrospinal fluid (CSF) samples prospectively tested by routine PCR assays between 2004 and 2012 in two university hospital centers (Toulouse and Reims, France) were retrospectively analyzed by PCR-MS analysis using primers targeted to adenovirus, human herpesviruses 1 to 8 (HHV-1 to -8), polyomaviruses BK and JC, parvovirus B19, and enteroviruses (EV). PCR-MS detected single or multiple virus infections in 190 (83%) of the 229 samples that tested positive by routine PCR analysis and in 10 (10.2%) of the 98 samples that tested negative. The PCR-MS results correlated well with herpes simplex virus 1 (HSV-1), varicella-zoster virus (VZV), and EV detection by routine PCR assays (kappa values [95% confidence intervals], 0.80 [0.69 to 0.92], 0.85 [0.71 to 0.98], and 0.84 [0.78 to 0.90], respectively), whereas a weak correlation was observed with Epstein-Barr virus (EBV) (0.34 [0.10 to 0.58]). Twenty-six coinfections and 16 instances of uncommon neurotropic viruses (HHV-7 [n = 13], parvovirus B19 [n = 2], and adenovirus [n = 1]) were identified by the PCR-MS analysis, whereas only 4 coinfections had been prospectively evidenced using routine PCR assays (P < 0.01). In conclusion, our results demonstrated that PCR-MS analysis is a valuable tool to identify common neurotropic viruses in CSF (with, however, limitations that were identified regarding EBV and EV detection) and may be of major interest in better understanding the clinical impact of multiple or neglected viral neurological infections.

INTRODUCTION

Viruses are the leading cause of central nervous system (CNS) infections, ahead of bacteria, parasites, and fungal agents (1, 2). They are responsible for a wide spectrum of neurological disorders, ranging from the frequent and benign aseptic meningitis due to enteroviruses (EV) to the rare but serious herpes simplex virus (HSV)-caused encephalitis (3–5). PCR has been recognized as the reference method for the diagnosis of viral CNS infections (6–10). Gene amplification by PCR allows both sensitive and specific virus detection in the cerebrospinal fluid (CSF). It also offers the rapid virological diagnosis required to improve therapeutic management by antiviral therapy in order to limit brain necrosis in cases of herpes simplex virus encephalitis or to increase cost savings in hospitalized cases of enterovirus-related aseptic meningitis during the epidemic season (7, 8, 11, 12). Moreover, it has been shown that quantitation of viral nucleic acid by real-time PCR assay of CSF is useful in monitoring the effectiveness of antiviral therapy, as well as for establishing the prognosis of the disease (13–15). However, the small volume of CSF available, combined with the wide number of DNA or RNA viruses potentially responsible for meningitis or encephalitis, makes it challenging to complete an exhaustive diagnostic profile. Currently, the virological diagnosis is determined through combinations of multiple PCR and reverse transcription-PCR assays, which lead to a virological detection in only 45% to 52% of clinically suspected CNS infections (16, 17). A comprehensive virologic diagnostic testing method is needed for rapid and broad viral detection and quantitation in CSF samples of patients hospitalized for CNS infections.

This study retrospectively assessed the clinical performance of a new technology coupling PCR amplification with electrospray ionization-time of flight mass spectrometry analysis (PCR-MS) for the diagnosis of viral CNS infections (18, 19). Three hundred twenty-seven CSF specimens taken from patients hospitalized in two French university hospital centers for clinically suspected CNS infections were retrospectively analyzed using the PCR-MS technique. The results obtained by PCR-MS analysis were compared to those of the qualitative routine PCR assays used prospectively and to those of real-time quantitative PCR assays performed retrospectively.

MATERIALS AND METHODS

Study design.

Cerebrospinal fluid samples collected by the virology laboratories of the university hospitals of Toulouse and Reims were retrospectively selected because (i) they had been categorized as positive or negative by routine PCR assays for detection of the main neurotropic viruses and (ii) a minimum volume of 300 μl remained for each of them. The selected samples were nucleic acid extracted, followed by PCR-MS analysis at the virology laboratory of the Saint-Louis hospital (Paris). The nucleic acid extracts corresponding to positive or discordant results obtained for routine PCR techniques and PCR-MS for CSF sample analyses were sent back to the virology laboratory of Reims in order to be tested using a third molecular assay that allows a viral load assessment.

Clinical specimens.

The 327 CSF samples were obtained from 136 children (sex ratio [male/female], 1.7; median age of 6 years, ranging from 2 to 18) and 182 adults (sex ratio [male/female], 0.9; median age of 54 years, ranging from 19 to 105) hospitalized between 2004 and 2012 and were routinely sent to the Toulouse and Reims laboratories for clinically suspected neurological virus infections. Cerebrospinal fluids were tested prospectively for human herpesviruses (HHV) (including HSV-1, HSV-2, varicella-zoster virus [VZV], cytomegalovirus [CMV], Epstein-Barr virus [EBV], and HHV-6), enterovirus, and JC virus (JCV) by in-house and commercially available PCR assays used in daily practice in the two laboratories (Table 1) (20–26). CSF samples had been routinely divided in aliquots and stored at −80°C since the date of collection.

TABLE 1.

Extraction methods and PCR tests used prospective CSF analysis and for retrospective analysis of results that were discordant between routine PCR tests and PCR-MS^a

Purpose of method, virus detected	Method used for indicated type of analysis by indicated laboratory
	Prospective analysis		Retrospective analysis of discordant results, Reims University Hospital laboratory
	Toulouse University Hospital laboratory	Reims University Hospital laboratory
Extraction	MagNA Pure LC Total Nucleic kit MagNA Pure LC instrument (Roche) (input sample volume, 200 μl; output elution volume, 100 μl)	Until 2009, QIAamp viral RNA MiniKit, QIAamp DNA MiniKit (Qiagen); Since 2009, NucliSens EasyMAG instrument (bioMérieux) (input sample volume, 300 μl; output elution volume, 75 μl)	PLEX-ID total nucleic acid isolation kit (Abbott) (input sample volume, 300 μl; output elution volume, 200 μl)
PCR assay
HSV-1 and -2	In-house method (20)	Herpes consensus generic (Argene-bioMérieux)	HSV1 HSV2 VZV R-Gene (Argene-bioMérieux)
VZV	In-house method (21)
CMV	In-house method (22)		CMV HHV6,7,8 R-Gene (Argene-bioMérieux)
EBV	In-house method (23)		EBV R-Gene (Argene-bioMérieux)
HHV-6	In-house method (24)		CMV HHV6,7,8 R-Gene (Argene-bioMérieux)
Enterovirus	In-house method (25)	Enterovirus consensus (Argene-bioMérieux), Enterovirus R-gene (Argene-bioMérieux)	In-house method (29)
JCV	In-house method (26)	JC/BK consensus (Argene-bioMérieux)	JC/BK consensus (Argene-bioMérieux)

^aThe extraction methods and PCR tests were used by the Toulouse and Reims virology laboratories for prospective CSF analysis and by the Reims virology laboratory for retrospective analysis of discordant results between routine PCR tests and PCR-MS. References for in-house methods are in parentheses.

Among these 327 CSF samples, 229 had prospectively tested positive for single (n = 225) and multiple (n = 4) viral infections with the routinely used PCR assays, whereas 98 were negative. The distribution of the virus-positive CSF samples selected for the present investigation is shown in Table 2.

TABLE 2.

Detection of neurotropic viruses in 327 CSF samples using routine PCR assays and PCR-MS

Parameter	Virus	No. of CSF samples with indicated result in indicated assay
		Positive (n = 229)		Negative (n = 98)
		Routine PCR	PCR-MS	Routine PCR	PCR-MS
Total no. of virus detections (%)		233 (100)	218 (93.6)	0	12 (12.2)
No. (%) of positive CSF samples		229 (100)	190 (83)	0	12 (12.2)
No. of monoinfections/no. of positive samples (%)		225/229 (98.3)	164/190 (86.3)	0	12/12 (100)
No. of coinfections/no. of positive samples (%)		4/229 (1.7)	26/190 (13.7)	0	0
No. of samples positive for:
Classical neurotropic viruses	HSV	25	27	0	4
	VZV	16	19	0	0
	CMV	1	1	0	0
	EBV	26	11	0	5
	HHV-6	6	11	0	0
	JCV	5	6	0	0
	Enterovirus	154	129	0	1
Viruses not included in routine diagnosis	HHV-7	NDa	11	ND	2
	ADV	ND	1	ND	0
	PVB19	ND	2	ND	0
Total		233	218	0	12

^aND, not done.

PCR-MS.

CSF samples were retrospectively analyzed with the PLEX-ID system (Abbott Molecular, Abbott Park, IL), as previously reported (18, 19). Briefly, RNA and DNA were extracted from 300 μl of the clinical specimens using the PLEX-ID total nucleic acid isolation kits on the PLEX-ID FH and SP instruments and recovered in 200 μl. For each clinical sample, 80 μl of nucleic acids was then distributed by the PLEX-ID fluid handler into 8 reaction wells of 96-well assay plates and amplified with the Viral IC Spectrum II assay (developed by Abbott Molecular, IL, but not yet commercially available), containing primers targeted to adenovirus, HHV-1 to -8, polyomaviruses BK and JC, parvovirus B19 (PVB19), and EV (Table 3). Cycling was done on the PLEX-ID TC (Mastercycler ProS; Eppendorf) according to the manufacturer's instructions. After desalting and purification, amplicons were analyzed by using the PLEX-ID analyzer. Methanol-based aerosols containing denatured ionized amplicons were sprayed into the mass spectrometer. The molecular weight of the amplicons was determined by an electrospray ionization-time of flight mass spectrometer and converted to base composition by database analysis. The virus was then identified by bioinformatics analysis of the base composition signatures produced by the virus target genes in each sample (27, 28).

TABLE 3.

Primers included in the viral IC spectrum II assay for virus detection in CSF samples using the PCR-MS

Virus targeted	Gene targeteda	Primer sequence
		Forward	Reverse
Adenovirus	Penton	TCGTTCCTGCCCTCACAGATCACG	TAGGTCCGGCGACTGGCGTCAGT
	Hexon	TTGCAAGATGGCCACCCCATCGAT	TGTGGCGCGGGCGAACTGCA
Polyomaviruses BK and JC	VP1	TGATGGCCCCAACCAAAAGAAAAG	TAGTTTTGGCACTTGCACGGG
	VP2	TGCCTTTACTTCTAGGGCTGTACGG	TAGTTTTGGCACTTGCACGGG
Enterovirus	5′ UTR	TTCCTCCGGCCCCTGAATG	TGAAACACGGGCACCGAAAGTAGT
	5′ UTR	TGGCTGCGTTGGCGGCC	TAGCCGCATTCAGGGGCCGGA
Erythrovirus (parvovirus B19)	NS1	TGGGCCGCCAAGTACTGGAAAAAC	TGTTTTCATTATTCCAGTTAACCATGCCATA
	VP1	TTACACAAGCCTGGGCAAGTTAGC	TCCTGAATCCTTGCAGCACTGTC
Alphaherpesvirinae (HHV-1 to -3)	DNA POL	TAAGCAGCAGCTGGCCATCAA	TGCCACCCCCGTGAAGCCGTA
Betaherpesvirinae (HHV-6 and -7)	DNA POL	TCGTCCCCATCGACATGTAC	TACTGTGTCCAGCTTGTAGTCTGA
Betaherpesvirinae (HHV-5 [CMV])	DNA POL	TGACTTTGCCAGCCTGTACCC	TCAGGGTGGAGTAGCACAGGTT
Gammaherpesvirinae	DNA POL	TCTGGAGTTTGACAGCGAATTCGAG	TGTTGTAACCGGTGGCGAACTCGGG
	DNA POL	TCCGCGCGGTATAATGCATGATGG	TAGAACATACGCGGTTCCGAGTCACAAA
	DNA POL	TCGCGCCCAGGTAGGC	TGGCCCCGGCCTCGTAGTG

^aUTR, untranslated region; POL, polymerase.

Viral load assessment was obtained by quantifying the total number of amplicons against an internal calibrant with a known copy number that is included in every well. The internal calibrant competes with the amplicon for the same primers and PCR reagents; it additionally serves as an internal PCR control to confirm negative results (19).

Discrepancy methods used for analysis of discordant results between routine PCR assays and PCR-MS.

CSF samples were tested with a third molecular technique if they displayed either (i) discordant results between prospective and retrospective analyses or (ii) additional viruses detected with PCR-MS that were not included in the prospectively used PCR panel. The same nucleic acid extracts as used for PCR-MS analysis were tested either with endpoint PCR (JC/BK consensus kit; Argene bioMérieux, Verniolle, France) or real-time quantitative PCR (RT-qPCR) assays for HSV-1 and -2, VZV, CMV, EBV, HHV-6, HHV-7, PVB19, adenovirus, and EV, allowing viral load assessment in virus-positive CSF samples, except for HHV-7 (HSV1 HSV2 VZV R-Gene, CMV HHV6,7,8 R-Gene, EBV R-Gene, and Adenovirus R-Gene; Argene bioMérieux, Verniolle, France) (Table 1) (29, 30).

The same real-time quantitative PCR techniques were also used to determine the viral load levels in positive samples by both routine PCR assays and PCR-MS in order to compare them with those assessed in the samples displaying discordant results.

Statistical analyses.

The accuracy of the PCR-MS in detecting neurotropic viruses in CSF samples was determined by sensitivity and specificity, considering the routine PCR assays used for prospective analysis of CSF samples as the reference method. The kappa test was also used for measuring the agreement between the PCR-MS system and routine PCR assay. The Spearman's rank correlation was used to evaluate linear associations between viral loads assessed with the PCR-MS and the RT-qPCR assays used as the third analytical technique. Statistical analyses were carried out with SAS software, version 8.2 (SAS Institute, Cary, NC). Results were considered significant for two-sided P values of <0.05.

RESULTS

Analysis of the 327 selected CSF samples by the PCR-MS system.

Among the 229 CSF samples that prospectively tested positive using routine PCR assays, the PCR-MS system identified 218 viruses in 190 specimens (Table 2). The PCR-MS assay showed results concordant with those of routine PCR techniques for 187 viruses (81.7%), corresponding to 183 positive CSF samples (79.9%). The kappa coefficients between the PCR-MS system and routine PCR assays appeared to be higher than 0.80 for HSV, VZV, and EV and 0.34 for EBV (Table 4). In comparison to the routine PCR assays, the PCR-MS demonstrated sensitivity of above 90% for HSV and VZV, 80% for EV, and about 30% for EBV. The specificity ranged from 97% for HSV and EBV to 99% for VZV and EV. The low number of samples positive for CMV, HHV6, and JCV did not allow us to assess the accuracy of the PCR-MS in detecting these neurotropic viruses in CSF samples (Table 4). Moreover, PCR-MS identified 26 coinfections, whereas only 4 had been detected by routine PCR techniques (P < 0.01). These coinfections consisted of 24 double infections and 2 triple infections. They involved (i) EV and human herpesviruses (n = 18 [70%], mainly EV and HHV-6 [n = 4] or HHV-7 [n = 7]), (ii) EV and other viral species (n = 4 [15%]), and (iii) different species of human herpesviruses (n = 4 [15%]) (not shown).

TABLE 4.

Comparison of neurotropic virus detection in CSF samples using routine PCR and PCR-MS assays^a

PCR-MS result	No. of samples with indicated result in routine PCR assay
	HSV		VZV		CMV		EBV		HHV-6		JCV		Enterovirus
	Pos	Neg	Pos	Neg	Pos	Neg	Pos	Neg	Pos	Neg	Pos	Neg	Pos	Neg
Pos	23	8c	15	4e	1	0	8	8g	6	5h	5	1i	129	1k
Neg	2b	294	1d	307	0	326	18f	293	0	316	0	321	25j	172
Kappa test (95% CI)	0.80 (0.69–0.92)		0.85 (0.71–0.98)		NC		0.34 (0.10–0.58)		NC		NC		0.84 (0.78–0.90)
Sensitivity (%)	92		94		NC		31		NC		NC		84
Specificity (%)	97		99		NC		97		NC		NC		99

^aA total of 327 CSF samples collected by the Reims and Toulouse virology laboratories between 2004 and 2012 were used. Pos, positive; Neg, negative; 95% CI, 95% confidence interval; NC, not calculated.

^bNeither sample was shown to be positive for HSV DNA by a third PCR method.

^cOne of these 8 samples was shown to be positive for HSV DNA by a third PCR method.

^dThis sample was shown to be negative for VZV DNA by a third PCR method.

^eOne of these 4 samples was shown to be positive for VZV DNA by a third PCR method.

^fNine of these 18 samples were shown to be positive for EBV DNA by a third PCR method.

^gTwo of these 8 samples were shown to be positive for EBV DNA by a third PCR method.

^hFour of these 5 samples were shown to be positive for HHV-6 DNA by a third PCR method.

ⁱThis sample was shown to be negative for JCV DNA by a third PCR method.

^jEight of these 25 samples were shown to be positive for EV RNA by a third PCR method.

^kThis sample was shown to be positive for EV RNA by a third PCR method.

Among the 98 negative CSF samples identified by routine PCR assays, 10 classical neurotropic viruses were detected with PCR-MS (10/98 [10.2%]). Interestingly, PCR-MS identified 4 HSV CNS infections that were undetected by routine PCR assays (Table 2).

Finally, among the 327 CSF samples tested, PCR-MS identified 16 instances of viruses that were not initially included in the panel of neurotropic viruses detected by routine PCR assays (Table 2). These newly detected viruses were HHV-7 (n = 13), PVB19 (n = 2), and adenovirus (n = 1) (Table 2).

Analysis of discrepant results between routine PCR techniques and PCR-MS using a third molecular technique.

Clinical samples demonstrating results that were discrepant between routine PCR techniques and PCR-MS were analyzed with a third molecular technique consisting of either endpoint (JCV) or RT-qPCR assays that allow virus detection and quantitation, except for HHV-7. Overall, 73 results were discordant between routine PCR techniques and PCR-MS (Table 4).

Forty-six viral infections prospectively diagnosed by routine PCR techniques were not detected retrospectively by PCR-MS. Among these 46 routine PCR-positive and PCR-MS-negative CSF samples, 17 (37%) were confirmed positive by a third PCR assay, whereas 29 (63%) remained undetectable (Table 4). Nine of the 18 EBV- and 8 of the 25 EV-positive CSF samples that were not detected retrospectively with PCR-MS were confirmed by RT-qPCR (Table 4). They demonstrated median viral loads significantly lower than those measured in positive samples by both routine PCR and PCR-MS techniques (112 copies/ml [range, 23 to 384] versus 2,620 [range, 93 to 10,000] [P < 0.01] for EBV and 217 copies/ml [range, 32 to 5,650] versus 3,695 [range, 97 to 212,000] [P < 0.01] for EV).

Conversely, 27 neurotropic viruses were detected by PCR-MS alone. Nine of these 27 viral detections (33.3%) were confirmed by the third molecular technique used to analyze discrepant samples (Table 4).

Comparison of viral load assessment between PCR-MS and real-time quantitative PCR assays.

Correlation analysis of the viral load levels obtained by PCR-MS and RT-qPCR assays was performed for CSF samples positive for HSV, VZV, EBV, and EV. The quantitative values obtained with RT-qPCR and PCR-MS were compared and showed poor correlation with r², ranging from 0.003 (P = 0.77) for HSV to 0.415 for EBV (P = 0.08) (data not shown).

DISCUSSION

The aim of this work was to assess the performance characteristics of a new test coupling multiplex PCR amplification with electrospray ionization-time of flight mass spectrometry analysis (PCR-MS) for the diagnosis of viral CNS infections in comparison with routine PCR assays used as reference methods. Among the 229 virus-positive CSF samples, the PCR-MS results were well correlated with the results of routine PCR techniques for HSV, VZV, and EV detection, demonstrating kappa coefficients greater than or equal to 0.8. The agreement was, however, weaker for EBV detection (kappa coefficient = 0.34), since only 8 of the 26 EBV-positive CSF specimens selected were retrospectively detected by PCR-MS. This suggests a sensitivity defect of the viral IC spectrum II kit regarding this virus (Table 4).

Among the 229 positive CSF samples tested, the PCR-MS system failed to detect 46 viruses identified by routine PCR techniques (Table 4). Three hypotheses can be proposed to explain these discrepancies, as follows. (i) Nucleic acid degradation occurred either during the storage of the samples from the date of collection or at the time of sample thawing—this hypothesis could be particularly true for samples with low viral loads (101 [44%] of the 229 positive CSF samples selected for this study displayed a viral load of <500 copies/ml by RT-qPCR) and RNA viruses, such as EV, which were not detected retrospectively in 25 of the 154 positive samples. (ii) The PCR-MS technology has a lower sensitivity than routine PCR assays—the quantitation of EBV and EV viral load levels revealed that they were significantly higher in the concordant CSF samples that were positive by both techniques than in discordant CSF samples not detected retrospectively by the PCR-MS. (iii) There was a higher concentration of nucleic acids in extracts used for prospective analysis by routine classical techniques than in those used for retrospective PCR-MS analysis—the ratio between the CSF volume subjected to extraction and the elution volume of nucleic acids ranged from 2 (Toulouse) to 4 (Reims) for routine PCR analyses, whereas it was only 1.5 for PCR-MS analysis, leading to signal loss, particularly for CSF samples with low viral loads.

Two main strengths of the PCR-MS technology have been highlighted in the present investigation on CSF samples. First, this new technology significantly improved the detection of multiple viral infections in comparison to the results of routine PCR assays (26 versus 4, P < 0.01). These coinfections mainly involved EV and human herpesviruses. The clinical relevance of viral coinfections of the CNS remains undefined, especially for those infections including HHV-7 and/or HHV-6 (10, 31). These viruses could be considered an innocent bystander of the immune response in the cerebrospinal fluid, passively carried in inflammatory cells to the central nervous system. In contrast, simultaneous infection of the brain by several viruses could also result in an increase of the inflammatory response, edema, or cerebral necrosis. Second, PCR-MS technology identified a significant number of viruses not included in the panel of classical neurotropic viruses detected by routine PCR techniques. Human herpesvirus 7 (n = 13) is currently considered an orphan virus with uncertain neuropathogenic properties, even though it is frequently detected in CSF samples, particularly of children (33, 34). The literature on PVB19 (n = 2) is much more extensive. Many neurological manifestations have been described in patients infected with PVB19, including encephalitis, aseptic meningitis, vasculitis, and peripheral neuropathy, involving mostly children and, in one-third of the cases, immunocompromised patients (35). Adenoviruses (n = 1) are opportunistic viruses responsible for severe disseminated infections in immunocompromised patients, particularly in bone marrow transplant recipients (36). Their involvement in immunocompetent patients is limited to rare cases of aseptic meningitis (37). Overall, the use of this innovative technology combining PCR and mass spectrometry could drastically increase the panel of pathogens detected in comparison with those detected by routine PCR assays. In a recent study, this new technology allowed the detection and semiquantitation in cardiac tissues of 84 different common or uncommon viruses, including confirmed neurological pathogens, such as dengue virus, West Nile virus, and Japanese or St. Louis encephalitis viruses (19). The addition of emerging viruses in neurological infectious disease, such as Nipah virus, Hendrah virus, or parechoviruses, to the viral IC spectrum II kit panel, as well as the combined detection of bacteria, fungi, and parasites, would make this new technology more attractive and could facilitate its implementation in clinical microbiology laboratories (38, 39).

The quantitative aspect of the PCR-MS technology is, however, more questionable. The quantitative values of RT-qPCR and PCR-MS were compared and showed poor correlation, with r² less than or equal to 0.415 for all targets. The ability of the PCR-MS technology to generate quantitative data in the CSF did not seem to be reliable through the results obtained in the present work.

The qualitative approach proposed by the PCR-MS therefore implicates this new technology as a rapid-screening, broad-spectrum method. In fact, the results concerning all 13 viral species detected by the viral IC spectrum II assay can be given to the clinician within 6 to 8 h after the arrival of the sample in the laboratory. A single technician can carry out this new technology using only one CSF aliquot of 300 μl. In this way, PCR-MS technology could find its place in clinical virology laboratories as a first-line test to perform a rapid and broad virological diagnosis in intensive care unit patients with severe disease or in immunocompromised patients at risk for developing a severe infection potentially induced by a wide range of viral species (40). Moreover, PCR-MS technology could be the second-line test used for neurological infections in immunocompetent patients when conventional molecular techniques detecting the most common pathogens, such as HSV, VZV, EBV, and EV, are negative. Following the qualitative analysis by PCR-MS, allowing a rapid implementation of appropriate medical care, an RT-qPCR would be performed in order to assess the viral load in the CSF for prognosis and quantitative monitoring of the disease under proper potential antiviral therapy (13–15). However, the sensitivity defect of the PCR-MS regarding EBV and EV detection identified in our study could be a limitation of its use in its present form as a first-line test in clinical practice, since sensitivity at least equivalent to that of the current molecular methods is mandatory. Moreover, the cost of the PLEX-ID platform could also limit its implementation in many clinical microbiology laboratories, either as a widely used first-line technique or as a second-line test whose cost would then add to the molecular techniques used routinely.

In conclusion, our results demonstrated that PCR-MS analysis is a valuable tool to identify the main common neurotropic viruses responsible for meningitis and encephalitis in hospitalized pediatric and adult patients. However, we observed lower sensitivity for the detection of EBV and EV by the PCR-MS method. The results of this study suggest that multiplex molecular platforms, such as PCR-MS, may allow for a greater understanding of the pathophysiology of CNS viral infections by allowing for a better assessment of the clinical implications of viral coinfections of the CNS and an increased understanding of the potential pathogenicity of neglected neurotropic viruses.