Detection by PCR of Enteroviruses in Cerebrospinal Fluid during a Summer Outbreak of Aseptic Meningitis in Switzerland

Abstract

Enteroviruses (EV) are among the most common causes of aseptic meningitis. Standard diagnostic techniques are often too slow and lack sensitivity to be of clinical relevance. EV RNA can be detected within 5 h by a commercially available reverse transcription-PCR (RT-PCR) test kit. Cerebrospinal fluid (CSF) samples from 68 patients presenting with aseptic meningitis during a summer outbreak in Switzerland were examined in parallel with cell culture and commercial RT-PCR. RT-PCR was positive in all 16 CSF specimens positive by cell culture (100%). In addition, 42 of 52 (80%) CSF samples negative by cell culture were PCR positive. In 26 of these 42 (62%) patients, viral culture from other sites (throat swab or stool) was also positive. The CSF virus culture took 3 to 7 days to become positive. Echovirus 30 was the type most often isolated in this outbreak. The sensitivity of CSF RT-PCR based on clinical diagnosis during this aseptic meningitis outbreak in patients with negative bacterial culture results was 85%, i.e., considerably higher than the sensitivity of CSF virus culture (24%). We conclude that this commercial RT-PCR assay allows a positive diagnosis with minimal delay and may thus influence clinical decisions.

Enteroviruses (EVs) currently account for 80 to 92% of all cases of aseptic meningitis for which an etiologic agent is identified (18). Infections due to EVs are often initially indistinguishable on clinical grounds from those due to bacterial pathogens (21), for which there is specific treatment, but the clinical course is usually self-limiting. Conventional laboratory diagnosis of EV meningitis relies on virus isolation from cerebrospinal fluid (CSF) in cell culture, followed by neutralization typing. This method is time-consuming and frequently unsuccessful, because of low titers of virus in CSF and difficulties in propagation of certain EV types in cell culture (3). Evidence for recent EV infection can also be gained from stool culture or demonstration of specific immunoglobulin M antibody (11), but this provides only circumstantial evidence of an etiologic role in current central nervous system (CNS) pathology. Improved techniques to demonstrate EV in CSF samples are therefore welcome. Detection of EV CNS infection through the amplification of EV RNA from CSF by a reverse transcription-PCR RT-PCR assay has been reported (17). Because almost all EV serotypes have a conserved 5′ nontranslated region, the use of PCR with primers from this region offers a means of identifying the majority of EVs that infect humans by using a single assay (6, 25). A number of reports have described the use of single or nested PCR for the amplification of EV RNA in clinical samples, and different detection systems have been employed (1, 2, 5, 9, 12, 13, 19, 20, 22). A new commercial test combining RT-PCR with a colorimetric microwell hybridization assay (19) has recently been developed (8, 24). The aim of the present study was to evaluate the diagnostic potential of this commercially available RT-PCR assay to detect EV in CSF samples from patients with a diagnosis of aseptic meningitis during a summer outbreak and to compare the results with those obtained by conventional virus isolation methods.

MATERIALS AND METHODS

Community outbreak.

Between June and September 1996, an outbreak of aseptic meningitis occurred in Fribourg, Switzerland. CSF and stool and/or throat samples from patients with clinically suspected aseptic meningitis were sent for diagnostic evaluation.

Patient selection.

During this outbreak, CSF from 83 patients was submitted to our diagnostic laboratory for virus culture. In 3 cases, there was insufficient material to perform virus culture and RT-PCR in parallel. Twelve patients were discharged from the hospital with another diagnosis. The case definition of aseptic meningitis was based on clinical and CSF findings compatible with viral meningitis and the lack of an alternative diagnosis (negative cultures for bacteria and fungi). Sixty-eight patients were discharged from the hospital with a final diagnosis of viral meningitis. Elevated CSF leukocyte (WBC) counts were not required for the diagnosis of aseptic meningitis; three patients whose CSF contained few or no cells were also included in the study. According to Chonmaitree and coworkers (3), CSF from such patients should also be subjected to virus culture if clinical signs and symptoms suggest EV CNS disease, particularly during an epidemic, because aseptic meningitis may be developing in the absence of initial CSF pleocytosis. The study population of 68 patients with aseptic meningitis comprised 27 females 1.6 to 60 years of age (median, 22 years) and 41 males 1 month to 42 years of age (median, 13.5 years). Sixty-five patients had pleocytosis of 6 to 2,200 (median, 120) WBC/mm³.

Viral culture.

CSF from 80 patients and stool samples from 46 patients were collected in sterile containers, and throat swabs from 36 patients were placed in viral transport medium (0.2 M sucrose-phosphate containing 1% bovine serum albumin, 0.1 mg of gentamicin per ml, and 2.5 mg of amphotericin B per ml. All specimens were transported to our institute at ambient temperature for 1 to 3 (median, 1.5) days. CSF was cultured directly, throat specimens and 10% stool suspensions (in phosphate-buffered saline) were centrifuged, and the supernatant was used for inoculation of virus culture. The culture was performed with three different cell lines (Vero, MRC 5, and Hep-2). Virus cultures were inoculated with a minimum of 0.2 ml of specimen and incubated at 35°C. The tubes were observed daily for cytopathic effect (CPE) characteristic for EV for the first 7 days and finally at day 10. Cultures showing CPE were subpassaged, and if positive, the virus was typed by neutralization with the Lim–Benyesh-Melnick equine antiserum pools (10). Multiple positive specimens from the same patient were typed separately.

RT-PCR.

One hundred microliters of CSF, kept frozen at −80°C after inoculation of the virus cultures, was used to perform RT-PCR with the Amplicor Enterovirus test (Roche Molecular Systems, Basel, Switzerland) according to the manufacturer’s instructions. Briefly, the primers EV1b and EV2b from the 5′ noncoding conserved region were used for amplification. Only echoviruses 22 and 23 were not amplified by these primers. RT and amplification were performed by a single enzyme. Carryover contamination was prevented by using dUTP and uracil N-glycosylase to inactivate any carryover products. Detection was performed colorimetrically in a microwell plate coated with the oligonucleotide probe EV3. Each assay run included one positive and three negative controls. Specimens were considered positive if they had an optical density greater than or equal to 0.35 at 450 nm. Samples were amplified in duplicate whenever sufficient material was available. Precautions to prevent laboratory contamination included specimen processing and master mix preparation in separate rooms with filter-plugged pipette tips and amplification and detection in a third room, with one-way traffic during each working day.

RESULTS

Laboratory results from the 68 patients selected for the evaluation are shown in Table 1. Echovirus 30 was the virus most frequently isolated in this outbreak. According to the results of RT-PCR and virus culture of CSF, stool, and throat swabs, the patients were divided retrospectively into four groups.

TABLE 1.

Results from virus isolation, RT-PCR, and characterization of 68 patients with negative bacterial CSF culture results and a final clinical diagnosis of aseptic meningitis during a summer outbreak

Group	Patient	Age (yr)	Sexa	CSF WBC count (no. of WBC/mm3)	Virus isolatedb			Result by CSF PCRc
					Stool	Throat	CSF
A	1	19	F	80	E30	E30	E30	+
	2	39	M	104	E30	E30	E30	+
	3	21	F	84	E30	—	E30	+
	4	20	M	40	E30	—	E30	+
	5	22	M	128	—	E30	E30	+
	6	6	M	300	E30	ND	E30	+
	7	14	F	206	E30	ND	E30	+
	8	13	M	85	E30	ND	E30	+
	9	8	M	625	E30	ND	E30	+
	10	0.1	M	11	E30	ND	E30	+
	11	0.2	M	275	CB5	ND	CB5	+
	12	28	M	164	ND	E30	E30	+
	13	42	M	18	ND	E30	E30	+
	14	42	F	497	—	—	E30	+
	15	27	F	90	ND	—	E30	+
	16	29	F	370	ND	ND	E30	+
B	17	16	M	182	E30	E30	—	+
	18	10	M	149	E30	E30	—	+
	19	60	F	586	E30	E30	—	+
	20	32	M	39	E30	E30	—	+
	21	33	M	50	E30	E30	—	+
	22	45	F	93	E30	—	—	+
	23	22	M	423	E30	—	—	+
	24	20	F	88	E30	—	—	+
	25	5	M	10	E30	ND	—	+
	26	12	F	84	E30	ND	—	+
	27	28	F	173	E30	ND	—	+
	28	2	M	6	E30	ND	—	+
	29	4	F	163	E30	ND	—	+
	30	9	M	770	E30	ND	—	+
	31	6	M	158	E30	ND	—	+
	32	9	M	103	E30	ND	—	+
	33	10	M	633	E30	ND	—	+
	34	8	F	36	E30	ND	—	+
	35	9	M	53	E30	ND	—	+
	36	6	M	250	E30	ND	—	+
	37	11	M	114	E30	ND	—	+
	38	21	M	370	ND	E30	—	+
	39	20	F	130	ND	E30	—	+
	40	16	M	245	ND	E30	—	+
	41	33	M	128	—	E30	—	+
	42	31	F	88	—	E9	—	+
C	43	33	F	150	—	—	—	+
	44	21	M	13	—	—	—	+
	45	35	F	1,450	—	—	—	+
	46	29	F	156	—	—	—	+
	47	34	M	21	—	—	—	+
	48	34	M	217	—	ND	—	+
	49	30	M	72	ND	—	—	+
	50	32	M	180	ND	—	—	+
	51	9	M	45	—	ND	—	+
	52	32	M	60	ND	—	—	+
	53	13	F	125	ND	ND	—	+
	54	8	F	42	ND	ND	—	+
	55	6	M	3	ND	ND	—	+
	56	8	M	143	ND	ND	—	+
	57	8	F	210	ND	ND	—	+
	58	9	F	271	ND	ND	—	+
D1	59	4	M	81	CB5	ND	—	−/−
	60	2	F	59	E30	ND	—	−
	61	14	F	2	ND	CB5	—	−/−
D2	62	38	F	103	—	—	—	−/−
	63	23	F	6	—	—	—	−
	64	28	F	2,250	ND	—	—	−/−
	65	9	M	310	ND	—	—	−
	66	0.1	M	133	ND	ND	—	−
	67	27	M	3	ND	ND	—	−/−
	68	22	M	399	ND	—	—	−/−

^aF, female; M, male. 

^bE30, echovirus 30; E9, echovirus 9; CB5, coxsackievirus B5; —, negative; ND, not done. 

^c+, positive; −, negative; −/−, repeatedly negative. 

(i) Group A (confirmed cases).

Group A contained 16 patients (6 females, 10 males) who had positive PCR and virus culture in CSF; virus culture from other sites was positive for 13 of 15 patients. The median age of this group was 20 years, and the median WBC count was 116 WBC/mm³ at a median of 1 day after onset of aseptic meningitis.

(ii) Group B (probable cases).

Group B contained 26 patients (9 females, 17 males) with positive PCR and negative virus culture in CSF; virus culture from other sites was positive for all patients of this group. The median age was 14 years, and the median WBC count was 129 WBC/mm³ at a median of 1.5 days after the onset of aseptic meningitis.

(iii) Group C (compatible cases).

Group C contained 16 patients (7 females, 9 males) with positive PCR and negative virus culture in CSF; virus culture from other sites was performed for 10 patients and remained negative. All 16 patients had typical clinical symptoms, and no other agent was isolated from CSF. The median age of this group was 25 years. The median WBC count was 134 WBC/mm³ at a median of 1 day after onset; all except one (patient 55 with only 3 WBC/mm³) of these CSF samples showed pleocytosis.

(iv) Group D.

Group D contained 10 patients with negative PCR and negative virus culture in CSF. The median age in this group was 18 years; the median WBC count in CSF was 92 WBC/mm³ at a median of 1.5 days after onset of aseptic meningitis. This group was divided into two subgroups, D1 and D2, according to the virus culture results from sites other than CSF. Subgroup D1 was made up of three patients with positive virus culture in stool (two patients) or throat swab (one patient). Patient 62 had only 2 WBC/mm³ in CSF. Subgroup D2 was made up of seven patients with negative virus culture from other sites in five of seven patients. Patient 67 had only 3 WBC/mm³ in CSF.

Assuming that patients with aseptic meningitis during a summer outbreak represent true-positive cases, the sensitivities of RT-PCR, virus culture from CSF, and virus culture from other sites are presented in Table 2. The sensitivities of RT-PCR were 100% (16 of 16) compared to that of CSF virus culture and 93% (42 of 45) compared to that of positive virus culture from any site. The sensitivity of CSF virus culture was 28% (16 of 58) compared to that of RT-PCR.

TABLE 2.

Calculation of the sensitivities of RT-PCR and virus culture for 68 patients with aseptic meningitis during a summer echovirus 30 outbreak

Method	No. of samples positive/no. tested	% Sensitivity
RT-PCR of CSF	58/68	85
Virus culture
CSF	16/68	24
Throat	16/36	44
Stool	33/46	72
Throat and stool	42/59	71
Any site	45/68	66

Virus culture of stool samples was positive for 33 of 46 patients (72%) after 2 to 11 days, virus culture of throat swabs was positive for 16 of 36 patients (44%) after 2 to 7 days, and virus culture from CSF was positive for 16 of 68 patients (24%) after 3 to 7 days. The cumulative time intervals required for positive culture results are shown in Fig. 1.

FIG. 1.

Time required to obtain positive cultures from different specimens from patients with EV meningitis.

All 12 patients discharged from the hospital with a diagnosis other than aseptic meningitis had negative virus culture and RT-PCR results in CSF. For one of them, bacterial culture grew Neisseria meningitidis.

DISCUSSION

PCR was compared with virus isolation to detect EVs in patients with suspected aseptic meningitis who were referred to the hospital. This study confirms that RT-PCR is more sensitive than culture (15, 16, 20). In addition to all CSF specimens from which an EV could be isolated, 80% of the virus culture-negative specimens were RT-PCR positive, which is in accordance with earlier findings (20). In two of three CSF virus culture-negative but RT-PCR-positive patients, virus culture from another site was positive, indicating an active viral infection. Although positive cultures from sites other than the CSF must be interpreted cautiously, since EV excretion may be prolonged and may coincide with another disease process (7), the increased pretest probability during outbreaks may justify the inclusion of such cultures as true positives. Sixteen patients without positive cultures at any site had EV RNA in their CSF detected by RT-PCR (20), including one CSF sample containing only three WBC, suggesting an early stage of the infection (15). All of these patients had a clinical illness characteristic of aseptic meningitis during an outbreak of EV meningitis, and no alternative diagnosis was made for any of them.

This study provides limited data on the specificity of RT-PCR. The 12 patients with a final diagnosis other than aseptic meningitis were RT-PCR negative, including one patient with meningococcal meningitis.

The turnaround time for RT-PCR was shorter than for virus culture in all cases, because virus culture of CSF typically requires 3 to 7 days for a CPE to develop, whereas RT-PCR results are available within 5 h. Since the costs are comparable, RT-PCR is preferable.

Because of the specimen transport and culture conditions used in this study, false-negative virus culture was more likely to occur than false-positive PCR results (13). The addition of primary cells for virus culture would probably have increased the recovery of EVs. The use of multiple cell lines can also increase sensitivity (4), although at an increased cost to the laboratory. In addition, transport of specimens at ambient temperature may lead to loss of viability of viruses. Prolonged transport will mainly affect the results of virus culture in samples with low viral titers, particularly in CSF. Because echovirus 30 was the predominant virus type isolated in this outbreak, and this virus type grows well in cell culture, negative results were probably not due to types that are generally difficult to propagate in cell culture (3). On the other hand, false-positive RT-PCR results can never be excluded, despite strict measures incorporated in this commercial test to avoid contamination and the one-way traffic through three separate rooms dedicated to the performance of RT-PCR.

The negative results of EV RT-PCR in three CSF samples in patients with typical aseptic meningitis and with positive virus culture from other sites may be due to very low virus concentrations in CSF. Alternatively, some EV infections of the CNS may not be accompanied by the presence of the virus in CSF (14). False-negative results may also be due to inhibitory substances which interfere with the function of the enzymes used in PCR. Retesting may be useful, but CSF is frequently available in limited amounts that do not allow this. The incorporation of an internal control into this commercial test would be a great improvement.

Because the outbreak described in this study was essentially due to echovirus 30, the inability of this PCR to amplify the genetically atypical echoviruses 22 and 23 (8) is unlikely to have contributed to false-negative results.

Since RT-PCR is more sensitive than virus culture, there is currently no single “gold standard” test against which the sensitivity and specificity of PCR can be assessed (12). It is therefore important when interpreting PCR results to take into account all laboratory results as well as clinical and epidemiological information.

In conclusion, in EV CNS infections, CSF is the most specific specimen available for diagnostic testing. This study clearly demonstrates that RT-PCR is superior to virus culture of CSF for the diagnosis of EV meningitis and that the clinical usefulness of virus culture from CSF is limited. RT-PCR is more sensitive and the workload is lower than for virus culture. The rapid availability of RT-PCR results increases their potential clinical impact, especially early in the course of the disease. A definitive diagnosis can protect the patient from unnecessary investigations and antibiotic treatment, allow early hospital discharge, and indicate a better prognosis (23).