Enteroviruses are the most frequent cause of aseptic meningitis in the paediatric age group.1 Cases of meningitis caused by enterovirus are usually benign and only require symptomatic treatment. However, the difficulty of differentiating them from bacterial meningitis in the initial assessment may lead to the unnecessary use of antibiotics, hospital admission and increases in the length of stay.2–4 Currently, molecular diagnostic tests for the detection of enterovirus in cerebrospinal fluid (CSF) specimens, which have a high sensitivity and specificity and offer quick results, are displacing viral culture,5,6 which is slower and whose results depend on the viability of the viruses contained in the clinical specimen. Some of these molecular tests are fully automated, as is the case of the Xpert EV assay (Cepheid®), which uses the GeneXpert® Dx system.

The aim of our study was to analyse changes in clinical practice since the introduction of this diagnostic test for the detection of enterovirus in CSF samples in children with suspected viral meningitis.

We conducted a retrospective descriptive study of children aged more than 1 year and less than 15 years that were previously healthy and received a diagnosis of meningitis by enterovirus at the Hospital Universitario de Getafe between November 2006 and February 2013. We collected epidemiological, clinical and laboratory data from the electronic health records of patients that tested positive for enterovirus, as well as data on the length of stay and the use of antibiotherapy. We calculated the modified Boyer score based on clinical and laboratory parameters for each patient, as previously described in the literature.8

Starting in September 2010, the diagnosis of meningitis due to enterovirus was based on a positive result of the Xpert® EV assay (Cepheid). This technique was available in our Department of Microbiology from that date. Prior to that, the diagnosis was made using conventional molecular techniques, submitting specimens to a reference laboratory and with a mean turnaround time of 7–10 days.

The Xpert® EV assay (Cepheid) was performed according to the directions of the manufacturer, and was available between 8:00 am and 8:00 pm everyday except Sundays, when it could be requested between 8:00 am and 1:00 pm.

Cerebrospinal fluid samples were stored in a refrigerator at 4°C until processing, and subsequently kept at −20°C. Paediatricians were notified immediately of results.

The Xpert® EV assay (Cepheid) uses the GeneXpert® Dx system, with integrated performance of sample processing, viral RNA extraction, reverse transcription (RT) and nucleic acid amplification, as well as the detection of the target sequence by real-time polymerase chain reaction (PCR), in a single reaction cartridge. This method simplifies microbiological diagnosis and can provide rapid and accurate results in 2.5h.

We compared patients in the period prior to the introduction of Cepheid's Xpert® EV method, group 1 (November 2006–August 2010) to patients in the period following its introduction, group 2 (September 2010–February 2013).

We calculated the health care costs generated by each patient based on the mean cost of staying in the paediatrics ward (337.55€/day) in the two periods under study, adding the cost of assessing for the presence of enterovirus in CSF samples with the Xpert EV test (Cepheid®) for patients in the second period (80€/specimen).

We have expressed quantitative variables as medians and ranges, and compared them by means of the Mann–Whitney U test. We have expressed qualitative variables as percentages and compared them by means of Fisher's exact test. We defined statistical significance as a P-value of less than .05. The statistical analysis was performed with SPSS version 20.0.

We included a total of 41 patients for the period under study, 60.9% male, and their median age was 64 months (interquartile range, 28–96 months). Twenty-six patients (63.4%) were included in the second period. Twenty-five patients (60.9%) had received a diagnosis between April and July. The most frequent presenting symptoms were headache (89.45%), vomiting (88.95%) and fever (71.25%)

We did not find statistically significant differences in epidemiological, clinical or laboratory characteristics between patients assessed in the first period and those assessed in the second period, save for the latter presenting less often with nuchal rigidity and/on meningeal signs in the initial evaluation (Table 1). We also found no differences between the two groups of patients when we computed the modified Boyer score (score ≥3, 0% in period 1 vs 7.6% in period 2, P>.05). Bacterial growth was not detected in any of the CSF cultures.

The comparative analysis of the two groups found a statistically significant decrease in the median length of stay of patients (48h in group 1 vs 40.5h in group 2; P=.039).

Only 2 patients (13.3%) in the first period and 2 patients (7.6%) in the second period received broad-spectrum antibiotherapy. We observed a decreasing trend in the use of antibiotherapy in the second period, but it was not statistically significant (4.7 days vs 1 day).

We observed a significant reduction in the costs of hospitalization per patient associated with length of stay: 779.77€ (98.35–1362.85) vs 656.05€ (0–1489.93) (P<.05).

In our study, the introduction of the Cepheid Xpert® EV assay was significantly associated with decreases in length of stay and associated health care costs.

Many studies published in recent years have shown the benefits of using PCR methods in the diagnosis of meningitis caused by enterovirus. Although publications on the subject started to appear in the 1990s, it was in year 2000 that 2 studies conducted in the United States directly evaluated their use and found a reduction in length of stay, antibiotic use and health care costs following the introduction of these techniques.3,9

In the past few years, the increased availability and accessibility of rapid, and in some cases automated, RT-PCR techniques have facilitated their widespread use. Data published in Europe and the United States demonstrate their benefits in terms of reduced costs and lengths of hospitalization.7,10–13

Two recent publications have assessed the use of such RT-PCR methods (Cepheid Xpert® EV) in Spain, with very good results.14,15 In the study by Menasalvas Ruiz et al.,14 as many as 25% of patients were discharged from the emergency department following a positive test result, and the median length of stay was 3.5 days (4.5 days in newborns vs 2.3 days in children aged more than 2 years). However, this study did not assess the impact of the introduction of the technique with a control group. The study conducted by Carrasco Fernández et al.,15 which compared 37 children with aseptic meningitis selected prospectively after the introduction of Cepheid's Xpert® EV assay with retrospective data for 29 children with a diagnosis of aseptic meningitis, found that the introduction of this technique was associated with a significant reduction in antibiotic use and duration of antibiotherapy, but did not find significant differences in length of stay.

The analysis of health care costs that we present is an analysis of direct costs relative to the length of hospitalization. One of the limitations of our study is that we did not perform a detailed economic analysis and our results are based on the costs of hospitalization averaged through time. This type of analysis, while very simple, has been used by other authors in the past.10,16 It may not adequately represent the actual savings associated with the use of these methods, although it does reflect the reduction of costs associated with hospitalization, which are probably more relevant in cases of aseptic meningitis. Studies conducted previously in other countries have suggested that the use of these techniques may be associated with reductions in length of stay and economic costs using a similar method for cost analysis.10,15,17

Length of stay in these patients has been associated with test turnaround times.18 One limitation in our study was that the technique was not always available, so that the stay of patients admitted at night or on weekends may have extended until results became available. Another possible limitation was that although there are no priori interpersonal differences in the way the staff of the paediatrics department discharge patients, there could have been variability in the applied discharge criteria, a factor that may have influenced the results we obtained. In spite of these circumstances, we were able to find evidence of the benefits derived from using this technique.

Our study also has other limitations characteristic for retrospective studies. The data used to calculate the length of stay were entered in the electronic records at the times of admission and discharge, so we believe that these data, although collected retrospectively, are reliable.

In conclusion, the introduction of the Xpert® EV assay developed by Cepheid has allowed a significant decrease in the length of stay and associated costs in children with meningitis by enterovirus in Spain.

The authors have no conflict of interests to declare.