Clinical management of patients with suspected CNS infections across the region is heterogeneous and sub-optimal, particularly for the investigation and treatment of viral encephalitis. In a previous study at the tertiary children’s hospital, we reported that the number of children with suspected CNS infection who had an LP (where no contraindications existed) was only 53% . In this study, of the 64 children who did not have a contra-indication to LP (both before and after imaging), 50 (80%) had an LP attempted. This figure is therefore encouraging and may partially reflect the recent implementation of LP guidelines and junior doctors training sessions at the specialist children’s hospital . However, in those children who had an LP and their CSF was suitable for analysis, a full set of standard investigations was only ordered in 24 (57%) with a paired plasma glucose being the main test missed. This is comparable to the findings in our single site study . Furthermore, no patient who had treatment started before their LP had a bacterium isolated from their CSF. This is similar to previous studies that have demonstrated a reduction in the sensitivity of bacterial CSF culture after antibiotics have been started . It seems possible in this study, that some children treated for suspected CNS infection will not have had a microbiological/virological diagnosis due to not having an LP at all, or failing to order specific tests such as CSF PCR. This makes the ongoing clinical decisions such as when to stop aciclovir or third generation cephalosporins very difficult and children may either be exposed to drugs they do not need and a longer hospital stay, or worse, miss out on necessary treatment.
The management of children started on aciclovir for suspected encephalitis was very variable with the indications for starting aciclovir not always clear. The ideal length of treatment recommended for adults and children with HSV encephalitis is still unknown. Initial trials suggested 10 days , but since then, relapses have been reported, so 14 – 21 days is now recommended [18, 19]. Additionally, many would recommend repeating the LP to ensure that the CSF is negative for HSV by PCR before stopping treatment and this is supported by a European consensus statement and recently published UK guidelines on the management of children with viral encephalitis . In this study, one child with HSV encephalitis was treated for only seven days and the other for 14. Neither case was discussed with either an infectious diseases or neurology specialist. In addition, neither child had a follow up LP to look for viral clearance.
We found an interesting subset of seven children who were not treated at all following a normal LP and/or cranial imaging. It is known that children with both meningitis and encephalitis can have a normal CSF white cell count in the early stages of the illness [20, 21] and in fact two of the children with proven encephalitis in this study had no white cells in the CSF (Table 2). Therefore if a decision is made not to treat a child with enough symptoms or signs to warrant an LP, it would seem sensible to observe them carefully and consider repeating the LP if there is any change in their clinical condition to make sure a CNS infection is not missed. Three of these children were discharged on the day of their LP.
As data was collected retrospectively, it is possible that children with suspected CNS infection in the study time period will not have been included due to incorrect discharge coding or incomplete data to search. One other possible limitation is in the classification of ‘encephalopathy’ (Table 1) as it is recognized that occasionally children with CNS infection may have an acellular CSF in the early stages of their illness . However, none of the children with normal CSF findings had a subsequent clinical course in keeping with a CNS infection. The numbers of patients in the study is small so statistical analysis has been limited and does not allow for comparison between management at the two different types of setting (secondary versus tertiary care).
However, it seems reasonable to conclude that the management of children with suspected CNS infections in our region was very variable and it also seems reasonable to conclude that these findings would be similar in other regions of the UK. Although the percentage of children with suspected CNS infection having an appropriate LP attempted is higher than our previous study, the number of children with a full set of CSF tests requested was disappointing. Over half the children started on aciclovir did not appear to have a clear indication. Of most concern, was the short treatment course for two children with proven HSV type 1 encephalitis. We therefore conclude that national guidelines for the management viral encephalitis are needed.