JLE

Epileptic Disorders

MENU

Prescription of emergency antiepileptic medication after a first childhood seizure: analysis of routine administrative data Volume 17, numéro 2, June 2015

Tableaux

When parents witness their child first having a febrile seizure (FS), which affects 2-5% of children under 5 years, many think their child is dying (Fetveit, 2008). Neither prophylactic antipyretics nor diazepam prevent recurrence (Fetveit, 2008). Most seizures are short-lived and long-term neurological, educational and social outcomes are good (Fetveit, 2008).

However, prolonged febrile seizures (PFS), those lasting at least 30 minutes, are associated with increased morbidity (Raspall-Chaure et al., 2006; Fetveit, 2008; Martinos et al., 2012; Yoong et al., 2013). With a 40% rate of recurrence of PFS, it would be reasonable to prescribe emergency antiepileptic drugs (AEDs) for the 0.7% of children whose FS are PFS, but not for all children with FS (Fetveit, 2008). Similarly, prolonged afebrile seizures lasting at least 30 minutes are associated not only with increased morbidity, but also with increased mortality (Raspall-Chaure et al., 2006). In a prospective study of new-onset seizures in children, 12% lasted at least 30 minutes (Shinnar et al., 2001) and the risk of recurrence of such seizures is reported to be 17% (Chin et al., 2006). Thus, it would be reasonable to consider prescribing emergency AEDs for such children.

In France, Germany and the Netherlands, there is routine prescription of emergency AEDs for all children with FS (Nabbout R, Bast T, Brouwera O; personal communication). UK guidelines indicate that emergency AEDs should not be routinely given for FS or for single afebrile seizures (SAS) (SIGN, 2005; NICE, 2012).

Benzodiazepines, the preferred emergency AEDs, can cause sedation and respiratory depression. Benefits of their usage must outweigh possible adverse effects. In Europe, rectal diazepam has been the gold standard emergency AED in children, but there has been diffidence towards its use because of the delivery route (McIntyre et al., 2005). Buccal/nasal midazolam is an alternative due to better effectiveness and social acceptability (Klimach, 2009). Since 2005, buccal midazolam has been included in UK guidelines as a preferred emergency AED for seizures, but guidelines are not always adhered to (Chin et al., 2004). Non-adherence by routine prescribing for FS/SAS could result in substantial cost to the health service.

In this retrospective cross-sectional study, we aimed to test the hypothesis that a high proportion of UK children with FS or SAS are prescribed emergency AEDs, with midazolam being prescribed more often than diazepam.

Methods

Scottish national hospital discharge (SMR01) records were used to identify children aged <5 years, admitted to hospital for a first FS or SAS between April 2009 to March 2012. Admissions for FS were defined as records with a diagnosis, either a principal or associated diagnosis, recorded as ICD10 code R56.0 (Febrile convulsions) and for SAS, ICD10 code R56.8 (Other and unspecified convulsions).

Only admission records for children aged ≥28 days were examined in order to exclude neonatal seizures. A first seizure was defined as no prior seizure-related admission with ICD10 codes R56.0, R56.8, G40 (epilepsy and recurrent seizures), or G41 (status epilepticus). A single seizure was defined as no subsequent seizure-related admission within 12 months of index admission. Rectal diazepam and buccal/nasal midazolam (emergency AED) prescriptions within 12 months of index admissions were identified on the national community prescribing (PIS) database by matching children's Community Health Index (CHI) numbers. CHI is the unique patient identifier used on all health records in Scotland. CHI completeness is very high for SMR01 records and is >90% complete for all community prescriptions from 2009 onwards, our chosen study period. All prescriptions included in the PIS database have an associated Gross Ingredient Cost (GIC) recorded. This is the amount (before any discounts) that a pharmacy can expect to receive after dispensing and reflects NHS price as per the British National Formulary or, for generic products, specified price in the Scottish Drug Tariff. Given that medicines for children were exempt from prescription charges in Scotland before the start of our study period, GIC reflects the cost of the prescriptions to the health service.

All data were analysed by NHS National Services Scotland, Information Services Division (ISD); the organisation managing routine data generated by NHS Scotland. Although it would have been ideal to be able to assess duration of seizures within our analysis, such data are not included in routine data collated by ISD. No individual level/identifiable data were made available to the authors and no new data were generated. The study was an audit of prescription practices for children with first seizures; ethical approval was not required.

Results

Tables 1 and 2 show the key findings. In total, 1,978 children aged <5 years admitted for a first, single FS were identified. FS was most common in the 12-23-month age group. A community prescription record of emergency AEDs within 12 months of admission was reported for 0.9% (18/1978). Total cost of emergency AED medication for FS was £570. Most prescriptions (15/18) were for rectal diazepam; midazolam prescriptions were uncommon. All children who received prescriptions for midazolam were 2 years or older, suggesting a particular preference for midazolam in older children compared to diazepam.

There were 663 children with a first SAS identified. Emergency AED prescriptions within 12 months of admission were reported for 1.7% (11/663); 8/11 (72%) were for rectal diazepam. Total AED cost was £340.

Discussion

There is high level of adherence to current guidance in order to avoid prescribing emergency AEDs for FS/SAS. Our findings are consistent with those of a recent national audit of management of childhood epilepsy that showed throughout the UK, but particularly in Scotland, a high level of adherence to 12 key areas of the guideline (RCPCH, 2012). It could be argued that since 12% of children with SAS have a prolonged seizure (Shinnar et al., 2001), then perhaps more children should be prescribed emergency AEDs rather than the 1.7% observed. However, individual social and medical circumstances, particularly the aetiology and duration of seizures, need to be considered before prescribing. The nature of our study did not allow us access to such data. Further study is needed in this population.

In general, adherence to clinical guidelines is not common. It can, however, be enhanced by clear evidence-based guidelines that are simple (and cheap) to implement in practice, and if the guidelines are supported by integrated networks comprising policy makers, managers, clinicians, health workers and service users (Lugtenberg et al., 2009; Garliardi et al., 2011). These features are evident in the high quality of guidance provided for the management of seizures in children (SIGN, 2005; NICE, 2012) and the training and monitoring activities of the Scottish Paediatric Epilepsy Network (SPEN) set up in 2007 to improve service provision and patient outcomes (National Services Division of NHS, 2010). We hypothesize that our finding of good adherence is at least partly attributable to the work of SPEN.

It was surprising, overall, that more prescriptions were for rectal diazepam even though midazolam is more effective, has fewer side effects, and is more socially acceptable (Fetveit, 2008; Klimach, 2009). Children were exempt from prescription charges; relative AED costs were unlikely to influence parental choice. However, midazolam incurs a substantially higher cost (up to 20 times more) to the NHS than diazepam, so this may have influenced prescribers’ choices. A licensed form of buccal midazolam only became available in 2012. Thus, prior to 2012, doctors may have preferred prescribing the more well-known and licensed rectal diazepam.

Our study demonstrates the utility of using routine administrative data in order to answer clinically relevant questions. However, it is not without its limitations. Only children admitted and assigned a relevant ICD10 diagnostic code were included. We acknowledge that not all children with first afebrile or febrile seizures may be admitted to hospital, despite national UK guidelines (https://www.nice.org.uk/guidance/cg160/chapter/1-recommendations#management-by-the-paediatric-specialist-2). If this were the case, our finding of a low inappropriate prescription rate would be an over-estimate and strengthens the conclusion that inappropriate overprescribing does not represent a major issue or NHS cost. It is possible though that clinicians are being conservative in prescribing emergency AEDS for children with prolonged SAS. Prescriptions dispensed from hospital pharmacies (e.g. given to patients at the point of discharge) are not recorded on the PIS database. However, in view of the shelf life of emergency AEDs, community-initiated repeat prescriptions would have been expected within a year and hence, picked up in our analysis.

Although, overall, CHI completeness on PIS records has been >90% since 2009, CHI completeness varies between specific medications and between areas. CHI completeness for rectal diazepam prescriptions from 2009 is around 55% and for midazolam prescriptions is around 86%. Since most prescriptions were for diazepam, this suggests that subsequent prescribing rates may be up to 2% rather than the 1% presented here for FS, and 3.5% rather than 1.7% for SAS, as prescriptions lacking a CHI number will not have been matched for the analysis. Finally, the PIS database records dispensed prescriptions, thus written prescriptions that were never cashed were excluded.

In summary, NHS Boards and policy makers in Scotland can be reassured that emergency AEDs are not being inappropriately overprescribed for FS and SAS. It would be of interest to compare and contrast results on adherence with other countries where there are nationally agreed guidelines for prescribing emergency AEDs.

Supplementary Data.

The STROBE checklist for this observational study and summary didactic slides are available on the www.epilepticdisorders.com website.

Acknowledgements and disclosures

Laura Marchbank and Colin Daly of the NHS National Services Scotland, Information Services Division undertook all data extraction and preliminary analysis. CJW was supported in this work by NHS Lothian via the Edinburgh Health Services Research Unit. Don Page (also from ISD) provided advice on interpretation of prescribing data and associated costs. Omotomilola Ajetunmobi is funded by a Muir Maxwell Trust PhD studentship. The Muir Maxell Epilepsy Centre receives support from the Muir Maxwell Trust (which had no role in the data collection, preparation of manuscript, or decision to submit).

None of the authors have any conflicts of interest to disclose.