Reevaluation of Meta-analysis on prophylactic drug management for recurrence of febrile seizures

Abstract

Objective To explore the efficiency and safety of drugs to prevent the recurrence of febrile seizures (FS). Methods Relevant literatures were searched via PubMed, EMBASE/SCOPUS, EBSCO-CINAHL, Web of Science, Cochrane Database of Systematic Reviews from December 1997 to November 2014 using the following keywords: febrile seizure OR febrile convulsion, recurrence, prevention OR prophylaxis, medicine OR medication. Publication type was limited to Meta-analysis. Extract the relevant information of Meta-analysis, such as characteristics of objects, types of study design, number of clinical trials, number of cases, search strategies, databases, information of methodology (methods of randomization, concealment, blinding, withdrawal and exit), follow-up time, heterogeneity analysis, subgroup analysis and outcome assessment, etc. Quality of Reporting of Meta-analyses (QUOROM) and Oxman-Guyatt Overview Quality Assessment Questionnaire (OQAQ) were used to assess the quality of included Meta-analyses. Jadad decision was used to assess inclusion and exclusion criteria, search strategies, effectiveness evaluation, data extraction and data analysis, to explore reliable evidence of evidence-based medicine. Results Eventually, four Meta-analyses were included after screening of all the literatures that can be searched out. Among those Meta-analyses, the Meta-analysis of Offringa and Newton (2012) was relatively more reliable. The results suggesed that no clinically important benefits were found in administering intermittent oral or rectal diazepam, oral phenobarbitone, phenytoin, valproate, pyridoxine, buprofen, diclofenac and acetominophen to children with FE. Only one clinical trial reported that intermittent oral clobazam could reduce the recurrence of FE in comparing with placebo at 6-month follow-up ( RR = 0.360, 95% CI: 0.200-0.640; P = 0.000), but it should be verified by more randomized controlled trials (RCTs). Among 4 Meta-analyses included in this study, the Meta-analyses of Rosenbloom (2013), Masuko (2003) and Rantala (1997) did not mention the problems of publication bias; in the Meta-analysis of Offringa and Newton (2012), a funnel plot was applied for evaluation of this problem, however, no quantitative analysis was conducted to evaluate the effect of publication bias on the validity of conclusions and no further processing was mentioned. Conclusions In view of good prognosis of recurrent FE, unsure efficacy of preventive medicine and high risk of adverse drug reactions, it should not be routinely recommended the use of antiepileptic drugs, antipyretic drugs or pyridoxine to prevent the recurrence of FE. DOI: 10.3969/j.issn.1672-6731.2015.08.011