Fever is a common childhood complaint with 20–40% of parents and carers reporting an episode each year (National Institute for Health and Care Excellence (NICE), 2021). The most common cause of fever is infection (El-Radhi et al, 2009). Febrile seizures (FS) occur in 2–5% of healthy preschool children and recur in approximately 40% of these patients (Whelan et al, 2017; National Institute of Neurological Disorders and Stroke, 2023). Using Whelan et al's (2017) statistics, it can be extrapolated that 76 000–190 000 children may experience a FS in their first 5 years.
Children who have complex FS—defined as such if they last longer than 15 minutes, recur within 24 hours or have associated focal neurology—may carry an increased likelihood of mortality in the subsequent two years and those affected are 3.6–5.6% more likely to develop epilepsy, a cause of mortality for 40–80 children per year (Smith et al, 2019; Royal College of Paediatrics and Child Health, 2013).
The NHS advice for a first FS is to call 999 or attend an emergency department, so it is likely that ambulance crews attend FS with some frequency (Ahmed El-Esrigy et al, 2021; NHS, 2023). As acknowledged by Daneman and McArthur in 1978, children who presented with FS may have been given anticonvulsants; however, treatment started to change around this time.
Thermoregulation maintains the body at the optimal temperature for normal cellular function but can be disrupted by infection and resultant fever (Osilla et al, 2023). The innate immune response stimulates the secretion of pro-inflammatory cytokines and prostaglandins, which act on thermosensitive neurons in the hypothalamus to effectively raise the thermostat of the hypothalamus (Osilla et al, 2023).
These cytokines may also directly affect neuronal excitability at a stage in a child's development where neurons may be in an enhanced excitatory state (Jensen and Sanchez, 2002). This may lower the threshold for seizure activity.
UK ambulance services use Joint Royal Colleges Ambulance Liaison Committee clinical practice guidelines produced by the Association of Ambulance Chief Executives (AACE) (2019) which advise against antipyretics for FS while noting that parental sensitivities should be considered in relation to fever. These are supported by NICE (2021) guidelines on the assessment and initial management of fever in under 5s.
This systematised literature review will analyse the body of evidence linked to the use of paracetamol and ibuprofen versus placebo or no antipyretic medication and explore factors influencing the management of FS.
Methods
A systematised literature review was chosen to critically appraise, consolidate and summarise the evidence base. First described by Grant and Booth (2009), a systematised review includes some of the elements of a systematic review but does not meet its methodological requirements nor usually include exhaustive searching. They are often conducted by a single reviewer.
The question that this review intends to answer is whether conventional antipyretics are effective in preventing the recurrence of febrile seizures. The PICO (participant(s); intervention(s); comparison/control; outcome) framework was used to develop the question (Table 1).
| Framework element | Titular question element | |
|---|---|---|
| P | Population/participant(s) | Children with febrile seizures |
| I | Intervention(s) | Oral paracetamol or ibuprofen |
| C | Comparison/control | No antipyretic medication |
| O | Outcome | Recurrence of febrile seizures |
A search of the literature was conducted using two databases: MEDLINE and CINAHL. MEDLINE is widely renowned for its high concentration of healthcare literature and benefits from Medical Subject Headings controlled vocabulary, which increases the likelihood of relevant results (National Library of Medicine, 2023). CINAHL is also healthcare focused and gives access to more than 1300 academic journals worldwide (Ebsco, 2023). The primary outcome measured was whether a febrile seizure recurred within the same fever episode.
The search terms used were FS and the alternative febrile convulsion, antipyretic and its alternative spelling anti-pyretic, ibuprofen, paracetamol and acetaminophen. Wildcard searches were used to search for plurals of the terms seizure, convulsion, antipyretic and anti-pyretic to capture results with these terms. Combinations of the search terms were used with the Boolean operator AND to return results that included both terms. The inclusion and exclusion criteria are summarised in Table 2.
| Inclusion | Exclusion | Rationale |
|---|---|---|
| Children aged ≤6 years | Children aged ≥7 years | These are the ages at which febrile seizures are prevalent |
| Publication year 2000–2021 | Publication pre 2000 | Studies are more recent and therefore perhaps more generalisable |
| Worldwide publications in English language | Language other than English | Reviewer is English speaking only |
| Publications in an academic journal | Not published in an academic journal or grey literature | Studies meet the standards required of academic journals and should therefore be ethically sound |
Snowball sampling was considered and reference lists were appraised but no studies in these lists met the inclusion and exclusion criteria. One author (VLD) was responsible for assessing whether studies met the inclusion and exclusion criteria.
Results
The literature search was conducted in November 2021 and 34 articles were identified. Sixteen were excluded once inclusion and exclusion limiters had been applied and three articles—Van Esch et al (2000), Strengell et al (2009) and Murata et al (2018)—were found to be suitable for inclusion in this review. Duplicate results were automatically excluded by the database search facility. The process is illustrated in Figure 1.
Two studies, Strengell et al (2009) and Murata et al (2018), were randomised controlled trials (RCTs) and the third, Van Esch et al (2000), compared the results of a previous randomised crossover trial (treatment versus treatment) with an observational study providing a control element. The hierarchy of evidence places RCTs above observational studies so RCTs can be held in higher regard (Wallace et al, 2022).
The studies were carried out in the Netherlands (Van Esch et al, 2000), Finland (Strengell et al, (2009) and Japan (Murata et al, 2018). The healthcare systems in these countries are similar to those of the UK, being either free or subsidised by the government and affordable for most of the population so any results could be considered generalisable to the UK population.
Quality review
All abstracts identified were reviewed by the primary author. A shortlist of abstracts was used to guide full-text reviews, and papers were included if they examined FS and antipyresis.
All included studies were assessed for methodological quality using the Mixed Methods Appraisal Tool (MMAT) developed by Hong et al (2018). MMAT was chosen for its versatility and adaptability to various methodologies, while focusing on the fundamental principles of methodological quality assessment in empirical studies. In line with the guidance provided by the authors of the MMAT, both authors carried out thematic synthesis and concluded the assessment with the same scores.
Following the systematised search and appraisal of the studies, five themes were identified. These were: the effectiveness of antipyretics; the definition of fever; the site at which body temperature should be measured; discrepancy in participant age range; and parental fear of FS.
Discussion
Effectiveness of antipyretics
One aim of this systematised review was to assess the effectiveness of paracetamol or ibuprofen when compared with no intervention in the prevention of FS.
The largest sample in the included studies was that of Murata et al (2018), which strengthens the validity of this study. Participants were followed up for only 24 hours after enrolment, however, so it could be considered that their results are most generalisable to this period only. Though Murata et al's (2018) results are statistically significant, they conflict with current JRCALC guidelines (AACE and JRCALC, 2019).
Van Esch et al (2000) also analysed the risk of FS recurrence by recording temperature before FS and first-degree relative history of FS. Their results are not statistically significant but suggest that antipyretics may be minimally beneficial to participants with a higher risk profile. However, Van Esch et al (2000) followed up the treatment group for a shorter period than the control group. This was corrected for in statistical analysis but could still represent potential bias affecting the results. It is also concerning that participants who were non-compliant with the treatment protocol were not excluded from analyses.
Strengell et al (2009) was the only study to also investigate the effectiveness of diclofenac. Participants in this study were randomised to either diclofenac or placebo and then again to paracetamol, ibuprofen or placebo. The effect of diclofenac on paracetamol and ibuprofen is unclear; however, a more recent study by Sharif et al (2016) found diclofenac to be more effective at antipyresis than paracetamol. The study design makes true synthesis of the results difficult, which potentially weakens the study's reliability and validity. However, the study's use of placebo may strengthen its validity.
Murata et al's (2018) study was the only one to analyse recurrence data stratified by age group. The study found paracetamol was more effective at preventing FS recurrence in the 22–60 month age group, which had a 4.1% recurrence rate compared with a 13.2% recurrence rate in the 6–21 month age group. Clinicians who decide to treat FS with an antipyretic could consider these findings and be aware that treatment might be more successful in those aged ≥22 months. However, no other supporting evidence could be found and this approach is not supported by the JRCALC clinical guidelines (AACE, 2019).
Definition of fever
A limitation of the studies is the discrepancy between the definitions of fever used, which makes the findings difficult to compare and calls into question the studies' validity. The body temperature given as the starting point for interventions varies by 0.5°C between the studies, which presents a question of whether findings would have been different had the same definition of fever been used. Multiple definitions of fever are used in the studies, as in practice.
| Study (year) | Objective | Design | Country | Sample size | Interventions | Outcomes |
|---|---|---|---|---|---|---|
| Van Esch et al (2000) | To compare the efficacy of ibuprofen versus paracetamol at antipyresis in children with febrile seizures when compared with children given no antipyresis | Combined experimental and descriptive patient series | Netherlands | 212 | Paracetamol versus ibuprofen versus no intervention | A 5.4% reduction in febrile seizure recurrence in groups treated with an antipyretic compared to no antipyretic (95% CI (0.2–1.2); P=0.74) |
| Strengell et al (2009) | To compare the efficacy of antipyretics in preventing recurrent febrile seizure | Randomised controlled trial | Finland | 231 | Diclofenac versus placebo; then ibuprofen versus paracetamol versus placebo | Antipyretics are ineffective at preventing febrile seizures and at antipyresis in fever episodes that led to febrile seizures (95% CI (−12.8–17.6); P=0.99) |
| Murata et al (2018) | To assess the efficacy of paracetamol in preventing recurrent febrile seizures | Randomised controlled trial | Japan | 423 | Paracetamol versus no intervention | Paracetamol may prevent recurrent febrile seizures in the same fever episode. Confidence interval 95% (2.3–13.3). P≤0.001 |
NICE (2023a), in a clinical knowledge summary, defines fever in children as ‘an elevation of body temperature above the normal variation’. Although for clarity it gives a figure of 38°C, it recognises that fever can be hard to define. Rather than focusing on body temperature as a diagnostic tool, the NICE guidelines for fever in under fives (2021) direct clinicians to consider body temperature in the context of other signs and symptoms such as respiratory rate, heart rate and urine output. This is supported by the work of El-Radhi et al (2009), who write that the appearance of a child is more important than body temperature alone.
A lack of consensus is apparent when searching literature, although many authors refer to institutions such as the NHS (2020), NICE (2023a), and the American Academy of Pediatrics and Stanford University as using 38°C as the definition of fever (Wallenstein et al, 2013). It is acknowledged that, in all cases, the study authors had to set a temperature as the indicator for starting intervention; failure to do so would likely have resulted in results with very low validity and reliability.
However, evidence-based guidelines and wider policy seem to de-emphasise considering one figure as the start of fever; they encourage the view that body temperature is a spectrum affected by multiple factors and that fever as a standalone symptom is not something to be treated but to be considered in a wider context.
Site of body temperature measurement
Differences between the included studies were found regarding the site used for thermometry.
Van Esch et al (2000) specified that participants' temperatures should be measured rectally but Murata et al (2018) and Strengell et al (2009) failed to specify a site, which calls into question the internal validity of those studies.
All studies involved temperatures being monitored at home by participants' parents and, in the absence of specific instruction, it is possible that parents would have chosen a variety of sites and may not have taken temperatures as accurately as clinicians.
These studies' results may have been impacted by these discrepancies as well as by the fact that the different sites are more or less reflective of true core temperature.
Rectal thermometry has long been considered the most accurate and reflective of true core temperature with oral temperatures lower on average and tympanic and axillar temperatures less accurate (Barbi et al, 2017). However, rectal thermometry is not recommended by current NICE (2021) guidelines as there is evidence to suggest that rectal temperatures are slow to change in response to a change in core temperature and are affected by the presence of stool (Leduc and Woods, 2000).
Moreover, different sites may have variable accuracy and practicality in children of different ages. Rectal thermometry may be considered inappropriate for the emergency clinician who must consider patient dignity, distress and the speed at which thermometry can be obtained.
Age discrepancy
The studies approached age differently, with Murata et al (2018) using an institutional definition to set age ranges for participants. Strengell et al (2009) refer to a textbook definition but then include children younger than this without explanation and Van Esch et al (2000) do not give a rationale for the age range of participants.
The youngest participants enrolled in any of the studies were aged 4 months in Strengell et al (2009), although there is a general consensus among the wider evidence that FS occur from 6 months. For example, this is stated in BMJ Best Practice (2022) and NICE (2018) guidelines, which are the basis for JRCALC guidelines used by ambulance services (AACE, 2019).
Van Esch et al (2000) specified that participants must be aged >10 months but the reason for this is unclear. It was considered that this could be because drug doses used were unsuitable for children aged <10 months; however, the study protocol called for participants to be given doses which are lower than those currently recommended by the British National Formulary for Children (NICE, 2023b; 2023c). This calls into question the study's internal validity; a higher dose may have produced different results.
The upper age limit for inclusion in the studies varied more widely, from 3–5 years, excluding follow-up periods of 2 years in the studies by Strengell et al (2009) and Van Esch et al (2000). Vestergaard and Christensen (2009) conducted a large study over 28 years and found that 93% of first FS occurred by the age of 4 years and all participants had experienced their first FS by the age of 5 years. In this study of 1.6 million participants, none had their first FS when aged over 5 years; this figure excludes participants having a second or subsequent FS after this age.
Ogino et al (2020) specifically studied late FS (i.e. occurring after age 5) and found that 90% of participants had one FS after the age of 5 but 10% had more than one, with a significant decrease in their occurrence after the age of 9. They also found that FS can occur as late as 14 years old. This may indicate that the included studies excluded some potential participants.
NICE (2018) clearly states that FS occurs between 6 months and 5 years, whereas BMJ Best Practice (2022) states they are ‘most common’ between the ages of 6 months and 5 years, while acknowledging that they can occur at other ages. This reflects that FS later in childhood are rare; it may also be the intention of these guidelines to ensure that seizures occurring after the age of 5 are subject to higher scrutiny regarding their aetiology.
Emergency clinicians should consider age, medical history and the history of the presentation when considering whether an FS has occurred and consider differential diagnoses, which may require different treatments.
Parental fear
All included studies acknowledge that FS can be frightening for parents and carers who witness them and incorporate this reasoning into their study rationales.
Interviews conducted on a small sample by Westin and Sund Levander (2018) regarding the parental experience of FS found that parents felt unsure of what was happening or anxious that the child could die. Anxiety during subsequent episodes of fever could persist for up to 2 years and some parents subsequently gave antipyretics in response to any fever episode. Millichap (1999), commenting on two studies, reported fear of FS and/or fever in almost 50% of parents of children after an FS, and found that those with lower levels of anxiety around fever and FS attributed this to information from health professionals. This evidences the impact that practitioner-led parental education can have. Prehospital clinicians, who may be present early in the timeline of FS, can contribute greatly to parental education during and after the emergency and influence how distressing parents find the situation.
There is evidence to suggest a wider issue of parental fever phobia. Purssell (2013) describes six studies that found parental anxiety around fever. The most recent, by Chiappini et al (2012), found that all parents surveyed believed that fever in itself could be harmful.
This study and another by Sakai et al (2012) found that parents consider health professionals their primary source of information about fever, again evidencing the potential impact of practitioner-led parental education.
Moreover, there are potential risks associated with fever phobia. These include giving higher than recommended doses of antipyretics, although this can be balanced against the advantage of increased vigilance (Purssell, 2008). Parental education could inform parents of the risks of accidental overdose as well as the immunological role of fever. Research on the benefits of parental education has focused more on behavioural rather than clinical aspects but it has been found to be effective (Morawska et al, 2015).
The JRCALC guidelines (AACE and JRCALC, 2019) advise that parental sensitivities should be considered when deciding whether to treat a fever with antipyretics. Shared decision-making was found to be essential for family-centred care based on respect and trust by Jordan et al (2020), who also identified little evidence of the perceived barriers to shared decision-making reported by health professionals. The evidence demonstrates that a balance should be struck between allaying parental fear and concern by treating a fever or FS with medication and educating parents as to the benign character of fever and FS.
Limitations and future research
This review is not without limitations.
One author conducted the searches, which may have subjected the review to article selection bias, though this was limited by the use of homogenous search criteria used across the databases.
The search did not identify any studies conducted in the UK; however; the healthcare systems of the countries where the included studies were conducted are similar to those in the UK.
Conducting searches of additional databases and literature not published online may have yielded a higher number of articles for consideration.
Similar, contemporary studies conducted on the current UK population may help to improve understanding of the factors influencing FS management and the effectiveness of antipyretics.
Further research into parental education surrounding febrile seizures may be beneficial.
Conclusion
Two of the included studies (Van Esch et al, 2000; Strengell et al, 2009) found that antipyretics were ineffective or not significantly effective at preventing FS recurrence and the study with the largest sample (Murata et al, 2018) found that paracetamol may be effective at preventing FS recurrence. Therefore, the findings of this review are not unanimous.
The aims and objectives of this systematised review were met but consensus on best practice in the management of FS and the effectiveness of conventional antipyretics in the prevention of recurrence has not been reached so further research is required.