Recruitment

Participants were recruited from a total of 30 sites, all EDs. Although 30 sites were formally opened, two subsequently closed with neither site having recruited any participants. The first participant was recruited from Alder Hey Children’s Hospital, Liverpool, on 22 July 2015 under version 3.0 of the protocol and the last participant was recruited from Southampton General Hospital, Southampton, on 7 April 2018 under version 5.0 of the protocol.

The five top-recruiting sites recruited 124 out of the 286 (43.4%) participants.

The observed site opening and participant recruitment rates closely followed those predicted (Figure 2).

FIGURE 2

Recruitment graph.

Study commencement was delayed by approximately 6 weeks because of contractual issues between the two co-sponsors of the study and, therefore, recruitment closed 6 weeks later than planned, on 10 April 2018.

Participant screening and throughput are summarised in Figure 3.

FIGURE 3

Consolidated Standards of Reporting Trials (CONSORT) flow diagram. ITT, intention to treat.

Of the 1432 children screened for eligibility, 404 were randomised (i.e. the randomisation pack was opened). A total of 1028 children were excluded before randomisation, including 972 children who did not meet the eligibility criteria (see Tables 1 and 2). Reasons for not randomising 53 children who were considered eligible included no trial-trained doctor available, loss of, or failure to, achieve i.v. access, clinical judgement (e.g. child too sick) and treatment given before random allocation.

Inclusion criterion 2 (see Table 1) combined seizure type with the need for second-line treatment. Table 2 looks at this inclusion criteria in greater detail to distinguish eligibility owing to seizure type and ongoing need for second-line treatment.

Of the 404 children randomised, 93 did not require a second-line treatment. Of those children randomised who did require a second-line treatment, 25 were excluded from the analysis as consent was either incompletely documented (n = 6) or declined (n = 19). Table 3 provides the reasons why consent was declined.

TABLE 3

Reasons for consent declined

Compliance with the intervention

Data on compliance (adherence) with treatment are shown in Tables 4 and 5.

TABLE 4

Treatment received by allocation

TABLE 5

Trial compliance data

Baseline characteristics

The baseline characteristics of the study participants were comparable (Table 6). However, the levetiracetam-treated group comprised more participants in whom the presenting episode of CSE [n = 69 (45%) vs. n = 49 (37%)] represented their first convulsive seizure ever, with a lower proportion of participants with a chronic epilepsy, as evidenced by participants taking oral maintenance antiepileptic drugs [n = 51 (34%) vs. n = 55 (41%)].

TABLE 6

Baseline demographic and seizure characteristics of the trial population

Protocol deviations

Protocol deviations were comparable across both treatment groups. Major and minor deviations are shown in Table 7.

TABLE 7

Major and minor protocol deviations

Major protocol deviations comprised:

• premature discontinuation of randomised treatment (none in the levetiracetam-treated group and two in the phenytoin-treated group)
• duration of infusion shorter than expected (none in the levetiracetam-treated group and one in the phenytoin-treated group)
• lower dose of the intervention administered (eight in the levetiracetam-treated group and four in the phenytoin-treated group)
• missing data for primary outcome (two in both the levetiracetam- and phenytoin-treated groups).

Primary outcome

Seizure cessation was achieved in 106 out of the 152 (70%) levetiracetam-treated participants and in 86 out of the 134 (64%) phenytoin-treated participants.

Table 8 provides the median time to seizure cessation from randomisation and Figure 4 shows the Kaplan–Meier curve and log-rank test. As the event of interest (i.e. seizure cessation) is positive, the lower curve indicates a shorter time to seizure cessation; however, there is no statistically significant difference between the treatment arms (log-rank p-value > 0.05).

TABLE 8

Time to cessation of all visible signs of convulsive seizure activity

FIGURE 4

Kaplan–Meier plot time from randomisation to seizure cessation. Product-limit survival estimates with number of subjects at risk. This figure been reused from Lyttle et al. This is an Open Access article distributed in accordance with the terms (more...)

The unadjusted HR was 1.2 (95% CI 0.91 to 1.6; p = 0.2) in favour of levetiracetam. The Schoenfeld residuals for the unadjusted model (p = 0.72) indicated the independency of time and the validity of the proportionality assumption. The Schoenfeld residuals for the adjusted model indicated that the assumption of proportionality for weight was not met (p = 0.05, p-value ranged from 0.27 to 0.71 for other variables). The data were subgrouped according to weight category as per the baseline table (i.e. < 12 kg, 12–36 kg and > 36 kg) and estimates from the adjusted model calculated (see Appendix 6, Tables 15–29). The proportionality assumption within each subgroup of data was supported by the Schoenfeld residuals. Direction of treatment effect was consistent across subgroups, CIs were wide and results were not statistically significant. The treatment effect was increased for children in the > 36 kg subgroup, but remained non-significant and numbers within this group are small.

A range of sensitivity analyses were considered to assess seizure cessation from the start of the infusion, rather than from randomisation (see Appendix 7, Table 32) and the impact censoring regarding RSI and death. The results demonstrated robustness of conclusions and are presented in Appendices 6 and 7.

Secondary outcomes

The study comprised four secondary outcomes. These assessed the efficacy and safety of the two trial interventions (Table 9).

TABLE 9

Secondary outcomes and 14-day follow-up

One participant who was allocated and received phenytoin experienced a SAR. This was profound hypotension, which was considered to be immediately life-threatening and responded to emergency treatment. This participant also experienced a severe unexpected serious adverse reaction (SUSAR), which manifested as a large increase in seizure frequency and marked sedation within 24 hours of receiving phenytoin. The SUSAR was considered medically significant and the participant required admission to the intensive care unit. The SUSAR resolved without complication.

Safety, tolerability and compliance

Two patients died in the study, but neither death was considered related to the randomised treatment. One participant presented to the ED in a generalised tonic–clonic seizure and unconscious. Resuscitation was immediate. The participant received levetiracetam (the randomised treatment) and then phenytoin, followed by RSI with thiopentone (Archimedes, Reading, UK) because of abnormal posturing. The participant died 36 hours following admission and a post-mortem examination revealed severe brain oedema secondary to encephalitis. Consent for recruitment into the study was subsequently obtained from the participant’s carers. The death was considered to be unrelated to the randomised treatment by the principal investigator and chief investigator.

The second participant received phenytoin. The participant died and the results of the post-mortem examination were not available prior to closure to recruitment to the study. The principal investigator and chief investigator considered the death to be unrelated to the randomised treatment. Consent was sought from the participant’s carers but had not been obtained by the time recruitment to the study closed on 10 April 2018 and, therefore, this participant’s data are not included in the analysis.

Adverse events and serious adverse events

A total of 51 AEs were reported in 39 out of the 286 participants. Some participants experienced more than one AE. Forty-one of the AEs were classified as mild, nine as moderate and one as severe. Sixteen out of 130 levetiracetam-treated participants, 18 out of 132 phenytoin-treated participants and four out of 24 participants who received both drugs experienced at least one AE. Each individual AE had a prevalence of < 10%. In the levetiracetam-treated group (20 AEs in 16 participants), a psychiatric AE was reported in 12 participants (agitation in 11 and hallucinations in one). In the phenytoin-treated group (23 AEs in 18 participants), a cardiovascular AE was reported in eight participants, an extravasation/administration site reaction AE in seven (severe in one) and an agitation AE in four. In the group that received levetiracetam and phenytoin (eight AEs in four participants), an extravasation/administration site reaction was reported in three. The full list of reported AEs is shown in Table 10.

TABLE 10

Adverse events

Five SAEs were reported in four participants [including one participant who experienced two SAEs (participant 00133027)]. In three participants, the SAE was considered unrelated to the intervention, and one each was considered to be possible and probable (Table 11).

TABLE 11

Serious adverse events

An additional follow-up questionnaire was completed by sites and the families of participants who had been randomised, treated and consented to take part in the study. The questionnaire was completed 2 weeks following randomisation. Results are shown in Table 12.

TABLE 12

Follow-up questionnaires

Only 74 (25.9%) families completed their 14-day follow-up questionnaire. As documented in the internal meeting minutes, 8 May 2018, details for these questionnaires are not presented within this report because of the low response rate.

Laboratory parameters (haematological, biochemical analysis and urinalysis)

Laboratory analyses in the trial protocol were limited to measurements of blood levels of levetiracetam and phenytoin in the randomised, treated and consented participants. Blood samples were obtained between 1 and 2 hours after completion of infusion of the two treatments. Biochemistry laboratories in each site processed the samples in accordance with their standard operating procedures. The reference ranges for blood levels of levetiracetam were provided by a single central laboratory that analysed all of the samples of levetiracetam-treated participants. The reference ranges for blood levels of phenytoin were provided by the biochemistry department of each participating site that analysed samples of its phenytoin-treated participants.

A total of 192 participants underwent measurement of a blood level [96 participants in the levetiracetam-treated group (63.2%) and 96 participants in the phenytoin-treated group (77.7%)] and the results are shown in Figures 5 and 6.

FIGURE 5

Levetiracetam blood levels.

FIGURE 6

Phenytoin blood levels.