Fármacos anticonvulsivos para el síndrome de Lennox‐Gastaut
Resumen
Antecedentes
El síndrome de Lennox‐Gastaut (LGS) es un síndrome epiléptico específico de la edad que se caracteriza por múltiples tipos de crisis. El LGS tiene un electroencefalograma característico, aparece antes de los ocho años y presenta farmacorresistencia.
Esta es una versión actualizada de la revisión Cochrane publicada en 2013.
Objetivos
Evaluar la eficacia y la tolerabilidad de los fármacos anticonvulsivos (FAC) para el LGS.
Métodos de búsqueda
El 2 de marzo de 2020 se realizaron búsquedas en el Registro Cochrane de estudios (CRS Web) y en MEDLINE (Ovid, 1946 hasta el 28 de febrero de 2020). El CRS Web incluye ensayos controlados aleatorizados (ECA) o cuasialeatorizados del Registro Cochrane central de ensayos controlados (CENTRAL); los Registros especializados de los Grupos de revisión Cochrane, incluido el Grupo Cochrane de Epilepsia (Cochrane Epilepsy Group); PubMed; Embase; ClinicalTrials.gov; y la Plataforma de registros internacionales de ensayos clínicos de la Organización Mundial de la Salud (ICTRP). No se impusieron restricciones de idioma. Se estableció contacto con empresas farmacéuticas y compañeros del área para buscar cualquier estudio no publicado o en curso.
Criterios de selección
Se consideraron los ECA, incluidos los ensayos cruzados (cross‐over), de FAC para el LGS en niños y adultos. Se incluyeron estudios de los FAC utilizados como monoterapia o como tratamiento complementario (adyuvante). Se excluyeron los estudios que compararon diferentes dosis del mismo FAC.
Obtención y análisis de los datos
Se utilizaron los procedimientos metodológicos estándar de Cochrane, que incluyen la evaluación doble e independiente del riesgo de sesgo y la aplicación del método GRADE para calificar la certeza de la evidencia de los desenlaces.
Resultados principales
No se encontraron ensayos de monoterapia con FAC. La revisión incluyó 11 ensayos (1277 participantes; aproximadamente 11 semanas a 112 semanas de seguimiento después de la asignación al azar) que utilizaron FAC complementarios para el LGS en niños, adolescentes y adultos.
Dos estudios compararon el cannabidiol complementario (dos dosis) con el placebo complementario en niños, adolescentes y adultos. No se proporcionó suficiente información para calcular las diferentes proporciones de la tasa de respuesta en todos las crisis. Se encontró evidencia de certeza alta de que 82 personas más por cada 1000 (intervalo de confianza [IC]: 19 más a 350 más) presentaron eventos adversos (EA) que provocaron la interrupción del estudio con cannabidiol complementario, en comparación con placebo complementario (dos estudios; 396 participantes; razón de riesgos [RR] 6,62; IC del 95%: 1,56 a 28,15).
Un estudio comparó la cinromida complementaria con el placebo complementario sólo en niños y adolescentes. Se encontró evidencia de certeza muy baja de que 35 personas más por cada 1000 (IC: 123 menos a 434 más) tuvieron una reducción del 50% o más en el promedio de las convulsiones generales con la cinromida complementaria en comparación con el placebo complementario (un estudio; 56 participantes; RR 1,15; IC del 95%: 0,47 a 2,86). Este estudio no informó sobre los participantes con EA que provocaron la interrupción del estudio.
Un estudio comparó el clobazam complementario (tres dosis) con el placebo complementario. Este estudio no informó sobre el cese ni la reducción general de las crisis. Se encontró evidencia de certeza alta de que 106 personas más por cada 1000 (IC: 0 más a 538 más) presentaron EA que provocaron la interrupción del estudio con clobazam complementario en comparación con el placebo complementario (un estudio; 238 participantes; RR 4,12; IC del 95%: 1,01 a 16,87).
Un estudio comparó el felbamato complementario con el placebo complementario. No hubo casos de cese de las crisis con ninguno de los dos regímenes durante la fase de tratamiento (un estudio; 73 participantes; evidencia de certeza baja). Hubo evidencia de certeza baja de que 53 personas más por cada 1000 (IC: 19 menos a 716 más) con felbamato complementario no presentaron crisis durante un registro de EEG al final de la fase de tratamiento, en comparación con el placebo complementario (RR 2,92; IC del 95%: 0,32 a 26,77). El estudio no informó sobre la reducción general de las crisis. Se encontró evidencia de certeza baja de que una persona menos por cada 1000 (IC: 26 menos a 388 más) con el felbamato complementario presentó EA que provocaron la interrupción del estudio en comparación con el placebo complementario (un estudio, 73 participantes; RR 0,97; IC del 95%: 0,06 a 14,97).
Dos estudios compararon la lamotrigina complementaria con el placebo complementario. Ningún estudio informó sobre el cese general de las crisis. Se encontró evidencia de certeza alta de que 176 personas más por cada 1000 (IC: 30 más a 434 más) tuvieron una reducción ≥ 50% en el promedio de las convulsiones con la lamotrigina complementaria en comparación con el placebo complementario (un estudio; 167 participantes; RR 2,12; IC del 95%: 1,19 a 3,76). Se encontró evidencia de certeza baja de que 40 personas menos por cada 1000 (IC: 68 menos a 64 más) presentaron EA que provocaron la interrupción del estudio con la lamotrigina complementaria en comparación con el placebo complementario (un estudio; 169 participantes; RR 0,49; IC del 95%: 0,13 a 1,82).
Dos estudios compararon la rufinamida complementaria con el placebo complementario. Ningún estudio informó sobre el cese de las crisis. Se encontró evidencia de certeza alta de que 202 personas más por cada 1000 (IC: 34 a 567 más) tuvieron una reducción ≥ 50% en el promedio de las crisis (un estudio; 138 participantes; RR 2,84; IC del 95%: 1,31 a 6,18). Se encontró evidencia de certeza baja de que 105 personas más por cada 1000 (IC: 17 menos a 967 más) presentaron EA que provocaron la interrupción del estudio con la rufinamida complementaria en comparación con el placebo complementario (un estudio; 59 participantes; RR 4,14; IC del 95%: 0,49 a 34,86). Un estudio comparó la rufinamida complementaria con otro FAC complementario. Este estudio no informó sobre el cese ni la reducción general de las crisis. Se encontró evidencia de certeza baja de que tres personas menos por cada 1000 (IC: 75 menos a 715 más) presentaron EA que provocaron la interrupción del estudio con la rufinamida complementaria en comparación con otro FAC complementario (un estudio; 37 participantes; RR 0,96; IC del 95%: 0,10 a 9,57).
Un estudio comparó el topiramato complementario con el placebo complementario. Este estudio no informó sobre el cese general de las crisis. Se encontró evidencia de certeza baja de una reducción ≥ 75% en el promedio de las crisis con el topiramato complementario (un estudio; 98 participantes; odds ratio de Peto [OR de Peto] 8,22; IC del 99%: 0,60 a 112,62) y poca o ninguna diferencia en los EA que provocaran la interrupción del estudio en comparación con el placebo complementario; ningún participante presentó EA que provocaran la interrupción del estudio (un estudio; 98 participantes; evidencia de certeza baja).
Conclusiones de los autores
En la actualidad se carece de ECA de monoterapia y de comparación directa de los FAC complementarios. Sin embargo, se encontró evidencia de certeza alta de la reducción general de las crisis con lamotrigina y rufinamida complementarias, con evidencia de certeza baja en el caso de los EA que provocan la interrupción del estudio en comparación con el placebo u otro FAC complementario. La evidencia para otros FAC complementarios en el cese o la reducción general de las crisis fue de certeza baja a muy baja, con evidencia de alta a baja certeza en el caso de los EA que provocan la interrupción del estudio.
Los estudios futuros deben considerar informar acerca de los desenlaces de reducción general de las crisis (aplicando dispositivos de detección automática de crisis), el impacto en el desarrollo, la cognición y el comportamiento; también deben investigar la eficacia específica de los FAC en función de la edad y dirigirse a las etiologías subyacentes.
PICO
Resumen en términos sencillos
¿Son los medicamentos anticonvulsivos tratamientos eficaces y seguros para el síndrome de Lennox‐Gastaut?
¿Por qué es importante esta pregunta?
El síndrome de Lennox‐Gastaut (LGS) es un tipo grave de epilepsia que afecta principalmente a los niños. El síntoma principal del LGS es la frecuencia y los múltiples tipos de crisis. Las crisis están causadas por picos repentinos e incontrolados de actividad eléctrica anómala en el cerebro. Las crisis son difíciles de tratar con medicamentos anticonvulsivos. Para intentar detener las crisis se administran muchos medicamentos anticonvulsivos diferentes. A menudo se administran dos o tres medicamentos anticonvulsivos al mismo tiempo, lo que se conoce como polifarmacia. No está claro qué medicamentos son más eficaces. La mayoría de las personas con LGS también presentan dificultades de aprendizaje y de conducta.
¿Cómo se identificó y evaluó la evidencia?
Se buscó en la literatura médica los ensayos controlados aleatorizados (ECA) que analizaran los efectos de los medicamentos anticonvulsivos para el tratamiento del LGS. Se incluyó cualquier ECA que comparara los medicamentos anticonvulsivos, ya sea como monoterapia o como tratamiento complementario (adyuvante), con placebo (tratamiento simulado), ningún tratamiento u otro tipo de tratamiento. A continuación se compararon los resultados de los ECA encontrados y se resumió la evidencia de todos los estudios. La confianza en la "certeza" de la evidencia se calificó sobre la base de factores como los métodos y el tamaño de los estudios, así como la consistencia de los hallazgos entre ellos.
Características de los estudios
Esta revisión incluyó 11 ensayos (1277 participantes, que incluyeron niños, adolescentes y adultos). Los ensayos duraron entre unas 11 semanas y 112 semanas después de la asignación al azar. Ninguno de los ensayos incluidos comparó un medicamento anticonvulsivo por sí solo con otro tratamiento. Dos ensayos compararon regímenes de cannabidiol (medicina con cannabis) con regímenes de placebo (396 niños, adolescentes y adultos). Un ensayo comparó un régimen de cinromida complementaria con un régimen de placebo complementario (56 niños y adolescentes solamente). Un ensayo comparó un régimen de clobazam complementario con un régimen de placebo complementario (238 participantes). Un ensayo comparó un régimen de felbamato complementario con un régimen de placebo complementario (73 participantes). Dos ensayos compararon regímenes de lamotrigina complementaria con regímenes de placebo complementario (186 participantes). Dos ensayos compararon regímenes de rufinamida complementaria con regímenes de placebo complementario (197 participantes). Un ensayo comparó un régimen de rufinamida complementaria con un régimen de otro medicamento anticonvulsivo (37 participantes). Un ensayo comparó un régimen de topiramato complementario con un régimen de placebo complementario (98 participantes).
La mayoría de la evidencia de esta revisión estaba relacionada con personas de países de ingresos medios o altos y, cuando se informó, con participantes de etnia blanca.
Resultados y certeza de la evidencia
Se encontró evidencia de certeza alta de que la lamotrigina complementaria aumentó el número de participantes con al menos una reducción del 50% en el número promedio de crisis notificadas. También se encontró evidencia de certeza baja de que la lamotrigina complementaria podría haber reducido el número de participantes con eventos adversos que provocaron la interrupción del estudio en comparación con el placebo complementario.
Se encontró evidencia de certeza alta de que la rufinamida complementaria aumentó el número de participantes con al menos una reducción del 50% en el número promedio de crisis notificadas, en comparación con el placebo complementario. También se encontró evidencia de certeza baja de que la rufinamida complementaria, en comparación con el placebo complementario u otro medicamento anticonvulsivo no especificado, podría dar lugar a poca o ninguna diferencia en el efecto sobre la reducción del número de participantes con eventos adversos que provocan la interrupción del estudio.
El topiramato complementario podría haber aumentado el número de participantes con al menos una reducción del 75% en el número promedio de crisis notificadas, y probablemente dio lugar a poca o ninguna diferencia en el número de eventos adversos que provocaron la interrupción del estudio, en comparación con el placebo complementario (evidencia de certeza baja).
El felbamato complementario (fase de tratamiento) podría haber dado lugar a poca o ninguna diferencia en cuanto a la notificación de ausencia de crisis y a los episodios adversos que provocaron la interrupción del estudio en comparación con el placebo complementario (evidencia de certeza baja). Sin embargo, se encontró que cuando las convulsiones se registraron en un ámbito de investigación, el felbamato complementario podría dar lugar a un aumento en la ausencia de convulsiones en comparación con el placebo complementario (evidencia de certeza baja).
No se sabe con certeza si otros tratamientos farmacológicos complementarios, incluidos el cannabidiol, la cinromida y el clobazam, redujeron todos los tipos de crisis, porque este desenlace no se informó o tenía evidencia de certeza muy baja. Se encontró evidencia de certeza alta de que el cannabidiol y el clobazam complementarios aumentaron el número de participantes con eventos adversos que provocaron la interrupción del estudio, en comparación con el placebo complementario. No se encontró evidencia de eventos adversos que provocaran la interrupción del estudio en la comparación de cinromida complementaria con placebo complementario.
La evidencia está actualizada hasta marzo de 2020.
Authors' conclusions
Summary of findings
| Cannabidiol (10 mg/kg and 20 mg/kg) + ASMs compared to placebo + ASMs | ||||||
| Patient or population: Lennox‐Gastaut syndrome, any age | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo + ASMs | Risk with cannabidiol + ASMs | |||||
| Number of participants free from all seizures follow up: after 14 weeks' treatment (titration and maintenance) | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. No participants were free from drop seizures (one study; 225 participants) Analysis 1.1 |
| Number of participants with ≥ 75% reduction in all seizures follow‐up: after 14 weeks' treatment (titration and maintenance) | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. Number of participants with ≥ 75% reduction in drop seizures: RR 3.51 (95% CI 1.24 to 9.92) in favour of the cannabidiol regimen (two studies; 396 participants) Analysis 1.2. |
| Number of participants with ≥ 50% reduction in all seizures follow‐up: after 14 weeks' treatment (titration and maintenance) | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. Number of participants with ≥ 50% reduction in drop seizures: RR 2.12 (95% CI 1.48 to 3.03) in favour of the cannabidiol regimen (two studies; 396 participants; Analysis 1.3). |
| Number of participants with adverse events leading to study discontinuation | Study population | RR 6.62 | 396 | ⊕⊕⊕⊕ | Intervention in Devinsky 2018 involved two doses of cannabidiol (10 mg/kg and 20 mg/kg); 6 of the 7 adverse events leading to study discontinuation occurred in the higher dose group; intervention in Thiele 2018 also involved the higher dose of cannabidiol (20 mg/kg). | |
| 12 per 1000 | 82 per 1000 | |||||
| *The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| Cinromide + ASMs compared to placebo + ASMs | ||||||
| Patient or population: Lennox‐Gastaut syndrome, any age1 | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo + ASMs | Risk with cinromide + ASMs | |||||
| Number of participants free from all seizures | Study population | not estimable | 56 | ⊕⊝⊝⊝ | ||
| 0 per 1000 | 0 per 1000 | |||||
| Number of participants with ≥ 75% reduction in mean weekly seizures | Study population | Peto OR 9.35 | 56 | ⊕⊝⊝⊝ | Cinromide + ASMs: 3/26 participants; placebo + ASMs: 0/30 participants. | |
| see comment | see comment | |||||
| Number of participants with ≥ 50% reduction in mean weekly seizures | Study population | RR 1.15 | 56 | ⊕⊝⊝⊝ | ||
| 233 per 1000 | 268 per 1000 | |||||
| Number of participants with adverse events leading to study discontinuation follow‐up: after 18 weeks | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. |
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1The evidence for this comparison included children and adolescents only. 2Downgraded twice for study limitations because there was a high risk of bias from incomplete data (study terminated prematurely). 3Downgraded twice for imprecision because the study was not powered to detect a between‐group difference in zero event outcomes. 4Downgraded twice for imprecision because the effect estimate has a very wide confidence interval. | ||||||
| Clobazam (low, medium and high doses) + ASMs compared to placebo + ASMs | ||||||
| Patient or population: Lennox‐Gastaut syndrome, any age | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo + ASMs | Risk with clobazam + ASMs | |||||
| Number of participants free from all seizures follow‐up: "from the 4‐week baseline period to the 12‐week maintenance period" (with 3 weeks titration) | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. Number of participants free from drop seizures: RR 4.10 (95% CI 1.00 to 16.83) in favour of the clobazam regimen (1 study; 217 participants) Analysis 3.1. |
| Number of participants with ≥ 75% reduction in all seizures follow‐up: "from the 4‐week baseline period to the 12‐week maintenance period" (with 3 weeks titration) | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. |
| Number of participants with ≥ 50% reduction in all seizures follow‐up: "from the 4‐week baseline period to the 12‐week maintenance period" (with 3 weeks titration) | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. |
| Number of participants with adverse events leading to study discontinuation | Study population | RR 4.12 | 238 | ⊕⊕⊕⊕ | Intervention in Ng 2011 involved three doses of clobazam (low, medium, high) and study authors state that "A dosage related trend was observed for the overall incidence of [adverse events] leading to discontinuation." | |
| 34 per 1000 | 140 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| Felbamate + ASMs compared to placebo + ASMs | ||||||
| Patient or population: Lennox‐Gastaut syndrome, any age | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo + ASMs | Risk with felbamate + ASMs | |||||
| Number of participants free from all seizures (recorded by closed‐circuit television and electroencephalography) ‐ | Study population | RR 2.92 | 73 | ⊕⊕⊝⊝ | ||
| 28 per 1000 | 81 per 1000 | |||||
| Number of participants free from all seizures | Study population | not estimable | 73 | ⊕⊕⊝⊝ | ||
| 0 per 1000 | 0 per 1000 | |||||
| Number of participants with ≥ 75% reduction in all seizures follow‐up: after the treatment phase, which "consisted of a 14‐day titration period and a 56‐day maintenance period" | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. |
| Number of participants with ≥ 50% reduction in all seizures follow‐up: after the treatment phase, which "consisted of a 14‐day titration period and a 56‐day maintenance period" | see comment | see comment | not estimable | see comment | see comment | According to a retrospective analysis "Approximately 50% of patients randomised to FBM obtained at least a 50% reduction in seizure frequency compared with about 15% receiving placebo." |
| Number of participants with adverse events leading to study discontinuation | Study population | RR 0.97 | 73 | ⊕⊕⊝⊝ | ||
| 28 per 1000 | 27 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1Downgraded twice for imprecision because the effect estimate has a very wide confidence interval. 2Downgraded twice for imprecision because the study was not powered to detect a between‐group difference in zero event outcomes. | ||||||
| Lamotrigine + ASMs compared to placebo + ASMs | ||||||
| Patient or population: Lennox‐Gastaut syndrome, any age | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo + ASMs | Risk with lamotrigine + ASMs | |||||
| Number of participants free from all seizures follow‐up: after 16 weeks' treatment | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. |
| Number of participants with ≥ 75% reduction in all seizures follow‐up: after 16 weeks' treatment | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. |
| Number of participants with ≥ 50% median reduction in all seizures | Study population | RR 2.12 | 167 | ⊕⊕⊕⊕ | ||
| 157 per 1000 | 333 per 1000 | |||||
| Number of participants with adverse events leading to study discontinuation | Study population | RR 0.49 | 169 | ⊕⊕⊝⊝ | ||
| 78 per 1000 | 38 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1Downgraded twice because the effect estimate has a very wide confidence interval. | ||||||
| Rufinamide + ASMs compared to placebo + ASMs | ||||||
| Patient or population: Lennox‐Gastaut syndrome, any age | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo + ASMs | Risk with rufinamide + ASMs | |||||
| Number of participants free from all seizures follow‐up: after 84 days/12 weeks' treatment (titration and maintenance) | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. |
| Number of participants with ≥ 75% reduction in all seizures follow‐up: after 84 days/12 weeks' treatment (titration and maintenance) | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. |
| Number of participants with ≥ 50% reduction in all seizures | Study population | RR 2.84 | 138 | ⊕⊕⊕⊕ | ||
| 109 per 1000 | 311 per 1000 | |||||
| Number of participants with adverse events leading to study discontinuation follow‐up: after 12 weeks | 33 per 1000 | 138 per 1000 (16 to 1,000) | RR 4.14 (0.49 to 34.86) | 59 (1 RCT) | ⊕⊕⊝⊝ | |
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1Downgraded twice because the effect estimate has a very wide confidence interval. | ||||||
| Rufinamide + ASMs compared to other ASM + ASMs | ||||||
| Patient or population: Lennox‐Gastaut syndrome, any age | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo + ASMs | Risk with rufinamide + ASMs | |||||
| Number of participants free from all seizures follow‐up: after 106 weeks' treatment (titration and maintenance) | see comments | see comments | not estimable | see comments | see comments | No studies measured this outcome. |
| Number of participants with ≥ 75% reduction in all seizures follow‐up: after 106 weeks' treatment (titration and maintenance) | see comments | see comments | not estimable | see comments | see comments | No studies measured this outcome. |
| Number of participants with ≥ 50% reduction in all seizures follow‐up: after 106 weeks' treatment (titration and maintenance) | see comments | see comments | not estimable | see comments | see comments | No studies measured this outcome. |
| Number of participants with treatment‐emergent adverse events leading to study discontinuation | Study population | RR 0.96 | 37 | ⊕⊕⊝⊝ | ||
| 83 per 1000 | 80 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1Downgraded twice because the effect estimate has a very wide confidence interval. | ||||||
| Topiramate + ASMs compared to placebo + ASMs | ||||||
| Patient or population: Lennox‐Gastaut syndrome, any age | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo + ASMs | Risk with Topiramate+ ASMs | |||||
| Number of participants free from all seizures follow‐up: after 11 weeks' treatment (titration and maintenance) | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. |
| Number of participants with ≥ 75% reduction in all seizures | Study population | Peto OR 8.22 | 98 | ⊕⊕⊝⊝ | Topiramate + ASMs: 4/48 participants; placebo + ASMs: 0/50 participants. | |
| see comment | see comment | |||||
| Number of participants with ≥ 50% reduction in all seizures follow‐up: after 11 weeks' treatment (titration and maintenance) | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. |
| Number of participants with adverse events leading to study discontinuation | Study population | not estimable | 98 | ⊕⊕⊝⊝ | ||
| 0 per 1000 | 0 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1Downgraded twice because the effect estimate has a very wide confidence interval. 2Downgraded twice because the study was not powered to detect a between‐group difference in zero event outcomes. | ||||||
Background
Description of the condition
This is an updated version of the Cochrane Review published in 2013 (Hancock 2013).
Lennox‐Gastaut Syndrome (LGS) is a severe chronic epilepsy syndrome with onset in early childhood (seizure onset between one and seven years of age, with a peak age of onset at three to five years) (ILAE 1989; ILAE 2020). The disorder commonly persists, with drug‐resistant seizures throughout life (Arzimanoglou 2009).
People with LGS often have ongoing care needs into adulthood because of significant comorbidities, including cognitive, behavioural, and motor impairments. Along with the drug‐resistant seizures, associated severe behavioural and psychiatric disorders can be particularly challenging to manage clinically, as well as by the person living with LGS, their families and caregivers. In addition to the consequences of the underlying aetiology, LGS is regarded as a developmental and epileptic encephalopathy, a concept that refers to the adverse and disruptive impact of frequent epileptiform activity on cerebral activity, even in the absence of clinical seizures.
LGS represents approximately 3% to 5% of all childhood onset epilepsies. In population‐based childhood epilepsy incidence cohorts, only up to 0.9% of patients are identified with LGS when epilepsy is diagnosed; other defining characteristics such as multiple seizure types and electroencephalogram (EEG) features evolve over time (Berg 1999; Berg 2018; Callenbach 1998; Wirrell 2011). The prevalence of LGS was estimated at 26 per 100,000 in a regional US study (Trevathan 1997). This syndrome presents with multiple seizure types and in some people, follows initial presentation with infantile spasms. For three in ten people, LGS may evolve following the initial presentation in infancy with West syndrome (infantile spasms) or Ohtahara syndrome (Cross 2017). However, there are also cases of late‐onset LGS with seizure onset after the age of 10 years, in adolescence and adulthood (Smith 2018).
Underlying aetiologies, identified in 60% to 75% of patients with LGS, are diverse, and include developmental structural brain abnormalities, chromosomal derangements, monogenetic conditions, and less frequently, metabolic disorders and acquired brain insults (e.g. perinatal hypoxic‐ischaemic brain injuries, perinatal CNS infections) (Asadi‐Pooya 2018). The diagnosis of LGS in early childhood can be challenging due to the overlap in clinical presentation with other epileptic syndromes, especially epilepsy with myoclonic‐atonic seizures’ (Cross 2017; Eschbach 2018; Kaminska 1999).
The occurrence of tonic seizures is mandatory for the diagnosis of LGS (ILAE 2020). While atypical absence seizures are the second most characteristic seizure type in LGS, other observed seizure types also include generalised tonic‐clonic, atonic, myoclonic, myoclonic‐atonic, focal seizures, and epileptic spasms (Crespel 2019). The EEG features that form part of the fully evolved electro‐clinical syndrome are slow spike‐wave discharges (< 2.5 Hz) and generalised paroxysmal fast activity in slow wave sleep. Prolonged periods of obtundation or episodes of non‐convulsive status epilepticus are other common epilepsy manifestations interfering with cognitive and developmental function. These episodes and injuries sustained with seizures can have a negative impact on the quality of life as well as increase mortality risk (Autry 2010; Berg 2018).
Description of the intervention
Eligible interventions included any type of anti‐seizure medication (ASM) as either monotherapy or add‐on (adjunctive) therapy.
How the intervention might work
ASMs provide symptomatic treatment with the intent to suppress seizure generation. To date, there is no evidence that ASMs can achieve disease modification and prevent the development of drug‐resistant seizures. Cellular targets for ASMs, in general, include voltage‐gated ion channels, receptors enhancing GABA inhibition or inhibit excitation mediated by glutamate receptors (Rogawski 2016). The anti‐seizure effects of cannabidiol are proposed to be mediated by interaction with diverse molecular targets including G protein‐coupled receptor‐55 (GPR55‐antagonist), transient receptor potential vanilloid 1 (TRPV1‐ activation) channels and adenosine 2A2 receptors (Alves 2020, Lattanzi 2020). Cinromide (3 brono‐N‐ethylcinnamide), an experimental agent, showed its anti‐seizure effect in animal models, through a mechanism of action that is not well understood (Group for the Evaluation of Cinromide 1989). Felbamate (FBM) inhibits glycine‐enhanced N‐methyl‐D‐aspartate (NMDA)‐induced intracellular calcium currents, and at high concentrations, potentiates GABA responses and inhibits excitatory NMDA responses (Shorvon 2010). Clobazam enhances the inhibitory effects of GABA, binding to benzodiazepine receptors at the GABAA ligand‐gated chloride channel complex and boosting chloride conductance through GABA‐regulated channels (Trinka 2015, Brigo 2021). Lamotrigine acts as a use‐dependent blocker of voltage‐sensitive sodium channels, interacts with the open‐channel conformation of voltage‐sensitive sodium channels, interacts at a specific site of the alpha pore‐forming subunit of voltage‐sensitive sodium channels, and inhibits the release of glutamate (Shorvon 2010). Rufinamide acts as a blocker of voltage‐sensitive sodium channels and prevents sodium channels from returning to an activated state, thereby preventing the generation of sustained bursts of high‐frequency action potentials (Shorvon 2010). Topiramate has multiple mechanisms of action: it enhances GABA‐mediated inhibition, inhibits voltage‐dependent sodium channels, enhances potassium channel conduction; it also inhibits L‐type high voltage‐activated calcium channels, decreases glutamate‐mediated excitatory neurotransmission, and can inhibit carbonic anhydrase (Shorvon 2010).
Why it is important to do this review
LGS is complex and one of the most medically refractory epilepsy syndromes. It impacts on individuals' learning, development and mental well‐being, as well as on the psychosocial and socio‐economic situation of families and caregivers. Furthermore, there is evidence to suggest that healthcare and medical treatment costs are higher for people with LGS compared to other types of epilepsy (Pina‐Garza 2017).
Objectives
To assess the efficacy and tolerability of anti‐seizure medications (ASMs) for LGS.
Methods
Criteria for considering studies for this review
Types of studies
We considered randomised controlled trials (RCTs), including cross‐over trials, of ASMs for LGS.
Types of participants
We included children and adults with a diagnosis of LGS. We did not apply age restrictions. As an electrochemical syndrome associated with specific types of epileptic seizures and a characteristic EEG pattern, the diagnosis of LGS was based on clinical criteria. These diagnostic criteria included but were not limited to those provided by the International League Against Epilepsy (ILAE 1989).
Types of interventions
We included any trial that compared ASMs (monotherapy or add‐on therapy) with placebo, no therapy or another therapy. We excluded trials evaluating ketogenic diet, vagus nerve stimulation or other non‐pharmaceutical treatments (including homeopathy or acupuncture) unless they were provided as a co‐intervention with ASMs. In this review update, we also excluded studies comparing different doses of the same drug, which is different to the previous version of this review, although dose comparisons were not pre‐specified in the methods of the review protocol.
Types of outcome measures
If a study reported usable continuous and dichotomous data for quantitative reduction in seizures or other outcomes, we would have reported the dichotomous data as number of participants with a quantitative reduction in the outcome. We included outcome measures reported at any time point. If a study reported multiple time points, we planned to report and meta‐analyse the outcome at the longest follow‐up. However, we found that outcomes reported phases of dose adjustment before and, or after dose maintenance; we retrospectively agreed to prioritise the reporting of results for the treatment period as described in studies.
Primary outcomes
-
Cessation of all seizures (defined as total cessation of all seizure types within the trial period)
-
Quantitative reduction of all seizure types (measured as the number of all seizures occurring before treatment was commenced compared with the total number of seizures occurring at the end of the trial period)
-
Adverse events leading to study discontinuation
Secondary outcomes
-
Quantitative reduction in the following types of seizures (measured as the number of seizures occurring before treatment was commenced compared with the number occurring at the end of the trial period): absence seizures; tonic seizures; atonic seizures; drop seizures (defined as any seizure type resulting in postural loss); myoclonic seizures; tonic‐clonic seizures; and focal onset seizures
-
Death (i.e. alive/deceased)
-
Any adverse events
Search methods for identification of studies
Electronic searches
Searches were run for the original review in March 2003. Subsequent searches were run in February 2009, March 2011, April 2012, October 2012, July 2014, October 2016, and September 2018. For the latest update, we searched the following databases on 2 March 2020. There were no language restrictions.
-
Cochrane Register of Studies (CRS Web), using the search strategy shown in Appendix 1.
-
MEDLINE (Ovid), 1946 to 28 February 2020, using the search strategy shown in Appendix 2.
CRS Web includes RCTs or quasi‐RCTs from the Cochrane Central Register of Controlled Trials (CENTRAL); the Specialised Registers of Cochrane Review Groups, including Cochrane Epilepsy; PubMed; Embase; ClinicalTrials.gov; and the World Health Organization's International Clinical Trials Registry Platform (ICTRP).
Searching other resources
One review author (FB) contacted pharmaceutical companies (Eisai, GW Pharmaceuticals, and UCB Pharma) and colleagues in the field on 25 November 2020 in an effort to identify unpublished data.
Data collection and analysis
At least two review authors (from among KJ, FB, CE, and SM) extracted data and resolved any discrepancies by discussion.
Selection of studies
Two review authors (KJ and FB) examined records identified by the search strategy for studies eligible for inclusion. The review authors independently confirmed that studies were RCTs of drug treatment for LGS.
Data extraction and management
At least two review authors (from among KJ, FB, CE, and SM) independently performed data extraction using a specially designed data extraction form. FB and KJ checked and entered data into the Cochrane authoring and statistical software, Review Manager 5 (Review Manager 2020); at least one other review author (CE or SM) checked the data entry.
For each trial, we sought the following information.
Participants
-
Inclusion criteria
-
Exclusion criteria
-
Total number randomised
-
Baseline imbalances
-
Withdrawals and exclusions
-
Age at onset
-
Age at diagnosis
-
Age at start of treatment
-
Sex
-
Race/ethnicity
-
Type of seizures
-
Seizure frequency during the baseline period
-
Number of background drugs
-
Co‐morbidities
Interventions
-
Type of intervention (description, including dose, frequency and route of administration)
-
Number of participants randomised to each intervention
-
Duration of baseline period
-
Duration of treatment period
-
Co‐interventions (if any)
-
Compliance
Outcome measures
For each outcome reported in the included studies we extracted the following data:
-
Outcome name and definition
-
Time point when each outcome was measured
-
Time point when each outcome was reported
-
Validation of the outcome (yes/no/unclear)
-
Imputation of missing data (e.g. assumptions made for intention‐to‐treat (ITT) analysis)
-
Assumed risk estimate (e.g. baseline or population risk noted in the Background)
-
Statistical power (e.g. power and sample size calculation, level of power achieved)
Assessment of risk of bias in included studies
At least two review authors (from among KJ, CE, and FB) independently assessed risk of bias for each of the included trials using the Cochrane 'Risk of bias' tool (Higgins 2011). We assessed risk of bias based on sequence generation, concealment of allocation, methods of blinding, incomplete outcome data, selective reporting, and other types of bias. For each of these categories, the review authors judged the domain to be at 'low', 'high', or 'unclear' risk of bias. We resolved any disagreements by discussion.
Measures of treatment effect
We analysed dichotomous data with a risk ratio (RR) and a 95% confidence interval (CI) or using the Peto odds ratio (Peto OR) and a 99% CI if there were less than 1% events. We analysed continuous data using the mean difference (MD) with a 95% CI where we found data provided as means and standard deviations (SDs).
Unit of analysis issues
For any unit of analysis issues, we planned to deal with them using the guidance in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2020). We planned to analyse randomised cross‐over studies in meta‐analyses, using the results from paired analyses of the first period only to account for carry‐over effect with sequential intervention (Higgins 2020).
For studies in which different doses of the same drug were reported separately, we combined these data into a single treatment group to avoid a duplicative error with multiple‐armed trials.
Dealing with missing data
For this review update, one review author (FB) attempted to contact the study authors of Arzimanoglou 2019 to obtain missing outcome data.
Assessment of heterogeneity
We assessed heterogeneity by visually inspecting forest plots, and by using the Chi² test and I² statistic as follows: 0% to 40% might not be important, 30% to 60% may represent moderate heterogeneity, 50% to 90% may represent substantial heterogeneity, and 75% to 100% indicates considerable heterogeneity (Deeks 2020).
Assessment of reporting biases
We would have used a funnel plot to explore small‐study biases in an outcome if data from sufficient studies (10 or more) had been pooled in a single meta‐analysis.
Data synthesis
We used a random‐effects model on the basis that drug trials assessed different but related intervention effects.
Subgroup analysis and investigation of heterogeneity
We found insufficient evidence to perform subgroup analysis for dosage, timing, and length of treatment.
Sensitivity analysis
No sensitivity analyses were pre‐specified or undertaken.
Summary of findings and assessment of the certainty of the evidence
For this review update, we created 'Summary of findings' tables and used the GRADE approach to evaluate the certainty of the evidence (Schünemann 2020).
We included the following outcomes:
-
Number of participants free from all seizures.
-
Number of participants with ≥ 75% reduction in all seizures.
-
Number of participants with ≥ 50% reduction in all seizures.
-
Number of participants with adverse events leading to study discontinuation.
Two review authors (from among KJ, FB, and CE) used the GRADE approach to judge the certainty of evidence for outcomes based on five criteria (risk of bias, inconsistency, indirectness, imprecision, and publication bias) (Schunemann 2011). If we had serious concerns regarding one of the five criteria, we downgraded the evidence from 'high quality' by one level; if we had very serious concerns, we downgraded the evidence by two levels. We resolved any discrepancies through discussion and reported our rationale for downgrading evidence in the 'Summary of findings' table footnotes.
We used GRADEpro GDT software (GRADEpro GDT 2015) to record our judgements and to create the 'Summary of findings' tables for each comparison included in the review.
Results
Description of studies
Results of the search
See Figure 1.
Study flow diagram.
We identified 40 studies, of which 11 met our inclusion criteria (1277 randomised participants). Four studies were published since the previous version of this review; see Characteristics of included studies. We excluded eight studies in total. We excluded two studies from the current review because they compared different doses of the same ASM (Conry 2009; Inanaga 1989), and three other studies were judged not to be RCTs (Oletsky 1996; Perry 2019; Vigevano 1994). We also excluded three studies that were excluded in a previous version of this review because the outcome data were not reported in a usable way (Battaglia 1991; Vajda 1985 ; Vassella 1978); see Characteristics of excluded studies.
We identified eight ongoing studies from ClinicalTrials.gov and the World Health Organization International Clinical Trials Registry Platform (CTRI/2010/091/001449; NCT00004776; NCT01370486; NCT02318537; NCT03355209; NCT03650452; NCT03808935; Wechsler 2017); see Characteristics of ongoing studies. Of these eight trials, one trial of adjunctive rufinamide versus lamotrigine was terminated, and another trial of adjunctive cannabidiol versus placebo had its sponsorship withdrawn. Other adjunctive drug therapies under investigation included topiramate, melatonin, fenfluramine hydrochloride and TAK‐935 versus placebo.
On 25 November 2020, one review author (FB) contacted pharmaceutical companies (Eisai, GW Pharmaceuticals, and UCB Pharma), and colleagues in the field to obtain information on any additional unpublished data. They identified no unpublished data relating to this review. On 4 December 2020, one review author (FB) also contacted the principal investigators of Arzimanoglou 2019 to obtain information on missing data, and the principal investigators of NCT01370486 to obtain information on study completion. We received no response by the time of review completion.
Despite its completion, we listed Ohtahara 2008 in Studies awaiting classification because we were unable to obtain a translation from the Japanese language into English.
Included studies
See Characteristics of included studies.
Two studies published protocols that we considered as part of the evidence synthesis (Devinsky 2018; Thiele 2018). Of the 11 included studies, 10 were randomised and double‐blinded, placebo‐controlled trials (Arzimanoglou 2019; Devinsky 2018; Felbamate Study Group 1993; Glauser 2008; Group for the Evaluation of Cinromide 1989; Motte 1997; Ng 2011; Ohtsuka 2014; Sachdeo 1999; Thiele 2018) and one was a randomised and double‐blinded, cross‐over, placebo‐controlled trial (Eriksson 1998).
Two studies involved children and adolescents only, with ages ranging from one year to 18 years, (Arzimanoglou 2019; Group for the Evaluation of Cinromide 1989) and eight studies involved children, adolescents and adults, with ages ranging from over one year to 55 years (Devinsky 2018; Felbamate Study Group 1993; Glauser 2008; Motte 1997; Ng 2011; Ohtsuka 2014; Sachdeo 1999; Thiele 2018). Another study of children, adolescents and young adults with refractory generalised epilepsy included a subgroup of participants with LGS who were all aged under 18 years (Eriksson 1998).
In ten studies, diagnosis of LGS was reported according to clinical and EEG criteria or the classification of the International League Against Epilepsy (ILAE 1989); (Arzimanoglou 2019; Devinsky 2018; Eriksson 1998; Felbamate Study Group 1993; Glauser 2008; Group for the Evaluation of Cinromide 1989; Ng 2011; Ohtsuka 2014; Sachdeo 1999; Thiele 2018). For one other study, the diagnosis was agreed by an international expert panel of child neurologists (Motte 1997). There was limited reporting of co‐morbidities in the study population but as many as 40% of participants from one study were reported to have underlying causes, such as tuberous sclerosis and cerebral palsy, cerebral dysgenesis, encephalitis and bacterial meningitis (Ohtsuka 2014).
The drug treatment regimens of the included RCTs were as follows: cannabidiol and ASMs (Devinsky 2018; Thiele 2018); cinromide and ASMs (Group for the Evaluation of Cinromide 1989); clobazam and ASMs (Ng 2011); felbamate and ASMs (Felbamate Study Group 1993); lamotrigine and ASMs (Eriksson 1998; Motte 1997); rufinamide and ASMs (Arzimanoglou 2019; Glauser 2008, and Ohtsuka 2014); and topiramate and ASMs (Sachdeo 1999). Aside from ASMs, the two most recent studies of add‐on cannabidiol reported non‐pharmacological co‐interventions including vagus nerve stimulation and ketogenic diet in a subset of participants (Devinsky 2018; Thiele 2018). All studies included a pre‐randomisation, baseline period of four to eight weeks. The follow‐up after randomisation ranged from approximately 11 weeks to 112 weeks, including variable duration of titration and maintenance periods, with or without tapering, and safety follow‐up.
In three studies, the primary outcome for treatment efficacy was overall seizure reduction, which was the main focus of the current review (Eriksson 1998; Felbamate Study Group 1993; Group for the Evaluation of Cinromide 1989). Two studies included primary efficacy outcomes for overall seizure reduction and reduction in drop attacks or tonic‐atonic seizures (Glauser 2008; Sachdeo 1999). Four later studies had a primary efficacy outcome that focused on drop seizure or tonic‐atonic seizure reduction (Devinsky 2018; Ng 2011; Ohtsuka 2014; Thiele 2018). The primary efficacy outcome was major motor seizures in one study (Motte 1997) and behavioural outcomes in another study (Arzimanoglou 2019).
Arzimanoglou 2019
Arzimanoglou 2019 was conducted in 19 centres across Canada, France, Greece, Italy, Poland, and the United States. It involved 37 randomised children (aged one year to less than four years) with inadequately controlled LGS. The participants had LGS with seizures uncontrolled by a fixed dose of one to three concomitant ASMs for a minimum of four weeks before randomisation. The study included an eight‐week pre‐randomisation phase (screening period and baseline visit), followed by a 106‐week randomised phase (including titration and maintenance). Participants were randomised to receive add‐on rufinamide (target maintenance dose: 45 mg/kg/day in two divided doses, given as oral suspension) or any other ASM chosen by the investigator and added to the existing regimen of one to three ASMs. Baseline characteristics were similar in the trial groups although the 'Any other ASM' group had a higher proportion of males. The categories for ethnicity (Hispanic/Latino; non‐Hispanic/Latino) differed from those reported in the 6‐month dataset (White; Black or African‐American). Collectively, the baseline characteristics indicated mostly non‐Hispanic participants (more than 75%) and a majority of white participants (more than 70%) in the study.
Devinsky 2018
Devinsky 2018 was conducted at 30 participating centres (20 in the United States, 5 in Spain, 3 in the United Kingdom, and 2 in France). It involved 225 randomised children and adults (aged 2 to 55 years). They took between one and four ASMs and had at least two drop seizures each week during the baseline period. Participants were randomised to receive add‐on cannabidiol at a dose of either 20 mg/kg/day or 10 mg/kg/day or matching add‐on placebo. The trial included a 4‐week baseline period, a 14‐week treatment period (2 weeks of dose escalation, followed by a maintenance phase of 12 weeks), a tapering period of up to 10 days, and a 4‐week safety follow‐up period after discontinuation of cannabidiol or placebo. Baseline characteristics were similar in the trial groups although a slightly higher proportion of the placebo group also received vagus nerve stimulation as a concomitant intervention (28%) as compared with the 10 and 20mg cannabidiol groups (21% each). The majority of study participants were of white ethnicity (more than 88%).
Eriksson 1998
Eriksson 1998 was conducted in Finland. The study initially involved 30 children (aged over two years), adolescents and two young adults with refractory generalised epilepsy. Twenty of the participants in the open phase had LGS, and all were aged under 18 years. Diagnostic criteria for LGS were based on the classification of the International League Against Epilepsy (ILAE 1989). All included participants experienced more than two seizures per month at baseline. The trial consisted of six phases. There was first an eight‐week baseline phase during which each child was observed on pre‐study medication, followed by an open phase, during which an attempt was made to find the optimal lamotrigine dose for each child. At the end of the open phase, children had been categorised as responders if they showed any improvement (in alertness, behaviour, motor skills, or seizures). The 'responders' (17/27 (13 with LGS); three excluded because of incomplete seizure diaries) were subsequently entered in a double‐blind phase of 12‐week periods during which, for each child, add‐on lamotrigine and placebo tablets were administered in random order. The treatment periods were separated by a three‐week washout phase. This study did not report participants' ethnicity and other characteristics for baseline comparison of treatment groups.
Felbamate Study Group 1993
Felbamate Study Group 1993 involved 73 randomised children and adults (aged 4 years to 36 years) and was conducted in the USA. Included participants had a history of multiple types of seizures and a minimum of 90 atonic seizures or atypical absence seizures per month during an eight‐week pre‐study phase; they took no more than two ASMs at baseline. The trial consisted of a 28‐day baseline period followed by a 14‐day titration phase and a 56‐day maintenance period. The initial dose of add‐on felbamate was 15 mg/kg/day, increased to 30 mg/kg after seven days and to either 14 mg/kg/day or 3600 mg/day (whichever was lower) after 14 days. Baseline characteristics were similar in the trial groups. The study largely involved people of white ethnicity (more than 89%).
Glauser 2008
Glauser 2008 was conducted at 36 sites in nine countries (Belgium, Brazil, Germany, Hungary, Italy, Norway, Poland, Spain, and the United States). It involved 138 randomised children and adults (aged four years to 37 years). Study participants had a history of multiple seizure types, including atypical absence seizures and drop attacks, and a minimum of 90 seizures in the month prior to trial entry. At baseline, included participants took a fixed‐dose regimen of one to three concomitant ASMs. The trial consisted of a 28‐day baseline period at the end of which, patients continuing to meet the study criteria entered an 84‐day double‐blind treatment phase of either add‐on rufinamide or add‐on placebo. This phase consisted of a 14‐day titration period followed by a 70‐day maintenance period. Doses were titrated according to a recommended schedule based on body weight, up to a maximum dose of 45 mg/kg/day.
Baseline characteristics of the two treatment groups were similar, including mostly male participants and with comparable usage of concomitant ASMs. However, a higher proportion of the add‐on placebo group was under 12 years of age, and a higher proportion of the add‐on rufinamide group was over 17 years of age. Both groups largely involved participants of white ethnicity.
Group for the Evaluation of Cinromide 1989
Group for the Evaluation of Cinromide 1989 involved 56 randomised children and adolescents (aged two years to 18 years) and was conducted in the USA. Included participants had seizures for at least six months prior to study entry. At baseline, no individual was receiving more than three marketed antiepileptic drugs and none had previous exposure to cinromide. The trial consisted of a six‐week baseline period, following which participants were randomised to receive either add‐on cinromide or add‐on placebo for a period of 18 weeks. Study medication was initiated at 20 to 40 mg/kg/day, divided into four equal doses. Further increases (to a total daily maximum of 83 to 109 mg/kg) were prescribed at weekly visits if each prior dose was tolerated and seizures continued. Baseline characteristics were comparable between groups, with the majority of study participants being of white ethnicity (more than 92%).
Motte 1997
Motte 1997 was conducted at 43 unspecified sites. The study involved 169 randomised children and young adults (aged three years to 25 years). Included participants experienced more than one type of predominantly generalised seizure for at least one year, and they had seizures at least every other day. They took up to three ASMs at baseline. The trial consisted of a four‐week baseline period during which all recipients received add‐on placebo. Participants were then randomised to receive either add‐on lamotrigine or add‐on placebo for a 16‐week treatment period. Participants were assigned to one of four dosing regimens according to concomitant valproate use and body weight based on paediatric dosing recommendations at that time. The characteristics of the two groups were similar at baseline although the lamotrigine group had a higher proportion of male participants. Both groups largely involved participants of white ethnicity (more than 90%).
Ng 2011
Ng 2011 was conducted at 51 sites in the United States, India, Europe, and Australia. It involved 238 randomised children and adults (aged 2 years to 54 years). They received one to three ASMs with stable dosages for at least 30 days before screening, and they experienced at least two drop seizures per week during the four‐week baseline period. Included participants were randomised to add‐on placebo or one of three doses of add‐on clobazam (0.25, 0.5, and 1.0 mg/kg/day), up to a maximum dosage of 40 mg/day. Treatment included a three‐week titration phase and a 12‐week maintenance phase followed by two to three weeks tapering or continuation in an open‐label extension. Baseline characteristics were comparable between groups, although the medium‐dose clobazam group had a lower mean baseline average weekly drop seizure rate (58.8; SD 119.6) compared with the low‐dose group (98.3; SD 198.5), high‐dose group (94.9; SD 152.2), and placebo group (95.6; SD 168.2). Most study participants in each group were of white ethnicity (57% to 71%).
We included a post‐hoc analysis of Ng 2011 as a subsidiary paper (Paolicchi 2015). This post‐hoc analysis aimed to determine potential drug‐related effects on four behaviour domains of the Child Behaviour Checklist (CBCL): aggressive behaviour, anxious/depressed, attention problems, and somatic complaints (Achenbach 1991).
Ohtsuka 2014
Ohtsuka 2014 included 59 randomised children and adults (aged between 4 years and 30 years), and was conducted in Japan. Included participants experienced at least 90 seizures during the 28 days before the baseline period. They took one to three background ASMs at baseline. The participants were randomised to either add‐on rufinamide (doses titrated according to a predetermined schedule based on body weight, with the target dose maintained during the maintenance period) or add‐on placebo. This study consisted of four phases: a four‐week baseline, a two‐week titration, a 10‐week maintenance, and either a follow‐up visit or entry into an open‐label extension. Baseline characteristics were comparable between groups although there was a relatively higher concomitant use of lamotrigine in the placebo group and relatively higher concomitant use of clobazam in the rufinamide group. Participants' ethnicity was not reported in this study.
Sachdeo 1999
Sachdeo 1999 was conducted at 12 centres in the USA. The study included 98 randomised children and young adults (aged over one year to under 30 years). Included participants experienced seizure types including drop attacks and atypical absence seizures; they had a frequency of at least 60 seizures during the month prior to the baseline phase, while being maintained on one or two standard ASMs. The trial consisted of a baseline phase of four weeks and an 11‐week treatment phase that included three weeks titration and eight weeks maintenance. The participants were titrated up to a dose of 6 mg/kg/day or their maximal tolerated dosage of either add‐on topiramate or add‐on placebo over the first three weeks of the treatment period. Of note, concomitant treatment with felbamate was prohibited part way through the study, due to adverse effects. Baseline characteristics were comparable between groups, although there was a slightly higher proportion of males in the topiramate group. Both groups largely involved participants of white ethnicity (more than 80%).
Thiele 2018
Thiele 2018 was conducted at 24 clinical sites in the USA (17 sites), the Netherlands (one site), and Poland (six sites). It included 171 children and adults (aged two years to 45 years). The study participants experienced more than one type of generalised seizure, including drop seizures, for at least six months. At baseline, they took one to four ASMs, and had at least two drop seizures per week during the four‐week baseline period. Participants were randomised to add‐on cannabidiol 20 mg/kg (given as oral solution) or matching add‐on placebo solution administered orally in two equally divided doses. The study duration was 14 weeks, which included two weeks of dose escalation (starting at a daily dose of 2.5 mg/kg, followed by a maintenance period of 12 weeks), a tapering period of up to 10 days, and a four‐week safety follow‐up period. Baseline characteristics were comparable between groups, although the placebo group had a higher proportion of participants on a ketogenic diet as a concomitant intervention (12% versus 5%). Both groups largely involved participants of white ethnicity (more than 87%).
Excluded studies
See Characteristics of excluded studies.
We excluded eight studies from the search results for this updated review. We excluded two previously included studies that compared different doses of the same ASM (Conry 2009 and Inanaga 1989). Three studies were judged not to be randomised trials (Oletsky 1996; Perry 2019; Vigevano 1994). We also considered studies excluded from the previous version of this review, and agreed on the exclusion of three further studies. We excluded these studies because incomplete reporting for participants with LGS meant that the data were not usable (Battaglia 1991; Vajda 1985 ; Vassella 1978). We did not attempt to contact authors for additional data from these studies, which were published between 29 and 42 years ago. We included one previously excluded study as a subsidiary publication relating to the Felbamate Study Group 1993.
Risk of bias in included studies
We have reported 'Risk of bias' assessments for each study in Characteristics of included studies. Figure 2 summarises the review authors' 'Risk of bias' assessments. Only two included studies of add‐on cannabidiol had published protocols (Devinsky 2018; Thiele 2018). Based on the available information, we identified a high risk of selective reporting bias in five of the 11 RCTs (Devinsky 2018; Eriksson 1998; Felbamate Study Group 1993; Group for the Evaluation of Cinromide 1989; Thiele 2018), three of which were included in the previous version of the review (Eriksson 1998; Felbamate Study Group 1993; Group for the Evaluation of Cinromide 1989). In three of the 11 RCTs, we identified a high risk of bias associated with incomplete outcome data (Arzimanoglou 2019; Eriksson 1998; Group for the Evaluation of Cinromide 1989), two of which were included in the previous version of this review (Eriksson 1998; Group for the Evaluation of Cinromide 1989); the more recent study had a relatively long follow up of two years (Arzimanoglou 2019). We identified a high risk of bias due to the exclusion of 'non‐responders' in the initial open phase of one study (Eriksson 1998).
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
For outcomes included in the 'Summary of findings' tables, we assessed risk of bias at the outcome level. However, we downgraded outcomes for study limitations in one comparison that was stopped prematurely and judged to have a very serious risk of bias due to a combination of selective reporting and incomplete data (Group for the Evaluation of Cinromide 1989).
Allocation
We assessed all eleven included studies to have a low risk of selection bias associated with random sequence generation. In terms of allocation concealment, we judged six of the eleven included studies to have a low risk of selection bias and five studies as having an unclear risk of selection bias.
Blinding
We judged two of the eleven included studies to have a low risk of performance bias and detection bias, and nine studies to have an unclear risk of performance bias and detection bias.
Incomplete outcome data
We judged five of the eleven included studies to have a low risk of attrition bias, three studies as having an unclear risk and three studies as having a high risk of attrition bias.
Selective reporting
We judged six of the eleven included studies to have a low risk of reporting bias and five as having a high risk of reporting bias.
Other potential sources of bias
We did not identify any other potential sources of bias in the eleven included studies.
Effects of interventions
See: Summary of findings 1 Cannabidiol (10 mg/kg and 20 mg/kg) plus ASMs compared to placebo plus ASMs; Summary of findings 2 Cinromide plus ASMs compared to placebo plus ASMs; Summary of findings 3 Clobazam (low, medium and high doses) plus ASMs compared to placebo plus ASMs; Summary of findings 4 Felbamate plus ASMs compared to placebo plus ASMs; Summary of findings 5 Lamotrigine plus ASMs compared to placebo plus ASMs; Summary of findings 6 Rufinamide plus ASMs compared to placebo plus ASMs; Summary of findings 7 Rufinamide plus ASMs compared to other ASM plus ASMs; Summary of findings 8 Topiramate plus ASMs compared to placebo plus ASMs
See: summary of findings Table 1; summary of findings Table 2; summary of findings Table 3; summary of findings Table 4; summary of findings Table 5; summary of findings Table 6; summary of findings Table 7; and summary of findings Table 8.
Cannabidiol (10 mg/kg and 20 mg/kg) plus ASMs versus placebo plus ASMs
Two studies contributed data for this comparison (Devinsky 2018; Thiele 2018).
Devinsky 2018 evaluated cannabidiol 10 mg/kg/day and 20 mg/kg/day, whereas Thiele 2018 evaluated only 20 mg/kg/day.
In Devinsky 2018, "A total of 13 patients (6%) discontinued either cannabidiol (11 patients) or placebo (2 patients); in 7 of the 11 patients who discontinued cannabidiol, the treatment was discontinued because of adverse events." In Thiele 2018, "14 patients in the cannabidiol group and one in the placebo group (9%) withdrew from the trial; in nine (60%) of these patients, adverse events were the primary reason for study discontinuation."
Number of participants free from drop seizures during the treatment phase
No participants were free from drop seizures (one study; 225 participants; Analysis 1.1).
Number of participants with ≥ 75% reduction in drop seizures during the treatment phase
The RR was 3.51 (95% CI 1.24 to 9.92) in favour of the cannabidiol regimen (two studies; 396 participants; Analysis 1.2).
Number of participants with ≥ 50% reduction in drop seizures during the treatment phase
The RR was 2.12 (95% CI 1.48 to 3.03) in favour of the cannabidiol regimen (two studies; 396 participants; Analysis 1.3).
Number of participants with ≥ 25% reduction in drop seizures during the treatment phase
The RR was 1.45 (95% CI 1.19 to 1.78) in favour of the cannabidiol regimen (two studies; 396 participants; Analysis 1.4).
Number of participants with > 0% to < 25% reduction in drop seizures during the treatment phase
The RR was 1.58 (95% CI 0.85 to 2.93) in favour of the cannabidiol regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 225 participants; Analysis 1.5).
Number of participants with > 0% to < 25% increase in drop seizures during the treatment phase
The RR was 1.40 (95% CI 0.66 to 3.00) in favour of the placebo regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 225 participants; Analysis 1.6).
Number of participants with > 25% increase in drop seizures during the treatment phase
The RR was 0.71 (95% CI 0.33 to 1.53) in favour of the cannabidiol regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 225 participants; Analysis 1.7).
Number of participants with improvement in the patient and caregiver Global Impression of Care scale
The RR was 1.52 (95% CI 1.22 to 1.89) in favour of the cannabidiol regimen (two studies; 392 participants; Analysis 1.8).
Number of participants free from drop seizures during the maintenance phase
The Peto OR was 7.76 (99% CI 0.75 to 79.85) in favour of the cannabidiol regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 170 participants; Analysis 1.9).
Number of participants with ≥ 75% reduction in drop seizures during the maintenance phase
The RR was 2.86 (95% CI 1.28 to 6.40) in favour of the cannabidiol regimen (one study; 170 participants; Analysis 1.10).
Number of participants with ≥ 50% reduction in drop seizures during the maintenance phase
The RR was 1.95 (95% CI 1.25 to 3.05) in favour of the cannabidiol regimen (one study; 170 participants; Analysis 1.11).
Number of participants with ≥ 25% reduction in drop seizures during the maintenance phase
The RR was 1.35 (95% CI 1.02 to 1.78) in favour of the cannabidiol regimen (one study; 170 participants; Analysis 1.12).
Number of participants with adverse events
The RR was 1.24 (95% CI 1.11 to 1.38) in favour of the placebo regimen (two studies; 396 participants; Analysis 1.13).
Number of participants with treatment‐related adverse events
The RR was 1.81 (95% CI 1.29 to 2.54) in favour of the placebo regimen (one study; 171 participants; Analysis 1.14).
Number of participants with serious adverse events
The RR was 4.94 (95% CI 1.76 to 13.85) in favour of the placebo regimen (one study; 171 participants; Analysis 1.15).
Number of participants with adverse events leading to dose reduction
The RR was 5.93 (95% CI 0.73 to 48.22) in favour of the placebo regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 171 participants; Analysis 1.16).
Number of participants with adverse events leading to study discontinuation
The RR was 6.62 (95% CI 1.56 to 28.15) in favour of the placebo regimen (two studies; 396 participants; Analysis 1.17). We assessed the certainty of the evidence for this outcome to be high (summary of findings Table 1).
Death
The Peto OR was 7.30 (99% CI 0.04 to 1261.58) in favour of the placebo regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 171 participants; Analysis 1.18).
Cinromide plus ASMs versus placebo plus ASMs
One study contributed data for this comparison (Group for the Evaluation of Cinromide 1989). This study was terminated prematurely "when it was clear to the sponsor that cinromide was not effective"; only efficacy data collected before study interruption were included in the analysis. This study also reported changes in adjuvant ASM dosage following adverse events for 62% of the cinromide group and 27% of the placebo group.
Number of participants free from all seizures
No participants were free from seizures (one study; 56 participants; Analysis 2.1). We downgraded the certainty of evidence for this outcome twice for study limitations because there was a high risk of bias for incomplete data and twice for imprecision because the study was not powered to detect a between‐group difference in zero event outcomes. Our certainty in the evidence for this outcome was very low (summary of findings Table 2).
Number of participants with ≥ 75% reduction in mean weekly seizures
The Peto OR was 9.35 (99% CI 0.45 to 194.96) in favour of the cinromide regime, but the CI included the possibility of an effect favouring either treatment regimen (one study; 56 participants; Analysis 2.2). We downgraded the certainty of evidence for this outcome twice for study limitations because there was a high risk of bias for incomplete data and twice for imprecision because the effect estimate has a very wide CI. Our certainty in the evidence for this outcome was very low (summary of findings Table 2).
Number of participants with ≥ 50% reduction in mean weekly seizures
The RR was 1.15 (95% CI 0.47 to 2.86) in favour of the cinromide regime, but the CI included the possibility of an effect favouring either treatment regimen (one study; 56 participants; Analysis 2.3). We downgraded the certainty of evidence for this outcome twice for study limitations because there was a high risk of bias for incomplete data and twice for imprecision because the effect estimate has a very wide CI. Our certainty in the evidence for this outcome was very low (summary of findings Table 2).
Number of participants with ≥ 25% reduction in mean weekly seizures
The RR was 1.07 (95% CI (95% CI 0.59 to 1.91) in favour of the cinromide regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 56 participants; Analysis 2.4).
Number of participants with ≥ 0% to < 25% reduction in mean weekly seizures
The RR was 1.15 (95% CI 0.42 to 3.14) in favour of the cinromide regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 56 participants; Analysis 2.5).
Number of participants with ≥ 0% to < 25% increase in mean weekly seizures
The RR was 0.87 (95% CI 0.21 to 3.52) in favour of the cinromide regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 56 participants; Analysis 2.6).
Number of participants with > 25% increase in mean weekly seizures
The RR was 0.82 (95% CI 0.30 to 2.29) in favour of the cinromide regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 56 participants; Analysis 2.7).
Number of participants with improvement in global evaluation (at week 12, 18 and 24)
Week 12: the RR was 0.99 (95% CI 0.56 to 1.74) in favour of the placebo regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 56 participants; Analysis 2.8).
Week 18: the RR was 0.80 (95% CI 0.41 to 1.56) in favour of the placebo regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 56 participants; Analysis 2.8).
Week 24: the RR was 1.28 (95% CI 0.62 to 2.66) in favour of the cinromide regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 56 participants; Analysis 2.8).
Number of participants with no change in global evaluation (at week 12, 18 and 24)
Weeks 12 and 18: the RR was 0.91 (0.50 to 1.64) in favour of the cinromide regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 56 participants; Analysis 2.9).
Week 24: the RR was 1.03 (95% CI 0.46 to 2.27) in favour of the placebo regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 56 participants; Analysis 2.9).
Number of participants with worsening in global evaluation (at week 12, 18 and 24)
Week 12: the RR was 3.46 (95% CI 0.38 to 31.28) in favour of the placebo regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 56 participants; Analysis 2.10).
Week 18: the RR was 4.62 (95% CI 0.55 to 38.74) in favour of the placebo regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 56 participants; Analysis 2.10).
Week 24: the Peto OR was 9.09 (99% CI 0.17 to 475.60) in favour of the placebo regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 56 participants; Analysis 2.11).
Death
No deaths were reported in the study.
Clobazam plus ASMs versus placebo plus ASMs
One study contributed data for this comparison (Ng 2011). According to study authors, "the most common reasons for discontinuing the study were lack of efficacy for placebo‐treated patients and adverse events for clobazam‐treated patients." In the placebo group, 30.5% of participants discontinued the study. Ten participants discontinued treatment due to lack of efficacy, four due to the request of the participant, parent or caregiver, two due to side effects, and two participants were lost to follow up. With clobazam, 13.8% of participants discontinued from the low‐dose group, 27.4% discontinued from the medium‐dose group, and 30.5% discontinued from the high‐dose group. Twenty‐four participants treated with add‐on clobazam discontinued due to side effects, five discontinued due to lack of efficacy, five discontinued due to the request of the participant, parent or caregiver, five discontinued for other reasons such as protocol violations; and four participants were lost to follow up.
A post‐hoc study investigated aggression‐related adverse events that occurred after the first dose of study drug and within 30 days after the last dose of study drug and found no difference between clobazam and placebo (RR 1.89, 95% CI 0.69 to 5.19). Among the 23 participants randomised to rufinamide who experienced aggression‐related adverse events, five had a history of aggression or behavioural problems. Among the four participants randomised to placebo who experienced aggression‐related adverse events, one person had a history of aggression or behavioural problems.
Number of participants free from drop seizures
The RR was 4.10 (95% CI 1.00 to 16.83) in favour of the clobazam regimen (one study; 217 participants; Analysis 3.1).
Number of participants with ≥ 75% reduction in drop seizures (from baseline to maintenance phase in average weekly rate)
The RR was 4.04 (95% CI 1.85 to 8.79) in favour of the clobazam regimen (one study; 217 participants; Analysis 3.2).
Number of participants with ≥ 50% reduction in drop seizures (from baseline to maintenance phase in average weekly rate)
The RR was 1.88 (95% CI 1.26 to 2.81) in favour of the clobazam regimen (one study; 217 participants; Analysis 3.3).
Number of participants with ≥ 25% reduction in drop seizures (from baseline to maintenance phase in average weekly rate)
The RR was 1.54 (95% CI 1.17 to 2.03) in favour of the clobazam regimen (one study; 217 participants; Analysis 3.4).
Number of participants with adverse events
The RR was 1.17 (95% CI.097 to 2.03) in favour of the placebo regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 238 participants; Analysis 3.5).
Number of participants with adverse events leading to dose reduction
The RR was 9.23 (95% CI 1.28 to 66.37) in favour of the placebo regimen (one study; 238 participants; Analysis 3.6).
Number of participants with aggression‐related adverse events
The RR was 1.89 (95% CI 0.69 to 5.19) in favour of the placebo regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 194 participants; Analysis 3.7).
Number of participants with serious adverse events
The RR was 2.31 (95% CI 0.54 to 9.86) in favour of the placebo regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 238 participants; Analysis 3.8).
Number of participants with adverse events leading to study discontinuation
The RR was 4.12 (95% CI 1.01 to 16.87) in favour of the placebo regimen (one study; 238 participants; Analysis 3.9). We assessed the certainty of the evidence for this outcome to be high (summary of findings Table 3).
Death
There were no deaths in this study (one study; 238 participants; Analysis 3.10).
Felbamate plus ASMs versus placebo plus ASMs
One study contributed data for this comparison (Felbamate Study Group 1993). One participant had treatment stopped because of somnolence and ataxia in the felbamate group and one because of pancreatitis in the placebo group. According to retrospective analysis, "Approximately 50% of patients randomised to FBM obtained at least a 50% reduction in seizure frequency compared with about 15% receiving placebo" (Jensen 1994).
Number of participants free from all seizures (recorded by closed‐circuit television and electroencephalography)
Treatment phase: the RR was 2.92 (95% CI 0.32 to 26.77) in favour of the felbamate regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 73 participants; Analysis 4.1). We downgraded the certainty of evidence for this outcome twice for imprecision (from high to low) because the effect estimate has a very wide CI (summary of findings Table 4).
Maintenance phase: the RR was 5.84 (95% CI 0.74 to 46.11) in favour of the felbamate regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 73 participants; Analysis 4.1).
Number of participants free from all seizures
Treatment phase: No participants were free from seizures (one study; 73 participants; Analysis 4.2). We downgraded the certainty of the evidence twice for imprecision (from high to low) because the study was not powered to detect a between‐group difference in zero event outcomes (summary of findings Table 4).
Maintenance phase: the RR was 3.89 (95% CI 0.46 to 33.17) in favour of the felbamate regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 73 participants; Analysis 4.3).
Number of participants free from atonic seizures
Treatment phase: the Peto OR was 6.43 (99% CI 0.30 to 137.10) in favour of the felbamate regime, but the CI included the possibility of an effect favouring either treatment regimen (one study; 50 participants; Analysis 4.4).
Maintenance phase: the Peto OR was 6.99 (99% CI 0.62 to 78.73) in favour of the felbamate regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 50 participants; Analysis 4.4).
Number of participants free from tonic‐clonic seizures
Treatment phase: the RR was 1.63 (95% CI 0.17 to 15.99) in favour of the felbamate regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 29 participants; Analysis 4.5.
Maintenance phase: the RR was 5.69 (95% CI 0.80 to 40.51) in favour of the felbamate regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 29 participants; Analysis 4.5).
Number of participants with severe side effects
The RR was 2.59 (95% CI 0.75 to 9.01) in favour of the placebo regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 73 participants; Analysis 4.6).
Number of participants with adverse events leading to study discontinuation
The RR was 0.97 (95% CI 0.06 to 14.97) in favour of the felbamate regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 73 participants; Analysis 4.7). We downgraded the certainty of evidence for this outcome twice for imprecision (from high to low) because the effect estimate has a very wide CI (summary of findings Table 4).
Death
No deaths were reported in the study.
Lamotrigine plus ASMs versus placebo plus ASMs
Two studies assessed this comparison (Eriksson 1998; Motte 1997) but only one contributed usable data for this review (Motte 1997).
Eriksson 1998 excluded "non‐responders" involved in an initial open phase of treatment from participating in the randomised study. The study also incompletely reported efficacy and tolerability outcomes for the subgroup of participants with LGS. The study reported that seven of the 20 children with LGS responded to treatment with > 50% seizure reduction, three children with > 75% seizure reduction, and two children became seizure free. However, individual data were only reported for 13 children with LGS, and no data were provided for the baseline number of seizures prior to crossover. In terms of adverse events, the study only reported data for the overall study population with refractory generalised epilepsy.
In this study, seven out of 13 children with LGS entered into the double‐blind phase of the trial showed improvement in the lamotrigine phase compared with the placebo phase, with one child showing a 100% reduction in their seizures. Three participants on lamotrigine had treatment withdrawn (23%), one had deterioration of seizure control, the other two developed a rash. Seven participants receiving placebo had treatment withdrawn (54%), six because of deterioration in seizure control and one other developed a rash.
In Motte 1997, 12% of participants discontinued from the study. "Seven patients in the lamotrigine group stopped treatment early, four because of protocol violations and three because of adverse events. Fourteen patients in the placebo group did not complete the study: seven had adverse events, three had protocol violations, two had a deterioration in the control of seizures, one failed to return for follow‐up, and the parents of one patient withdrew their consent."
Number of participants with ≥ 50% median reduction in all seizures
The RR was 2.12 (95% CI 1.19 to 3.76) in favour of the lamotrigine regimen (one study; 167 participants; Analysis 5.1). We assessed the certainty of the evidence for this outcome to be high (summary of findings Table 5).
Number of participants with > 25% to < 50% median reduction in all seizures
The RR was 1.41 (95% CI 0.80 to 2.47) in favour of the lamotrigine regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 167 participants; Analysis 5.2).
Number of participants with either 0 to ≤ 25% median reduction or an increase in all seizures
The RR was 0.61 (95% CI 0.45 to 0.83) in favour of the lamotrigine regimen (one study; 167 participants; Analysis 5.3).
Number of participants with ≥ 50% median reduction in drop attacks
The RR was 1.66 (95% CI 1.02 to 2.70) in favour of the lamotrigine regimen (one study; 164 participants; Analysis 5.4).
Number of participants with > 25% to < 50% median reduction in drop attacks
The RR was 1.61 (95% CI 0.87 to 2.99) in favour of the lamotrigine regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 164 participants; Analysis 5.5).
Number of participants in with either ≤ 25% median reduction or an increase in drop attacks
The RR was 0.60 (95% CI 0.43 to 0.85) in favour of the lamotrigine regimen (one study; 164 participants; Analysis 5.6).
Number of participants with ≥ 50% median reduction in tonic‐clonic seizures
The RR was 2.13 (95% CI 1.21 to 3.75) in favour of the lamotrigine regimen (one study; 124 participants; Analysis 5.7.
Number of participants with > 25% to < 50% median reduction in tonic‐clonic seizures
The RR was 0.91 (95% CI 0.33 to 2.57) in favour of the placebo regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 124 participants; Analysis 5.8.
Number of participants with either 0 to ≤ 25% median reduction or an increase in tonic‐clonic seizures
The RR was 0.68 (95% CI 0.49 to 0.93) in favour of the lamotrigine regimen (one study; 124 participants; Analysis 5.9).
Number of participants with adverse events leading to study discontinuation
The RR was 0.49 (95% CI 0.13 to 1.82) in favour of the lamotrigine regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 169 participants; Analysis 5.10). We downgraded the certainty of evidence for this outcome twice for imprecision because the effect estimate has a very wide CI. Our certainty in the evidence for this outcome was low (summary of findings Table 5).
Death
No deaths were reported in the study.
Rufinamide plus ASMs versus placebo plus ASMs
Two studies contributed data for this comparison (Glauser 2008; Ohtsuka 2014). In Glauser 2008, 15 participants (11%) discontinued from the study. "Ten patients in the rufinamide group discontinued therapy during the double‐blind phase because of adverse events (n = 6), unsatisfactory therapeutic effect (n = 3), or withdrawal of consent (n = 1). Five patients in the placebo group did not complete the study because of protocol violations (n = 2), unsatisfactory therapeutic effect (n = 1), administrative problems (n = 1), or withdrawal of consent (n = 1)."
In Ohtsuka 2014, "One patient assigned to the rufinamide group was excluded from the efficacy analysis due to inappropriate diagnosis."
Number of participants with ≥ 50% reduction in all seizures
The RR was 2.84 (95% CI 1.31 to 6.18) in favour of the rufinamide regimen (one study; 138 participants; Analysis 6.1). We assessed the certainty of the evidence for this outcome to be high (summary of findings Table 6).
Number of participants with ≥ 75% reduction in tonic‐atonic seizures
The RR was 10.71 (95% CI 1.46 to 78.39) in favour of the rufinamide regimen (one study; 58 participants; Analysis 6.2).
Number of participants with ≥ 50% reduction in tonic‐atonic seizures
The RR was 2.70 (95% CI 1.52 to 4.81) in favour of the rufinamide regimen (two studies; 191 participants; Analysis 6.3).
Number of participants with ≥ 25% reduction in tonic‐atonic seizures
The RR was 1.88 (95% CI 0.93 to 3.77) in favour of the rufinamide regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 58 participants; Analysis 6.4).
Number of participants 'unchanged' (< 25% reduction in tonic‐atonic seizures)
The RR was 1.38 (95% CI 0.59 to 3.20) in favour of the rufinamide regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 58 participants; Analysis 6.5).
Number of participants with increased tonic‐atonic seizures
The RR was 0.36 (95% CI 0.15 to 0.85) in favour of the rufinamide regimen (one study; 58 participants; Analysis 6.6).
Number of participants with improvement in seizure severity rating
The RR was 1.74 (95% CI 1.13 to 2.68) in favour of the rufinamide regimen (one study; 135 participants; Analysis 6.7).
Number of participants with adverse events
The RR was 1.13 (95% CI 0.86 to 1.50) in favour of the placebo regimen but the CI included the possibility of an effect favouring either treatment regimen (two studies; 197 participants; Analysis 6.8).
Number of participants with adverse events suspected to be treatment‐related
The RR was 1.27 (95% CI 0.90 to 1.79) in favour of the placebo regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 138 participants; Analysis 6.9).
Number of participants with serious adverse events
The RR was 0.86 (95% CI 0.13 to 5.97) in favour of the rufinamide regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 138 participants; Analysis 6.10).
Number of participants with adverse events leading to study discontinuation
In Ohtsuka 2014, four of 29 participants in the rufinamide group did not complete the study due to adverse events, and 1 of 30 participants in the placebo group discontinued due to an adverse event. It is unclear if a participant with an inappropriate diagnosis was one of the few withdrawals experiencing adverse events. RR 4.14 (CI 0.49 to 34.86) in favour of the placebo regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 59 participants; Analysis 6.11). We downgraded the certainty of evidence for this outcome twice for imprecision because the effect estimate has a very wide CI. Our certainty in the evidence for this outcome was low (summary of findings Table 6).
Death
No deaths were reported in the study.
Rufinamide plus ASMs versus other ASM plus ASMs
One study contributed data for this comparison (Arzimanoglou 2019). Ten of 25 people (40%) in the rufinamide group discontinued treatment, primarily due to adverse events (n = 3), patient choice (n = 2), inadequate therapeutic effect (n = 2), or withdrawal of consent (n = 3). Eight of 12 people (67%) in the 'Any other ASM' group discontinued treatment, primarily due to loss to follow‐up (n = 1), patient choice (n = 1), inadequate therapeutic effect (n = 1), withdrawal of consent (n = 4) or other reason (n = 1).
Number of participants with treatment‐emergent adverse events
The RR was 1.06 (95% CI 0.79 to 1.41) in favour of the other ASM regime, but the CI included the possibility of an effect favouring either treatment regimen (one study; 37 participants; Analysis 7.1).
Number of participants with severe treatment‐emergent adverse events
The RR was 0.96 (95% CI 0.20 to 4.53) in favour of the rufinamide regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 37 participants; Analysis 7.2).
Number of participants with serious treatment‐emergent adverse events
The RR was 0.96 (95% CI 0.42 to 2.19) in favour of the rufinamide regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 37 participants; Analysis 7.3).
Number of participants with treatment‐emergent adverse events leading to study‐drug dose adjustment
Reduction: the Peto OR was 5.91 (99% CI 0.61 to 57.64) in favour of the other ASM regime, but the CI included the possibility of an effect favouring either treatment regimen (one study; 37 participants).
Interruption: the Peto OR was 0.04 (99% CI 0.00 to 2.17) in favour of the rufinamide regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 37 participants; Analysis 7.4).
Number of participants with treatment‐emergent adverse events leading to study discontinuation
The RR was 0.96 (95% CI 0.10 to 9.57) in favour of the rufinamide regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 37 participants; Analysis 7.5). We downgraded the certainty of evidence for this outcome twice for imprecision (from high to low) because the effect estimate has a very wide CI (summary of findings Table 7).
Death
The Peto OR was 4.39 (99% CI 0.02 to 1077.58) in favour of the other ASM regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 37 participants; Analysis 7.6).
Child Behaviour Checklist Questionnaire
The MD was 2.60 points (95% CI ‐10.30 to 15.50) in favour of the other ASM regime, but the CI included the possibility of an effect favouring either treatment regimen (one study; 19 participants; Analysis 7.7).
Topiramate + ASMs versus placebo + ASMs
One study contributed data for this comparison (Sachdeo 1999). Of the randomised participants, only one person was withdrawn (due to patient choice), from the topiramate group.
Number of participants with ≥ 75% reduction in all seizures
The Peto OR was 8.22 (99% CI 0.60 to 112.62) in favour of the topiramate regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 98 participants; Analysis 8.1). We downgraded the certainty of evidence for this outcome twice for imprecision (from high to low) because the effect estimate has a very wide CI (summary of findings Table 8).
Number of participants free from major seizures (drop attacks and tonic‐clonic seizures)
The RR was 4.08 (95% CI 1.46 to 11.39) in favour of the topiramate regimen (one study; 96 participants; Analysis 8.2).
Number of participants with ≥ 75% reduction in major seizures (drop attacks and tonic‐clonic seizures)
The RR was 4.35 (95% CI 0.97 to 19.42) in favour of the topiramate regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 96 participants; Analysis 8.3).
Number of participants with ≥ 50% reduction in major seizures (drop attacks and tonic‐clonic seizures)
The RR was 4.08 (95% CI 1.46 to 11.39) in favour of the topiramate regimen (one study; 96 participants; Analysis 8.4).
Number of participants free from drop attacks
The Peto OR was 8.06 (99% CI 0.05 to 1398.41) in favour of the topiramate regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 96 participants; Analysis 8.5.
Number of participants with ≥ 75% reduction in drop attacks
The RR was 2.84 (95% CI 0.80 to 10.06) in favour of the topiramate regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 95 participants; Analysis 8.6).
Number of participants with ≥ 50% reduction in drop attacks
The RR was 1.98 (95% CI 0.87 to 4.52) in favour of the topiramate regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 95 participants; Analysis 8.7).
Number of participants free from drop attacks during the maintenance phase
The RR was 1.33 (95% CI 0.38 to 4.66) in favour of the topiramate regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 95 participants; Analysis 8.8).
Number of participants with severe adverse events
The RR was 2.29 (95% CI 0.86 to 6.11) in favour of the placebo regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 98 participants; Analysis 8.9).
Number of participants with adverse events leading to dose reduction or temporary discontinuation
The RR was 3.13 (95% CI 0.90 to 10.85) in favour of the placebo regimen but the CI included the possibility of an effect favouring either treatment regimen (one study; 98 participants; Analysis 8.10).
Number of participants with adverse events leading to study discontinuation
No adverse events led to study discontinuation (one study; 98 participants; Analysis 8.11). We downgraded the certainty of the evidence twice for imprecision (from high to low) because the study was not powered to detect a between‐group difference in zero event outcomes (summary of findings Table 8).
Death
No deaths were reported in the study.
Discussion
LGS is a complex and severe developmental and epileptic encephalopathy that manifests as multiple seizure types, which often evolve over time. Drop seizures (typically tonic, atonic or myoclonic‐atonic), is the characteristic seizure type. It is a syndrome that is difficult to treat, with many people receiving combination therapy (polypharmacy) but frequently without achieving seizure‐freedom. The aim of this review was to assess the efficacy and tolerability of anti‐seizure medications in the treatment of LGS.
LGS is thought to account for up to 5% of all childhood epilepsies,yet we found only 11 RCTs for the pharmaceutical treatment of this syndrome. There are several possible explanations for this lack of evidence. LGS is primarily a syndrome of childhood; long‐term prognosis is expected to be poor with regard to the cognitive outcome, but seizures tend to become less troublesome into adulthood (Ferlazzo 2010; Vignoli 2017). Preferably, drug treatments would be evaluated earlier in the course of the condition, closer to seizure onset. There has traditionally been reluctance to set up trials in the paediatric age group as it is both difficult and expensive, and continues to be so, despite the early exclusivity clause set out by EU Regulations (European Union 2006). Due to the low incidence of LGS (two cases per 100,000 children), a multicentre collaborative study design to enrol the numbers of individuals is required for sufficient power with consensus on the selection of individuals (diagnostic criteria), drug therapy, and outcome measures (Arzimanoglou 2009). Investigation of adjunctive and especially monotherapy ASM regimens is difficult because LGS characteristics usually evolve over time, including multiple seizure types and typical EEG appearances. As such, many individuals will have already received several ASMs at diagnosis. The natural history of the syndrome shows that the frequency and type of seizures often fluctuate over time, which means the observed improvement or deterioration might not be an effect of the study intervention. Most challenging is the adequate observation and reporting of seizure types with more subtle clinical manifestations, such as atypical absence and myoclonic seizures. This variation may account for the heterogeneous seizure outcome measures applied across the studies included in this review. Some more recent RCTs use pragmatic 'seizure definitions', such as drop seizures with motor manifestations resulting in loss of posture (Devinsky 2018; Ng 2011; Thiele 2018); this approach might be more quantifiable and have more clinical impact. Studies included in this review have applied different seizure outcome measures, with more recent studies focusing on drop seizures, a seizure type that is measurable and is expected to have the most negative clinical impact. Observation and recording of the different seizure types, in particular, those with more subtle clinical manifestations, such as atypical absence and myoclonic seizures, is challenging for carers of participating patients. This factor may be one reason for the variation in seizure outcome measures applied across studies in this review. Furthermore, LGS is an aetiologically heterogeneous condition, and the different aetiologies of LGS could affect the effectiveness outcomes of the different ASMs. For example, selected patients with an underlying monogenetic disorder such as a SCN2A mutation (resulting in gain or loss of sodium channel function) and LGS phenotype and may respond differently to sodium channel‐blocking ASMs (Brunklaus 2020; Wolff 2017).
We found only one trial that included behaviour as a primary outcome domain, even though behaviour and cognition are frequently cited by families as being the most difficult features of the syndrome to accept and manage. The primary focus of most trials of ASMs is on seizure control and it is expected that other types of medications would be indicated for modifying cognitive and behavioural function rather than ASMs.
Recent clinical practice guidelines have recommended the use of some ASMs for the treatment of LGS. The 2004 guidelines of the American Academy of Neurology and American Epilepsy Society for pharmacological management of treatment‐resistant epilepsy (French 2004) found felbamate, lamotrigine, and topiramate to be effective ASMs in treating LGS. The 2018 update of these guidelines (Kanner 2018) listed also rufinamide as an established treatment, reporting that clobazam should be considered to decrease seizure frequency. According to the 2012 National Institute for Health and Care Excellence (NICE) guidelines (NICE 2012), add‐on ASMs that may be considered for children, young people and adults with LGS are rufinamide and topiramate; if they are ineffective or not tolerated, felbamate can be offered. Carbamazepine, gabapentin, oxcarbazepine, pregabalin, tiagabine or vigabatrin should be avoided. According to a 2019 NICE technology appraisal guidance (NICE 2019), cannabidiol with clobazam is recommended as an option for patients aged two years and older with LGS, if their drop seizures are not controlled after 2 or more ASMS, and provided that "the frequency of drop seizures is checked every 6 months, and cannabidiol is stopped if the frequency has not fallen by at least 30% compared with the 6 months before starting treatment."
Summary of main results
We found that the evidence for drug treatment of LGS related specifically to the use of anti‐seizure medications. Some of the included studies involved the treatment of children only, and others involved both children and adults with LGS. All 11 included trials (1277 participants) used ASMs as add‐on therapy and had variable periods of dose adjustment and maintenance; we found no trials of ASMs used as monotherapy and no head‐to‐head comparisons of specific add‐on drugs. We identified one trial registry record for a study that planned to compare add‐on rufinamide with lamotrigine but the study was reported to have been terminated because those supporting the study did not want to proceed with marketing the product.
Three included RCTs reported outcomes for both overall seizure reduction and adverse events leading to study discontinuation. We found high‐certainty evidence that add‐on lamotrigine increased the number of participants with at least 50% reduction in the average number of reported seizures and low‐certainty evidence that add‐on lamotrigine may have reduced the number of participants with adverse events leading to study discontinuation when compared to add‐on placebo. We also found high‐certainty evidence that add‐on rufinamide increased the number of participants with at least 50% reduction in the average number of reported seizures when compared with add‐on placebo and low‐certainty evidence that add‐on rufinamide may have made little or no difference compared to add‐on placebo or another unspecified ASM in terms of reducing the number of participants with adverse events leading to study discontinuation. Add‐on topiramate may have increased the number of participants with at least 75% reduction in the average number of reported seizures, and probably made little or no difference to the number of adverse events leading to study discontinuation when compared to add‐on placebo (low‐certainty evidence). Add‐on felbamate (treatment phase) may have made little or no difference in terms of complete seizure freedom and adverse events leading to study discontinuation when compared to add‐on placebo (low certainty evidence). However, we found that when seizures were recorded under EEG monitoring, add‐on felbamate may have increased seizure freedom compared to add‐on placebo. We remain uncertain whether other add‐on drug therapies, including cannabidiol, cinromide and clobazam, reduced all types of seizures because this outcome was not reported or had very low‐certainty evidence. Despite therapeutic potential for reducing drop seizures, we found high‐certainty evidence that add‐on cannabidiol and add‐on clobazam increased the number of participants with adverse events leading to study discontinuation compared to add‐on placebo. We did not find any evidence for adverse events leading to study discontinuation in the comparison of add‐on cinromide with add‐on placebo.
There were insufficient data to perform subgroup analysis for the pre‐specified variables, dosage, timing, and length of treatment. Most of the evidence included in this review related to people from middle‐ or high‐income countries and, where reported, participants of white ethnicity.
Overall completeness and applicability of evidence
We excluded three studies on the basis that data were not usable. However, we did not attempt to obtain additional data from the authors because the studies were published more than 25 years ago. We acknowledge this as a limitation of the review. In addition, we listed one potentially relevant study as awaiting classification in the absence of a translation into English. We also acknowledge this as a limitation of the review.
All included trials used ASMs as add‐on therapy. Hence, the data on the treatment of LGS from RCTs refer to ASMs given as adjunctive treatments. The earliest trial of an add‐on therapy, cinromide, is not currently used in clinical practice and evidence relates only to patients up to the age of 18 years. We found limited reporting of outcomes for cognition and behaviour. We also found that some of the more recent studies focused on drop seizure reduction as a primary outcome. Our 'Summary of findings' tables focused on overall seizure cessation, reduction and adverse events in accordance with the previous version of this review and because multiple types of seizures are reported in LGS.
Quality of the evidence
For outcomes included in the Summary of Findings table, we found the certainty of evidence to be of moderate‐ to very low‐certainty. We primarily downgraded outcomes because the evidence was not powered to detect a between‐group difference (study limitations). We also downgraded outcomes due to imprecise estimates with wide or very wide CIs. In addition, we downgraded outcomes from one study that was judged to have a very serious risk of bias due to attrition and selective reporting (Group for the Evaluation of Cinromide 1989).
Potential biases in the review process
We made every effort to identify all RCTs on the use of ASMs for LGS through a comprehensive search of the literature, and it is unlikely that we failed to identify large relevant studies. However, despite our efforts, there remains the possibility that we have missed small studies published in the less accessible literature. We also contacted drug companies and experts in the field to obtain information on ongoing trials or unpublished results. Finally, we contacted authors of one included study, published in 2019, to obtain information on missing data, and we contacted trialists to determine the completion of another study. We acknowledge a potential bias in our review process because we did not attempt to contact all study authors for additional data. Unavoidably, some of the authors of this review were familiar with most of the included trials before updating this review. However, data extraction was undertaken blind to the results of the prior version of the review.
Agreements and disagreements with other studies or reviews
We identified a narrative review of drug therapies for LGS that concluded a lack of evidence of the efficacy of any one drug over another. The review reported that drug therapies are difficult to compare even in RCTs, and suggested that non‐medical therapies may be considered after the failure of two to three drugs (Borrelli 2019). We found some evidence in favour of the use of add‐on lamotrigine compared with add‐on placebo in terms of both overall reduction in average reported seizures and adverse events leading to study discontinuation. However, we also found a lack of RCTs assessing the therapeutic potential of monotherapy and head‐to‐head comparison of specific add‐on ASMs. It is beyond the scope of this review to make recommendations on the use of non‐medical therapies for LGS.
We identified several reviews of individual add‐on ASMs for the treatment of LGS. In 2018, a meta‐analysis of add‐on cannabidiol versus add‐on placebo was performed, based on two studies included in this review (Devinsky 2018 and Thiele 2018). In addition to the assessment of different types of seizures, this meta‐analysis also included data for all seizure frequency. The review authors reported that "rates of > 50% reduction of all seizures were also higher among patients randomised to the active drug rather than placebo (37.2% vs 21.2%; RR 1.76 (95% CI 1.07‐2.88); P = 0.025)" (Lattanzi 2018). However, we did not find outcome data reported for greater than 50% reduction in all seizures in Devinsky 2018 or Thiele 2018; these data were not included in our analysis of the evidence.
Previously, an indirect comparison was undertaken of add‐on felbamate, lamotrigine, topiramate, rufinamide and clobazam (Cramer 2013). Although the authors found some evidence in favour of add‐on clobazam, they acknowledged several limitations including overlapping 95% CIs, omission of adverse event outcomes and studies often using dose ranges rather than a specific dosage. The authors of this analysis were also employed by the manufacturer of clobazam. We did not identify publication bias in the included studies in this review but we did find that some study investigators declared financial interests in their study sponsor.
In 2010, one study evaluated the cost effectiveness of rufinamide in the treatment of LGS. Based on data from the UK, this study concluded that, for managing drop seizures, "if society is willing to pay an additional £250 or more for a 1% increase in patients achieving at least a 50% reduction in seizure frequency over 3 years, rufinamide has more than an 80% probability of being cost effective compared with either lamotrigine or topiramate" (Benedict 2010). We did not assess cost‐effectiveness in this review, but recognise a need for up‐to‐date modelling to help inform decision‐making.
Study flow diagram.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Comparison 1: Cannabidiol (10 mg/kg and 20 mg/kg) + ASMs versus placebo + ASMs, Outcome 1: Number of participants free from drop seizures during the treatment phase
Comparison 1: Cannabidiol (10 mg/kg and 20 mg/kg) + ASMs versus placebo + ASMs, Outcome 2: Number of participants with ≥ 75% reduction in drop seizures during the treatment phase
Comparison 1: Cannabidiol (10 mg/kg and 20 mg/kg) + ASMs versus placebo + ASMs, Outcome 3: Number of participants with ≥ 50% reduction in drop seizures during the treatment phase
Comparison 1: Cannabidiol (10 mg/kg and 20 mg/kg) + ASMs versus placebo + ASMs, Outcome 4: Number of participants with ≥ 25 % reduction in drop seizures during the treatment phase
Comparison 1: Cannabidiol (10 mg/kg and 20 mg/kg) + ASMs versus placebo + ASMs, Outcome 5: Number of participants with > 0% to < 25% reduction in drop seizures during the treatment phase
Comparison 1: Cannabidiol (10 mg/kg and 20 mg/kg) + ASMs versus placebo + ASMs, Outcome 6: Number of participants with > 0% to < 25% increase in drop seizures during the treatment phase
Comparison 1: Cannabidiol (10 mg/kg and 20 mg/kg) + ASMs versus placebo + ASMs, Outcome 7: Number of participants with > 25 % increase in drop seizures during the treatment phase
Comparison 1: Cannabidiol (10 mg/kg and 20 mg/kg) + ASMs versus placebo + ASMs, Outcome 8: Number of participants with improvement in the patient and caregiver Global Impression of Care scale
Comparison 1: Cannabidiol (10 mg/kg and 20 mg/kg) + ASMs versus placebo + ASMs, Outcome 9: Number of participants free from drop seizures during the maintenance phase
Comparison 1: Cannabidiol (10 mg/kg and 20 mg/kg) + ASMs versus placebo + ASMs, Outcome 10: Number of participants with ≥ 75% reduction in drop seizures during the maintenance phase
Comparison 1: Cannabidiol (10 mg/kg and 20 mg/kg) + ASMs versus placebo + ASMs, Outcome 11: Number of participants with ≥ 50% reduction in drop seizures during the maintenance phase
Comparison 1: Cannabidiol (10 mg/kg and 20 mg/kg) + ASMs versus placebo + ASMs, Outcome 12: Number of participants with ≥ 25 % reduction in drop seizures during the maintenance phase
Comparison 1: Cannabidiol (10 mg/kg and 20 mg/kg) + ASMs versus placebo + ASMs, Outcome 13: Number of participants with adverse events
Comparison 1: Cannabidiol (10 mg/kg and 20 mg/kg) + ASMs versus placebo + ASMs, Outcome 14: Number of participants with treatment‐related adverse events
Comparison 1: Cannabidiol (10 mg/kg and 20 mg/kg) + ASMs versus placebo + ASMs, Outcome 15: Number of participants with serious adverse events
Comparison 1: Cannabidiol (10 mg/kg and 20 mg/kg) + ASMs versus placebo + ASMs, Outcome 16: Number of participants with adverse events leading to dose reduction
Comparison 1: Cannabidiol (10 mg/kg and 20 mg/kg) + ASMs versus placebo + ASMs, Outcome 17: Number of participants with adverse events leading to study discontinuation
Comparison 1: Cannabidiol (10 mg/kg and 20 mg/kg) + ASMs versus placebo + ASMs, Outcome 18: Death
Comparison 2: Cinromide + ASMs versus placebo + ASMs, Outcome 1: Number of participants free from all seizures
Comparison 2: Cinromide + ASMs versus placebo + ASMs, Outcome 2: Number of participants with ≥ 75% reduction in mean weekly seizures
Comparison 2: Cinromide + ASMs versus placebo + ASMs, Outcome 3: Number of participants with ≥ 50% reduction in mean weekly seizures
Comparison 2: Cinromide + ASMs versus placebo + ASMs, Outcome 4: Number of participants with ≥ 25% reduction in mean weekly seizures
Comparison 2: Cinromide + ASMs versus placebo + ASMs, Outcome 5: Number of participants with ≥ 0% to < 25% reduction in mean weekly seizures
Comparison 2: Cinromide + ASMs versus placebo + ASMs, Outcome 6: Number of participants with ≥ 0% to < 25% increase in mean weekly seizures
Comparison 2: Cinromide + ASMs versus placebo + ASMs, Outcome 7: Number of participants with > 25% increase in mean weekly seizures
Comparison 2: Cinromide + ASMs versus placebo + ASMs, Outcome 8: Number of participants with improvement in global evaluation
Comparison 2: Cinromide + ASMs versus placebo + ASMs, Outcome 9: Number of participants with no change in global evaluation
Comparison 2: Cinromide + ASMs versus placebo + ASMs, Outcome 10: Number of participants with worsening in global evaluation
Comparison 2: Cinromide + ASMs versus placebo + ASMs, Outcome 11: Number of participants with worsening in global evaluation ‐ Week 24
Comparison 3: Clobazam (low, medium and high doses) + ASMs versus placebo + ASMs, Outcome 1: Number of participants free from drop seizures
Comparison 3: Clobazam (low, medium and high doses) + ASMs versus placebo + ASMs, Outcome 2: Number of participants with ≥ 75 % reduction in drop seizures (from baseline to maintenance phase in average weekly rate)
Comparison 3: Clobazam (low, medium and high doses) + ASMs versus placebo + ASMs, Outcome 3: Number of participants with ≥ 50 % reduction in drop seizures (from baseline to maintenance phase in average weekly rate)
Comparison 3: Clobazam (low, medium and high doses) + ASMs versus placebo + ASMs, Outcome 4: Number of participants with ≥ 25 % reduction in drop seizures (from baseline to maintenance phase in average weekly rate)
Comparison 3: Clobazam (low, medium and high doses) + ASMs versus placebo + ASMs, Outcome 5: Number of participants with adverse events
Comparison 3: Clobazam (low, medium and high doses) + ASMs versus placebo + ASMs, Outcome 6: Number of participants with adverse events leading to dose reduction
Comparison 3: Clobazam (low, medium and high doses) + ASMs versus placebo + ASMs, Outcome 7: Number of participants with aggression‐related adverse events
Comparison 3: Clobazam (low, medium and high doses) + ASMs versus placebo + ASMs, Outcome 8: Number of participants with serious adverse events
Comparison 3: Clobazam (low, medium and high doses) + ASMs versus placebo + ASMs, Outcome 9: Number of participants with adverse events leading to study discontinuation
Comparison 3: Clobazam (low, medium and high doses) + ASMs versus placebo + ASMs, Outcome 10: Death
Comparison 4: Felbamate + ASMs versus placebo + ASMs, Outcome 1: Number of participants free from all seizures (recorded by closed‐circuit television and electroencephalography)
Comparison 4: Felbamate + ASMs versus placebo + ASMs, Outcome 2: Number of participants free from all seizures ‐ Treatment phase
Comparison 4: Felbamate + ASMs versus placebo + ASMs, Outcome 3: Number of participants free from all seizures ‐ Maintenance phase
Comparison 4: Felbamate + ASMs versus placebo + ASMs, Outcome 4: Number of participants free from atonic seizures
Comparison 4: Felbamate + ASMs versus placebo + ASMs, Outcome 5: Number of participants free from tonic‐clonic seizures
Comparison 4: Felbamate + ASMs versus placebo + ASMs, Outcome 6: Number of participants with severe side effects
Comparison 4: Felbamate + ASMs versus placebo + ASMs, Outcome 7: Number of participants with adverse events leading to study discontinuation
Comparison 5: Lamotrigine + ASMs versus placebo + ASMs, Outcome 1: Number of participants with ≥ 50% median reduction in all seizures
Comparison 5: Lamotrigine + ASMs versus placebo + ASMs, Outcome 2: Number of participants with > 25% to < 50% median reduction in all seizures
Comparison 5: Lamotrigine + ASMs versus placebo + ASMs, Outcome 3: Number of participants with either 0 to ≤ 25% median reduction or an increase in all seizures
Comparison 5: Lamotrigine + ASMs versus placebo + ASMs, Outcome 4: Number of participants with ≥ 50% median reduction in drop attacks
Comparison 5: Lamotrigine + ASMs versus placebo + ASMs, Outcome 5: Number of participants with > 25% to < 50% median reduction in drop attacks
Comparison 5: Lamotrigine + ASMs versus placebo + ASMs, Outcome 6: Number of participants with either ≤ 25% median reduction or an increase in the number of drop attacks
Comparison 5: Lamotrigine + ASMs versus placebo + ASMs, Outcome 7: Number of participants with ≥ 50% median reduction in tonic‐clonic seizures
Comparison 5: Lamotrigine + ASMs versus placebo + ASMs, Outcome 8: Number of participants with > 25% to < 50% median reduction in tonic‐clonic seizures
Comparison 5: Lamotrigine + ASMs versus placebo + ASMs, Outcome 9: Number of participants with 0 to ≤ 25% median reduction or an increase in the number of tonic‐clonic seizures
Comparison 5: Lamotrigine + ASMs versus placebo + ASMs, Outcome 10: Number of participants with adverse events leading to study discontinuation
Comparison 6: Rufinamide + ASMs versus placebo + ASMs, Outcome 1: Number of participants with ≥ 50% reduction in all seizures
Comparison 6: Rufinamide + ASMs versus placebo + ASMs, Outcome 2: Number of participants with ≥ 75 % reduction in tonic‐atonic seizures
Comparison 6: Rufinamide + ASMs versus placebo + ASMs, Outcome 3: Number of participants with ≥ 50% reduction in tonic‐atonic seizures
Comparison 6: Rufinamide + ASMs versus placebo + ASMs, Outcome 4: Number of participants with ≥ 25% reduction in tonic‐atonic seizures
Comparison 6: Rufinamide + ASMs versus placebo + ASMs, Outcome 5: Number of participants 'unchanged' (< 25% reduction in tonic‐atonic seizures)
Comparison 6: Rufinamide + ASMs versus placebo + ASMs, Outcome 6: Number of participants with increased tonic‐atonic seizures
Comparison 6: Rufinamide + ASMs versus placebo + ASMs, Outcome 7: Number of participants with improvement in seizure severity rating
Comparison 6: Rufinamide + ASMs versus placebo + ASMs, Outcome 8: Number of participants with adverse events
Comparison 6: Rufinamide + ASMs versus placebo + ASMs, Outcome 9: Number of participants with adverse events suspected to be treatment‐related
Comparison 6: Rufinamide + ASMs versus placebo + ASMs, Outcome 10: Number of participants with serious adverse events
Comparison 6: Rufinamide + ASMs versus placebo + ASMs, Outcome 11: Number of participants with adverse events leading to study discontinuation
Comparison 7: Rufinamide + ASMs versus other ASM + ASMs, Outcome 1: Number of participants with treatment‐emergent adverse events
Comparison 7: Rufinamide + ASMs versus other ASM + ASMs, Outcome 2: Number of participants with severe treatment‐emergent adverse events
Comparison 7: Rufinamide + ASMs versus other ASM + ASMs, Outcome 3: Number of participants with serious treatment‐emergent adverse events
Comparison 7: Rufinamide + ASMs versus other ASM + ASMs, Outcome 4: Number of participants with treatment‐emergent adverse events leading to study‐drug dose adjustment
Comparison 7: Rufinamide + ASMs versus other ASM + ASMs, Outcome 5: Number of participants with treatment‐emergent adverse events leading to study discontinuation
Comparison 7: Rufinamide + ASMs versus other ASM + ASMs, Outcome 6: Death
Comparison 7: Rufinamide + ASMs versus other ASM + ASMs, Outcome 7: Child Behaviour Checklist Questionnaire
Comparison 8: Topiramate + ASMs versus placebo + ASMs, Outcome 1: Number of participants with ≥75% reduction in all seizures
Comparison 8: Topiramate + ASMs versus placebo + ASMs, Outcome 2: Number of participants free from major seizures (drop attacks and tonic‐clonic seizures)
Comparison 8: Topiramate + ASMs versus placebo + ASMs, Outcome 3: Number of participants with ≥ 75% reduction in major seizures (drop attacks and tonic‐clonic seizures)
Comparison 8: Topiramate + ASMs versus placebo + ASMs, Outcome 4: Number of participants with ≥ 50% reduction in major seizures (drop attacks and tonic‐clonic seizures)
Comparison 8: Topiramate + ASMs versus placebo + ASMs, Outcome 5: Number of participants free from drop attacks
Comparison 8: Topiramate + ASMs versus placebo + ASMs, Outcome 6: Number of participants with ≥ 75% reduction in drop attacks
Comparison 8: Topiramate + ASMs versus placebo + ASMs, Outcome 7: Number of participants with ≥ 50% reduction in drop attacks
Comparison 8: Topiramate + ASMs versus placebo + ASMs, Outcome 8: Number of participants free from drop attacks during the maintenance phase
Comparison 8: Topiramate + ASMs versus placebo + ASMs, Outcome 9: Number of participants with severe adverse events
Comparison 8: Topiramate + ASMs versus placebo + ASMs, Outcome 10: Number of participants with adverse events leading to dose reduction or temporary discontinuation
Comparison 8: Topiramate + ASMs versus placebo + ASMs, Outcome 11: Number of participants with adverse events leading to study discontinuation
| Cannabidiol (10 mg/kg and 20 mg/kg) + ASMs compared to placebo + ASMs | ||||||
| Patient or population: Lennox‐Gastaut syndrome, any age | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo + ASMs | Risk with cannabidiol + ASMs | |||||
| Number of participants free from all seizures follow up: after 14 weeks' treatment (titration and maintenance) | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. No participants were free from drop seizures (one study; 225 participants) Analysis 1.1 |
| Number of participants with ≥ 75% reduction in all seizures follow‐up: after 14 weeks' treatment (titration and maintenance) | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. Number of participants with ≥ 75% reduction in drop seizures: RR 3.51 (95% CI 1.24 to 9.92) in favour of the cannabidiol regimen (two studies; 396 participants) Analysis 1.2. |
| Number of participants with ≥ 50% reduction in all seizures follow‐up: after 14 weeks' treatment (titration and maintenance) | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. Number of participants with ≥ 50% reduction in drop seizures: RR 2.12 (95% CI 1.48 to 3.03) in favour of the cannabidiol regimen (two studies; 396 participants; Analysis 1.3). |
| Number of participants with adverse events leading to study discontinuation | Study population | RR 6.62 | 396 | ⊕⊕⊕⊕ | Intervention in Devinsky 2018 involved two doses of cannabidiol (10 mg/kg and 20 mg/kg); 6 of the 7 adverse events leading to study discontinuation occurred in the higher dose group; intervention in Thiele 2018 also involved the higher dose of cannabidiol (20 mg/kg). | |
| 12 per 1000 | 82 per 1000 | |||||
| *The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| Cinromide + ASMs compared to placebo + ASMs | ||||||
| Patient or population: Lennox‐Gastaut syndrome, any age1 | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo + ASMs | Risk with cinromide + ASMs | |||||
| Number of participants free from all seizures | Study population | not estimable | 56 | ⊕⊝⊝⊝ | ||
| 0 per 1000 | 0 per 1000 | |||||
| Number of participants with ≥ 75% reduction in mean weekly seizures | Study population | Peto OR 9.35 | 56 | ⊕⊝⊝⊝ | Cinromide + ASMs: 3/26 participants; placebo + ASMs: 0/30 participants. | |
| see comment | see comment | |||||
| Number of participants with ≥ 50% reduction in mean weekly seizures | Study population | RR 1.15 | 56 | ⊕⊝⊝⊝ | ||
| 233 per 1000 | 268 per 1000 | |||||
| Number of participants with adverse events leading to study discontinuation follow‐up: after 18 weeks | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. |
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1The evidence for this comparison included children and adolescents only. 2Downgraded twice for study limitations because there was a high risk of bias from incomplete data (study terminated prematurely). 3Downgraded twice for imprecision because the study was not powered to detect a between‐group difference in zero event outcomes. 4Downgraded twice for imprecision because the effect estimate has a very wide confidence interval. | ||||||
| Clobazam (low, medium and high doses) + ASMs compared to placebo + ASMs | ||||||
| Patient or population: Lennox‐Gastaut syndrome, any age | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo + ASMs | Risk with clobazam + ASMs | |||||
| Number of participants free from all seizures follow‐up: "from the 4‐week baseline period to the 12‐week maintenance period" (with 3 weeks titration) | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. Number of participants free from drop seizures: RR 4.10 (95% CI 1.00 to 16.83) in favour of the clobazam regimen (1 study; 217 participants) Analysis 3.1. |
| Number of participants with ≥ 75% reduction in all seizures follow‐up: "from the 4‐week baseline period to the 12‐week maintenance period" (with 3 weeks titration) | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. |
| Number of participants with ≥ 50% reduction in all seizures follow‐up: "from the 4‐week baseline period to the 12‐week maintenance period" (with 3 weeks titration) | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. |
| Number of participants with adverse events leading to study discontinuation | Study population | RR 4.12 | 238 | ⊕⊕⊕⊕ | Intervention in Ng 2011 involved three doses of clobazam (low, medium, high) and study authors state that "A dosage related trend was observed for the overall incidence of [adverse events] leading to discontinuation." | |
| 34 per 1000 | 140 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| Felbamate + ASMs compared to placebo + ASMs | ||||||
| Patient or population: Lennox‐Gastaut syndrome, any age | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo + ASMs | Risk with felbamate + ASMs | |||||
| Number of participants free from all seizures (recorded by closed‐circuit television and electroencephalography) ‐ | Study population | RR 2.92 | 73 | ⊕⊕⊝⊝ | ||
| 28 per 1000 | 81 per 1000 | |||||
| Number of participants free from all seizures | Study population | not estimable | 73 | ⊕⊕⊝⊝ | ||
| 0 per 1000 | 0 per 1000 | |||||
| Number of participants with ≥ 75% reduction in all seizures follow‐up: after the treatment phase, which "consisted of a 14‐day titration period and a 56‐day maintenance period" | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. |
| Number of participants with ≥ 50% reduction in all seizures follow‐up: after the treatment phase, which "consisted of a 14‐day titration period and a 56‐day maintenance period" | see comment | see comment | not estimable | see comment | see comment | According to a retrospective analysis "Approximately 50% of patients randomised to FBM obtained at least a 50% reduction in seizure frequency compared with about 15% receiving placebo." |
| Number of participants with adverse events leading to study discontinuation | Study population | RR 0.97 | 73 | ⊕⊕⊝⊝ | ||
| 28 per 1000 | 27 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1Downgraded twice for imprecision because the effect estimate has a very wide confidence interval. 2Downgraded twice for imprecision because the study was not powered to detect a between‐group difference in zero event outcomes. | ||||||
| Lamotrigine + ASMs compared to placebo + ASMs | ||||||
| Patient or population: Lennox‐Gastaut syndrome, any age | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo + ASMs | Risk with lamotrigine + ASMs | |||||
| Number of participants free from all seizures follow‐up: after 16 weeks' treatment | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. |
| Number of participants with ≥ 75% reduction in all seizures follow‐up: after 16 weeks' treatment | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. |
| Number of participants with ≥ 50% median reduction in all seizures | Study population | RR 2.12 | 167 | ⊕⊕⊕⊕ | ||
| 157 per 1000 | 333 per 1000 | |||||
| Number of participants with adverse events leading to study discontinuation | Study population | RR 0.49 | 169 | ⊕⊕⊝⊝ | ||
| 78 per 1000 | 38 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1Downgraded twice because the effect estimate has a very wide confidence interval. | ||||||
| Rufinamide + ASMs compared to placebo + ASMs | ||||||
| Patient or population: Lennox‐Gastaut syndrome, any age | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo + ASMs | Risk with rufinamide + ASMs | |||||
| Number of participants free from all seizures follow‐up: after 84 days/12 weeks' treatment (titration and maintenance) | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. |
| Number of participants with ≥ 75% reduction in all seizures follow‐up: after 84 days/12 weeks' treatment (titration and maintenance) | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. |
| Number of participants with ≥ 50% reduction in all seizures | Study population | RR 2.84 | 138 | ⊕⊕⊕⊕ | ||
| 109 per 1000 | 311 per 1000 | |||||
| Number of participants with adverse events leading to study discontinuation follow‐up: after 12 weeks | 33 per 1000 | 138 per 1000 (16 to 1,000) | RR 4.14 (0.49 to 34.86) | 59 (1 RCT) | ⊕⊕⊝⊝ | |
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1Downgraded twice because the effect estimate has a very wide confidence interval. | ||||||
| Rufinamide + ASMs compared to other ASM + ASMs | ||||||
| Patient or population: Lennox‐Gastaut syndrome, any age | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo + ASMs | Risk with rufinamide + ASMs | |||||
| Number of participants free from all seizures follow‐up: after 106 weeks' treatment (titration and maintenance) | see comments | see comments | not estimable | see comments | see comments | No studies measured this outcome. |
| Number of participants with ≥ 75% reduction in all seizures follow‐up: after 106 weeks' treatment (titration and maintenance) | see comments | see comments | not estimable | see comments | see comments | No studies measured this outcome. |
| Number of participants with ≥ 50% reduction in all seizures follow‐up: after 106 weeks' treatment (titration and maintenance) | see comments | see comments | not estimable | see comments | see comments | No studies measured this outcome. |
| Number of participants with treatment‐emergent adverse events leading to study discontinuation | Study population | RR 0.96 | 37 | ⊕⊕⊝⊝ | ||
| 83 per 1000 | 80 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1Downgraded twice because the effect estimate has a very wide confidence interval. | ||||||
| Topiramate + ASMs compared to placebo + ASMs | ||||||
| Patient or population: Lennox‐Gastaut syndrome, any age | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo + ASMs | Risk with Topiramate+ ASMs | |||||
| Number of participants free from all seizures follow‐up: after 11 weeks' treatment (titration and maintenance) | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. |
| Number of participants with ≥ 75% reduction in all seizures | Study population | Peto OR 8.22 | 98 | ⊕⊕⊝⊝ | Topiramate + ASMs: 4/48 participants; placebo + ASMs: 0/50 participants. | |
| see comment | see comment | |||||
| Number of participants with ≥ 50% reduction in all seizures follow‐up: after 11 weeks' treatment (titration and maintenance) | see comment | see comment | not estimable | see comment | see comment | No studies measured this outcome. |
| Number of participants with adverse events leading to study discontinuation | Study population | not estimable | 98 | ⊕⊕⊝⊝ | ||
| 0 per 1000 | 0 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1Downgraded twice because the effect estimate has a very wide confidence interval. 2Downgraded twice because the study was not powered to detect a between‐group difference in zero event outcomes. | ||||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1.1 Number of participants free from drop seizures during the treatment phase Show forest plot | 1 | 225 | Peto Odds Ratio (Peto, Fixed, 99% CI) | Not estimable |
| 1.2 Number of participants with ≥ 75% reduction in drop seizures during the treatment phase Show forest plot | 2 | 396 | Risk Ratio (M‐H, Random, 95% CI) | 3.51 [1.24, 9.92] |
| 1.3 Number of participants with ≥ 50% reduction in drop seizures during the treatment phase Show forest plot | 2 | 396 | Risk Ratio (M‐H, Random, 95% CI) | 2.12 [1.48, 3.03] |
| 1.4 Number of participants with ≥ 25 % reduction in drop seizures during the treatment phase Show forest plot | 2 | 396 | Risk Ratio (M‐H, Random, 95% CI) | 1.45 [1.19, 1.78] |
| 1.5 Number of participants with > 0% to < 25% reduction in drop seizures during the treatment phase Show forest plot | 1 | 225 | Risk Ratio (M‐H, Random, 95% CI) | 1.58 [0.85, 2.93] |
| 1.6 Number of participants with > 0% to < 25% increase in drop seizures during the treatment phase Show forest plot | 1 | 225 | Risk Ratio (M‐H, Random, 95% CI) | 1.40 [0.66, 3.00] |
| 1.7 Number of participants with > 25 % increase in drop seizures during the treatment phase Show forest plot | 1 | 225 | Risk Ratio (M‐H, Random, 95% CI) | 0.71 [0.33, 1.53] |
| 1.8 Number of participants with improvement in the patient and caregiver Global Impression of Care scale Show forest plot | 2 | 392 | Risk Ratio (M‐H, Random, 95% CI) | 1.52 [1.22, 1.89] |
| 1.9 Number of participants free from drop seizures during the maintenance phase Show forest plot | 1 | 170 | Peto Odds Ratio (Peto, Fixed, 99% CI) | 7.76 [0.75, 79.85] |
| 1.10 Number of participants with ≥ 75% reduction in drop seizures during the maintenance phase Show forest plot | 1 | 170 | Risk Ratio (M‐H, Random, 95% CI) | 2.86 [1.28, 6.40] |
| 1.11 Number of participants with ≥ 50% reduction in drop seizures during the maintenance phase Show forest plot | 1 | 170 | Risk Ratio (M‐H, Random, 95% CI) | 1.95 [1.25, 3.05] |
| 1.12 Number of participants with ≥ 25 % reduction in drop seizures during the maintenance phase Show forest plot | 1 | 170 | Risk Ratio (M‐H, Random, 95% CI) | 1.35 [1.02, 1.78] |
| 1.13 Number of participants with adverse events Show forest plot | 2 | 396 | Risk Ratio (M‐H, Random, 95% CI) | 1.24 [1.11, 1.38] |
| 1.14 Number of participants with treatment‐related adverse events Show forest plot | 1 | 171 | Risk Ratio (M‐H, Random, 95% CI) | 1.81 [1.29, 2.54] |
| 1.15 Number of participants with serious adverse events Show forest plot | 1 | 171 | Risk Ratio (M‐H, Random, 95% CI) | 4.94 [1.76, 13.85] |
| 1.16 Number of participants with adverse events leading to dose reduction Show forest plot | 1 | 171 | Risk Ratio (M‐H, Random, 95% CI) | 5.93 [0.73, 48.22] |
| 1.17 Number of participants with adverse events leading to study discontinuation Show forest plot | 2 | 396 | Risk Ratio (M‐H, Random, 95% CI) | 6.62 [1.56, 28.15] |
| 1.18 Death Show forest plot | 1 | 171 | Peto Odds Ratio (Peto, Fixed, 99% CI) | 7.30 [0.04, 1261.58] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 2.1 Number of participants free from all seizures Show forest plot | 1 | 56 | Peto Odds Ratio (Peto, Fixed, 99% CI) | Not estimable |
| 2.2 Number of participants with ≥ 75% reduction in mean weekly seizures Show forest plot | 1 | 56 | Peto Odds Ratio (Peto, Fixed, 99% CI) | 9.35 [0.45, 194.96] |
| 2.3 Number of participants with ≥ 50% reduction in mean weekly seizures Show forest plot | 1 | 56 | Risk Ratio (M‐H, Random, 95% CI) | 1.15 [0.47, 2.86] |
| 2.4 Number of participants with ≥ 25% reduction in mean weekly seizures Show forest plot | 1 | 56 | Risk Ratio (M‐H, Random, 95% CI) | 1.07 [0.59, 1.91] |
| 2.5 Number of participants with ≥ 0% to < 25% reduction in mean weekly seizures Show forest plot | 1 | 56 | Risk Ratio (M‐H, Random, 95% CI) | 1.15 [0.42, 3.14] |
| 2.6 Number of participants with ≥ 0% to < 25% increase in mean weekly seizures Show forest plot | 1 | 56 | Risk Ratio (M‐H, Random, 95% CI) | 0.87 [0.21, 3.52] |
| 2.7 Number of participants with > 25% increase in mean weekly seizures Show forest plot | 1 | 56 | Risk Ratio (M‐H, Random, 95% CI) | 0.82 [0.30, 2.29] |
| 2.8 Number of participants with improvement in global evaluation Show forest plot | 1 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 2.8.1 Week 12 | 1 | 56 | Risk Ratio (M‐H, Random, 95% CI) | 0.99 [0.56, 1.74] |
| 2.8.2 Week 18 | 1 | 56 | Risk Ratio (M‐H, Random, 95% CI) | 0.80 [0.41, 1.56] |
| 2.8.3 Week 24 | 1 | 56 | Risk Ratio (M‐H, Random, 95% CI) | 1.28 [0.62, 2.66] |
| 2.9 Number of participants with no change in global evaluation Show forest plot | 1 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 2.9.1 Week 12 | 1 | 56 | Risk Ratio (M‐H, Random, 95% CI) | 0.91 [0.50, 1.64] |
| 2.9.2 Week 18 | 1 | 56 | Risk Ratio (M‐H, Random, 95% CI) | 0.91 [0.50, 1.64] |
| 2.9.3 Week 24 | 1 | 56 | Risk Ratio (M‐H, Random, 95% CI) | 1.03 [0.46, 2.27] |
| 2.10 Number of participants with worsening in global evaluation Show forest plot | 1 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 2.10.1 Week 12 | 1 | 56 | Risk Ratio (M‐H, Random, 95% CI) | 3.46 [0.38, 31.28] |
| 2.10.2 Week 18 | 1 | 56 | Risk Ratio (M‐H, Random, 95% CI) | 4.62 [0.55, 38.74] |
| 2.11 Number of participants with worsening in global evaluation ‐ Week 24 Show forest plot | 1 | 56 | Odds Ratio (M‐H, Fixed, 99% CI) | 9.09 [0.17, 475.60] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 3.1 Number of participants free from drop seizures Show forest plot | 1 | 217 | Risk Ratio (M‐H, Random, 95% CI) | 4.10 [1.00, 16.83] |
| 3.2 Number of participants with ≥ 75 % reduction in drop seizures (from baseline to maintenance phase in average weekly rate) Show forest plot | 1 | 217 | Risk Ratio (M‐H, Random, 95% CI) | 4.04 [1.85, 8.79] |
| 3.3 Number of participants with ≥ 50 % reduction in drop seizures (from baseline to maintenance phase in average weekly rate) Show forest plot | 1 | 217 | Risk Ratio (M‐H, Random, 95% CI) | 1.88 [1.26, 2.81] |
| 3.4 Number of participants with ≥ 25 % reduction in drop seizures (from baseline to maintenance phase in average weekly rate) Show forest plot | 1 | 217 | Risk Ratio (M‐H, Random, 95% CI) | 1.54 [1.17, 2.03] |
| 3.5 Number of participants with adverse events Show forest plot | 1 | 238 | Risk Ratio (M‐H, Random, 95% CI) | 1.17 [0.97, 1.42] |
| 3.6 Number of participants with adverse events leading to dose reduction Show forest plot | 1 | 238 | Risk Ratio (M‐H, Random, 95% CI) | 9.23 [1.28, 66.37] |
| 3.7 Number of participants with aggression‐related adverse events Show forest plot | 1 | 194 | Risk Ratio (M‐H, Random, 95% CI) | 1.89 [0.69, 5.19] |
| 3.8 Number of participants with serious adverse events Show forest plot | 1 | 238 | Risk Ratio (M‐H, Random, 95% CI) | 2.31 [0.54, 9.86] |
| 3.9 Number of participants with adverse events leading to study discontinuation Show forest plot | 1 | 238 | Risk Ratio (M‐H, Random, 95% CI) | 4.12 [1.01, 16.87] |
| 3.10 Death Show forest plot | 1 | 238 | Peto Odds Ratio (Peto, Fixed, 99% CI) | Not estimable |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 4.1 Number of participants free from all seizures (recorded by closed‐circuit television and electroencephalography) Show forest plot | 1 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 4.1.1 Treatment phase | 1 | 73 | Risk Ratio (M‐H, Random, 95% CI) | 2.92 [0.32, 26.77] |
| 4.1.2 Maintenance phase | 1 | 73 | Risk Ratio (M‐H, Random, 95% CI) | 5.84 [0.74, 46.11] |
| 4.2 Number of participants free from all seizures ‐ Treatment phase Show forest plot | 1 | 73 | Odds Ratio (M‐H, Fixed, 99% CI) | Not estimable |
| 4.3 Number of participants free from all seizures ‐ Maintenance phase Show forest plot | 1 | 73 | Risk Ratio (M‐H, Random, 95% CI) | 3.89 [0.46, 33.17] |
| 4.3.1 Maintenance phase | 1 | 73 | Risk Ratio (M‐H, Random, 95% CI) | 3.89 [0.46, 33.17] |
| 4.4 Number of participants free from atonic seizures Show forest plot | 1 | Peto Odds Ratio (Peto, Fixed, 99% CI) | Subtotals only | |
| 4.4.1 Treatment phase | 1 | 50 | Peto Odds Ratio (Peto, Fixed, 99% CI) | 6.43 [0.30, 137.10] |
| 4.4.2 Maintenance phase | 1 | 50 | Peto Odds Ratio (Peto, Fixed, 99% CI) | 6.99 [0.62, 78.73] |
| 4.5 Number of participants free from tonic‐clonic seizures Show forest plot | 1 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 4.5.1 Treatment phase | 1 | 29 | Risk Ratio (M‐H, Random, 95% CI) | 1.62 [0.17, 15.99] |
| 4.5.2 Maintenance phase | 1 | 29 | Risk Ratio (M‐H, Random, 95% CI) | 5.69 [0.80, 40.51] |
| 4.6 Number of participants with severe side effects Show forest plot | 1 | 73 | Risk Ratio (M‐H, Random, 95% CI) | 2.59 [0.75, 9.01] |
| 4.7 Number of participants with adverse events leading to study discontinuation Show forest plot | 1 | 73 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.97 [0.06, 14.97] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 5.1 Number of participants with ≥ 50% median reduction in all seizures Show forest plot | 1 | 167 | Risk Ratio (M‐H, Random, 95% CI) | 2.12 [1.19, 3.76] |
| 5.2 Number of participants with > 25% to < 50% median reduction in all seizures Show forest plot | 1 | 167 | Risk Ratio (M‐H, Random, 95% CI) | 1.41 [0.80, 2.47] |
| 5.3 Number of participants with either 0 to ≤ 25% median reduction or an increase in all seizures Show forest plot | 1 | 167 | Risk Ratio (M‐H, Random, 95% CI) | 0.61 [0.45, 0.83] |
| 5.4 Number of participants with ≥ 50% median reduction in drop attacks Show forest plot | 1 | 164 | Risk Ratio (M‐H, Random, 95% CI) | 1.66 [1.02, 2.70] |
| 5.5 Number of participants with > 25% to < 50% median reduction in drop attacks Show forest plot | 1 | 164 | Risk Ratio (M‐H, Random, 95% CI) | 1.61 [0.87, 2.99] |
| 5.6 Number of participants with either ≤ 25% median reduction or an increase in the number of drop attacks Show forest plot | 1 | 164 | Risk Ratio (M‐H, Random, 95% CI) | 0.60 [0.43, 0.85] |
| 5.7 Number of participants with ≥ 50% median reduction in tonic‐clonic seizures Show forest plot | 1 | 124 | Risk Ratio (M‐H, Random, 95% CI) | 2.13 [1.21, 3.75] |
| 5.8 Number of participants with > 25% to < 50% median reduction in tonic‐clonic seizures Show forest plot | 1 | 124 | Risk Ratio (M‐H, Random, 95% CI) | 0.91 [0.33, 2.57] |
| 5.9 Number of participants with 0 to ≤ 25% median reduction or an increase in the number of tonic‐clonic seizures Show forest plot | 1 | 124 | Risk Ratio (M‐H, Random, 95% CI) | 0.68 [0.49, 0.93] |
| 5.10 Number of participants with adverse events leading to study discontinuation Show forest plot | 1 | 169 | Risk Ratio (M‐H, Random, 95% CI) | 0.49 [0.13, 1.82] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 6.1 Number of participants with ≥ 50% reduction in all seizures Show forest plot | 1 | 138 | Risk Ratio (M‐H, Random, 95% CI) | 2.84 [1.31, 6.18] |
| 6.2 Number of participants with ≥ 75 % reduction in tonic‐atonic seizures Show forest plot | 1 | 58 | Risk Ratio (M‐H, Random, 95% CI) | 10.71 [1.46, 78.39] |
| 6.3 Number of participants with ≥ 50% reduction in tonic‐atonic seizures Show forest plot | 2 | 191 | Risk Ratio (M‐H, Random, 95% CI) | 2.70 [1.52, 4.81] |
| 6.4 Number of participants with ≥ 25% reduction in tonic‐atonic seizures Show forest plot | 1 | 58 | Risk Ratio (M‐H, Random, 95% CI) | 1.88 [0.93, 3.77] |
| 6.5 Number of participants 'unchanged' (< 25% reduction in tonic‐atonic seizures) Show forest plot | 1 | 58 | Risk Ratio (M‐H, Random, 95% CI) | 1.38 [0.59, 3.20] |
| 6.6 Number of participants with increased tonic‐atonic seizures Show forest plot | 1 | 58 | Risk Ratio (M‐H, Random, 95% CI) | 0.36 [0.15, 0.85] |
| 6.7 Number of participants with improvement in seizure severity rating Show forest plot | 1 | 135 | Risk Ratio (M‐H, Random, 95% CI) | 1.74 [1.13, 2.68] |
| 6.8 Number of participants with adverse events Show forest plot | 2 | 197 | Risk Ratio (M‐H, Random, 95% CI) | 1.13 [0.86, 1.50] |
| 6.9 Number of participants with adverse events suspected to be treatment‐related Show forest plot | 1 | 138 | Risk Ratio (M‐H, Random, 95% CI) | 1.27 [0.90, 1.79] |
| 6.10 Number of participants with serious adverse events Show forest plot | 1 | 138 | Risk Ratio (M‐H, Random, 95% CI) | 0.86 [0.13, 5.97] |
| 6.11 Number of participants with adverse events leading to study discontinuation Show forest plot | 1 | 59 | Risk Ratio (M‐H, Random, 95% CI) | 4.14 [0.49, 34.86] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 7.1 Number of participants with treatment‐emergent adverse events Show forest plot | 1 | 37 | Risk Ratio (M‐H, Random, 95% CI) | 1.06 [0.79, 1.41] |
| 7.2 Number of participants with severe treatment‐emergent adverse events Show forest plot | 1 | 37 | Risk Ratio (M‐H, Random, 95% CI) | 0.96 [0.20, 4.53] |
| 7.3 Number of participants with serious treatment‐emergent adverse events Show forest plot | 1 | 37 | Risk Ratio (M‐H, Random, 95% CI) | 0.96 [0.42, 2.19] |
| 7.4 Number of participants with treatment‐emergent adverse events leading to study‐drug dose adjustment Show forest plot | 1 | Peto Odds Ratio (Peto, Fixed, 99% CI) | Subtotals only | |
| 7.4.1 Reduction | 1 | 37 | Peto Odds Ratio (Peto, Fixed, 99% CI) | 5.91 [0.61, 57.64] |
| 7.4.2 Interruption | 1 | 37 | Peto Odds Ratio (Peto, Fixed, 99% CI) | 0.04 [0.00, 2.17] |
| 7.5 Number of participants with treatment‐emergent adverse events leading to study discontinuation Show forest plot | 1 | 37 | Risk Ratio (M‐H, Random, 95% CI) | 0.96 [0.10, 9.57] |
| 7.6 Death Show forest plot | 1 | 37 | Peto Odds Ratio (Peto, Fixed, 99% CI) | 4.39 [0.02, 1077.58] |
| 7.7 Child Behaviour Checklist Questionnaire Show forest plot | 1 | 19 | Mean Difference (IV, Random, 95% CI) | 2.60 [‐10.30, 15.50] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 8.1 Number of participants with ≥75% reduction in all seizures Show forest plot | 1 | 98 | Peto Odds Ratio (Peto, Fixed, 99% CI) | 8.22 [0.60, 112.62] |
| 8.2 Number of participants free from major seizures (drop attacks and tonic‐clonic seizures) Show forest plot | 1 | 96 | Risk Ratio (M‐H, Random, 95% CI) | 4.08 [1.46, 11.39] |
| 8.3 Number of participants with ≥ 75% reduction in major seizures (drop attacks and tonic‐clonic seizures) Show forest plot | 1 | 96 | Risk Ratio (M‐H, Random, 95% CI) | 4.35 [0.97, 19.42] |
| 8.4 Number of participants with ≥ 50% reduction in major seizures (drop attacks and tonic‐clonic seizures) Show forest plot | 1 | 96 | Risk Ratio (M‐H, Random, 95% CI) | 4.08 [1.46, 11.39] |
| 8.5 Number of participants free from drop attacks Show forest plot | 1 | 96 | Peto Odds Ratio (Peto, Fixed, 99% CI) | 8.06 [0.05, 1398.41] |
| 8.6 Number of participants with ≥ 75% reduction in drop attacks Show forest plot | 1 | 95 | Risk Ratio (M‐H, Random, 95% CI) | 2.84 [0.80, 10.06] |
| 8.7 Number of participants with ≥ 50% reduction in drop attacks Show forest plot | 1 | 95 | Risk Ratio (M‐H, Random, 95% CI) | 1.98 [0.87, 4.52] |
| 8.8 Number of participants free from drop attacks during the maintenance phase Show forest plot | 1 | 95 | Risk Ratio (M‐H, Random, 95% CI) | 1.33 [0.38, 4.66] |
| 8.9 Number of participants with severe adverse events Show forest plot | 1 | 98 | Risk Ratio (M‐H, Random, 95% CI) | 2.29 [0.86, 6.11] |
| 8.10 Number of participants with adverse events leading to dose reduction or temporary discontinuation Show forest plot | 1 | 98 | Risk Ratio (M‐H, Random, 95% CI) | 3.12 [0.90, 10.85] |
| 8.11 Number of participants with adverse events leading to study discontinuation Show forest plot | 1 | 98 | Peto Odds Ratio (Peto, Fixed, 99% CI) | Not estimable |
