Melatonina como tratamiento complementario para la epilepsia
Resumen
Antecedentes
Ésta es una versión actualizada de la revisión Cochrane original publicada en el número 6, 2012.
La epilepsia es uno de los trastornos neurológicos crónicos más frecuentes. A pesar de la gran cantidad de fármacos antiepilépticos disponibles actualmente el 30% de los pacientes aún presenta crisis convulsivas. Este grupo de pacientes requiere un tratamiento más intensivo debido a que la monoterapia, el esquema de primera elección, no logra controlar las crisis convulsivas. No obstante, con frecuencia la politerapia provoca algunos efectos indeseables que incluyen trastornos neurológicos (somnolencia, ataxia, mareos), síntomas psiquiátricos y conductuales y alteración metabólica (osteoporosis, inducción o inhibición de las enzimas hepáticas, etc.). La necesidad de fármacos antiepilépticos que muestren una mejor tolerabilidad es aun más urgente en este grupo de pacientes. Los informes han indicado una función antiepiléptica de la melatonina con un buen perfil de seguridad.
Objetivos
Evaluar la eficacia y la tolerabilidad de la melatonina como tratamiento complementario para la epilepsia.
Métodos de búsqueda
Para la última actualización, se realizaron búsquedas en el Registro especializado del Grupo Cochrane de Epilepsia (Cochrane Epilepsy Group) (12 de enero de 2016), en el Registro Cochrane Central de Ensayos Controlados (CENTRAL) a través del Registro Cochrane de Estudios en Línea (CRSO, 12 de enero de 2016) y en MEDLINE (Ovid, 11 de enero de 2016). Para obtener referencias adicionales se realizaron búsquedas en las bibliografías de cualquier estudio identificado. Se hicieron búsquedas manuales en revistas seleccionadas y resúmenes de congresos. No se aplicaron restricciones de idioma. Además, se estableció contacto con los fabricantes de melatonina (es decir, Nathura) y con investigadores originales para identificar cualquier estudio no publicado.
Criterios de selección
Ensayos controlados aleatorizados; ensayos doble ciego, simple ciego o sin cegamiento; estudios de grupos paralelos o cruzados. Pacientes con epilepsia, independientemente de su edad o sexo, con la inclusión de niños y adultos con discapacidades. Administración de melatonina como tratamiento complementario a cualquier fármaco antiepiléptico en comparación con placebo complementario o ningún tratamiento complementario.
Obtención y análisis de los datos
Los autores de la revisión, de forma independiente, seleccionaron los ensayos para inclusión según criterios predefinidos, extrajeron los datos pertinentes y evaluaron la calidad metodológica de los ensayos. Se evaluaron los siguientes resultados: al menos el 50% de reducción de las convulsiones, períodos libres de convulsiones, eventos adversos y calidad de vida.
Resultados principales
Se incluyeron seis publicaciones con 125 participantes (106 menores de 18 años). Hubo dos comparaciones diferentes disponibles: melatonina versus placebo y melatonina 5 mg versus melatonina 10 mg. A pesar de la intención primaria de la revisión, y debido a la información insuficiente sobre los resultados, no fue posible realizar un metanálisis, pero los datos se resumieron de forma narrativa. Cuatro estudios fueron ensayos aleatorizados doble ciego, cruzados (crossover) y controlados con placebo y dos fueron ensayos aleatorizados doble ciego, paralelos y controlados con placebo. Sólo dos estudios proporcionaron el número exacto de convulsiones durante el ensayo en comparación con el inicio: ninguno de los participantes con convulsiones durante el ensayo tuvo un cambio en la frecuencia de las convulsiones en comparación con el inicio. Dos estudios evaluaron los efectos adversos de forma sistemática (se informó de un empeoramiento de la cefalea en un niño con migraña bajo tratamiento con melatonina). Solo un estudio evaluó de forma sistemática la calidad de vida y no mostró mejoras estadísticamente significativas en la calidad de vida del grupo de melatonina complementaria.
Conclusiones de los autores
Los estudios incluidos fueron de calidad metodológica baja y no evaluaron de forma sistemática la frecuencia de las crisis convulsivas ni los eventos adversos, por lo que no fue posible resumir los datos en un metanálisis. No es posible establecer conclusiones acerca de la función de la melatonina para la reducción de la frecuencia de las crisis convulsivas ni la mejora en la calidad de vida de los pacientes con epilepsia.
PICO
Resumen en términos sencillos
Uso de melatonina como tratamiento complementario para la epilepsia
Antecedentes
La epilepsia es uno de los trastornos a largo plazo más comunes del sistema nervioso y, a pesar de que se dispone de varios fármacos antiepilépticos, el 30% de los pacientes continúan con convulsiones (crisis). Los informes han indicado que la melatonina puede funcionar en la epilepsia con un buen perfil de seguridad. La melatonina es producida por el cuerpo y es recetada por los médicos para tratar los trastornos del sueño y problemas como el desfase horario (“jet lag”).
Características de los estudios
Se buscaron en las bases de datos médicas ensayos clínicos de melatonina agregada a otro fármaco antiepiléptico (tratamiento complementario) en comparación con el fármaco antiepiléptico más un tratamiento complementario simulado (placebo) o ningún tratamiento complementario en pacientes con epilepsia. Los participantes fueron de cualquier edad o sexo e incluyeron niños y adultos con discapacidades. Los estudios midieron la reducción de la frecuencia de las convulsiones a la mitad, la proporción de pacientes sin convulsiones (períodos libres de convulsiones), los efectos secundarios y la mejora de la calidad de vida.
Resultados clave
En la presente revisión se encontraron seis ensayos con 125 participantes. Estos ensayos informaron de dos comparaciones diferentes: melatonina versus placebo y melatonina 5 mg versus melatonina 10 mg.
Los ensayos incluidos no evaluaron de manera metódica la frecuencia de las convulsiones, los períodos libres de convulsiones ni los eventos adversos. Sólo un estudio informó sobre la frecuencia de las convulsiones y ninguno de los participantes tuvo un cambio en la frecuencia durante el ensayo en comparación con antes del mismo. Sólo un ensayo evaluó el efecto de la melatonina en la calidad de vida y no encontró mejora con el agregado de melatonina en comparación con el agregado de placebo.
Calidad de la evidencia
Los ensayos incluidos fueron de calidad metodológica baja y no fue posible establecer conclusiones definitivas sobre la función de la melatonina en la reducción de la frecuencia de las convulsiones o en la mejora de la calidad de vida de los pacientes con epilepsia.
La evidencia está actualizada hasta enero de 2016.
Authors' conclusions
Background
This review is an update of a previously published review in The Cochrane Database of Systematic Reviews (Issue 6, 2012) on 'Melatonin as add‐on treatment for epilepsy' (Brigo 2012).
Epilepsy is defined as the occurrence of at least two unprovoked epileptic seizures (Commission ILAE 1989). It is one of the most common neurological disorders: in Western countries the incidence in adults is 50/100,000 population per year (Hauser 1998) with a prevalence of 5/1000 to 10/1000 people (Goodridge 1983; McDonald 2000). In children from birth to 15 years of age, the incidence is 5/10,000 to 7/10,000 children per year and prevalence 5/1000 children, the differences being mainly due either to benign epilepsy syndromes that remit spontaneously or to severe pathologies with neurological involvement causing death.
Despite the plethora of antiepileptic drugs (AEDs) developed since the introduction of phenobarbital in 1912, 30% of people continue having seizures (Annegers 1979; Camfield 1996; Cockerell 1995; Elwes 1984; Goodridge 1983; McDonald 2000; Shorvon 1982). This group of people requires a more aggressive treatment, since monotherapy, the first choice scheme, fails to control seizures. Nevertheless, polytherapy often results in a number of unwanted effects, including neurological disturbances (somnolence, ataxia, dizziness), psychiatric and behavioural symptoms, and metabolic alteration (osteoporosis, inducement or inhibition of hepatic enzymes, etc.). The need for better tolerated AEDs is even more urgent in this group of people.
Research has suggested an antiepileptic role of melatonin, an indolamine synthesized from tryptophan in the pineal gland and released in a circadian pattern (Brzezinski 1997). In clinical practice, melatonin is used to treat sleep‐wake cycle disorders, mainly jet lag syndrome and shift worker disturbances (Herxheimer 2003; Herxheimer 2005; Revell 2006), as well as for the treatment of sleep disorders in children with neurological and developmental problems (Cortesi 2010; Weiss 2010).
Both in vitro and in vivo studies have suggested an antiepileptic activity of melatonin (Anton‐Tay 1974; Je 1996; Molina‐Carballo 1997; Mevissen 1998; Fauteck 1999), mediated by an antioxidant effect (Kabuto 1998), an increase in γ‐aminobutyric acid (GABA) concentration (Niles 1987) and GABA receptor affinity (Acuna‐Castroviejo 1986), or a reduction of the N‐methyl‐D‐aspartate (NMDA) excitatory effect (Munoz‐Hoyos 1998). In contrast, one author has reported a proconvulsant effect of melatonin (Sheldon 1998). Although studies investigating the long‐term effects of melatonin are still lacking, high melatonin doses have so far proved safe (Seabra 2000).
Objectives
To evaluate the efficacy and tolerability of melatonin as add‐on treatment for epilepsy.
Methods
Criteria for considering studies for this review
Types of studies
-
Randomized controlled trials.
-
Double, single, or unblinded trials.
-
Parallel group or cross‐over studies.
Types of participants
People with epilepsy, defined as with at least two unprovoked seizures, diagnosed by a physician, regardless of age, sex, ethnicity, and diagnosis, including children and adults with disabilities.
Types of interventions
Administration of melatonin as add‐on treatment to any AED(s) compared to add‐on placebo or no add‐on treatment.
Types of outcome measures
Primary outcomes
At least a 50% reduction in frequency of seizures of any type
The proportion of people with a 50% or greater reduction in seizure frequency during the treatment period compared to the pre‐randomization baseline period.
Seizure freedom
The proportion of people with cessation of seizures during the treatment period.
Adverse events
The proportions of people with adverse events.
Secondary outcomes
Improvement in quality of life
The proportion of people reported to have a better quality of life according to validated questionnaires (e.g. Quality of Life in Childhood Epilepsy (QOLCE), United States Quality of life in Childhood Epilepsy (USQOLCE), 36‐item Short Form (SF‐36) health profile including quality of life, etc.)
Search methods for identification of studies
Electronic searches
Searches were run for the original review in March 2008. Subsequent searches were run in May 2011, May 2012, July 2014, and January 2016.
For the latest update, we searched the following databases:
-
Cochrane Epilepsy Group Specialized Register (12 January 2016) using the search strategy outlined in Appendix 1;
-
the Cochrane Central Register of Controlled Trials (CENTRAL) via the Cochrane Register of Studies Online (CRSO, 12 January 2016) using the search strategy outlined in Appendix 2;
-
MEDLINE (Ovid, 1946 to 11 January 2016) using the search strategy outlined in Appendix 3.
Searching other resources
We searched the bibliographies of any included studies for further references. We handsearched selected journals and conference proceedings. We applied no language restrictions. In addition, we contacted melatonin manufacturers (i.e. Nathura) and original investigators to identify any unpublished study.
Data collection and analysis
Selection of trials
Both review authors (FB and SCI) independently assessed trials for inclusion and resolved any disagreements by discussion.
Assessment of methodological quality
Both review authors (FB and SCI) independently assessed the methodological quality of all the included studies and recorded the findings. We noted the following methodology aspects: study design, type of control, method of allocation, concealment, and completeness of follow‐up. We used preprinted selection forms to evaluate methodological quality.
Data extraction
One review author (FB) extracted the data onto a pre‐specified data extraction form, and the other review author (SCI) independently checked the data. We pilot tested the data collection forms to improve reliability. Data reported by published sources were used in this trial.
Data analysis
We extracted the following data for trials meeting our inclusion criteria.
Methodological/trial design
-
Method of generation of random list.
-
Method of concealment of randomization.
-
Blinding methods.
Participants' covariates
-
Age.
-
Sex.
-
Seizure type.
-
Epileptic syndrome.
-
Presence of neurological signs/intellectual disabilities.
-
Electroencephalogram.
-
Neuroradiology (computed tomography, magnetic resonance imaging).
-
Duration of disease prior to treatment.
-
Monotherapy versus polytherapy before randomizations.
Outcomes data
-
Fifty per cent or greater reduction in seizure frequency: proportion of participants with at least 50% or greater reduction in seizure frequency at the end of the study (numerator)/number of participants at pre‐randomization baseline period (denominator).
-
Seizure freedom: proportion of participants achieving cessation of seizures (numerator)/number of participants at pre‐randomization baseline period (denominator).
-
Incidence of adverse events of any type: number of adverse events (numerator)/total number of participants at pre‐randomization baseline period (denominator).
-
Improvement in quality of life as assessed by validated and reliable rating scales (quality of life rating scores).
Data analysis plan
We sought data on the number of participants in the treatment groups and with each outcome, irrespective of compliance or completeness of follow‐up, in the articles to undertake an intention‐to‐treat analysis.
Despite our primary intention, due to insufficient information on outcomes, we were unable to perform any meta‐analyses.
Therefore, we planned to:
-
extract data from the trials;
-
summarize efficacy (seizure frequency and seizure freedom) data narratively;
-
document tolerability (incidence of individual adverse events) narratively;
-
summarize quality of life data narratively.
Results
Description of studies
See: Characteristics of included studies and Characteristics of excluded studies tables.
Results of the search
See Figure 1.
Study flow diagram. This diagram refers only to the updated version of the review.
The update of searches for this review yielded 24 results (three from Cochrane Epilepsy Group Specialized Register, 11 from CENTRAL, and 10 from MEDLINE). After removing seven duplicates, we identified 17 articles for possible inclusion. After careful evaluation of titles and abstracts, we excluded 14 articles. We added two studies (Goldberg‐Stern 2012; Jain 2015), to the four studies already included in the previous version of this review (Coppola 2004; Gupta 2004a; Gupta 2004b; Hancock 2005).
Hence, six publications with 125 participants (mostly children) fulfilled the inclusion criteria of the present review. Four publications were randomized, double‐blind, cross‐over, placebo‐controlled trials (Coppola 2004; Hancock 2005; Goldberg‐Stern 2012; Jain 2015), and two studies were randomized, double‐blind, parallel, placebo‐controlled trials (Gupta 2004a; Gupta 2004b).
Included studies
See: Characteristics of included studies table.
Coppola 2004
Coppola 2004 was a randomized, double‐blind, cross‐over, placebo‐controlled trial conducted in people aged over 12 months, with mental retardation with/without seizures, and diagnosed with wake‐sleep disorders. The study enrolled 32 participants, seven participants (28%) were lost to the study. Twenty‐five participants (16 males, nine females), aged from 3.6 to 26 years (mean 10.5 years), mostly (18/25) with epileptic seizures, completed both melatonin and placebo phases. Participants were randomized to oral synthetic fast‐release melatonin or placebo. Phase one (melatonin or placebo) lasted four weeks and, after a cross‐over period of one week, each participant entered phase two (melatonin or placebo), which lasted four weeks. Melatonin was initiated at a daily dose of 3 mg, at nocturnal bedtime. In case of inefficacy, melatonin dose could be titrated up to 9 mg during the following two weeks in increments of 3 mg/week, unless the participant was unable to tolerate it. The dose of pre‐existing medication was maintained throughout the trial. At the end of phase two, responders to melatonin entered an open‐label phase of two months. The study reported seizure freedom and seizure frequency without clearly specifying the exact number of seizures; authors stated only that out of the 11 seizure‐free participants before starting the study, nine remained unchanged on melatonin treatment. There was no overall significant change with regards to seizure control with melatonin. Melatonin was well tolerated in all participants and no adverse effects were reported.
Gupta 2004a
Gupta 2004a was a randomized, double‐blind, parallel, placebo‐controlled trial. It was conducted in children with epilepsy aged 3 to 12 years on carbamazepine monotherapy to evaluate the effects of add‐on melatonin administration on quality of life using a parental questionnaire (Sleep Behavior Questionnaire). Children had to be seizure‐free for at least the last six months before enrolment. Of the 31 children enrolled, 13 (mean age 8.3 years) randomly received add‐on melatonin, and 15 (mean age 8.1 years) received add‐on placebo. Two children in the placebo group and one child in the melatonin group were lost to follow‐up. The questionnaire was administered before add‐on melatonin/placebo and four weeks after (28 to 32 days). No data on seizure freedom or seizure frequency were reported. No adverse events warranting discontinuation of the therapy were reported.
Gupta 2004b
Gupta 2004b was a subsequent study performed by the same investigators (Gupta 2004b). It was a randomized, double‐blind, placebo‐controlled trial. It was also conducted in children with an epilepsy, aged 3 to 12 years on valproate monotherapy to evaluate the effect of add‐on melatonin on quality of life using a parental questionnaire (QOLCE). Children had to be seizure‐free for six months before enrolment. Of the 31 children enrolled, 16 (mean age 7.4 years) randomly received add‐on melatonin, and 14 (mean age 6.6 years) received add‐on placebo. One child in the placebo group was lost to follow‐up. The questionnaire was administered before add‐on melatonin/placebo and four weeks after (28 to 32 days). No data on seizure freedom or seizure frequency were reported. No adverse events warranting discontinuation of therapy were reported.
Hancock 2005
Hancock 2005 was a randomized, double‐blind, controlled, cross‐over trial investigating the response to oral melatonin using two dose regimens in people with sleep disorders associated with tuberous sclerosis complex (Hancock 2005). Eight outpatients with tuberous sclerosis complex and sleep disorder (aged 18 months to 31 years) received either 5 or 10 mg of melatonin. All participants had epilepsy and received concurrent AEDs, and all had mental retardation and behavioural difficulties. The trial consisted of following phases: an initial two‐week baseline period to confirm the sleep disorder and familiarize the participant or carer with the requirements of the trial with no treatment; a two‐week period of treatment with either 5 or 10 mg of melatonin; a two‐week washout period with no treatment; and a two‐week period of treatment with the alternative dose of melatonin (i.e. 5 mg after 10 mg or 10 mg after 5 mg). Sleep latency, total sleep time, number of awakenings, and seizure frequency were recorded in sleep and seizure diaries. There was no evidence of a dose effect between 5 and 10 mg with respect to any outcome measure. The exact number of seizures occurring during the trial, compared to baseline, was given. None of the children who had seizures during the trial had a change in seizure frequency compared with the baseline period before melatonin treatment at either dose. Trial duration (including follow‐up) was six weeks. During the study, three out of seven participants who completed the study remained seizure‐free. There were no adverse effects reported or observed. Quality of life was not considered as an outcome.
Goldberg‐Stern 2012
Goldberg‐Stern 2012 was a randomized, double‐blind, cross‐over, placebo‐controlled trial conducted in people aged four years or over with drug‐resistant epilepsy, defined as failure of three or more AEDs to control seizures. The inclusion criteria were no changes in AEDs dosage (therapeutic stability) for two months prior to the study and four or more seizures in the three weeks prior to the initiation of the study. Twelve participants (five male, seven female) aged between 9 and 32 years were included. The analysis excluded two participants because of lack of compliance in completing the diaries to assess seizure occurrence, and behaviour and sleep features. Five out of the 10 participants analyzed were younger than 18 years. All adults had childhood‐onset epilepsy. Participants were randomly assigned to receive treatment with melatonin (10 mg daily at bedtime) or placebo for three weeks (stage one). After a one‐week washout period, participants were switched to the other group (either placebo or melatonin) for three weeks (stage two). Trial duration (including follow‐up) was seven weeks. The mean number of diurnal seizures was 7.75 during placebo treatment and 4.6 during melatonin treatment (P value = 0.034). Three participants showed a decrease of 50% or greater in diurnal seizures during melatonin treatment compared to placebo. There were no adverse effects reported or observed. Quality of life was not considered as an outcome.
Jain 2015
Jain 2015 was a randomized, double‐blind, cross‐over study conducted on 11 children aged 6 to 11 years with epilepsy and without developmental delay. Participants were randomized to receive placebo or a 9 mg sustained release melatonin formulation given about 30 minutes before bedtime for four weeks, followed by a one‐week washout and a four‐week cross‐over condition. Seizure frequency was assessed by seizure diary reported by parents or carers. Thirteen participants were enrolled, 11 were randomized, and 10 (91%) completed the entire study and served the cohort for the analysis. Adherence was assessed at greater than 93% by counting tablets remaining at each treatment visit. Data were available for all 10 participants.
Four participants reported adverse events while taking melatonin (increased severity of headache in a child with history of migraine, bronchitis and ear infection, agitation, and increased urinary frequency) as compared to two participants taking placebo (agitation and increased urinary frequency).
Regarding antiepileptic efficacy, only two participants had ongoing epilepsy, although they had 50% reduction in seizure frequency with melatonin. Eight participants who were seizure‐free remained seizure‐free during the study. There was no increase in seizure frequency or difference in epilepsy frequency between melatonin and placebo.
Participants
We included six studies with 125 participants (106 participants aged under 18 years). One study was conducted in children, adolescents, and young adults with wake‐sleep disorder and mental retardation, most of them receiving chronic AED therapy (Coppola 2004). The principal investigator of Coppola 2004 indicated by mail (December 2011) that 21/25 participants who completed the study were aged under 18 years. The same group of investigators conducted two studies on children (Gupta 2004a; Gupta 2004b). When contacted by mail (29 December 2011), authors of one trial that also included adults specified that all participants enrolled (included participants lost to follow‐up) except one were aged under 18 years old (Hancock 2005). One study was conducted in children and adults with drug‐resistant epilepsy (Goldberg‐Stern 2012). One study was conducted in prepubertal children aged 6 to 11 years (Jain 2015).
One study was conducted in participants with or without epilepsy (Coppola 2004). We included this study but we analyzed only data from people with epilepsy. One study was conducted in people with tuberous sclerosis complex and sleep disorder (children and adults) (Hancock 2005). In this trial, all participants had epilepsy and were on concurrent AEDs.
Five studies compared the efficacy of melatonin versus placebo (Coppola 2004; Gupta 2004a; Gupta 2004b; Goldberg‐Stern 2012; Jain 2015), whereas Hancock 2005 compared two different doses of melatonin (5 mg versus 10 mg).
Adverse events
Only one study systematically reported adverse events (Hancock 2005).
Quality of life
Based on the protocol's requirement and the data collected from the included studies, the only validated quality‐of‐life scale adopted in the included studies was the QOLCE (Gupta 2004b). One study included the evaluation of sleep quality as an outcome; however, we excluded this trial for this outcome as the questionnaire used assessed sleep quality alone, and not the overall quality of life (Gupta 2004a).
Excluded studies
None of the seven articles obtained by the search strategy and eventually excluded (see Results of the search) appeared to meet the eligibility criteria. For studies excluded by the previous version of this systematic review please see: Brigo 2012.
Risk of bias in included studies
See: Characteristics of included studies table.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Three trials explicitly reported and adequately performed sequence generation (Gupta 2004a; Gupta 2004b; Jain 2015). In all three studies, a statistician, not connected to the study, prepared randomization code lists and the studies used computer‐generated permutation of code numbers for the treatment groups. In three publications, the authors did not explicitly state how they created sequence generation (Coppola 2004; Goldberg‐Stern 2012; Hancock 2005). However, after contact (by mail, 29 December 2011), the principal investigators of Hancock 2005 and Coppola 2004 indicated that allocation sequence was computer generated. We also contacted the principal investigator of Goldberg‐Stern 2012 (by mail, 27 April 2015), but he provided no further information.
Allocation concealment was adequate in five trials (Gupta 2004a; Gupta 2004b; Goldberg‐Stern 2012; Hancock 2005; Jain 2015), whereas in one study it was unclear due to lack of information (Coppola 2004). In four studies, the placebo tablets were identical in shape, size, colour, and packaging (Gupta 2004a; Gupta 2004b; Goldberg‐Stern 2012; Jain 2015). In Hancock 2005, participants received identical capsules (principal investigator specified that the placebo was made by the company who supplied the melatonin). We contacted the principal investigator of the fourth study by mail (29 December 2011), but he provided no additional information to evaluate allocation concealment (Coppola 2004). All studies except Jain 2015 reported reasons for loss of follow‐up, although no study performed an intention‐to‐treat analysis. Only two studies explicitly reported or systematically evaluated the presence of adverse events (Hancock 2005; Jain 2015). In two trials, the completeness of outcome reporting was unclear: each participant received a daily diary with the instruction to record any adverse events or unusual symptoms observed immediately; however, in the results sections, the authors stated only that no adverse events warranting discontinuation of the therapy were observed (Gupta 2004a; Gupta 2004b).
Effects of interventions
There were two different comparisons available (melatonin versus placebo; melatonin 5 mg versus melatonin 10 mg), but only the melatonin versus placebo comparison included more than one study.
Comparison 1: melatonin (any dose) versus placebo
Three trials evaluated melatonin versus placebo (Coppola 2004; Gupta 2004a; Gupta 2004b). The melatonin dose varied between trials. Despite our primary intention, due to insufficient information on each outcome, we were unable to perform a meta‐analysis.
Primary outcomes
Fifty per cent or greater reduction in seizure frequency
In one study, the authors stated that seizures occurred, without further specifying the exact number of seizures occurring in each group (melatonin/placebo) (Coppola 2004). We contacted the principal investigator of this study by mail (29 December 2011), but no additional information was provided. In one trial, which selected only participants who were seizure‐free for at least six months before the beginning of the trial, all participants (13/13 receiving add‐on melatonin, 16/16 receiving add‐on placebo) remained seizure‐free, so that a 50% seizure reduction rate during the treatment period from baseline could not be calculated (Gupta 2004a). One study gave no data regarding seizure reduction (Gupta 2004b); this trial selected only participants who were seizure‐free for at least six months before the beginning of the trial, so that 50% seizure reduction during the treatment period from baseline could not be calculated. In one cross‐over placebo‐controlled study conducted in 12 participants, the mean number of diurnal seizures was 7.75 during placebo treatment and 4.6 during melatonin treatment (P value = 0.034) (Goldberg‐Stern 2012). In this trial, three participants showed a decrease of 50% or greater in diurnal seizures during melatonin treatment compared to placebo. In one study conducted in prepubertal children with epilepsy without developmental delay, only two children had ongoing seizures, although had 50% reduction in seizure frequency with melatonin (Jain 2015). There was no increase in seizure frequency or difference in epilepsy frequency observed between melatonin and placebo.
The duration of trials (including follow‐up) varied from four weeks (Gupta 2004a) to nine weeks (Coppola 2004; Jain 2015).
Seizure freedom
Three studies reported data regarding seizure freedom (Coppola 2004; Gupta 2004a; Jain 2015). We requested, but did not receive, additional data on seizure freedom by mail for the fourth included study (Gupta 2004b). In one trial, all participants (13/13 receiving add‐on melatonin, 15/15 receiving add‐on placebo) remained seizure‐free at eight week' follow‐up (Gupta 2004a). One study described seizure freedom and seizure frequency without clearly specifying the exact number of seizures; authors stated only that out of the 11 seizure‐free participants before starting the study, nine remained unchanged on melatonin treatment (Coppola 2004). They did not report the number on placebo (if any) who remained seizure free. We contacted the principal investigator of this study by mail (29 December 2011), but he provided no additional information to clarify such an aspect (Coppola 2004). In the study by Jain 2015, eight participants who were seizure‐free remained seizure‐free during the study.
Incidence of adverse events of any type
Only one study reported adverse events systematically (Jain 2015). In two trials, no adverse effects were reported (Coppola 2004; Goldberg‐Stern 2012). In two trials, the completeness of outcome reporting was unclear (Gupta 2004a; Gupta 2004b). The trial provided a daily diary for each participant with the instruction to record any adverse events or unusual symptoms observed immediately. However, in the results section the authors stated only that no adverse events warranting discontinuation of the therapy were observed. In the study by Jain 2015, four children reported adverse events while taking melatonin (increased severity of headache in a child with history of migraine, bronchitis and ear infection, agitation, and increased urinary frequency) as compared to two participants taking placebo (agitation and increased urinary frequency); the authors of the study considered only the worsening of headache in the child with migraine to be related to melatonin treatment.
Secondary outcomes
Improvement in quality of life
Only one study systematically evaluated quality of life, showing no statistically significant improvement in quality of life in either group (valproic acid plus melatonin group: intragroup P value = 0.08; valproic acid plus placebo group: intragroup P value = 0.16) (Gupta 2004b). The authors performed no intergroup statistical evaluation between valproic acid plus melatonin and valproic acid plus placebo.
One study evaluated sleep quality alone, without focusing on global quality of life (Gupta 2004a). One study did not provide quantitative or validated data (Coppola 2004). The authors stated only that "half the parents/caregivers preferred improved behavior and alertness in children who appeared more quiet and better disposed to rehabilitation treatment; familial environment improved concomitantly, as a consequence of a better quality of night time" (Coppola 2004).
Comparison 2: melatonin 5 mg versus melatonin 10 mg
One study evaluated melatonin 5 mg versus melatonin 10 mg (Hancock 2005).
Primary outcomes
Fifty per cent or greater reduction in seizure frequency
The study reported the exact number of seizures occurring during the trial, compared to the baseline. None of the four participants who had seizures during the trial (on either melatonin 5 mg or 10 mg) had a 50% seizure reduction during the treatment (on either melatonin 5 mg or 10 mg) period compared to baseline. The duration of the trial (including follow‐up) was six weeks.
Seizure freedom
During the study three out of seven participants who completed the study remained seizure‐free (on either melatonin 5 mg or 10 mg).
Incidence of adverse events of any type
The study authors asked the carers to record any illness the child had or any possible adverse effects experienced during the trial period. No adverse events were reported.
Secondary outcomes
Improvement in quality of life
The study did not report improvement in quality of life.
Discussion
This systematic review included six trials (two studies (Goldberg‐Stern 2012; Jain 2015) were added to the four RCTs already included in the previous version of the review (Brigo 2012)). They did not systematically evaluate seizure frequency and adverse events, so it was impossible to summarize data in a meta‐analysis. Only two studies systematically examined seizure frequency occurring during the trial, compared to baseline (Hancock 2005; Jain 2015). With the exception of two trials (Hancock 2005; Jain 2015), none of the included trials systematically evaluated seizure freedom and adverse events. Two trials selected only people who were seizure‐free for at least six months before the beginning of the trial, so that 50% seizure reduction during the treatment period from baseline could not be calculated (Gupta 2004a; Gupta 2004b). From a clinical point of view, the reported follow‐up duration of the included studies was not long enough to evaluate the antiepileptic efficacy of add‐on melatonin, maybe because none of the trials was primarily designed to evaluate seizure freedom/reduction. Nevertheless, the relative high number of seizure‐free participants reported in three included studies may be attributed to the fact that a large proportion of them was already recruited as seizure‐free (three out of seven participants on either melatonin 5 mg or 10 mg in Hancock 2005, and all participants in Gupta 2004a and Gupta 2004b). Conversely, in one study conducted in 12 people with drug‐resistant epilepsy only three people experienced a decrease of at least 50% in diurnal seizures and no person was reported to be seizure free during melatonin treatment compared to placebo (Goldberg‐Stern 2012). Only two studies explicitly reported the presence of adverse events (Hancock 2005, none was reported; Jain 2015), and only one trial evaluated the direct influence of melatonin on quality of life showing to significant improvement (Gupta 2004b).
Five out of six studies included in the present review were primarily aimed to evaluate the effect of melatonin on wake‐sleep disorders, thus not focusing on efficacy and tolerability of add‐on melatonin as treatment for epilepsy. The fact that primary outcomes evaluated in most studies did not focus on seizure control is responsible for the lack of information regarding the efficacy of melatonin as add‐on treatment for epilepsy. As a consequence of this lack of data, it was not possible to draw definite conclusions concerning efficacy and tolerability of melatonin as add‐on treatment for epilepsy. Furthermore, it was impossible to establish whether a possible beneficial role of melatonin in epilepsy treatment was direct or indirect (i.e. seizure reduction as a consequence of an improvement in sleep quality).
Study flow diagram. This diagram refers only to the updated version of the review.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
